Radioiodine Therapy in Benign Thyroid Diseases: Effects, Side Effects, and Factors Affecting Therapeutic Outcome

Radioiodine (¹³¹I) therapy of benign thyroid diseases was introduced 70 yr ago, and the patients treated since then are probably numbered in the millions. Fifty to 90% of hyperthyroid patients are cured within 1 yr after ¹³¹I therapy. With longer follow-up, permanent hypothyroidism seems inevitable in Graves' disease, whereas this risk is much lower when treating toxic nodular goiter. The side effect causing most concern is the potential induction of ophthalmopathy in predisposed individuals. The response to ¹³¹I therapy is to some extent related to the radiation dose. However, calculation of an exact thyroid dose is error-prone due to imprecise measurement of the ¹³¹I biokinetics, and the importance of internal dosimetric factors, such as the thyroid follicle size, is probably underestimated. Besides these obstacles, several potential confounders interfere with the efficacy of ¹³¹I therapy, and they may even interact mutually and counteract each other. Numerous studies have evaluated the effect of ¹³¹I therapy, but results have been conflicting due to differences in design, sample size, patient selection, and dose calculation. It seems clear that no single factor reliably predicts the outcome from ¹³¹I therapy. The individual radiosensitivity, still poorly defined and impossible to quantify, may be a major determinant of the outcome from ¹³¹I therapy. Above all, the impact of ¹³¹I therapy relies on the iodine-concentrating ability of the thyroid gland. The thyroid ¹³¹I uptake (or retention) can be stimulated in several ways, including dietary iodine restriction and use of lithium. In particular, recombinant human thyrotropin has gained interest because this compound significantly amplifies the effect of ¹³¹I therapy in patients with nontoxic nodular goiter.

• Introduction

• The Story of ¹³¹I Therapy

• Cellular Effects of Radiation

• Effects of ¹³¹I Therapy

• Adverse Effects of ¹³¹I Therapy

  • Posttherapy hyperthyroidism

  • Transient hypothyroidism

  • Thyroid swelling

  • Sialadenitis

  • Immunogenic effects

  • Carcinogenicity

  • Teratogenicity

• Modulators of ¹³¹I Therapy

  • Applied dose

  • Thyroid size

  • Age and gender

  • Thyroid autoantibodies

  • Severity of disease

  • Effect of stunning

  • Thyroid radioiodine uptake

  • Smoking

  • Beta-blockers

  • Corticosteroids

  • Antioxidants

  • Antithyroid drugs

• Enhancers of ¹³¹I Uptake

  • Low-iodine diet

  • Stable iodine and lithium

  • Diuretics

  • Recombinant human TSH

  • Other compounds

• Closing Remarks, Remaining Questions, and Directions for Future Research

I Introduction

¹³¹I therapy has been used for seven decades, and an overwhelming body of data regarding effects and side effects has accumulated. However, the field is characterized by a lack of consensus regarding indications, logistics, procedures, and other aspects related to the clinical care of the patients considered for ¹³¹I therapy. The reason for this is undoubtedly the slowly evolving use of ¹³¹I, beginning in the 1940s. At the time of introduction, concepts like “Evidence Based Medicine” and “Good Clinical Practice” were either unknown or were given little attention, and national and international guidelines were yet to appear on the horizon. It follows that each center implemented ¹³¹I therapy according to local facilities and practice, traditions, and personal experience—routines that inherently are difficult to change. It is unknown how many ¹³¹I treatments have been administered worldwide through the years, but they are probably numbered in the millions. It is therefore striking that large well-designed prospective trials resolving fundamental questions in relation to ¹³¹I therapy are very sparse, despite the high prevalence of thyroid diseases. Viewed cynically, this discrepancy may in part rely on the limited commercial interest within the thyroid field and the difficulties in financing such trials.

After a brief description of ¹³¹I therapy in a historical perspective and of the biological effects of radiation, this review covers the effects and side effects of ¹³¹I therapy in benign thyroid diseases. Much focus will be on factors with potential influence on the outcome from ¹³¹I therapy and on the various methods used for enhancing the efficacy. Reviews covering ¹³¹I therapy of thyroid cancer patients can be found elsewhere (1–4), but data from studies in cancer patients will be discussed if relevant in this context.

II The Story of ¹³¹I Therapy

In 1895, it was discovered that iodine was a constituent of the thyroid gland, and 20 yr later it was demonstrated that the gland could take up iodine actively. In 1923, Henry Plummer introduced iodine as adjunctive treatment to surgery, which was at that time the only treatment for Graves' disease. Soon thereafter, iodine became an important component in the treatment of Graves' disease. György Hevesy established, also in 1923, the principle of using radioactive substances (“tracers”) to assess biological processes, but further research in that field was limited by the lack of naturally occurring radioactive substances. A breakthrough came in 1934 when Enrico Fermi described the artificial production of 22 new radioactive elements, among these iodine isotopes, by irradiating aluminum foil with an α-emitting source.

The historical background of the use of radioactive iodine was described in detail by Sawin and Becker (5) in 1997, and the story in their beautiful paper deserves a short revival in the present review. The idea of using radioactive iodine for thyroid research emerged on November 12, 1936, during a luncheon held at Harvard Medical School by Dr. Karl Compton, president of the Massachusetts Institute of Technology (MIT). The topic was “What Physics Can Do for Biology and Medicine,” and among the attendants were Professor Howard Means, head of the Department of Medicine at the Massachusetts General Hospital (MGH), and Saul Hertz, head of the Thyroid Clinic. At the end of Compton's lecture, the physicians asked whether there was a radioactive isotope of iodine. The physicists recalled Fermi's paper, which the thyroid physicians had never heard of. From then on, a MIT-MGH joint venture developed to produce and use radioactive iodine. The MIT part was delegated to Arthur Roberts, a young physicist, whereas Hertz took charge of the MGH part. Hertz and Roberts did their first experiment with ¹²⁸I on a series of rabbits in late 1937. Their experiments were limited by the small amounts of radioactivity and a half-life of ¹²⁸I of 25 min. However, their animal experiments confirmed that the thyroid gland was able to concentrate ¹²⁸I, and their report, the first to show radioiodine uptake (RAIU) by the thyroid gland, appeared in May 1938 (6). The researchers also realized that the RAIU was dependent on the intake of stable iodide, much used at that time to treat hyperthyroidism. The potential for a wider use of isotopes was evident, but stronger sources and isotopes with longer half-lives were needed. Only a cyclotron could meet these demands. A Californian group, having access to the cyclotron at Berkeley, had shown in humans that the thyroid gland took up radioiodine, although the result was not quantified (7). A new cyclotron, the first built exclusively for medical purposes, started operating in September 1940 at the MIT-MGH. Then, on November 4, 1940, Roberts and Hertz gave a longer-lived radioiodine to a woman with hyperthyroidism and estimated that her thyroid gland took up about 80% of the radioiodine. This was the first quantitative thyroid RAIU at MGH.

The intention was initially to study human iodine physiology, not to treat the disease. Most of the radioiodine produced by the cyclotron was ¹³⁰I, which has a half-life of 12 h, and about 10% of the cyclotron product was ¹³¹I. Hertz and Roberts gave radioiodine to a patient with the intention to treat hyperthyroidism for the first time on March 31, 1941. They used ¹³⁰I because its radiation was delivered rapidly to the thyroid cells over a day or two. The two co-workers continued to treat about one new patient per month for the rest of 1941. The total estimated radioiodine given to each of the eight patients ranged from 55 to 230 MBq, with an average of 144 MBq. Most of the radioiodine was taken up by the patients' thyroid glands, and the patients did in fact get better. A confounder was that Hertz gave each patient a fairly large amount of stable iodine beginning 1–3 d after the radioiodine. He did this at the insistence of his chief, Prof. Means, to protect the patients against thyrotoxicosis, should the radioiodine therapy not succeed. However, it was unknown whether the stable iodine interfered with the destructive effect of the radioiodine. At the American Society for Clinical Investigation meeting in May 1942, Hertz presented a series of eight patients treated with radioiodine and followed for at least 3 months; according to the abstract, there were both “failures and successes.” The Berkeley team in California also presented a small series of three patients at the same meeting. Apparently, the abstracts received little attention, but these were the first public reports of the successful treatment of hyperthyroidism with radioiodine. Neither paper was presented to American thyroidologists because their annual meetings were cancelled for the duration of the war. The difference between the two research teams was that Hertz from the MIT-MGH team was enthusiastic about this new treatment, whereas the Berkeley team had much concern about its use.

Hertz continued to treat hyperthyroid patients with ¹³⁰I throughout 1942. In January 1943, Hertz joined the U.S. Navy, and he asked Earle Chapman to follow up the 24 patients treated so far. Hertz was reluctant to publish his results until he was more confident of the long-term outcome of the treatment. When Chapman took over, he realized that it was impossible to settle whether any improvement in the patients was due to ¹³⁰I or to the stable iodine given afterward. Although there was some initial reluctance from his chief, Prof. Means, Chapman started to give ¹³⁰I without stable iodine. In the fall of 1945, Chapman prepared a paper for the Journal of the American Medical Association (JAMA) on 22 patients treated in the period 1943–1945. Hertz, returning from the war, was very disappointed when he learned that Chapman had prepared a paper on his own patients, and this resulted in much dispute between the two colleagues. Hertz quickly completed the follow-up of his patients treated in the period 1941–1943, and he submitted a separate paper, also to JAMA. In Hertz's paper, 21 of 29 patients were euthyroid. The addition of stable iodine probably had little effect because most of the radiation of the thyroid cells from the ¹³⁰I occurred in the first day. Chapman had essentially the same results in 22 patients, also treated with ¹³⁰I but without the addition of stable iodine. The dose of ¹³⁰I was larger, and although follow-up was shorter, four patients had developed hypothyroidism. The editor of JAMA, Morris Fishbein, was puzzled by receiving two different manuscripts completely independent of each other from the same institution. Reassured by Prof. Means, the head of the institution, that the two manuscripts described entirely different patients, Fishbein decided to publish both papers. These appeared in the same issue of JAMA on May 11, 1946 (8, 9).

After the war, the U.S. Atomic Energy Commission was allowed to supply the fission products for peaceful medical use. After this service began in August 1946, no one used ¹³⁰I anymore because ¹³¹I was much cheaper. From early on, physicians were aware of the possible risk of radioiodine-induced malignancies, and it was therefore the policy to make yearly examinations of each treated patient. Because it appeared that radioiodine therapy was not associated with any greater cancer risk, Prof. Means, within a few years, changed his attitude and now considered radioactive iodine to be the best agent for treating Graves' disease (10). At that time, ¹³¹I therapy was about to become well established worldwide.

III Cellular Effects of Radiation

Ionizing radiation has profound effects on living cells. In high doses, it leads to genetic damage, mutations, or cell death. At all times, living organisms have been exposed to potentially harmful radiation from natural sources such as sunlight and cosmic radiation. In addition, the intracellular aerobic metabolism produces reactive oxygen species and lipid peroxidation. Thus, it has been estimated that 10⁹ free radicals are produced in every cell each day due to the endogenous oxygen turnover, mainly by the mitochondria (11). Exogenous and endogenous antioxidants (such as vitamin C, glutathione peroxidase, and superoxide dismutase) to a high degree protect against the damaging effects of these reactive species. Nevertheless, mammalian cells undergo about 10,000 measurable DNA modification events per cell per hour, in part due to the spontaneous hydrolysis of nucleotide residues (12). To maintain DNA integrity and gene function, the organism is armed with a variety of strategies to respond to and repair these damages. Thereby, the number of persistent DNA changes declines to less than 100 per cell per day, and apoptosis and immunological responses reduce the number of persistent mutations even further (11). DNA damage escaping this array of defense mechanisms is carcinogenic and constitutes a main feature of cell aging.

The yearly background radiation is around 3 mSv, and this dose has much less impact on the DNA compared with endogenous factors (11). By increasing the radiation, the antioxidative defense and the repair mechanisms are challenged. It is believed that exposure to all radiation, including that from ¹³¹I therapy, has a negative impact on biological systems in a cumulative manner. However, the dose-response relationship is very complex and involves adaptive processes that differ between in vivo and in vitro conditions (13, 14). In fact, cells exposed to low doses of radiation (1–100 mGy) might become relatively resistant to the DNA-damaging effects of a subsequent high-dose radiation due to the preconditioning of the defense systems and an increased capacity for DNA repair (14, 15).

Insight into the cellular response to ionizing radiation and the repair mechanisms is rapidly expanding (12, 16–18). DNA damage from radiation occurs either by a direct effect—through breakage of molecular bands—or indirectly through the formation of free radicals (19). The DNA damage is diverse and comprises single-strand breaks, double-strand breaks (DSB), sugar and base modifications, oxidative damage of bases, inter-strand cross-links, DNA-protein cross-links, and locally multiply damaged sites (LMDS; clusters of ionizations). The most harmful forms of damage are probably DSB and LMDS (12, 16, 18). More than 100 genes are involved in DNA repair, and two main repair pathways have been recognized: homologous recombination repair (HRR) and nonhomologous end-joining (NHEJ), each employing separate protein complexes (16, 18). Irradiation typically generates LMDS, including single-strand breaks, DSB, and base damage. These LMDS, not produced endogenously in a significant number, are long lived and constitute the most biologically relevant DNA damage induced by radiation (16). NHEJ seems to be the dominant pathway for the removal of these lesions. Because the NHEJ repair system ligates two DSB without the use of a homologous DNA template, this pathway is much more error-prone than the HRR pathway (12, 18). In cell lines deficient in HRR, NHEJ, or both pathways, radiation-induced chromosome aberrations are increased manifold in a complementary fashion (12, 18).

Besides the DNA damage, irradiation causes a spectrum of other lesions in the cellular compartments due to the production of reactive oxygen species and lipid peroxidation of the plasma membrane. A range of protein complexes and signal transduction pathways—here among protein kinase C, c-Jun N-terminal kinases, ceramide, and MAPK—are involved in the downstream events after exposure to irradiation (12). The activation of these pathways leads to the translocation of different proteins from the cytoplasm to the nucleus. A key protein seems to be p53, which coordinates the two principal DNA repair pathways, HRR and NHEJ, with the fate of the cell. After radiation exposure, p53 is transported to the nucleus, where it mediates the transcription of specific genes directing the cell toward apoptosis or cell cycle arrest (12). Radiation-induced cell death is usually considered the result of apoptosis. Strictly speaking, apoptosis is interphase cell death occurring before the first postirradiation mitosis. In most cases of tissue irradiation, however, the mode governing cell death early after irradiation should be termed “mitotic-linked death” rather than apoptosis because the event usually occurs after the first mitosis due to fatal DNA damage (17). Radiation-induced apoptosis/mitotic-linked death can be influenced by extracellular factors. Thus, growth factors and cytokines like IGF-I, TNF-α, and IL-1 have antiapoptotic properties, and molecularly targeted agents like cetuximab, gefitinib, and erlotinib, all of which interfere with the repair of DSB, have been used to enhance the response to radiotherapy (17).

It has become apparent that the action of radiation is not limited to the target cell (12). Nontargeted tissue, in which the DNA is not directly exposed to radiation, is able to develop a wide range of responsiveness, including genotoxic effects. Genomic instability covers the accumulation of the multiple changes required to convert a normal stable genome into an unstable genome characteristic of a tumor cell. Radiation-induced genomic instability is usually associated with chromosomal aberrations and becomes manifest in the progeny of exposed cells many generations after the initial insult (12). Another nontargeted effect is the “bystander effect,” which concerns a signal generated from irradiated cells that convey manifestations of damage to nonirradiated cells, either through gap junctions or secreted soluble factors (12). For example, internally deposited α-emitters result in chromosome aberrations in cells not traversed by the particles (12). Reactive free radicals, in combination with lower basal activities of the detoxifying enzymes, are involved in these delayed effects. In addition, free radical species as well as cytokines are features of inflammatory responses having the potential for both bystander-mediated and persistent cell damage (12).

Some cancer patients who undergo external radiotherapy develop adverse effects in the surrounding tissue, indicative of an increased radiosensitivity (20). The occurrence and severity of these reactions are probably genetically determined. Indeed, a number of genetic disorders (e.g., xeroderma pigmentosum and ataxia-telangiectasia) are associated with a higher incidence of cancer and an increased susceptibility to radiation. Many of these diseases are due to defects in the enzymes responsible for DNA repair. Normal and tumor cell radiosensitivity is to some extent correlated, and several methods have been developed to assess the individual radiosensitivity (20). The micronucleus test, a simple cytogenetic test that detects chromosome fragility, correlates with the radiation toxicity in some studies (20). More sophisticated tests involve DNA microarrays to depict individual genetic patterns that are associated with clinical radiosensitivity (20). Yet, no ideal method exists that can predict the clinical response to radiation, including that from ¹³¹I therapy. The advent of such a test would have great clinical implications because the radiation dose then could be tailored much more accurately (Fig. 1).

Figure 1

Relationship between in vitro radiosensitivity, estimated by different tests, and the clinical effects of radiation. Ideally, results from in vitro tests are transferred into a tailored ¹³¹I therapy so that each patient is given a dose that optimally balances the effects and side effects. [Modified from M. H. Bourguignon et al.: Genetic and epigenetic features in radiation sensitivity. Part II: implications for clinical practice and radiation protection. Eur J Nucl Med Mol Imaging 32:351–368, 2005 (20), with permission. © Springer.]

The ¹²³I isotope, with a half-life of 13 h, is primarily used for diagnostic imaging due to its γ-radiation, which has a photon energy ideal for recording by a γ-camera. The ¹³¹I isotope emits β-particles as well as γ-rays, and with a half-life of 8 d, this isotope is very suitable in a therapeutic setting. The effect on the thyrocytes is primarily the result of the β-particle radiation, which has a path length of 1–2 mm. Exposure to other parts of the body occurs. Thus, after a single administration of 3700 MBq to thyroid cancer patients, a mild decline in white blood cells and platelet counts is seen that persists for at least 1 yr (21). As repeatedly shown (22–25), chromosomal damage in peripheral lymphocytes is also induced by ¹³¹I therapy, including the activities used for treating hyperthyroidism. Although the chromosomal damage, estimated by the micronucleus test in lymphocytes, correlates to some extent with the bone marrow dose (25), the impact is much lower after ¹³¹I therapy than after external beam radiotherapy for Hodgkin's disease or cervix uteri carcinoma (26). Cells with chromosomal damage after ¹³¹I therapy are eliminated with time, and full recovery seems to be achieved within 24 months after ¹³¹I therapy (13, 15, 27). Interestingly, supplementation of certain agents, like ginkgo biloba extract, neutralizes the genotoxic effect induced by ¹³¹I therapy for Graves' disease as well as for thyroid cancer, without affecting the clinical outcome (25, 28).

More exact methods for detection of DNA damage have become available. The histone variant γH2AX and the p53-binding protein 1 (53BP1) are markers of DSB formation (and repair), and γH2AX has been demonstrated in thyroid cell lines after external irradiation (29). Moreover, γH2AX as well as 53BP1, measured in peripheral blood leukocytes, peak at 2 h after ¹³¹I administration to thyroid cancer patients, in whom the blood receives an absorbed dose of less than 1 Gy (30). Pronounced interindividual variations are seen, but the markers are still elevated at 144 h, reflecting either a slower rate of DNA repair or a continuous de novo formation of DSB (30). Hershman et al. (31) recently developed an in vitro model in FRTL-5 cells by which γH2AX and 53BP1, as indicators of DSB, can be measured after internal radiation from ¹³¹I (Fig. 2). In their study, 4.3% of nonirradiated cells and 91% of cells incubated with 0.37 MBq/ml for 90 min were positive for 53BP1. Partial recovery occurred spontaneously (DSB < 50% after 24 h), and the viability of the cells was not affected by the irradiation. The radioprotective effect of iodide, perchlorate, and thiocyanate, all blockers of the Na+/I− symporter (NIS), were also studied. Perchlorate was 6-fold more potent than the two other agents (31), probably due to a higher affinity of NIS for perchlorate. Delayed addition of the agents to the cell cultures resulted in a time-dependent loss of the inhibitory effect, reflecting that the intracellular concentration of ¹³¹I was decreased by the blockade of NIS. Interestingly, natural organic compounds, like Yel-001 and Yel-002, showed additive radioprotective features in combination with perchlorate (31), which indicates involvement of mechanisms other than a blockade of NIS.

Figure 2

Induction of DSB in vitro in FRTL-5 cells after incubation with ¹³¹I and recovery after cessation of exposure. Top, ¹³¹I-induced DSB detected by γH2AX immunostaining, 90-min incubation. The solid line shows the relationship of DSB with concentration of ¹³¹I in the incubation medium; there was no increase of DSB at ¹³¹I concentrations greater than 10 μCi/ml. The dashed line shows the ¹³¹I taken up by cells at each concentration of ¹³¹I. Bottom, Recovery from DSB induced by exposure to ¹³¹I. The DSB of cells not incubated with ¹³¹I was 10% (basal DSB). The DSB at time zero is the maximum DSB after the 90-min incubation. After the removal of the ¹³¹I, the incubation was continued for 24 h, and immunostaining for 53BP1 was performed at various times. [Reproduced from J. M. Hershman et al.: Prevention of DNA double-strand breaks induced by radioiodide-131I in FRTL-5 thyroid cells. Endocrinology 152:1130–1135, 2011 (31), with permission. © The Endocrine Society.]

Thus, by its impact on the machinery of the thyrocyte, ¹³¹I therapy results in irreparable DNA damage and initiates a destructive process in the thyroid gland, which eventually leads to the clinical goals, i.e., a decrease in the thyroid function and/or a reduction of the thyroid volume. Late posttherapy histological findings include adenomatous nodules with cystic changes, oxyphilic cell changes, and various degrees of chronic thyroiditis, but there may be differences between ¹³¹I therapy of Graves' disease and other types of thyroid disorders (32, 33).

IV Effects of ¹³¹I Therapy

According to surveys done 20 yr ago, ¹³¹I serves, in most parts of the world, as a second-line treatment for Graves' disease in case of recurrence after a treatment course with antithyroid drugs (ATD) (34, 35). Also in the United States, recent data suggest that ATD seem to replace ¹³¹I as a first-line choice (36). The concept of ¹³¹I therapy evolved in the 1940s and 1950s, in parallel with the introduction of ATD. Both ¹³¹I therapy and ATD soon became established treatments for Graves' disease, to some extent replacing thyroidectomy. Although very different in their mode of action, it is evident that all available treatments—¹³¹I therapy, ATD, and thyroidectomy—are highly effective (37, 38). This is probably the reason why only one randomized controlled trial (RCT) (with a rather complex design) has been conducted to compare the treatments head-to-head (39). That study included 179 patients with Graves' disease, and after stratification for age patients were randomized to ATD in a block-replacement regimen for 18 months, subtotal thyroidectomy, or ¹³¹I therapy. The follow-up period was at least 48 months. Although the risk of relapse was highest in the ATD group, there were no significant differences in sick leave or satisfaction with the treatment (39). Without clear-cut differences in endpoints, other factors became (and still are) important for the choice of treatment, such as access to skilled surgeons and the ¹³¹I isotope, in addition to patient preference. Thyroidectomy seems more cost-effective than ¹³¹I therapy or lifelong ATD treatment in patients with Graves' disease who fail to achieve euthyroidism after an initial course of ATD (40). A recent study from the United States found that subtotal thyroidectomy is more cost-effective than ATD or ¹³¹I therapy when the initial postoperative euthyroid rate is greater than 49.5% and the total cost is less than $7391 (41). The results of such studies are, of course, highly dependent on the cost and set-up of each of the treatment options.

Numerous papers on this topic have been published (37, 42, 43). Retrieved via PubMed, and using the terms [hyperthyroidism/Graves'] and [radioiodine/¹³¹I], more than 2000 papers (including approximately 200 reviews) written in English appear during the period 1965–2010. Summarized, relief of hyperthyroidism is achieved within 3–12 months after ¹³¹I therapy in 50–90% of patients with Graves' disease. However, the success rate varies considerably between studies and is dependent on a range of more or less well-characterized factors. It appeared from the very start of the ¹³¹I era that many patients end up hypothyroid. Initially, this was considered to be due to overtreatment, and much effort has focused on an optimal dose algorithm, by which euthyroidism is rapidly obtained and preserved. Disappointingly, long-term follow-up studies have disclosed that hypothyroidism is an inevitable consequence of ¹³¹I therapy in the majority of patients with Graves' disease, even with low doses (44–46) (Fig. 3). Thus, the incidence rate of hypothyroidism is 5–50% within the first year, followed by a yearly rate of 3–5% (47). Using ¹³¹I therapy in toxic thyroid nodules offers a much more favorable ratio between success rate and risk of hypothyroidism than that seen with Graves' disease, mainly due to reduced ¹³¹I uptake in partly suppressed paranodular thyroid tissue. Thus, 75–95% of patients with a solitary toxic nodule achieve euthyroidism within 3–12 months, whereas less than 10% develop permanent hypothyroidism, even after long-term follow-up (48, 49). Similar success rates are seen in patients treated for a toxic multinodular goiter, but the risk of hypothyroidism seems to be higher, ranging from approximately 20% to as high as 75% within 8 yr, depending on the applied thyroid radiation dose (50, 51).

Figure 3

The cumulative incidence of hypothyroidism after ¹³¹I therapy for hyperthyroidism in patients with Graves' disease and in patients with toxic nodular goiter. The risk of hypothyroidism in the former group is very high, approaching 100% with long-term follow-up. [Reproduced from S. Metso et al.: Long-term follow-up study of radioiodine treatment of hyperthyroidism. Clin Endocrinol (Oxf) 61:641–648, 2004 (45), with permission. © John Wiley & Sons, Inc.]

Multinodular nontoxic goiter (MNG), causing compression symptoms or cosmetic complaints, is a very prevalent disease in iodine-deficient regions, and prevention of goiter is a major goal for iodine fortification programs. For many years, treatment of MNG was confined to either thyroid hormone suppressive therapy or thyroidectomy. From the experience with ¹³¹I therapy in Graves' disease, it appeared that shrinkage of the thyroid gland was pronounced (52), more so in Graves' disease than in toxic nodular goiter (53, 54), and that the impact on the thyroid volume paralleled the effect on the thyroid function (55, 56). Inspired by these observations, ¹³¹I therapy was introduced about three decades ago for treatment of MNG (57). Subsequent studies have unanimously proved the concept and demonstrated that ¹³¹I therapy reduces the MNG volume by approximately 40% after 1 yr and by 50–60% after 2–5 yr (58–66) (Table 1). In very large goiters (>100 ml), the reduction is less, around 35%, despite the application of equivalent thyroid ¹³¹I doses (64), but a nearly 75% reduction 3 yr after ¹³¹I therapy with higher thyroid doses has been reported (66). Patients with large substernal goiters have also been treated with beneficial results (67, 68). The individual response to ¹³¹I therapy is difficult to predict, as with Graves' disease, but symptoms most often improve, and patient satisfaction is high (60, 63, 69, 70). The treatment can be repeated if further goiter reduction is required (59). ¹³¹I therapy of diffuse nontoxic goiter, a condition much less prevalent than MNG, is even more effective with a more than 50% goiter volume reduction after 1 yr (69, 71, 72).

Table 1

Studies of ¹³¹I therapy in nontoxic and/or very large multinodular goiter employing an imaging method for monitoring

First author, year (Ref.)	n	Baseline goiter size (ml)	Design	Goiter size estimation	131I-activity per gram thyroid (MBq)	Goiter reduction
Hegedüs, 1988 (57)	25	73 ± 6 sem	Observational noncontrolled	US scan	3.7	41% after 1 yr
Verelst, 1990 (58)	15	175 ± 13 sem	Observational noncontrolled	Scintigraphy	2.8–3.7	39% after 2.5 yr
Nygaard, 1993 (59)	69	74 (21–296)	Observational noncontrolled	US scan	3.7	55% after 2 yr
Huysmans, 1994 (60)	19	269 ± 153 sd	Observational noncontrolled	MRI	3.7	40% after 1 yr
de Klerk, 1997 (62)	27	194 ± 138 sd	Observational noncontrolled	CT scan	1.1–4.8	34% after 1 yr
Le Moli, 1999 (63)	50	82 (17–325)	Observational noncontrolled	US scan	4.4	49% after 2 yr
Bonnema, 1999 (64)	23	311 ± 133 sd	Observational noncontrolled	MRI	3.7–5.5	34% after 1 yr
Bachmann, 2009 (66)	38	127 ± 39 sem	Observational noncontrolled	US scan	6–24	66% after 1 yr
Wesche, 2001 (65)	29	56 (17–198)	Randomized not blinded	US scan	4.4	44% after 2 yr

The baseline goiter size is presented as median (range) or mean ± sd/sem. In the study by Wesche et al. (65), patients were randomized to either ¹³¹I therapy or l-T₄ suppressive therapy. US, Ultrasound.

In some European countries, ¹³¹I therapy has now replaced surgery and thyroid hormone suppressive therapy as the treatment of choice in many patients with benign nontoxic goiter (73), and the use of this modality is now supported by current guidelines (74, 75). Thyroid hormone suppressive therapy, widely used until recently, has fallen out of favor due to poor efficacy and potential side effects from the induction of subclinical hyperthyroidism (76). In a randomized 2-yr follow-up trial, thyroid hormone—in comparison with ¹³¹I therapy—had no effect on goiter shrinkage (65). No randomized trial has compared surgery with ¹³¹I therapy. In many countries, ¹³¹I therapy is an outpatient treatment, and it seems superior to surgery as regards cost-effectiveness (77). In addition, most patients probably prefer nonsurgical treatment if offered. However, a major limitation of ¹³¹I therapy for MNG is a low thyroid RAIU, a problem that is rarely encountered in hyperthyroid diseases. Facing a patient with a low thyroid RAIU, treatment of MNG is usually restricted to surgery—unless the target tissue can regain its ability to trap ¹³¹I. For the latter purpose, recent data using recombinant human TSH (rhTSH) prestimulation offer intriguing prospects (78), and the topic will be discussed comprehensively later in this review.

V Adverse Effects of ¹³¹I Therapy

A few studies have evaluated the overall risk of ¹³¹I therapy. Franklyn et al. (79) found an increased mortality among 7209 hyperthyroid patients treated with ¹³¹I between 1950 and 1989. The excess mortality—related to thyroid, cardiovascular, and cerebrovascular diseases, and fractures of the femur—was most evident in the first year after therapy and declined thereafter. In a population-based study of 2668 individuals aged 40 yr or older treated between 1984 and 2002, the same group reported that ¹³¹I therapy was associated with an increased mortality, mostly due to cardiovascular diseases (80). Interestingly, the mortality was not increased among patients receiving levothyroxine therapy for ¹³¹I-induced hypothyroidism (80). Studies by Metso et al. (81, 82) demonstrated that the rates of death and vascular diseases were increased by 12%, and the risks remained elevated 35 yr after the ¹³¹I therapy. However, it is very difficult to make any conclusion on causality from population-based studies because any increased risk associated with ¹³¹I therapy may be caused by the hyperthyroid disorder rather than the treatment (83). It needs further clarification by well-controlled studies whether ¹³¹I therapy per se causes excess morbidity and mortality.

¹³¹I therapy is generally well tolerated, whether it is given for a hyperthyroid disorder or a compressive goiter. However, adverse effects may occur related to: 1) thyroid function; 2) thyroid size; 3) an immunological response; and 4) the consequences of extrathyroidal irradiation. Each of these topics is discussed in Sections V.A. to V.G.

A Posttherapy hyperthyroidism

In the early period after ¹³¹I therapy, a transient elevation of the thyroid hormone levels can be seen, probably due to release of stored hormone from the thyroid gland. More than 15 cases with ¹³¹I-induced thyroid storm have been reported (84), some of which have had a fatal outcome. In hyperthyroid diseases, the short-term impact of ¹³¹I therapy on the thyroid hormone levels depends on the pretreatment status. ¹³¹I therapy of overtly hyperthyroid and non-pretreated Graves' patients does not in general lead to an exacerbation of the hyperthyroidism, and the thyroid hormone levels decline a few days after therapy (85–88). The few patients showing a transient rise of the thyroid hormones are usually asymptomatic (87). As for toxic nodular goiter, the risk of a transient worsening of the thyrotoxicosis is apparently greater than in Graves' disease (87). In a study of ¹³¹I therapy of nontoxic goiter, the thyroid hormone levels were increased by 20% on d 7 and by 13% on d 14 after therapy, respectively, but had no clinical impact (89). These observations support that the pool of stored thyroid hormone is lower in a high turnover gland, such as Graves' disease, as compared with toxic and nontoxic nodular goiter.

To lower the risk of transient hyperthyroidism, many physicians pretreat hyperthyroid patients with ATD before ¹³¹I therapy. Although the intrathyroidal hormone stores are depleted by ATD, randomized trials (85, 86, 90) found no protection from transient hyperthyroidism, unless the drug was resumed after the ¹³¹I therapy (91) (Fig. 4). High levels of TSH-receptor antibodies (TRAb) at diagnosis are associated with such a posttherapy flare-up of the thyrotoxicosis (86). Thus, by cessation of ATD before ¹³¹I therapy, the hormone production quickly resumes until the destructive effect of the irradiation sets in. The period with an unsuppressed thyroid hormone production may last several weeks, occasionally leading to severe hyperthyroidism (92).

Figure 4

Effect of adjunctive ATD on the level of serum T₄ in the early period following ¹³¹I therapy illustrated by data from the randomized studies by Bonnema et al. (88, 91). Top, Hyperthyroid patients in the −PTU group were given ¹³¹I therapy naive of ATD. In patients randomized to the +PTU group, PTU was stopped 8 d before therapy with no resumption afterward. Serum T₄ was transiently increased in this group. The shaded area depicts the normal range. Bottom, Patients who were randomized to resume methimazole 7 d after ¹³¹I administration given for hyperthyroidism showed a steady level of serum T₄, whereas a transient increase was seen in those patients who did not resume methimazole (same situation as in the +PTU group in top panel). Methimazole was stopped 4 d before ¹³¹I therapy in both groups. The figure shows the relative changes in serum free T₄ index.

After ¹³¹I therapy, it is difficult to differentiate between treatment failure and transient hyperthyroidism. A retrospective study (93) indicates that treatment failure can be anticipated if the thyrotoxicosis has not improved within 3 months. Rarely, persistent and severe hyperthyroidism after ¹³¹I therapy of Graves' disease may be due to an immunological exacerbation caused by the therapy (94).

B Transient hypothyroidism

The majority of patients with Graves' disease show some degree of temporary pituitary dysfunction within the first months after ¹³¹I therapy, reflected by an insufficient TSH secretion (95). Moreover, hypothyroidism (defined as increased serum TSH and/or low levels of thyroid hormones) of transient character is seen among 3–20% of patients treated for hyperthyroidism (96–103). This condition is also encountered after subtotal thyroidectomy (104). In general, it is very difficult to predict the development of transient hypothyroidism and to establish its nature, whether transient or permanent. In fact, transient hypothyroidism may be a marker of treatment failure. In a Japanese study (102), hyperthyroidism recurred in six of 39 patients (15%) in whom transient hypothyroidism was observed. Alexander et al. (97) found that young age and a large thyroid gland were predictors for developing transient hypothyroidism succeeded by recurrent hyperthyroidism. In the study of Gómez et al. (101), a 2-h thyroid RAIU greater than 70% was a risk factor for developing transient hypothyroidism, but the risk was very low when activities exceeding 370 MBq were administered (101). At the time of diagnosis, the thyroid hormone pattern does not significantly differ from patients who develop permanent hypothyroidism, but recovery is unlikely to occur with a serum TSH above 45 mU/liter (101). Many patients with transient hypothyroidism are asymptomatic, and to avoid chronic medication, observation for some months without thyroid hormone replacement may be justified (96, 101, 103).

Suggested mechanisms behind transient hypothyroidism include a delayed recovery of the hypothalamic-pituitary axis after ¹³¹I therapy and depletion of the intrathyroidal iodine stores by ATD pretreatment. Subsequent studies have not supported these assumptions (101). Instead, transient hypothyroidism may have an immunological origin, mediated by a shift between TSH receptor blocking and stimulating antibodies, as indicated by some studies (102, 105, 106). Another explanation is that ionizing radiation in sublethal doses may lead to a hibernation-like condition (or prolonged stunning), which temporarily impairs the thyroid function. In support of this theory, Connell et al. (99) demonstrated that some (9%) patients with transient hypothyroidism showed evidence of an impaired, but short-lived, iodine organification and a normal iodine trapping by the thyroid, in contrast to the marked diminished iodide trapping with no recovery among patients who developed permanent hypothyroidism. The precise mechanisms behind the development of transient hypothyroidism are still unclarified, but immunological as well as cellular factors are likely to be involved and may vary between individuals. Importantly, when hypothyroidism develops after ¹³¹I therapy for toxic nodular goiter, this is unlikely to be transient (100), and thyroid hormone replacement can therefore be initiated without delay.

C Thyroid swelling

¹³¹I therapy may result in thyroid pain and a sensation of thyroid growth. The symptoms most likely reflect actinic thyroiditis (inflammation resulting from radiation) and usually vanish within a short time period with intervention rarely being needed. The condition is probably accompanied by an acute and variable temporary thyroid enlargement, asymptomatic in most patients. However, quantitative data are sparse and exclusively obtained from patients treated for a large nontoxic goiter, whereas there are no data from patients with Graves' disease. Indeed, considering the relatively high prevalence of tracheal compression and compromised inspiration in patients with a large goiter, even in the absence of symptoms (107, 108), an acute goiter enlargement may become critical, leading to respiratory distress (109). With the advent of rhTSH-stimulated ¹³¹I therapy, the topic has become even more relevant (110). Tracheal compression is related to the goiter size (108), but it is likely that the exact topo-anatomical relationship between the goiter and the trachea is also important. Thus, the benefit resulting from goiter shrinkage/removal should theoretically be greater if the trachea is encircled by the thyroid rather than just being displaced from the midline, but no studies have evaluated this aspect. It is worth noting that inspiration is more affected than expiration by upper airway obstruction. During inspiration, the higher air flow through a stenotic passage induces a negative transmural pressure gradient across the tracheal wall, and this may cause a partial collapse of the tracheal cartilage. During expiration, the drop in the transmural pressure is less critical because the prestenotic (i.e., intrathoracic) air pressure is above the atmospheric level.

Three studies (64, 70, 89) have measured the acute changes in thyroid volume after ¹³¹I therapy of nontoxic goiter, whereas two studies (64, 111) evaluated the acute impact on the upper airways. On average, the thyroid volume was unchanged during ¹³¹I therapy, but an enlargement by 15–25% was observed in a few patients (64, 70, 89). The smallest cross-sectional area of the trachea [measured by magnetic resonance imaging (MRI)] was decreased by 5–9% 1 wk after therapy. However, great interindividual variations with reductions by as much as 60% were observed (64, 111). The respiratory function was not significantly affected (64, 111). Although these data are reassuring, the series are relatively small, and a critical thyroid enlargement may occasionally result from ¹³¹I therapy of a large goiter (112). Corticosteroids are probably beneficial for preventing goiter swelling during ¹³¹I therapy and should be considered in case of tracheal compression, although such a strategy remains to be explored by a controlled trial (68).

D Sialadenitis

Because other cells besides the thyrocyte have the capability to accumulate ¹³¹I, some organs receive unintended irradiation during ¹³¹I therapy. The impact from radiation on the salivary glands has been studied exclusively in thyroid cancer patients. Despite the fact that the absorbed dose to the salivary glands is less than 0.5 Gy/GBq, as assessed by ¹²⁴I-positron emission tomography (PET)/computed tomography (CT) (113, 114), many patients develop a parotitis-like condition after ¹³¹I therapy. Salivary gland side effects, like swelling, pain, dry mouth, or altered taste, occurred in 39% of 262 thyroid cancer patients within the first year after high-dose ¹³¹I therapy, and they correlated with the administered radioactivity (115). Another study found a compromised salivary secretion in approximately 30% of the patients (116). It is mainly the composition of the saliva and the content of antioxidants and prostaglandins, rather than the saliva flow, which are affected (117, 118). The parotids are seemingly more vulnerable than the submandibular glands (119, 120) despite similar absorbed doses (113). Persistent side effects seem restricted to only a minority of patients (115), but some studies report that up to 20% of thyroid cancer patients suffer from xerostomy, even several years after ¹³¹I therapy (118, 121, 122). The compromised function of the salivary glands is of consequence. In a study of 176 thyroid cancer patients followed for 6.6 yr after ¹³¹I therapy, the risk of tooth extraction was significantly increased, and it correlated to the radiation dose and presence of xerostomy (123). The development of salivary gland cancer after exposure to ¹³¹I therapy has been reported (124).

Various methods to diminish the radiation damage to the salivary glands have been evaluated. Stimulation of the salivary glands, for example by lemon juice, leads to a faster ¹³¹I excretion from the salivary glands and a 40% reduction of the radiation dose (125). Such stimulation should not be commenced until 24 h after ¹³¹I administration (126) because too early stimulation, in fact, increases the absorbed doses to the salivary glands due to a rebound effect (127), and repeat stimulation with lemon juice may be necessary to avoid ¹³¹I reaccumulation (125). The effect of vitamin C stimulation seems limited (128), but a randomized study comparing vitamin C with lemon juice on the salivary absorbed dose has not been performed. Inhibitors of the salivary secretion, like amifostine and pilocarpine, have no effect on the salivary gland function beyond what can be obtained by acid-stimulating agents (129, 130).

Patients with autoimmune thyroid disorders may suffer from other autoimmune diseases, including Sjögren's syndrome. This disease affects the salivary glands, and the histopathological appearance has several similarities with that of autoimmune thyroiditis. Patients with Sjögren's syndrome can have subtle symptoms, and the disease may remain undiagnosed for years. Nevertheless, some studies indicate that as many as 30% of patients with autoimmune thyroiditis also have signs or symptoms of Sjögren's disease, a prevalence nearly 10 times higher than expected (131–133). If patients with autoimmune thyroid diseases, including Graves' disease, suffer from xerostomy, the salivary secretion is significantly reduced compared with patients without such complaints (133–135). Unfortunately, no prospective series are available in which the orodental status is evaluated systematically in patients with benign thyroid diseases. Neither has the impact on the salivary glands been studied in patients treated with ¹³¹I for a benign thyroid disorder.

E Immunogenic effects

The damage of the thyroid gland resulting from ¹³¹I therapy leads to an immunological response that may be disadvantageous—or beneficial. The TSH-receptor ectodomain on the thyrocyte is relatively unstable, and by the irradiation of the cell this part of the receptor may be shed into the circulation, acting as an epitope for the immune system (136). The immunoreactivity in Graves' disease can be assessed by the measurement of autoantibodies. The assays for these have been refined over time (137), and differences in affinity, sensitivity, and specificity may exist between studies. Nevertheless, it has long been recognized that ¹³¹I therapy results in a rise in thyroid autoantibodies, peaking approximately 3–6 months after therapy (55, 138–145). In a direct comparison with ATD and thyroidectomy, ¹³¹I therapy in Graves' disease leads to an initial surge of TRAb, in contrast to a steady decline of this antibody seen with the two other treatments (140, 146) (Fig. 5). Serum TRAb returns to the baseline level after 1 yr but is detectable even many years after ¹³¹I therapy (146, 147), and with higher levels than observed in patients treated with either ATD or by surgery (146). The increment in serum TRAb after ¹³¹I therapy seems dependent on the initial immunoreactivity (143), but even patients with Graves' disease who are TRAb-negative mostly turn TRAb-positive after ¹³¹I therapy (148). As for thyroglobulin antibodies (TgAb), the response seems biphasic with an initial rapid decline and a subsequent increase above baseline levels (149).

Figure 5

The immunogenic effect of ¹³¹I therapy. Variations in TRAb in serum after randomly assigning patients with Graves' hyperthyroidism to radioiodine, surgery, or ATD. Values before therapy were not different between groups. After therapy, all values were significantly higher for the ¹³¹I group when compared with values for the medical or surgical therapy groups. The stapled line indicates the upper normal reference for the TRAb assay. [Reproduced from P. Laurberg et al.: TSH-receptor autoimmunity in Graves' disease after therapy with anti-thyroid drugs, surgery, or radioiodine: a 5-year prospective randomized study. Eur J Endocrinol 158:69–75, 2008 (146), with permission. © European Society of Endocrinology.]

In parallel with the response in autoantibodies, the number of suppressor T cells declines after ¹³¹I therapy (150, 151), whereas serum levels of total Ig seem unaffected (149, 151, 152). A transient increase in both proinflammatory and antiinflammatory cytokines has been reported (153), but results have been conflicting (154). The most serious adverse effect of the ¹³¹I-induced (re)activation of thyroid autoimmunity is the de novo development of thyroid-associated ophthalmopathy (TAO), seen in approximately 15–33% of patients with Graves' disease (155–157). According to a Swedish randomized study, in which ¹³¹I was compared with ATD, worsening of TAO in patients who already had ophthalmopathy was not more common among patients receiving ¹³¹I therapy (157). However, previous (or preexisting) TAO is usually regarded as a risk factor for developing TAO after ¹³¹I therapy, among other predisposing factors like smoking, and a high disease activity reflected by either a T₃-dominated hyperthyroidism or a high titer of TRAb (157–160). Moreover, the risk of TAO resulting from ¹³¹I therapy is associated with a surge of TRAb and development of hypothyroidism (158). This immunological disturbance can be modified by intervention. Thus, whereas ATD are of no benefit for the prevention of TAO after ¹³¹I therapy (161), a large randomized study (156) has proved corticosteroids to be very effective, particularly when used iv (162). These drugs are now widely used for this purpose, although there is no consensus on the regimen used (163).

After ¹³¹I therapy of toxic or nontoxic nodular goiter, the transition into Graves' disease due to the de novo development of TRAb, and occasionally including TAO, is well described (164–171). This adverse effect occurs in 1–5% of patients, often with some latency, and the risk is strongly correlated with the presence of pretherapy thyroid peroxidase antibodies (TPOAb). The condition usually remits spontaneously within some months.

F Carcinogenicity

An important question, addressed by several large studies, is whether ¹³¹I therapy carries a risk of development of malignant diseases. In a recent large cohort study (172), secondary cancers were observed in 3,223 of 14,589 patients who had received ¹³¹I therapy for low-risk thyroid cancer, corresponding to an excess absolute risk of 4.6 cases per 10,000 patient-years. In particular, the risk of salivary gland malignancies and leukemia was significantly increased and correlated inversely with age. A recent Chinese study also found an increased risk of secondary malignancy in thyroid cancer patients previously receiving ¹³¹I therapy (173).

The risk related to ¹³¹I therapy of benign thyroid diseases—using lower amounts of radioactivity than for the treatment of thyroid cancers—is less clear. In a study from the United Kingdom (174) including nearly 7500 patients given ¹³¹I therapy for hyperthyroidism, there was a slightly increased incidence of thyroid cancer (nine cases vs. an expected three), but the absolute risk was small, and the overall standardized cancer mortality rate was in fact reduced. In another study by the same group (80), the long-term mortality was increased after ¹³¹I therapy, but this was not related to cancer deaths. In a Swedish study (175), including more than 10,000 patients treated with ¹³¹I therapy for hyperthyroidism (51% with Graves' disease) and a mean follow-up period of 15 yr, the overall cancer risk was only slightly greater than expected, and only among patients with toxic nodular goiter. Most of the increased mortality was due to thyroid cancer deaths, but the mortality diminished during prolonged follow-up, and the relation to the administered radioactivity was not evident (175). No significantly increased incidence of thyroid cancers was found in an Italian study (176) in which the majority of the patients had toxic nodular goiter. A Finnish single-center study (81) found, among 2793 ¹³¹I-treated patients, an increased mortality in both men and women, mainly from cerebrovascular disease. The mortality from cancer in the stomach, kidney, and breast was also increased (177), compared with an age- and sex-matched control group. The incidence was highest among patients with toxic nodular goiter, comprising 43% of the study population, and it correlated with the administered ¹³¹I activity. The number of thyroid cancers was insignificant. Interestingly, there was a lower mortality in subjects who developed hypothyroidism during follow-up (81), confirming the earlier findings of Franklyn et al. (80). In the largest study of its kind (178), 21,000 hyperthyroid patients (91% had Graves' disease) were followed after ¹³¹I therapy given between 1946 and 1964, corresponding to more than 385,000 patient-years. No overall increase of cancer mortality was found among patients with Graves' disease, whereas a 31% increase in overall cancer mortality was observed in the 1089 patients treated for toxic nodular goiter. This increase was exclusively attributable to thyroid malignancy, but a similar prevalence was seen in patients not treated with ¹³¹I (178). In the United States, children with Graves' disease have for many years been treated with ¹³¹I, and long-term follow-up studies have not revealed any development of thyroid cancer or leukemia in these subjects (179). However, because of a possible increased risk of thyroid cancer associated with low-dose thyroid irradiation, ablative thyroid doses of ¹³¹I are advocated in children (180).

No long-term data are available with regard to the cancer risk after ¹³¹I therapy for nontoxic goiter. By choosing ¹³¹I rather than surgery for treatment of this disease, it is obviously of paramount importance to rule out preexisting thyroid malignancy, primarily by performing biopsy in dominant and/or suspicious thyroid nodules (75). The administered ¹³¹I activity is usually higher when treating a large compressive goiter than in Graves' disease. Data from thyroid cancer patients treated with high activities of ¹³¹I cannot uncritically be extrapolated because the biokinetics differ with the thyroid gland in situ and the patient being euthyroid. During ¹³¹I therapy of very large goiters with large amounts of radioactivity, Huysmans et al. (181) found the absorbed doses in various organs to be considerably higher than received by ¹³¹I therapy of smaller goiters. The theoretical lifetime risk of an extrathyroidal cancer was estimated to be 1.6% and a little lower, 0.5%, when given to subjects older than 65 yr (181). Although these estimates are based on theoretical calculations, the level of radioactivity should be aimed at the lowest amount possible, while still achieving an adequate goiter reduction.

Although the observational studies (80, 174–178) include several hundred thousand patient-years, it is difficult to conclude whether any increased cancer risk is due to the ¹³¹I therapy or the disease per se. It seems that patients with toxic nodular goiter have a higher incidence of thyroid and extrathyroidal malignancy after ¹³¹I therapy than do patients with Graves' disease, but this may be due to the confounding influence of smoking and age. Probably the most important factor is that thyroid malignancy may have coexisted (but been overlooked) in a nodular goiter at the time of the ¹³¹I treatment—a problem that might exist also with Graves' disease (182–184). A large randomized study with long-term follow-up is necessary to resolve the issue of ¹³¹I-induced malignancy, but it can be questioned whether such a trial will ever be conducted. Nevertheless, the absolute risk of developing cancer after ¹³¹I therapy for benign thyroid diseases seems low or negligible. If present, the risk should be balanced against the fact that long-term use of ATD as well as thyroidectomy occasionally leads to potentially fatal complications.

G Teratogenicity

¹³¹I therapy may have long-term consequences for the gonadal function. When an ¹³¹I activity of 3700 MBq is given to thyroid cancer patients, the ovaries receive a radiation dose of 140 mGy (185), which may give rise to ovarian failure and/or adversely affect subsequent pregnancies. In a review including 16 observational studies with more than 3000 women treated with ¹³¹I for thyroid cancer (186), transient absence of menstrual periods varied from 8 to 27% within the first year of treatment, and menopause occurred at a slightly younger age than in women not treated with ¹³¹I (186, 187). More importantly, ¹³¹I therapy was generally not associated with a significantly increased risk of long-term infertility, miscarriage, stillbirths, or congenital defects in the offspring (186). A large European study (188) found increased rates of abortions in the first year after ¹³¹I therapy, but by a further 10-yr follow-up this finding became insignificant (185). These results from cancer patients are reassuring. Furthermore, among the offspring of subjects treated with ¹³¹I for hyperthyroidism or thyroid cancer during childhood and adolescence, the incidence of congenital anomalies was not increased (180).

Pregnancy is an absolute contraindication for ¹³¹I therapy, and conception should be postponed until at least 4 months after the therapy, according to recent guidelines (38). The fetal thyroid begins to develop at 5–6 wk, with follicle and colloid production at 10–12 wk gestation. ¹³¹I administration at this time or later during pregnancy will result in a high thyroid radiation dose (20–600 Gy) and risk of fetal thyroid ablation (189–192). In case of ¹³¹I therapy for thyroid cancer, a total fetal dose of approximately 700 mGy together with hypothyroidism is likely to be fatal for the fetus (189). However, when ¹³¹I in doses used for hyperthyroidism is administrated before the 10th week of gestation, a normal outcome has been reported in several cases (189–192).

The radiation dose to the male gonads is slightly less than to the ovaries (193). Nevertheless, abnormalities in testicular function are common after ¹³¹I therapy of thyroid cancer patients. Within 2 to 6 months, an elevation of serum concentrations of FSH and LH is found, whereas serum testosterone is unaffected (194). The gonadal function is completely recovered 18 months after ¹³¹I therapy, although repeated or high cumulative ¹³¹I activities carry a risk of persistent gonadal dysfunction (194). The rates of infertility, pregnancy loss, and offspring congenital malformations do not seem to be elevated, but data are based on small studies and limited by self-reported outcomes (193–195). ¹³¹I therapy given for hyperthyroidism is associated with only marginal and a reversible decrease of the testosterone level (196). It should be noted, however, that sperm motility is reduced by hyperthyroidism per se (197).

VI Modulators of ¹³¹I Therapy

It has become increasingly clear that a number of factors influence the effect of ¹³¹I therapy, whether the indication is hyperthyroidism or a compressive goiter. Despite abundant publications—some of which are shown in Table 2—most studies are inadequately designed or have other limitations, i.e., are retrospective, lack a control group, and suffer from selection bias or a too low sample size. However, based on the present knowledge, it seems clear that no single factor reliably predicts the outcome of ¹³¹I therapy. In addition, the divergent results between studies may be due to the influence of important confounders not taken into consideration. A number of factors with known or potential influence on ¹³¹I therapy are discussed below and shown in Fig. 6. Because these factors probably interact mutually in a complex manner—making it even more difficult to segregate the impact of each single factor—the need for randomized trials becomes even more imperative.

Figure 6

Cure from hyperthyroidism after ¹³¹I therapy is affected by a range of factors. Some of these act directly on the thyroid ¹³¹I uptake. Other factors modify the impact of radiation on the thyroid gland, either positively or negatively. Many factors are probably interrelated in a complex fashion. For example, the adjunctive use of ATD may negatively affect the cure rate, but this may be counterbalanced by the concomitant influence on serum TSH and the thyroid RAIU, both of which per se may be associated with the cure rate.

Table 2

Studies evaluating factors influencing outcome from ¹³¹I therapy

First author, year (Ref.)	n	Design	Dose adjusted for RAIU/volume	Thyroid RAIU	Thyroid volume	Thyroid dose	Age	Male gender	TRAb	TPOAb	TAO	Levels of thyroid hormones	Serum TSH	Smoking	Adjunctive use of ATD
Alexander, 2002 (97)	261	Retrospective	Yes/no	Neg	Neg	Pos	Pos	Absent	ND	ND	Neg	Neg	ND	ND	Neg
Boelaert, 2009 (291)	1278a	Cohort	No/no	ND	Neg	Posb	Absentc	Neg	ND	ND	Pos	Neg	ND	Absent	Absent
Bogazzi, 2010 (276)	651	Retrospective	Yes/yes	ND	Neg	ND	Absent	Absent	Neg	ND	ND	Neg	Absent	Absent	ND
Catargi, 1999 (241)	100	Retrospective	Yes/yes	Neg	Absent	Absent	Absent	ND	Absent	Absent	ND	Absent	ND	ND	ND
Hancock, 1997 (391)	116	Retrospective	Yes/yes	Absent	Absent	Absent	Absent	ND	ND	ND	ND	Neg	ND	ND	Neg
Jensen, 2005 (280)	207	Retrospective	Yesd/yes	ND	Nege	ND	Neg	Absent	ND	Absent	Absent	Absent	Pos	Absent	Absent
Kobe, 2008 (244)	571	Prospective	Yes/yes	ND	Neg	Pos	Absent	Absent	Absent	ND	Absent	ND	Absent	ND	ND
Körber, 2001 (279)	GD, 144 TNG, 563	Prospective	Yes/yes	ND	Neg	Pos Absent	Absent	Absent	ND	ND	ND	Neg	Absent Pos	ND	Absent Neg
Metso, 2004 (45)	GD, 1086 TNG, 957	Cohort	No/no	Neg	ND	Absentb	Pos	Neg Absent	ND	ND	ND	ND	ND	ND	Neg Absent
Reinhardt, 2002 (249)	224	Prospective	Yes/yes	Absent	Neg	Pos	Absent	ND	Posf	ND	ND	ND	Absent	ND	Absent
Sabri, 1999 (252)	207	Prospective	Yes/yes	ND	Absent	Pos	ND	ND	Absent	ND	ND	Absent	ND	ND	Neg
Zantut-Wittmann, 2005 (278)	82	Retrospective	No/no	Neg	Neg	ND	Absent	ND	ND	Absent	ND	Absent	Absent	ND	Neg

The influence of various factors on the outcome from ¹³¹I therapy has been evaluated by numerous studies. Data from 12 of these (freely selected by the authors) are presented in the table. Nine studies included exclusively Graves' disease, whereas three studies also included toxic nodular goiter. The correlation between each factor and the outcome are given (positive, negative, or absent).Outcome covers either cure from hyperthyroidism or development of hypothyroidism. Thyroid dose, smoking, and adjunctive use of ATD are external factors, and the impact of these could in theory be investigated in randomized trials. ND, Not determined; GD, Graves' disease; TNG, toxic nodular goiter; Pos, positive; Neg, negative.

^a

GD, TNG, and indeterminate etiology.

^b

Amount of radioactivity rather than dose.

^c

No/Negative association with cure/hypothyroidism, respectively.

^d

In a subset of patients.

^e

In patients receiving RAIU unadjusted dose.

^f

Positive association with hypothyroidism.

Cells, organs, and whole organisms vary in their sensitivity to ionizing radiation due to differences in the cell cycle, intracellular protective systems, and ability to recover. At present, we have little knowledge of this aspect of ¹³¹I therapy. A yet unknown “factor X” might exist that quantitatively mirrors the radiosensitivity of each individual and that could be taken into account when scheduling ¹³¹I therapy.

A Applied dose

In the beginning of the ¹³¹I era, the applied thyroid dose was empirically chosen. By 1950, the standard ¹³¹I dose in the United States was 6 MBq/g of estimated thyroid weight (10). A high incidence of hypothyroidism, reaching 50–80% after 10 yr (198), was observed using this dose. To avoid this side effect, a low-dose protocol was employed, by which the administered radioactivity was reduced by 50% or more. Although a lower incidence of early hypothyroidism was seen with this dose, being 12% 1 yr after treatment, it increased steadily to 76% after more than 10 yr of follow-up (44). In line with the lessons from other studies (45, 47, 199), it soon became evident that permanent hypothyroidism is inevitable in the majority of patients treated with ¹³¹I therapy for Graves' disease, if followed long enough. There have been a few studies reporting lower rates of thyroid failure. Thus, in a recent Chinese study including 600 individuals with Graves' disease, 72% of patients remained euthyroid after 12 yr if adjusted ¹³¹I therapy was given according to a clinical score system (200).

Regarding radiation protection, ALARA (as low as reasonably achievable) is an important principle. This means that the goal of the ¹³¹I treatment should be achieved with the lowest radiation burden to the patient as well as to the environment—an issue that particularly applies to ¹³¹I therapy of thyroid cancer patients, in whom administration of a high ¹³¹I activity carries a risk of red bone marrow suppression. The whole-body ¹³¹I retention in thyroidectomized cancer patients varies considerably, and by the administration of a fixed amount of ¹³¹I radioactivity, the maximum tolerated blood dose of 2 Gy (201) is exceeded in some individuals (202). The radiation dose to various organs is usually estimated according to the medical internal radiation dose (MIRD) biokinetics models based on a standardized situation (203). A recent study found that the ablation of tumor tissue correlates significantly with the absorbed dose to the blood (a surrogate marker for the ¹³¹I available for iodine-avid tissue uptake) and not with the administered activity (204). Some authors (205) advocate for the use of an individual and more accurate dose calculation, employing three-dimensional internal dosimetry and ¹²⁴I- and ¹⁸F-fluoro-2-deoxy-d-glucose-PET (114), to optimize the response rate to the ¹³¹I therapy (i.e., ablation of tumor tissue) while minimizing the radiation to the rest of the body.

As for the treatment of benign thyroid diseases, the radioactivity is much less than used for cancer patients, but nevertheless, superfluous radiation should be avoided. ¹³¹I is usually given as a single dose, but another approach, minimizing the use of radioactivity, is the administration of 75 MBq ¹³¹I at intervals of 6 months until the patient becomes euthyroid (206). When comparing a high-dose vs. a low-dose regimen, it is evident that a significant fraction of patients are overtreated (207–209). The effect of the ¹³¹I therapy generally correlates with the applied thyroid dose, but the dose-response relationship is not linear, and huge interindividual variations exist. In a retrospective study from the United Kingdom (207), two fixed activities of ¹³¹I (185 vs. 370 MBq) were compared. Patients given 370 MBq had a higher cure rate (85 vs. 67%) but also a higher rate of hypothyroidism at 1 yr (61 vs. 41%). In a randomized study (210) comparing a low-dose with a high-dose regime, the outcome in the two groups was similar, but the difference between the activities was modest (235 vs. 350 MBq), and the study may have been underpowered. Another randomized study in patients with Graves' disease also found similar cure rates by use of 370 or 555 MBq, respectively (211). It seems clear that the choice of dose must be balanced between the need for rapid relief of the hyperthyroidism on the one hand and postponement of the ensuing hypothyroidism on the other hand. Different policies exist, and the issue is still under debate (212, 213). Accordingly, there is no consensus on the algorithm for dose calculation, the applied dose, and the safety practice related to ¹³¹I administration (214). Admittedly, variations in dealing with ¹³¹I therapy to some extent also rely on differences in the official radiation regulations between countries.

1 Algorithm for dose calculation

Usually, the dose calculation is based on the old formula of Marinelli (215), which incorporates the thyroid ¹³¹I retention time and the thyroid volume. The thyroid retention is often assessed by a single ¹³¹I RAIU measurement while keeping the ¹³¹I washout rate constant. Mostly, the 24-h thyroid RAIU is measured, but an earlier time point can be used instead (and is of practical advantage) unless the patient has a very rapid ¹³¹I turnover (4 or 6 h RAIU >75–80%) (216). Some authors have proposed a formula that predicts the 24-h RAIU based on the 3-h RAIU, serum free T₃, and the ATD discontinuation period (217). A thyroid RAIU measured at 96 h or later may be a more reliable estimate of the ¹³¹I kinetics (218), but it cannot be neglected that simplified protocols misjudge the accumulated ¹³¹I thyroid retention (219–221). Indeed, significant intraindividual differences in the calculated activity are found when comparing different algorithms (222–225). Moreover, the kinetics depends on the type of disease and the iodine status of the patient. The effective ¹³¹I half-life is shorter in Graves' disease than in toxic nodular goiter, whereas the maximal thyroid RAIU uptake is higher (226). By using the most simplified protocols, the applied thyroid dose is imprecise and may lead to the administration of unnecessarily high activities (223–225). A special and challenging situation is treating patients with end-stage renal disease because the ¹³¹I cannot be cleared by the kidneys. Special precautions have to be taken in such cases, but ¹³¹I therapy is feasible by an approximately 5-fold reduction in the amount of radioactivity (227).

A valid estimation of the thyroid volume is also of crucial importance for calculating the radiation dose correctly. Because a clinical assessment is notoriously inaccurate (228), thyroid imaging is needed, preferably by MRI, CT, or ultrasonography (229). Scintigraphic measures, obtained by planar images or by single-photon emission CT, are less accurate (230). This is a shortcoming of most studies performed before the era of ultrasound.

Realizing that the thyroid dose is difficult to apply with sufficient accuracy, and that a number of other factors (to some extent unknown or uncontrollable) interfere with the effect of ¹³¹I therapy, a simpler approach is used by many centers. In general, two algorithms are used: 1) standardized activity estimation, using a fixed activity or roughly adapted to the individual thyroid size; and 2) dosimetric calculation, based on the ¹³¹I biokinetics and the thyroid volume. Nine controlled trials have been conducted (209, 210, 231–237), among which three were randomized (209, 210, 231). The studies varied in the pre- and posttherapy use of ATD, length of follow-up, type of hyperthyroid disease, and method for thyroid volume estimation, and only one study (209) included an assessment of the ¹³¹I turnover. A meta-analysis (238) including eight of these trials concluded that the two methods (estimated vs. calculated algorithm) result in similar cure rates. The application of a more simple regimen is supported by a retrospective study (239), which demonstrated that in most patients with hyperthyroid diseases the outcome after ¹³¹I therapy is unrelated to the scintigraphic diagnosis and the thyroid RAIU. However, considering the limitations and dissimilarities of the studies and lacking large and well-designed trials, it is still too early to dismiss an individually based and precise dose calculation, particularly when keeping the ALARA principle in mind.

2 Correlation between dose and outcome

The limitations concerning dose calculation may partly explain the conflicting results among studies on the dose-response relationship. Thyroid doses from below 50 Gy to more than 300 Gy have been used, and great differences in the response rate are seen within this dose range. Applying 70–120 Gy to patients with Graves' disease results in a high cure rate of 90% after 1–5 yr, according to some studies (218, 240). In the study of Catargi et al. (241), two thirds of patients were cured after 1 yr when aiming at a thyroid dose of 50 Gy, but responders and nonresponders did not differ in their applied doses, as measured during therapy. Similarly, a randomized study (242), comparing 90 Gy with 60 Gy, found an almost identical failure rate (around 60% at 6 months) in the two groups. In the large study by Alevizaki et al. (243), the rate of hypothyroidism was positively correlated with the dose in the range 15–150 Gy, whereas this association was very vague with higher doses. It seems clear that a dose of 250–300 Gy is necessary to achieve rapid relief from hyperthyroidism (244, 245), but at the expense of a high incidence of hypothyroidism, approximately 95% after 18 months (245). However, even with these high doses, repeat treatment is needed in some cases (97, 208, 246), indicating that pronounced differences in radiosensitivity exist among patients.

Despite meticulous pretherapeutic calculations, the true applied thyroid dose is often far from the target, and wide individual variations exist (51, 219, 221, 241, 247–250). This is probably one of the main reasons for the effect of ¹³¹I therapy—at the individual level—being poorly related to the thyroid dose based on pretherapeutic calculations, and the actual therapeutic dose being a better predictor (251). According to German studies (209, 244, 249, 252) in which the applied dose was measured, 200–250 Gy or more was usually needed to obtain acceptable cure rates. Thus, in the randomized study of Graves' patients by Peters et al. (209), the cure rate was only 11% for a target dose of 50 Gy, but increased progressively and reached 80% with doses above 200 Gy. In an Italian study (208), Graves' patients were divided into three groups receiving 150, 300, or more than 300 Gy (applied dose), respectively. The treatment failure rate at 1 yr did not differ between groups, whereas the incidence of hypothyroidism correlated with the dose.

The applied dose may be higher than intended in some cases, depending on the administered activity (241) and the size of the gland (209). Other studies (221, 250) found the therapeutic dose to be 15% below target, mainly due to a lower RAIU during therapy, whereas the clearance rate on average seems unaffected (221). “Self-stunning” and/or a reduced capacity for thyroid ¹³¹I trapping under therapeutic conditions may be possible explanations for the lower RAIU seen during therapy. An imprecise dose calculation may also be due to other factors, often being overlooked, like the considerable day-to-day variation of the thyroid ¹³¹I RAIU (253) and the early thyroid size reduction occurring during the ¹³¹I therapy (254). It has been postulated that ¹³¹I in capsule form has a lower bioavailability than in liquid form (255). However, ¹³¹I administered orally in capsules or iv results in similar therapeutic doses, as shown in a randomized study (256).

Importantly, dose calculations are based on MIRD reference models. If the isotope retention and excretion pathways differ significantly from standard values, the dose estimates are subject to pronounced error (257). Thyrocytes lining larger follicles receive less radiation than cells in smaller follicles due to self-absorption of ¹³¹I from the follicular colloid, as shown recently (258). Because the variation in radiation can be as high as 30% (258), the size of the thyroid follicles may turn out to be of crucial importance, but this aspect of thyroid dosimetry has not been addressed in previous studies.

Another issue that has received very little attention is the influence of the dose rate. For a given amount of radiation accumulated in the target tissue, the biological effect might depend on the time period during which the dose is delivered. It is unsettled whether ¹³¹I half-life, reflecting the ¹³¹I turnover, is related to the outcome of ¹³¹I therapy (249, 259), but the predictive value of this variable has not been evaluated to any great extent. However, the half-life as a marker of the dose rate has shortcomings, partly because the assumption of ideal first-order ¹³¹I kinetics is rarely fulfilled. In the study by Puille et al. (260), patients with autonomous nodules were stratified, retrospectively, into two groups according to the dose rate. There were no differences in the total applied radiation dose. In patients with the high dose rate, the ¹³¹I therapy seemed slightly more effective, reflected by significantly higher levels of serum TSH after therapy (260). These findings certainly need to be confirmed, but it may be speculated that the application of the radiation dose within a narrow time window leads to more cellular damage because the cellular defense systems are not fully mobilized. If this holds true, the biological effect of ¹³¹I therapy could be amplified by increasing the dose rate, independently of how this is achieved.

3 Dose calculation in the toxic nodular goiter

As for the toxic nodular goiter, the dose-response issue is even more complex because the target volume (i.e., autonomous nodules) most often differs from the total thyroid volume. In the study by Allahabadia et al. (207), using fixed activities, no difference in the cure rates between Graves' disease and toxic nodular goiter was found. However, the thyroid dose was lower in the latter group due to larger goiters, and the thyroid RAIU probably also differed in the two diseases. By taking these key parameters (goiter size and thyroid RAIU) into account, the cure rate is usually higher and the incidence of hypothyroidism lower than seen in patients with Graves' disease (47, 48, 50, 88, 261, 262). Thus, the long-term constant incidence rate of hypothyroidism seen when treating patients with Graves' disease is much lower in patients treated for toxic nodular goiter (48–50), although higher rates have been reported, depending on the applied thyroid dose or adjunctive use of ATD (51, 263).

The determination of the target volume of a solitary toxic nodule can be done easily by ultrasound. In the case of a toxic multinodular goiter, this is much more complicated, or even impossible, because the gland consists of autonomous and nonautonomous nodules as well as paranodular tissue. To overcome this problem, it has been suggested (264) that the dose should be reduced by 50% when considering the entire thyroid volume as the target tissue. Studies using such an approach have shown high failure rates with doses less than 200 Gy (264). Single-photon emission CT or PET have the potential to determine the functional autonomous volume more precisely (265, 266), but these methods have not gained widespread use. German researchers (267, 268) have suggested that dose selection should be based on the technetium thyroid uptake under exogenous or endogenous suppression (TcTUs). The value of TcTUs correlates well with the sonographically determined volume of autonomous nodules, and five times TcTUs can be regarded as a measure of the functional volume (267, 269). Accordingly, German guidelines (270) recommend target doses of 300–400 and 150 Gy to patients with a solitary toxic nodule and toxic multinodular goiter, respectively. Application of such a dosimetric approach was evaluated in a large series of patients with autonomous nodules, showing a normalization of serum TSH in 92% of the patients, whereas hypothyroidism occurred in merely 0.9% after 1 yr (268). A modification of the TcTUs-based algorithm has been proposed because the failure rate seems inversely correlated with the TcTUs (271). In comparison with a conventional algorithm that includes the entire goiter volume as the target tissue, the modified algorithm based on TcTUs does not yield better results (264), and it can be questioned whether estimation of this parameter is worthwhile.

4 Dose and goiter volume reduction

The relationship between the radiation dose and the goiter reduction after ¹³¹I therapy has been much less explored. In a randomized study (52), patients with Graves' disease were stratified according to three dose categories—less than 100 Gy, 100–200 Gy, and more than 200 Gy; the median thyroid reduction obtained at 6 months was clearly dose-dependent, being 45, 56, and 67%, respectively. As for the compressive multinodular goiter, the issue has been addressed by post hoc analyses (62–64, 66, 70, 272). The administered ¹³¹I activity was in the range 3.7–14 MBq/g, corrected for the thyroid RAIU, which corresponded to a thyroid dose in the range 100–175 Gy. Four studies have assessed the applied thyroid dose (64, 66, 70, 272). The mean goiter reduction was most pronounced (66% after 1 yr) in the study (66) using the highest thyroid dose (175 Gy), but generally results have been conflicting, most likely due to lack of power and an inaccurate dose calculation in some studies. A head-to-head comparison of different doses was not done in any of the previous studies (62–64, 66, 70, 272), and future research should explore whether a thyroid dose lower than usually applied leads to a similar goiter reduction or, conversely, whether better results could be obtained by using higher doses. Recently, we showed, in a randomized trial, that application of 50 Gy preceded by rhTSH-stimulation resulted in a similar goiter reduction as that obtained by 100 Gy without rhTSH prestimulation (273). Whether this observation reflects that 50 and 100 Gy are equally effective for goiter reduction or whether rhTSH prestimulation has a preconditioning effect on the ¹³¹I therapy is unclarified.

B Thyroid size

The pretreatment thyroid volume might well be an independent predictor of the response to ¹³¹I therapy. Unfortunately, the findings in most studies are highly confounded by the applied thyroid dose not being taken sufficiently into account. Many centers use a fixed or semi-fixed ¹³¹I activity, as previously discussed. Based on such simplified algorithms, the thyroid dose will decrease with an increase in thyroid volume—a relationship further reinforced by the fact that small glands have a lower absolute RAIU but a significantly higher relative RAIU, compared with larger glands (274). Another serious limitation in many studies is the application of a clinical or scintigraphic estimate of the thyroid volume, disregarding that both methods are more imprecise than ultrasound and CT/MRI (228, 230, 275).

Differences in methods and the lack of differentiation between patients with toxic nodular goiter and Graves' disease in some of the studies are probably the main reasons for the inconsistent findings. Thyroid volume is inversely correlated with the effect of ¹³¹I therapy in some studies (55, 97, 207, 218, 242, 249, 276–283), whereas others find no such correlation (51, 88, 209, 210, 252, 261, 284). Studies in which the applied thyroid dose was assessed during therapy also showed conflicting results (51, 209, 249, 252). In the randomized prospective study by Peters et al. (209), the success rate at 1 yr in patients randomized to a standard ¹³¹I activity of 555 MBq was inversely related to the baseline thyroid volume. In contrast, those patients receiving a targeted dose of 100 Gy had a success rate of around 50%, covering the entire volume range except for volumes less than 15 ml, in which case the success rate was 80%. Sabri et al. (252) found no influence of the thyroid volume on the cure rate, whereas this variable was the only determinant of the outcome, apart from the applied dose, in a study by Reinhardt et al. (249). Thus, the role of the thyroid size remains unclear. One obstacle for this issue to be resolved relies on the thyroid volume being associated with a range of other factors, e.g., disease severity, immunoreactivity, age, etc., which potentially affect the outcome of ¹³¹I therapy in a complex interaction.

As regards ¹³¹I therapy of nontoxic goiter, most (63, 64, 272) but not all studies (70) have demonstrated an inverse relationship between goiter reduction and the pretreatment thyroid volume, despite the application of similar thyroid doses. Such a correlation seems logical because the relative amount of degenerated fibrous tissue and hibernating thyrocytes—not being affected by radiation to any greater extent—probably increases with goiter size. A quantitative analysis of how the amount of degenerated thyroid tissue affects the goiter reduction after ¹³¹I therapy remains to be performed.

C Age and gender

The impact of age on the effect of ¹³¹I therapy is also controversial. A number of studies, mainly including patients with Graves' disease, found no influence of age on the efficacy of ¹³¹I (51, 88, 91, 244, 246, 261, 279, 284–289). In children with Graves' disease, a standard dose of 3.7 MBq/g thyroid tissue adjusted for the thyroid RAIU resulted in a high failure rate (290), suggesting that younger individuals are more radio-resistant than adults. In support, other studies (97, 207, 210) have shown that young age was associated with a lower success rate, although this was partly reversed after adjusting for other variables (207). In contrast, some have found an inverse correlation between age and the cure rate (280, 281), in line with a large retrospective study including nearly 1300 patients (291). In that study, young individuals more frequently developed hypothyroidism, reflecting a higher degree of radiosensitivity than in older patients.

The discrepant findings may rely on the possibility that young and old patients differ in other aspects confounding the results, i.e., smoking status, severity of disease, use of ATD and corticosteroids, and goiter size. Interactions between these factors can only to some extent be adjusted for in post hoc analyses. Furthermore, young age as well as male gender may be negatively associated with spontaneous remission of Graves' disease (292). Thus, it remains unanswered whether age is an independent predictor of the outcome of ¹³¹I therapy in hyperthyroid diseases. As for the goiter reduction after ¹³¹I therapy, the issue is blurred by the strong correlation between age and nodularity/goiter size (293), the latter being one of the major determinants of the efficacy of ¹³¹I therapy in multinodular goiter (63, 64).

The literature is more consistent regarding the influence of gender because the majority of studies, performed mainly in patients with Graves' disease, found no relation between this variable and the outcome of ¹³¹I therapy (51, 88, 91, 244, 246, 261, 279, 280, 284, 286–289, 294). In two large series from the United Kingdom and Greece, respectively, male sex was associated with a poorer response to ¹³¹I therapy, and females more often developed hypothyroidism (243, 291). However, from a biological point of view, there is no reason to believe that cells in males and females differ in their radiosensitivity, and clinically different outcomes may be due to other factors with a known association to gender, as for example, the thyroid size. It follows that there is, at present, no justification for the use of gender-dependent protocols in male and female thyroid patients treated with ¹³¹I.

D Thyroid autoantibodies

TRAb is a highly specific and sensitive marker of Graves' disease, and the serum level of this antibody is thought to reflect the magnitude of the immune attack on the organism (295). It could be speculated that patients with clinically or serologically more pronounced autoimmunity react differently to ¹³¹I therapy, compared with patients with less pronounced distortion of the immune system. Results on the predictive value of serum TRAb levels in relation to the outcome of ¹³¹I therapy are unclear. In some studies (55, 56, 276, 296), the cure rate was related to low pretreatment TRAb/thyroid-stimulating autoantibody (TSAb) levels, whereas other studies have not found any correlation (244, 246, 282). More consistently, it was demonstrated (55, 143, 150) that the rise in serum levels of TRAb/TSAb in the posttherapy period is related to the development of hypothyroidism. However, interference by ATD, shown to attenuate the ¹³¹I-induced rise in serum TRAb/TSAb (143, 152, 161, 297), cannot be excluded. The immunological impact on the thyroid gland, rather than the serum level of TRAb, may be a more reliable predictor. Thus, Markovic and Eterovic (283) reported that thyroid glands with a hypoechogenic texture, seen with ultrasound, are more radiosensitive than normoechogenic glands, but the ultrasonic findings were not related to the level of thyroid antibodies. The possibility of ultrasonic parameters being related to the outcome of ¹³¹I-therapy is intriguing and deserves further attention.

The significance of TPOAb for the outcome of ¹³¹I therapy is less explored than the role of TRAb. Presence of TPOAb (or microsomal antibodies) in serum has been reported to predict the development of hypothyroidism after ¹³¹I therapy of hyperthyroid diseases (298), but not consistently (88, 91, 261, 280). As for ¹³¹I therapy of nontoxic goiter, the presence of TPOAb is associated with a higher risk of developing permanent hypothyroidism (161, 167), perhaps reflecting the coexistence of autoimmune thyroiditis. A more close surveillance in the posttherapy period of patients harboring these antibodies may thus be justified.

E Severity of disease

The severity of hyperthyroid diseases can be assessed in several ways, and the most adequate definition of disease severity is a matter for discussion. From a clinical point of view, the thyroid hormone levels seem to be relevant markers of disease activity/severity due to their correlation with the symptoms of hyperthyroidism. However, it could be argued that for monitoring Graves' disease, indicators of the autoimmune status—like the serum level of TRAb, the presence of TAO, and the size of the thyroid gland—could be just as relevant to monitor. Use of ATD is an important confounder in this context, particularly because the dose of ATD needed to obtain euthyroidism parallels the disease severity. Although the impact of external factors, like ATD or other kinds of intervention, can relatively easily be investigated by randomized trials, it is much more difficult to explore the impact of the intrinsic features of the disease. Usually, regression analyses are used to segregate any independent effect of “disease severity” (independent of variable chosen) from a range of other factors that might affect the outcome of ¹³¹I therapy. Post hoc regression analyses have their limitation, and this is probably one of the main reasons for the conflicting findings.

As regards serum T₄ and T₃, several studies (97, 207, 276, 277, 282, 287, 291, 296, 298, 970) have found the levels of these hormones to correlate with the rate of treatment failure after ¹³¹I therapy. Importantly, whereas no such relationship was found in a range of other studies (241, 249, 252, 278, 279, 284, 285, 288, 289), a reverse relationship has never been reported. It is unknown whether the most suitable marker in this respect is the thyroid hormone level before any intervention or immediately before ¹³¹I therapy. The majority of studies focus on the baseline hormone levels, but it can be argued that the status at the time of therapy is equally or even more relevant. Thus, with radioprotection in focus, a mobilization of the intracellular antioxidative defense (300), which is potentially higher in the metabolically active thyrocyte, is more important at the day of ¹³¹I administration than at initial diagnosis. A potential pitfall is the high ¹³¹I turnover seen with a high disease activity. This implies that if the assessment of the ¹³¹I biokinetics does not include ¹³¹I half-life, the applied dose to the thyroid may be less than intended (220), as discussed earlier. This problem pertains especially to those studies that find a high disease activity associated with treatment failure (97, 207, 282, 287, 291, 296, 298, 299).

A few studies (88, 276, 280, 285, 301) have indicated that serum TSH at the time of ¹³¹I therapy correlates positively with the effect of ¹³¹I therapy. Because serum TSH is fully suppressed in overt hyperthyroidism, such analyses are only meaningful in patients who are pretreated with ATD or in those with nodular goiter and marginal hyperthyroidism (301). In the randomized study by Bonnema et al. (88), the rise in serum TSH, induced by propylthiouracil (PTU), attenuated the radioprotective effect of the drug. It may be argued, however, that a measurable serum TSH merely reflects a less hyperthyroid state and, consequently, an increased radiosensitivity of the thyrocytes, in line with the above reasoning. Other studies found no such influence of serum TSH (244, 249, 252, 278, 279, 284, 289). Nevertheless, randomized studies on rhTSH-augmented ¹³¹I therapy of nontoxic goiters support the hypothesis that TSH may act as a preconditioning factor whereby the thyrocyte is rendered more radiosensitive (70, 272). This feature of TSH is still speculative and needs further study.

Whether the presence of TAO influences the effect of ¹³¹I therapy is unclear. Some studies found no correlation between pretherapy TAO and the cure rate (244, 245, 278, 280), whereas Alexander and Larsen (97) reported that the presence of TAO predicted a greater risk of treatment failure after ¹³¹I therapy. In contrast, two large studies (291, 302), each including more than 1000 patients, showed that post-¹³¹I hypothyroidism developed more frequently among patients with TAO. In addition to the pitfalls already discussed, the conflicting results may rely on differences in the grading of TAO and in the number and nature of potential confounders adjusted for. Furthermore, because the presence of TAO to many clinicians is considered an absolute contraindication for ¹³¹I therapy, potential selection bias across studies also makes it difficult to compare the findings.

F Effect of stunning

Thyroid stunning is the phenomenon where the ¹³¹I uptake by the thyrocyte is attenuated due to prior ionizing irradiation (303). It has been fiercely debated by many experts whether stunning constitutes a real clinical problem or is just a matter of theoretical concern. If stunning exists, the problem is most relevant for thyroid cancer patients in whom ablative ¹³¹I therapy is preceded by a diagnostic scan using a relatively high amount of radioactivity of up to 200 MBq ¹³¹I. In vitro studies have shown that stunning is associated with decreasing levels of NIS-mRNA and signs of cell cycle arrest but not cell death (304, 305). Stunning is dose-related and detectable, even at a very low level of radiation (0.15 Gy), but to some extent is dependent on the isotope used for irradiation (304, 306). Interestingly, external irradiation of thyroid cell cultures with high-energy photons increases the RAIU in a dose-dependent manner, despite slowing down cell proliferation (307). It is unknown why internal and external irradiation act so differently.

In a clinical setting, the focus has been on thyroid cancer patients treated with ¹³¹I therapy. The discrepancy between a positive diagnostic ¹³¹I scan and a subsequent negative imaging during the ¹³¹I therapy has been attributed to the impact of stunning (308). To avoid this, some authors have recommended the use of ¹²³I (more effective with less radiation) instead of ¹³¹I for diagnostic purposes (309–311). In a large series of cancer patients, ¹²³I whole-body scan findings were highly concordant with the corresponding posttreatment ¹³¹I scans (312). The clinical relevance can, however, be questioned because patients show similar ablation rates after ¹³¹I therapy, independent of the preceding use of either ¹²³I or ¹³¹I (303, 313, 314). In addition, it has been suggested that what seems to be stunning is in fact the result of the early destructive effects of ablative ¹³¹I therapy (315).

Stunning has been less studied in benign thyroid diseases. Thyroid scintigraphy and measurement of the thyroid RAIU require only little radioactivity compared with what is needed for diagnostic scans in cancer patients. However, considering the small doses by which stunning can occur (304), this could have clinical implications also in benign thyroid diseases. Induction of stunning by therapeutic doses of ¹³¹I seems apparent. By fractioning ¹³¹I therapy in hyperthyroid patients to whom two doses were administered with an interval of a few days, significant stunning was seen during the second therapy, and the degree of stunning correlated with the applied dose at the first ¹³¹I administration (316, 317). A major challenge, when studying stunning, is the fact that in patients with Graves' disease, the thyroid RAIU shows large intraindividual variations within short time intervals (253), which obviously makes it difficult to evaluate the impact of stunning. Furthermore, the applied thyroid dose is in general less than anticipated based on pretherapy measurements of the ¹³¹I biokinetics (247, 248). Besides methodological difficulties in a precise measurement of the thyroid dose during therapy, the reason for this phenomenon might be induction of “self-stunning” occurring during the first few hours after administration of the therapeutic ¹³¹I dose. This feature of ¹³¹I therapy is supported by studies in benign as well as in malignant thyroid diseases (221, 310). However, stunning seems to appear with some latency because a time interval of 2 d between exposure and dose administration results in less stunning than seen with a 4-d interval (316). Thereafter, stunning increases in severity until 3–4 wk, followed by some recovery (310). In patients with nontoxic goiter (318), two uptake measurements performed 4 wk apart showed similar values, indicating that stunning after this period has vanished.

Although stunning seems to have clinical implications, the extent of the problem is uncertain. Stunning may be more pronounced in vitro than in vivo (319), and the complexity of the issue is further magnified by a number of additional factors with potential effects on the ¹³¹I biokinetics, dosimetry, and other aspects of dose calculation. This underlines the observation that despite application of meticulous dosimetry a precise thyroid dose is very difficult to achieve at an individual level in patients treated with ¹³¹I.

G Thyroid radioiodine uptake

Several studies, most observational, have indicated that the cure rate after ¹³¹I therapy is inversely correlated with the 24-h thyroid RAIU (97, 241, 261, 282, 285, 320, 321). This contrasts what might intuitively be expected. In the randomized study by Bonnema et al. (261), the thyroid RAIU was reduced by the uninterrupted use of methimazole during the ¹³¹I therapy. This intervention was associated with a more favorable outcome, even after adjustment for the possible radioprotection induced by the drug. There may be several explanations for this observation. A high 24-h RAIU reflects a high metabolic state of the thyrocyte resulting in a high ¹³¹I turnover (i.e., short half-life), which has been associated with a poor response to ¹³¹I therapy (259). If the dose calculation does not include the ¹³¹I half-life, there is a risk of an applied thyroid dose less than intended, as discussed earlier. The metabolically active thyrocyte has a high production of hydrogen peroxide, necessary for thyroid hormone production. A lowering of the thyrocyte's metabolism might have a preconditioning effect by altering the balance within the intracellular environment between the generation of free oxygen radicals and the capacity of the antioxidant defense (300), and this may move the cell toward a more radiosensitive state. Probably the thyroid RAIU is merely a surrogate marker, but an inverse relationship between the thyroid RAIU and the radiosensitivity of the gland is supported by a theoretical model (322). Another potentially important factor is stunning due to the diagnostic RAIU performed before ¹³¹I therapy, which in theory has more impact with higher thyroid RAIU. Awaiting more prospective studies, it remains speculative whether a lowering of the thyroid RAIU (or the thyrocyte's metabolism), achieved by use of ATD or by other means, leads to a higher efficacy of ¹³¹I therapy. Clarifying whether the thyroid RAIU at 24 h, or any other time point, is the most appropriate remains to be done.

H Smoking

The link between smoking and some thyroid diseases is well accepted (323, 324), but the underlying biological mechanisms are incompletely understood. Thiocyanate, one of many components contained in tobacco smoke, has antithyroid properties that may lead to goiter development (325). Other plausible mechanisms include tissue hypoxia, increased immunogenic reactivity through HLA-DR expression, and modulation of cytokines involved in autoimmune thyroid disorders. In a meta-analysis (326), it was calculated that current smokers have an odds ratio of 3.3 for developing Graves' disease compared with never-smokers. The risk seems abolished in ex-smokers (326). Whether smoking is a risk factor for relapse after a course of ATD for Graves' disease is uncertain (292, 327). However, smoking increases the risk of TAO beyond the risk associated with Graves' disease per se (326), and cessation of smoking is recommended in patients with TAO because this seems to accentuate the response to immunosuppressive therapy (328–330).

Contrasting the impact on Graves' disease, smoking as a risk factor for developing autoimmune thyroiditis is not evident (326). In fact, smoking might protect against the development of TPOAb (331), and cessation of smoking may precipitate the development of TPOAb and/or TgAb (332). As for goiter development, epidemiological as well as twin studies suggest smoking to be a risk factor in borderline iodine-deficient areas (333, 334), whereas such a correlation seems to be absent or may even be reversed in iodine-replete areas (335).

In relation to ¹³¹I therapy, it is well established that smoking is a risk factor for the development or worsening of TAO in patients with Graves' disease (328). Smoking per se might be an indication for prophylactic treatment with corticosteroids, although such a strategy is not supported by any randomized trial. Whether smoking has any independent influence on the cure rate in patients with hyperthyroid diseases treated with ¹³¹I is only sparsely investigated, probably because exact data on smoking status and pack-year history are difficult to retrieve in retrospective studies. Obviously, for ethical reasons, no randomized study where smoking is used for intervention will ever be conducted. In the large retrospective study by Allahabadia et al. (207), including more than 800 patients with hyperthyroid diseases treated with ¹³¹I, the influence of smoking was not reported. In the prospective study by Andrade et al. (282), the effect of smoking on the cure rate was neutral, and this finding was supported by the retrospective study by Jensen et al. (280). No study has investigated whether smoking interferes with the volume reduction of nontoxic goiters treated with ¹³¹I.

I Beta-blockers

Propranolol, an unselective beta-blocker, constitutes the first-line treatment in severe hyperthyroidism, particularly in the rare cases of thyroid storm (336), due to its prompt symptomatic effect on tremor, heart rate, and cardiac output. Caution is recommended in patients with heart output failure in whom β-blockade may lead to cardiac arrest (337). Propranolol is also employed preoperatively before surgery for hyperthyroid diseases. A beta-blocker as the only preoperative agent is safe (338, 339), also when compared head-to-head with ATD (340, 341). A selective β-1-blocker like metoprolol seems as effective as propranolol (342). The benefits of β-blockade, as opposed to ATD, are a lower frequency of adverse drug effects and shortening of the time from diagnosis to operation. However, if propranolol is combined with PTU, this results in a lower peri-operative pulse rate and a lower risk of postoperative fever, compared with β-blockade alone (343).

In the majority of studies, patients suffering from hyperthyroid diseases are rendered euthyroid by ATD before ¹³¹I administration due to the fear of posttherapy thyroid storm. Beta-blockers are rarely necessary when euthyroidism is obtained, and the combination of these drugs with ATD before ¹³¹I therapy is usually not part of the routine. It is the prevailing opinion that propranolol does not lead to a decrease of the serum level of TRAb in Graves' disease, unless euthyroidism is achieved by other means. Although this remains to be proven in a clinical trial, earlier in vitro studies demonstrated that propranolol, in contrast to ATD (344, 345), has no immunosuppressive effect. It was indicated (346), however, that propranolol may be a weak scavenger of free oxygen radicals, which theoretically attenuates the effect of ¹³¹I therapy. A few studies compared the pre-¹³¹I-therapy use of methimazole with propranolol, without finding any difference in the cure rate between the two groups (347, 348). Because β-blockade has not been tested against placebo, a final conclusion on the possible impact of propranolol—or of other beta-blockers—cannot be drawn. Generally, these drugs are considered to cause no harm as regards the long-term outcome of ¹³¹I therapy.

J Corticosteroids

Corticosteroids have significant immunosuppressive properties and are used in the treatment of many disorders with an autoimmune etiology. Corticosteroids also affect thyroid parameters. The most significant effect is a decrease of serum T₃ and an increase of serum rT₃ (349), but also serum T₄, TSH, and thyroglobulin concentrations decrease slightly, depending on the dose and route of administration (350–353). In patients with Graves' disease, TPOAb and TgAb titers are reduced within a few weeks, whereas the impact on the level of TRAb is less evident (144).

The risk of developing or worsening TAO after ¹³¹I therapy is very likely to have an immunogenic origin. Based on the results from randomized studies (155, 156, 354), corticosteroids are recommended during ¹³¹I therapy to hinder any induction of orbital inflammation in patients with Graves' disease predisposed for developing TAO (355). However, there is consensus on neither the dose nor the length of treatment. According to a European survey among clinicians (163), the starting dose of prednisone is approximately 40 mg, but with a huge range (15–80 mg), and the mean duration of treatment is 16 d (range, 2–60 d).

As for the impact on the thyroid gland, corticosteroids might interfere with the effect of ¹³¹I therapy in several ways. Earlier studies suggested that glucocorticoids impair the thyroid ¹³¹I uptake (356, 357). In addition, a recent study indicates that corticosteroids in a high dose reduce the effective thyroid ¹³¹I half-life (358). In the study by Gamstedt et al. (349) in patients with Graves' disease, a high dose of betamethasone for 3 wk resulted in nearly a halving of the 24-h ¹³¹I thyroid uptake, but this may have been due to a lower activity of the disease because a significant decline of the thyroid hormone levels was seen as well. Within the area of tumor research, recent studies suggest that the killing of tumor cells after radiotherapy is immune-mediated (359–361), and improved activation of effector T cells by CTLA-4 blocking antibodies might enhance the proimmunogenic effects of radiotherapy (362). It is unknown to what extent the thyroid destruction after ¹³¹I therapy is immune-mediated. Because corticosteroids decrease the level of serum IgG after ¹³¹I therapy (144), this might reflect a decrease of any potential immune-mediated cytotoxicity elicited by the radiation. Furthermore, corticosteroids, under some circumstances, have a direct radioprotective effect by reducing the intracellular oxidative stress (363–366), but results have been conflicting, and the significance of corticosteroids as a possible scavenger of free oxygen radicals in a clinical context remains to be clarified.

Thus, corticosteroids might attenuate the effect of ¹³¹I therapy either by altering the ¹³¹I kinetics inadvertently or by immunoregulatory or cytoprotective mechanisms. However, clinical studies addressing this issue are very sparse, albeit highly relevant given the widespread adjunctive use of corticosteroids with ¹³¹I therapy. Gamstedt et al. (144, 152) investigated the effect of betamethasone in two randomized studies in patients with Graves' disease. No effect of betamethasone was found on the level of TSH-binding inhibitory Ig after ¹³¹I therapy (152). In a succeeding study, 40 patients were randomized to receive placebo or 6 mg betamethasone from 3 wk before until 4 wk after ¹³¹I therapy (144). Betamethasone delayed but did not prevent the ¹³¹I-induced rise in thyroid antibodies. Furthermore, betamethasone resulted in a significantly lower rate of hypothyroidism at 1 yr, despite the finding that hypothyroid patients had lower levels of thyroid antibodies (144). In the much larger study by Bartalena et al. (156), no significant difference in the thyroid function, 1 yr posttherapy, was observed between patients treated with or without prednisolone concomitantly with ¹³¹I therapy, but the cure rate was not a primary end-point in that study. In our retrospective study (280), 200 patients with Graves' disease were treated with ¹³¹I, half of whom received prednisolone prophylaxis due to present or previous TAO (Fig. 7). The cure rate at 1 yr was identical in the two groups. Based on the assumption that patients with TAO suffer from more severe disease, which may be associated with a poorer cure rate after ¹³¹I therapy as discussed earlier, it seems unlikely that corticosteroids per se interfere with the radiation-mediated thyroid destruction. A randomized study, specifically addressing this issue, is unlikely to be performed for ethical reasons.

Figure 7

Data from the retrospective study of Jensen et al. (280) showing the neutral influence of glucocorticoids on the cure rate in patients with Graves' disease treated with ¹³¹I therapy. Indication for prednisolone (25 mg/d for 4 wk) was previous or present mild/moderate Graves' ophthalmopathy or risk factors for developing this complication. The outcome 1 yr after ¹³¹I therapy was almost identical in the two groups. [Reproduced from B. E. Jensen et al.: Glucocorticoids do not influence the effect of radioiodine therapy in Graves' disease. Eur J Endocrinol 153:15–21, 2005 (280), with permission. © European Society of Endocrinology.]

K Antioxidants

The reactive oxygen species, required for thyroid hormone synthesis, are detoxified either during the hormone synthesis or by the endogenous antioxidant apparatus within the thyrocyte. The level of oxidative stress seems dependent on the iodine content (300, 367). Furthermore, hyperthyroidism per se leads to an increase of the oxidative stress load in general, and this is reversed by treatment (368–375). The damaging effect on cellular DNA is, however, very limited, probably due to an equivalent counteraction by the antioxidative defense system within the cell (376, 377). Also, the hypothyroid state may increase the oxidative stress load with potential clinical implications, although results have been conflicting (378–381).

The administration of ¹³¹I markedly increases the oxidative stress in the thyroid gland, but the mobilization of free radical scavengers to some extent limits the damage to cellular components. In rat thyroids exposed to subablative doses of ¹³¹I, reactive oxygen species are generated, leading to an increased lipid peroxidation, but this is opposed by an up-regulation of the protective enzyme glutathione peroxidase (382). Administration of antioxidants may reduce the cytotoxic effect of ionizing radiation in some radiosensitive organs, thereby allowing a higher radiation dose, and animal studies support such a role of antioxidants (383). In mice exposed to high doses of ¹³¹I, pretreatment with extracts of the antioxidant Ocimum sanctum reduces the lipid peroxidation in both kidneys and salivary glands, and less depletion of glutathione is observed (384). Furthermore, zinc supplementation in rats has beneficial effects on the levels of antioxidative enzymes in erythrocytes irradiated with ¹³¹I (385), and vitamin C given to ¹³¹I-exposed mice reduces the inhibition of spermatogenesis (386).

Only a few clinical studies, primarily in thyroid cancer patients, have focused on oxidative stress and the antioxidant status after ¹³¹I therapy. The content of antioxidants and the superoxide dismutase enzyme were significantly reduced in saliva after high-dose ¹³¹I therapy (118), reflecting that the salivary glands receive a significant amount of radiation. In erythrocytes examined 2 d after ¹³¹I therapy, markers of the oxidative stress and the antioxidant defense are increased (387). However, the slight but persistent DNA damage, as assessed by cytogenetic assays, that is found after high-dose ¹³¹I therapy is not clearly related to oxidative stress indicators like thiobarbituric acid-reactive substances and serum uric acid levels (388). That negative finding may rely on the nature of the markers studied. A more suitable substance seems to be 8-epi prostaglandin F2α (an isoprostane), which reflects the lipid peroxidation resulting from oxidative stress. Indeed, 8-epi-PGF2α is increased in both plasma and urine in patients undergoing ¹³¹I therapy for hyperthyroidism or thyroid cancer, and correlates significantly with the applied ¹³¹I activity (389).

No study has investigated a possible influence of exogenous antioxidants in thyroid patients treated with ¹³¹I. In patients with Graves' disease, euthyroidism obtained by methimazole normalized a range of peripheral oxidative stress parameters, whereas euthyroidism obtained by ¹³¹I therapy only partly reversed these parameters even several months after therapy (390). The authors of that study (390) speculated on the adjunctive use of antioxidants to correct for this imbalance. Considering the possible radioprotective effect of such agents, as indicated by the above studies (382, 384–386), such a strategy may counteract the intended destructive effect on the thyroid gland.

L Antithyroid drugs

Many clinicians prefer to render their patients euthyroid before ¹³¹I therapy to minimize the risk of an exacerbation of the hyperthyroidism or even a thyroid storm. Early in the era of ¹³¹I therapy, the question soon emerged whether adjunctive use of ATD counteracts the effect of ¹³¹I therapy. The debate has been ongoing since then, and the issue has been addressed by numerous studies. The prevailing opinion, based primarily on retrospective case-control studies (97, 207, 223, 279, 297, 391–400), is that use of ATD attenuates the effect of ¹³¹I therapy, although not all reports support this view (259, 278, 284, 291). Retrospective studies are difficult to compare due to huge differences in a number of important variables. Other disadvantages of this design are the risk of selection bias and the lack of adequate adjustment for confounders, some of which are even unknown.

Prospective RCT are needed for a more systematic and unbiased investigation of the role of adjunctive ATD. By 2011, 15 RCT had been conducted (88, 91, 152, 261, 262, 282, 347, 348, 401–406, 408). If a RCT results in a negative finding, the question arises whether the study was underpowered. To overcome this problem, Walter et al. (408) performed a meta-analysis based on the available 14 RCT and comprising a total of 1306 patients suffering from Graves' disease or toxic nodular goiter. The meta-analysis supported the hypothesis that treatment failure is associated with ATD administration in the period before, during, or after ¹³¹I therapy and, furthermore, that the risk of hypothyroidism is reduced if ATD are given during or after ¹³¹I therapy (408). Adverse events in the 14 trials were few, but they were probably underreported. Among patients randomized to adjunctive ATD, fewer cardiac events and deaths were noticed, although this was not statistically significant (408).

Meta-analyses, including the study by Walter et al. (408), have their limitations. The design of the individual trials differs in several aspects, such as the inclusion of patients with Graves' disease and toxic nodular goiter, different criteria defining the thyroid status, and differences in the methods for thyroid size estimation, in type and doses of ATD, and in the ¹³¹I dose applied. A major problem is that all trials were pooled in the meta-analysis (408), disregarding the fact that intervention in the trials was focused on different periods in relation to the ¹³¹I therapy. If an ATD has radioprotective properties, different modes of action may be in play. Thus, when used in the pretherapy period, the euthyroid state induced by ATD might change the radiosensitivity of the thyrocyte. When ATD are continued without interruption during the ¹³¹I therapy, the impact on the ¹³¹I biokinetics becomes a crucial point, and if this is ignored in the dose calculation, the patient will be undertreated. Finally, the role as a free radical scavenger might be the most important characteristic of ATD in the posttherapy period. In addition to these considerations, other important variables—not accounted for in the majority of studies—may be significantly affected by the intervention with ATD. Serum TSH and the thyroid RAIU are examples of potential confounders, as discussed earlier, and other factors may exist (88, 261, 262).

1 Temporal relation to ¹³¹I administration

The outcome from the meta-analysis by Walter et al. (408) opposes the results from the individual trials. The most recent RCT on methimazole in the pretherapy period were carried out by Braga et al. (348) and Andrade et al. (282), and no radioprotection by methimazole was found in those studies. In the posttherapy period, resumption of ATD prevents the transient hyperthyroid phase that follows ¹³¹I therapy (91). Although earlier retrospective studies suggested that such a strategy reduces the efficacy of ¹³¹I therapy (259, 297, 397), RCT found no increase in the treatment failure by the resumption of either methimazole (91, 404) or PTU (406) within 7 d after ¹³¹I therapy. However, a lower rate of hypothyroidism was found by Kung et al. (404), and the goiter volume reduction was attenuated in the study by Bonnema et al. (91). In the study by Bazzi and Bagchi (401), the higher failure rate among patients randomized to PTU 5 d after ¹³¹I therapy was explained by the greater prevalence of goiters in that group. Thus, with the possible exception of the small study by Gamstedt et al. (152), no single RCT (282, 347, 348, 400, 402) has been able to demonstrate an increased failure rate by the use of methimazole (or the pro-drug carbimizole) in the pre-¹³¹I therapy period. Similarly, a range of RCT (282, 347, 348, 400, 402, 403, 405, 406) found a neutral effect of ATD (of any kind) in the early post-¹³¹I therapy period.

The issue is more complex with a regimen where ATD are continued during the ¹³¹I therapy. The blockade of the intrathyroidal iodine organification results in a reduced intrathyroidal iodine pool and a depletion of the thyroid hormone stores. Indeed, it is well documented that the thyroid RAIU is reduced during an ongoing ATD treatment (252, 261, 262, 409–411). However, the refilling of the thyroid hormone pool by suspending the thyroid blockade may to some extent be counteracted by the destruction of the thyrocyte after the ¹³¹I therapy, and other factors derived from ATD treatment may also influence the final outcome. Urbannek et al. (409) showed that cessation of ATD at least 2 d before ¹³¹I administration increased the applied dose by about 50% due to an increase in the thyroid RAIU as well as in the ¹³¹I half-life. The thyroid ¹³¹I residence time also increased if the ATD were discontinued in the post-¹³¹I therapy period (409). Another German study (410), also nonrandomized, measured in detail the diagnostic and therapeutic ¹³¹I kinetics in patients with Graves' disease with or without thyroid blockade by thiamazole. Without thiamazole, the ¹³¹I kinetics fitted perfectly a simple in- and output function. During thyroid blockade, the absorbed thyroid dose was reduced due to a lower RAIU and a shorter ¹³¹I half-life. In addition, the ¹³¹I biokinetics indicated that a heterogeneous thyroid ¹³¹I uptake took place, probably due to blockade of thyroid follicles with varying degrees of hormone synthesis (410). It should be noted that radiation from the follicular colloid constitutes a significant fraction of the applied dose to the thyrocyte (258), as previously discussed. Because blockade of iodine organification by ATD affects the ratio between the height of the follicular epithelium and the size of the follicle, this may significantly interfere with the effect of ¹³¹I therapy. This aspect of ATD treatment and its impact on thyroid dosimetry has hitherto been neglected, and taking the variation in follicle size and the resulting radiation into account is unfortunately impossible with the available technology. As regards the outcome after ¹³¹I therapy, nonrandomized studies (223, 252, 284) reported that the cure rate was negatively affected by an ongoing ATD regimen, despite adjustments for the change in the ¹³¹I kinetics. A randomized trial supported such an association, but only after adjustment for the concomitant decrease of the thyroid RAIU (261). Another randomized trial confirmed that blockage of the iodine organification to some extent counteracts the radioprotective feature of ATD in patients with Graves' disease (262). Although simultaneous use of ATD during ¹³¹I therapy decreases the ¹³¹I trapping within the thyroid gland, this regimen also has its clear advantages. More stable levels of the thyroid hormones during and after the ¹³¹I therapy are seen than if ¹³¹I therapy is given without ATD (261, 262). As discussed earlier, the reduction in the thyroid RAIU by blocking the thyroid peroxidase may render the thyrocyte more susceptible to radiation, and finally, this also seems to decrease the extrathyroidal irradiation in patients with a high thyroid ¹³¹I turnover (412).

2 Methimazole vs. PTU

The meta-analysis by Walter et al. (408) found that methimazole and PTU have similar radioprotective features, but this contrasts with some retrospective studies suggesting that the latter drug has the most pronounced influence (289, 392, 413). In the study by Imseis et al. (392), the reduction in cure rate was observed even when PTU was discontinued for as long as 55 d before ¹³¹I therapy. It is difficult to explain how PTU can have such a long-lasting effect on the ¹³¹I therapy outcome. In that study, methimazole had a neutral effect (392). Although randomized studies (282, 348) found no influence of methimazole pretreatment, the randomized study by Bonnema et al. (88)—comparing PTU pretreatment with no pretreatment—demonstrated a significant reduction in the cure rate among PTU-treated patients, but only after an adjustment for the concomitant increase in serum TSH. Because PTU and methimazole have not been compared head-to-head in a randomized set-up, it remains unsolved whether these drugs interfere differently with ¹³¹I therapy. It is highly questionable whether such a trial will ever be conducted, considering the recent debate on the disadvantages of using PTU (414). A few studies (391, 395, 415) have found that the length of the period off ATD, before ¹³¹I administration, to some extent correlated with the cure rate, but this issue has not been addressed by any randomized trial.

3 Graves' disease vs. toxic nodular goiter

In relation to the use of ATD, dissimilarities between Graves' disease and toxic nodular goiter may exist. In toxic nodular goiter, ATD treatment diverts the ¹³¹I isotope from the autonomous nodules (i.e., the true target tissue) to the paranodular tissue, in parallel to the increase of serum TSH and the alleviation of the hyperthyroidism. Theoretically, such a “steal phenomenon” should increase the risk of hypothyroidism, and results from a retrospective study indeed support this (263). However, no study has evaluated this phenomenon in greater detail, and furthermore, a differentiation between target and nontarget tissue in toxic nodular goiter is probably rarely done at most centers. Because the hyperthyroidism in Graves' disease and toxic nodular goiter has a very different origin, additional aspects of ATD in relation to ¹³¹I therapy may be of importance in the two diseases. Indeed, the study by Körber et al. (279) indicated that the continuous use of ATD affects the cure rate only in patients with toxic nodular goiter, and not in those with Graves' disease. In the three RCT by Bonnema et al. (88, 261, 262), the difference in outcome was also more pronounced among patients with toxic nodular goiter. In one of these trials (262), patients with toxic nodular goiter randomized to a block-replacement regimen had a significantly lower cure rate than patients who discontinued a titrated methimazole regimen. In contrast, the cure rate was almost identical in the two groups among patients suffering from Graves' disease (262). Awaiting larger trials, it remains a possibility that adjunctive use of ATD in relation to ¹³¹I therapy is more harmful when given for toxic nodular goiter than for Graves' disease. Because the hyperthyroidism due to autonomous nodules is often modest and symptoms are sparse, ¹³¹I therapy can probably be given to ATD-naive hyperthyroid patients without any greater risk, at least in the absence of cardiovascular comorbidity.

4 Mechanisms of radioprotection

Although ongoing ATD treatment markedly suppresses the thyroid RAIU, this rapidly reverses when the drug is discontinued. Thus, the thyroid RAIU measured 7 d after ATD withdrawal is similar to the RAIU measured 2 d after withdrawal (416). Apart from the change in the ¹³¹I biokinetics, there are several mechanisms whereby ATD could attenuate the effect of ¹³¹I therapy, should such a feature of these drugs exist. Rats given either PTU or methimazole are more resistant to organ damage from the oxidative stress after ischemia, inflammation, whole-body irradiation, or chemotherapeutic drugs (417–420). The clinical relevance of these observations remains unknown. ATD inhibit the thyroid peroxidase and thereby the cellular capacity of producing free oxygen radicals. Additionally, the drugs may act as free radical scavengers themselves. The reduced erythema after UV radiation by the topical application of methimazole was attributed to this feature (421). Einhorn and Saterborg (422) proposed that the radioprotection associated with thiourea results from the presence of a sulfhydryl group. However, the scavenger-like property, when examined in human neutrophils, applies to PTU as well as to methimazole (423), of which the latter drug possesses no sulfhydryl groups. The postulated difference between methimazole/carbimazole and PTU, in their radioprotective potential, might instead be explained by the higher doses of PTU necessary to achieve euthyroidism. Another possibility is that ATD interfere with a ¹³¹I-mediated immune attack on the thyroid gland, considering the observation that pretreatment with ATD attenuates the ¹³¹I-induced rise of serum levels of TRAb/TSH-binding inhibitory Ig (143, 152, 161, 297). In addition, ATD inhibit MHC gene expression in the thyrocyte, and this might potentially lead to an altered immune response to the thyroid (424). Furthermore, the Fas ligand expression in thyrocytes is highly up-regulated by ATD treatment, which in turn induces Fas-mediated apoptosis in lymphocytes (425), but such a feature of ATD may be disadvantageous if the thyroid destruction is immune-mediated. Thus, the possible immunomodulatory effect of ATD seems to be complex and may involve several mechanisms. It is also a matter of debate whether these drugs exert their action directly on the immune system (426) or through the maintenance of euthyroidism (427).

In vitro rather than in vivo studies are more suitable for exploring the isolated effect of ATD. In the recent study by Kahmann et al. (428), the DNA damage after internal radiation (using Re-188-perrhenate as a β-emitter) of isolated FRTL5 cells was reduced to 60% by the coincubation with methimazole. Although a significant part of the radioprotection may be caused by the concomitant inhibition of NIS, less DNA damage by methimazole was seen also after external radiation (428). These results support the notion that ATD attenuate the cellular damage from radiation, an observation that in fact was made 60 yr ago (429). However, more confirmatory studies are needed, and a limitation of the study by Kahmann et al. (428) is that the concentration of methimazole was higher than that used in a clinical setting. The novel in vitro model presented recently by Hershman et al. (31), whereby ¹³¹I-induced DSB in thyroid cells can be quantified, is very promising, and future experiments might finally establish whether ATD have intrinsic radioprotective properties.

Despite the results from the meta-analysis by Walter et al. (408) and from the recent in vitro study by Kahmann et al. (428), supporting the possibility that ATD are radioprotective, it is still unsettled whether these drugs in vivo reduce the efficacy of ¹³¹I therapy. Randomized trials (88, 261, 262) have demonstrated that a number of key factors, some of which are modified by ATD, interact with the effect of ¹³¹I therapy in a complex manner (Fig. 6). By not taking this problem into account, studies may easily be misinterpreted. It should be recognized that ATD have properties that are obviously beneficial, and the risk of serious side effects such as agranulocytosis or liver failure are low (414, 430). These drugs result in rapid restoration of euthyroidism in newly diagnosed hyperthyroid patients, and they ensure a more stable level of thyroid hormones during the ¹³¹I therapy (91, 261, 262). Based on the present knowledge, some recommendations can be made, including: 1) most patients who are only slightly hyperthyroid and have no severe comorbidity do not necessarily need ATD before ¹³¹I therapy; 2) if obtaining euthyroidism is indicated before ¹³¹I therapy, methimazole rather than PTU should be used; 3) ATD should be discontinued 2–7 d before ¹³¹I therapy to increase the intrathyroidal ¹³¹I trapping; and 4) ATD should be resumed at about 1 wk after ¹³¹I therapy if the aim is to prevent temporary hyperthyroidism, particularly when the hyperthyroidism is caused by autonomous thyroid nodules.

Only further large-scale RCT can clarify whether the use of ATD is associated with an overall unfavorable profile in the context of ¹³¹I therapy. Including aspects of quality of life (QoL) and cost in such studies may well reveal aspects hitherto given insufficient attention.

VII Enhancers of ¹³¹I Uptake

The radiation dose delivered to the thyroid depends on the resident time and the type of isotope. Much effort has aimed at boosting the thyroid RAIU and the ¹³¹I half-life to increase the efficacy of the ¹³¹I therapy and/or to minimize the amount of radioactivity, while still attaining the thyroid target dose. However, it should be emphasized that an intervention advantageous to the thyroid resident time may have inadvertent effects on the whole-body radiation, or may interfere with the effect of the ¹³¹I therapy by altering the radiosensitivity of the target tissue or through other mechanisms, as previously mentioned. In the following, various methods for enhancing the thyroid RAIU are discussed.

A Low-iodine diet

It has long been recognized that the thyroid RAIU in normal subjects is not a static parameter but changes with time in parallel with the dietary iodine content (431, 432). Thus, in regions exposed to iodine fortification, a decline in the thyroid RAIU was observed over time in patients with nontoxic goiter (433) or Graves' disease (434). The dietary iodine intake is cumbersome to quantify, but the urinary iodine excretion can be used as an indirect estimate. A spot urine sample rather than a 24-h urine collection makes the procedure even more simple (435). Although wide intraindividual variation exists (436, 437), the urinary iodine excretion correlates inversely with the thyroid RAIU in patients with toxic goiter, and it has been estimated that a 2-fold increase of the iodine excretion corresponds to a decrease of the thyroid RAIU by 25% (438). Of importance, in thyroid cancer patients, the iodine pool and the iodine excretion are around 50% higher after rhTSH stimulation than after thyroid hormone withdrawal, due to the deiodination of levothyroxine (439).

To increase the thyroid RAIU and to optimize the effect of ¹³¹I therapy, iodine restriction has been recommended for many years, as evident from current guidelines (440). A high iodine pool rarely hinders ¹³¹I therapy of hyperthyroid diseases, and the dietary iodine intake is mainly of concern in patients with nontoxic goiter or thyroid cancer. Although a 2-wk period of iodine restriction is more effective than 1 wk (441, 442), a recent study indicates that 1 wk may be enough (443). Longer periods of iodine restriction are of no further benefit (444). Whether iodine restriction before ¹³¹I therapy has any clinical significance in thyroid cancer patients can be questioned because two retrospective studies, evaluating the rate of successful ablation, showed conflicting results (445, 446). A recent systematic review based on eight studies concluded that a low-iodine diet appears to increase the thyroid RAIU and possibly the efficacy of ¹³¹I therapy of thyroid cancer patients, but the impact on the long-term recurrence rate is unknown (447).

Iodine restriction has a price. The decrease in the iodine clearance results in a higher ¹³¹I blood concentration and a higher whole-body radiation (448, 449). Moreover, patients may have difficulties in complying with a low-iodine diet. A less stringent and more simple diet has been proposed by some authors (441, 450, 451), seemingly without affecting outcome (452). Most studies on iodine restriction were performed in iodine-sufficient or iodine-replete areas, and the results cannot uncritically be extrapolated to iodine-deficient areas. In fact, it remains to be proven whether iodine restriction has any relevance in relation to ¹³¹I therapy of MNG—a disorder related to iodine deficiency.

B Stable iodine and lithium

Although the administration of stable (elementary) iodine before¹³¹I therapy obviously reduces the thyroid RAIU (453), it may be beneficial in the posttherapy period for several reasons. Ross et al. (454) found that patients with Graves' disease more rapidly obtain euthyroidism by the adjunctive use of potassium iodine after ¹³¹I therapy compared with ¹³¹I alone, and without adversely affecting the outcome at 1 yr. Another potential benefit is that elementary iodine may prolong the thyroid ¹³¹I residence time. In patients with a high ¹³¹I turnover rate due to Graves' disease, administration of 600 mg elementary iodine for 2–4 d after ¹³¹I therapy increases the thyroid ¹³¹I half-life and the applied thyroid dose (455, 456). Similar observations have been made by others (457). This application of elementary iodine is interesting, but transient hypothyroidism, probably reflecting a Wolff-Chaikoff effect, is a drawback (454). Unfortunately, there are no RCT investigating the short- and long-term effects of elementary iodine as an adjunct to ¹³¹I therapy.

Lithium, used for many years in the management of psychiatric disorders, has short-term as well as long-term effects on the thyroid (458). In predisposed individuals, long-term lithium treatment may result in hypothyroidism and goiter development (458–460). Hypothyroidism typically appears during the first years of lithium treatment and in the presence of thyroid autoimmunity. In patients treated with lithium for several years, the prevalence of thyroid disorders is not much increased compared with the background population (459). Whether the thyroid enlargement seen with lithium is due to a rise in TSH is incompletely understood because results have been conflicting (461–463). Long-term lithium therapy has also been associated with thyrotoxicosis, mainly due to silent thyroiditis (464, 465), but other thyroid disorders have been described (466).

Short-term lithium treatment has antithyroid properties in a manner similar to that seen with elementary iodine, and lithium has been used in the treatment of thyrotoxicosis (467–469). Lithium inhibits the release of iodine from the thyroid and suppresses the thyroid hormone synthesis (462, 468). These properties seem ideal in relation to ¹³¹I therapy. Turner et al. (470) demonstrated in the 1970s that low-dose lithium (400 mg/d) for 2 wk increases the retention of ¹³¹I in the thyroid gland without affecting the thyroid RAIU. A closer evaluation of the ¹³¹I kinetics during lithium therapy was performed in patients with a short thyroid ¹³¹I half-life due to Graves' disease (471). Administration of 885 mg lithium carbonate daily for 2 wk did not affect the thyroid RAIU but prolonged the ¹³¹I half-life by 60%, and the applied thyroid radiation dose increased by 39% (471). The major benefit of lithium-supported ¹³¹I therapy in benign disorders seems to be a more prompt control of hyperthyroidism. Two randomized studies by Bogazzi et al. (90, 472) investigated the effect of a short-term course of adjunctive lithium in patients with Graves' disease. Lithium was given either at the time of ¹³¹I therapy or 5 d before, as methimazole was withdrawn (90). ¹³¹I-induced transient hyperthyroidism was prevented by lithium, and a prompter control of the hyperthyroidism was achieved, probably due to the increased thyroid ¹³¹I retention. A more pronounced reduction of the thyroid volume was seen in the lithium-treated group (472), whereas the long-term cure rate was not affected significantly in either study. However, the same authors demonstrated retrospectively (276) that the cure rate, as well as time to cure, was significantly improved in 298 patients given adjunctive lithium, compared with 353 patients given ¹³¹I therapy alone (Fig. 8). These results are in line with another recent retrospective study from the United Kingdom (473), in which 101 patients with either Graves' disease or toxic nodular goiter were given lithium 800 mg/d for 10 d starting 3 d before the ¹³¹I therapy. Compared with patients not receiving lithium, the cure rate was increased by 60%, and the posttherapy serum levels of thyroid hormones were also lower than observed in the control group. The outcomes from the retrospective studies (276, 473) contrast the findings in a large randomized trial (474), including 350 hyperthyroid patients, in which adjunctive lithium failed to affect the cure rate or the number of additional ¹³¹I treatments. The applied thyroid dose and the thyroid hormones in the early posttherapy period were not monitored.

Figure 8

The effect of the adjunctive use of lithium on the cure rate after ¹³¹I therapy illustrated by Kaplan-Meier curves from the retrospective cohort study of Bogazzi et al. (276). A total of 651 patients with Graves' disease were included, of whom 298 were treated with ¹³¹I plus lithium (900 mg/d for 12 d) and 353 with ¹³¹I alone. The cure rate at the end of the follow-up period was significantly higher among patients given adjunctive lithium. [Reproduced from F. Bogazzi et al.: Impact of lithium on efficacy of radioactive iodine therapy for Graves' disease: a cohort study on cure rate, time to cure, and frequency of increased serum thyroxine after antithyroid drug withdrawal. J Clin Endocrinol Metab 95:201–208, 2010 (276), with permission. © The Endocrine Society.]

One randomized trial on the effect of lithium has been conducted in patients with nontoxic goiter treated with ¹³¹I (475). The levels of the thyroid hormones in the lithium-treated group were marginally but significantly lower in the early period after ¹³¹I therapy, compared with the group treated with ¹³¹I alone. The goiter reduction after 2 yr was similar in the two groups. The thyroid ¹³¹I retention and the applied thyroid dose were not measured in that study (475). As regards thyroid cancer patients, results have been conflicting. Some authors found no effect of lithium therapy (476), whereas in an American study (477) an increase of the ¹³¹I half-life by 50% in tumors and by 90% in remnants was observed by lithium therapy, corresponding to a 2.3-fold increase in the radiation dose. In another study, lithium improved the ablation rate after ¹³¹I therapy in low-risk cancer patients (478), but this may have been due to the simultaneous use of furosemide. As with iodine restriction, the whole-body radiation may be significantly increased by lithium-augmented ¹³¹I therapy (479).

Thus, several studies have confirmed that adjunctive lithium may prolong the thyroid ¹³¹I half-life, thereby increasing the radiation dose. In addition, lithium seems to prevent the transient thyrotoxicosis occurring after ¹³¹I therapy, and it might replace ATD in the early posttherapy period (91). Side effects by a short-term lithium administration are few and insignificant (276). The reduction of nontoxic goiter after ¹³¹I therapy seems unaffected by adjunctive lithium, and as regards the application in hyperthyroid patients, it remains to be proven in a randomized trial that the long-term cure rate is improved. Whether adjunctive lithium allows a reduction of the radioactivity without loss of efficacy also remains to be proven. Because lithium acts quite similar to elementary iodine in this context, the effect might depend on the iodine status of the patient—an issue not addressed in previous studies.

C Diuretics

Diuretics increase the excretion of many minerals, including iodine (480), and these drugs can thereby augment the thyroid RAIU by depleting the iodine pool. More than 40 yr ago, it was shown that an acute iodine excretion by osmotic diuretics increases the thyroid 24-h RAIU (481), but subsequent studies showed that prolonged administration of thiazides in euthyroid patients has no effect on the thyroid RAIU (482, 483).

Different diuretics, such as mannitol, ethacrynic acid, thiazide, and furosemide (478, 484–487), have been studied, and the concept has been applied to cancer patients (448, 478, 485, 486) as well as to patients with benign thyroid disorders. In a randomized study (488), in patients with Graves' disease and with a low thyroid RAIU, hydrochlorothiazide 100 mg/d for 5 d in combination with a low-iodine diet resulted in a significant improvement of the 24-h RAIU (from baseline 29 to 51%), compared with the control group treated with a low-iodine diet alone (from baseline 30 to 37%). Another study (489), performed in an iodine-depleted area, investigated the effect of 40 mg furosemide for 5 d in 40 patients with a nontoxic goiter and with a baseline thyroid 24-h RAIU of 27%. Furosemide administration resulted in an increase of the thyroid RAIU by 58%, which was significantly higher than the 17% increase in the control group given a low-iodine diet. The decrease in the urinary ¹³¹I excretion and iodine clearance was similar in both groups.

Although use of diuretics before ¹³¹I therapy seems attractive, there are drawbacks. By depleting the iodine pool, by diuretics or by other means, also the iodine clearance is decreased, and this leads to a higher total-body radiation during ¹³¹I therapy, as discussed previously. In fact, some authors advocate against the use of diuretics in cancer patients previously iodine depleted by a low-iodine diet (448, 449). Use of diuretics carries some other potential risks. Thiazides increase the T₃ level (482), and furthermore, these drugs can cause thyroid enlargement in rats (487). Whether these findings have any clinical relevance is questionable. Nevertheless, concern about the side effects is probably the main reason why iodine depletion by diuretics before ¹³¹I therapy has not gained widespread use.

D Recombinant human TSH

Within 1 yr after ¹³¹I therapy, the volume reduction of a nodular goiter is approximately 40–45%, but with pronounced interindividual variation (59). The effect in diffuse nontoxic goiters is even more pronounced (69). Despite sufficient volume reduction, a high degree of patient satisfaction, and relief of the cervical compression and cosmetic complaints, many patients with goiter are referred to thyroidectomy rather than ¹³¹I therapy. It should be noted, however, that vocal symptoms and a decrease in swallowing performance are frequent occurrences even late after total thyroidectomy, despite the absence of immediate operative complications (490). Particularly elderly patients are at risk of having surgical complications (491). No trial has compared the two treatments head to head, and the choice of observation, surgery, or ¹³¹I therapy must be based on a range of individual factors in a dialogue with the patient.

There are several reasons for the limited use—applying a global perspective—of ¹³¹I therapy for nontoxic goiters. In some countries, there is no tradition for using ¹³¹I for treatment of this condition. According to surveys made 5–10 yr ago, thyroid hormone was widely used at that time (73, 492–494), but this treatment is probably on the wane for several reasons, leaving surgery and ¹³¹I therapy as the remaining options. Many patients with MNG are referred to surgery without being considered for ¹³¹I therapy because the goiter is very large, because of a low thyroid RAIU, or because of fear of overlooking thyroid malignancy. Recommending surgery in patients with a very large goiter is rational because the relative goiter reduction after ¹³¹I therapy diminishes with increasing goiter size (64), despite application of an equivalent thyroid dose.

Obviously, the effect of ¹³¹I therapy is dependent on the thyroid being capable of concentrating and organifying ¹³¹I sufficiently. In many MNG, the RAIU is inhomogeneous and relatively low due to inactive (scintigraphically “cold”) nodules and partly suppressed paranodular tissue. As discussed earlier, the iodine fortification of salt, in some countries, has resulted in a lower thyroid RAIU (433, 434), which decreases the efficacy and occasionally hinders the use of ¹³¹I therapy. If the thyroid scintigram shows the tracer confined to a few autonomous nodules, the overall thyroid RAIU may seem low on visual inspection. The same applies to very large goiters. However, the thyroid RAIU is correlated positively with serum T₄ and negatively with the age of the patient, while a correlation to serum TSH and the goiter size is absent (293). A subnormal TSH, often seen in MNG, indicates that the RAIU is confined to a few “hot spots” surrounded by dormant thyroid tissue. These dormant areas of the goiter only weakly concentrate ¹³¹I and thus diminish the efficacy of ¹³¹I therapy. If suppressed thyroid tissue were to be reactivated, this might lead to an increase in goiter volume reduction after ¹³¹I therapy. Indeed, using rhTSH, many obstacles associated with conventional ¹³¹I therapy can be overcome, as demonstrated by several recent studies (78).

1 Evolution of rhTSH

Many patients with differentiated thyroid cancer are given adjunctive high-dose ¹³¹I after thyroidectomy. The RAIU in thyroid tumor cells is increased by TSH stimulation (4, 495). For this purpose, exogenous TSH (bovine or human cadaver) has been used, but was abandoned due to possible allergic reactions, loss of potency, development of TSH antibodies, and fear of Creutzfeldt-Jakob disease (496–500). Instead, a thyroid hormone withdrawal regimen is employed by which endogenous TSH increases, at the cost of reduction in QoL due to the resulting hypothyroidism. Characterization of the human TSH α- and β-subunit in the late 1980s (501, 502) and cloning of the human TSH-β gene in the late 1980s allowed highly purified rhTSH production in Chinese hamster ovary cells (503) and made the extensive use of exogenous TSH possible. Although rhTSH has an amino acid structure identical to native TSH (503), its glycosylation is different, with higher sialic acid content (504, 505). As a consequence, rhTSH exhibits lower in vitro bioactivity than native TSH, but because the metabolic clearance rate is significantly lower for rhTSH than for native TSH, rhTSH in vivo may be equivalent to or maybe more potent than native TSH (504). Estimated by immunoassays, the specific activity of rhTSH is between 5.51 and 7.63 IU/mg, based on the second International Reference Preparation (80/558) of native TSH as the standard (506). The biological activity of rhTSH is preserved after dilution under different durations and temperatures of storage (507). A new formulation of rhTSH has been made with a sialylation more similar to the native hormone (508). Furthermore, new TSH analogs have been designed that are much more potent than rhTSH for the stimulation of FRTL-5 cells and mice thyroids (509).

Iodine uptake, across the basolateral membrane of the thyroid follicular cells, is catalyzed by the NIS, which under normal conditions is only minimally expressed (510). rhTSH stimulation occurs with some latency. In FRTL-5 cells, rhTSH induces a significant increase in NIS-mRNA after 3–6 h, reaching a maximum at 24 h approximately six times above the basal level. NIS protein increases in parallel, and a rise in cellular RAIU is detected 12 h after rhTSH stimulation and peaks at 72 h, approximately 27 times above basal levels (511). In vivo studies in rhesus monkeys found that the half-life of rhTSH after iv administration is approximately 63 min for the rapid phase and 326 min for the slow phase (512). Furthermore, im injections of 0.43 mg rhTSH for 3 d double the 6- and 20-h thyroid iodine uptake (512). Subsequent human trials have confirmed that rhTSH profoundly stimulates the thyrocyte (78). After im injection of rhTSH, the serum level of TSH increases significantly within 2 h and peaks 4 h after injection (513) with a clear dose-response relationship (514). The suppressed serum TSH observed 7 d after injection is probably best explained by the increased levels of serum T₄ and serum T₃ at this time (514). As shown in thyroid cancer patients, peak serum TSH levels correlate with age after rhTSH administration (515), whereas other studies found body size and especially lean body mass to be the major determinants (516–518).

2 Application in thyroid cancer patients

For several years, rhTSH has been labeled for diagnostic and therapeutic use in patients with differentiated thyroid cancer, and it is routinely used in many centers (4, 495). The hypothyroid state during a thyroid hormone withdrawal regimen results in a longer effective ¹³¹I half-life due to a reduced renal clearance, as compared with the rhTSH stimulation regimen (519). It has been questioned whether this leads to a diminished efficacy of the rhTSH stimulation regimen (520, 521). However, the effective half-life in remnant thyroid tissue is longer after rhTSH than during hypothyroidism (522), and the residence time is similar in the two regimens, as shown in randomized studies (522, 523). If the dosimetry is based on ¹²⁴I-PET/CT, it has been indicated that rhTSH stimulation results in a lower radiation dose to metastases than obtained by thyroid hormone withdrawal (524), and detection of metastases is poorer compared with a thyroid hormone withdrawal regimen ¹²⁴I-PET (525). However, opposite results have been found when ¹⁸F-fluoro-2-deoxy-d-glucose-PET is used for imaging (526).

Although the ablation rate seems dependent on the absorbed dose (527), it is reassuring that follow-up studies have documented that the two regimens (rhTSH stimulation vs. thyroid hormone withdrawal) are equally effective (528–530), even in the context of a low-dose regimen (531). Importantly, the extrathyroidal irradiation as well as the impact on the chromosomes is lower during rhTSH stimulation (521, 522, 532–535), the patient's well-being is improved (536), other consequences of acute thyroid hormone deficiency are avoided (537–540), and it is economically superior to the withdrawal regimen (540–544), although this depends on the cost of rhTSH, rates of remnant ablation, and time off work.

3 Impact on the thyroid RAIU

Although rhTSH usage is restricted to thyroid cancer patients, experience with the use of rhTSH in subjects with an intact thyroid gland has accumulated during the last 10 yr, and there are several comprehensive reviews on this topic (78, 545–550). A core issue is the effect of rhTSH on the thyroid RAIU. In healthy subjects, 0.9 mg rhTSH increased the 6- and 24-h thyroid RAIU from 12.5 and 23%, respectively, at baseline, to 27 and 41%, respectively (551). In healthy but iodine-loaded subjects, 0.9 mg rhTSH administered 32 h before ¹³¹I increased the thyroid RAIU by 97%, but far from normalized the thyroid RAIU, because the absolute RAIU increased only from 3 to 6% (552). Another study found that in patients exposed to iodinated contrast media, 0.1 mg rhTSH increased the 4-h thyroid RAIU from 7 to 19% (553). In the case of amiodarone-induced thyrotoxicosis, the effect is more pronounced (554, 555). In these individuals, 0.1 mg rhTSH given twice increased the thyroid 24-h RAIU from around 6 to 30%, enabling subsequent administration of ¹³¹I therapy. However, caution is warranted because one of the subjects experienced a sustained rise in serum thyroid hormones resulting in cardiac arrhythmias (555).

Many studies have demonstrated that rhTSH has a marked effect in patients with MNG, increasing the thyroid RAIU 2- to 4-fold (248). There are, however, large interindividual variations, as seen also in healthy subjects (551). Thus, a remarkable increase from a very low baseline of 3 to 39% after rhTSH has been described (556). Such a variation indicates that individual factors, most of which are yet unidentified, are involved. Much of the variation is explained by differences in the baseline thyroid RAIU because the effect is highly inversely correlated with this variable (248, 318, 557) (Fig. 9). This means that patients with MNG, in whom therapy previously was restricted to thyroidectomy due to a low RAIU, may benefit from ¹³¹I therapy if combined with rhTSH prestimulation. Also, the serum TSH level seems of importance because the increase in the thyroid 24-h RAIU is inversely correlated with the baseline serum TSH (318). This can be explained by the down-regulation of normal TSH-responsive thyrocytes, comprising the major part of the thyroid, which upon rhTSH stimulation becomes reactivated as for ¹³¹I uptake. A similar mechanism may be responsible for the important observation that rhTSH stimulation in patients with MNG causes a more homogeneous distribution of ¹³¹I (558, 559).

Figure 9

The positive effect of rhTSH stimulation on the thyroid ¹³¹I uptake in patients with MNG is clearly inversely related to the initial ¹³¹I uptake. With a high thyroid ¹³¹I uptake, very little is gained by rhTSH stimulation. [Reproduced from S. Fast et al.: Optimizing 131I uptake after rhTSH stimulation in patients with nontoxic multinodular goiter: evidence from a prospective, randomized, double-blind study. J Nucl Med 50:732–737, 2009 (318), with permission. © The Society of Nuclear Medicine.]

4 Application in patients with benign goiter

During the last decade, several studies have evaluated the goiter reduction after rhTSH-stimulated ¹³¹I therapy in MNG (70, 272, 273, 556, 559–569) (Table 3). In nine trials (70, 272, 273, 556, 563–565, 568, 569), a control group was included, and four of these (70, 272, 273, 565) were double-blinded. In these studies comparing rhTSH with placebo prestimulation, the concept of rhTSH-stimulated ¹³¹I therapy was convincingly proved (Fig. 10). In goiters below 100 ml, the 12-month goiter reduction was improved from 46 to 62%, using 0.3 mg rhTSH-stimulated ¹³¹I therapy, corresponding to a relative increase of 34% (272). An even greater improvement, from 34 to 53% (relative increase of 56%), was found in goiters above 100 ml (70). Thus, the inverse correlation between goiter reduction and the initial goiter size, as seen with conventional ¹³¹I (63, 64), does not seem to apply for rhTSH-augmented ¹³¹I therapy (70, 272). In the study by Bonnema et al. (111), rhTSH-stimulated ¹³¹I therapy increased the smallest cross-sectional area of the trachea by 31%, and the inspiratory flow improved by 25%, whereas no change was found in either parameter after placebo-stimulated ¹³¹I therapy. The implication was that rhTSH-stimulated ¹³¹I therapy is more effective in relieving tracheal compression, and mainly in those patients who are most needy (111).

Figure 10

Goiter volume reduction after rhTSH-stimulated ¹³¹I therapy according to a “superiority approach” (top) or an “equality approach” (bottom). Data are from the two randomized studies by Nielsen et al. (272) and Fast et al. (273), respectively. The control groups received placebo-stimulated ¹³¹I therapy. Following the “superiority approach,” the usual amount of radioactivity was given, resulting in a more pronounced goiter volume reduction. Following an “equality approach,” the goiter reduction was the same as in the control group, but the amount of radioactivity could be reduced by 70% due to the increase in the ¹³¹I thyroid uptake.

Table 3

Studies on the effect of rhTSH-stimulated ¹³¹I therapy in patients with benign multinodular goiter

First author, year (Ref.)	n	Dose of rhTSH (mg)	Design	Goiter size estimation	131I activity or intended dose	Goiter reduction
Studies with an equality approach
Nieuwlaat, 2003 (561)	12	0.01	Observational	MRI	100 Gy	35% at 1 yr
	10	0.03	Noncontrolled			41% at 1 yr
Fast, 2010 (273)	60	0.1	Randomized	US or MRI	50 Gy	35% at 1 yr
	30	Placebo	Double-blinded		100 Gy	35% at 1 yr
Ceccarelli, 2011 (568)	11	0.03	Observational	CT scan	100 Gy	47% at 1 yr
	7	Controls	Matched controls			35% at 1 yr
Studies with a superiority approach
Duick, 2003 (566)	6	0.3	Observational	Palpation	Fixed 1110 MBq	30–40% at 7 months
	10	0.9	Noncontrolled			30–40% at 7 months
Albino, 2005 (559)	18	2 × 0.1	Observational noncontrolled	CT scan	Fixed 1110 MBq	39% at 6 months
Cohen, 2006 (560)	17	0.03	Observational noncontrolled	CT scan	Fixed 1110 MBq	34% at 6 months
Paz-Filho, 2007 (562)	17	0.1	Observational noncontrolled	CT scan	Fixed 1110 MBq	46 & 53% at 1 & 2 yr
Romão, 2009 (567)	42	0.1	Observational noncontrolled	CT scan	Fixed 1110 MBq	From 153 ml to 32 ml at 3 yr
Giusti, 2006 (563)	12	2 × 0.2	Observational	CT scan	Fixed 370–555 MBq	44% at 20 months
	8	Controls	Matched controls			25% at 22 months
Giusti, 2009 (569)	19	2 × 0.1	Observational	US scan	Restricted to 600 MBq	60% at 3 yr
	21	Controls	Matched controls			44% at 3 yr
Cubas, 2009 (564)	9	0.1	Observational	CT scan	Fixed 1110 MBq	33 & 37% at 1 & 2 yr
	9	0.005	Placebo-controlled			33 & 39% at 1 & 2 yr
	10	Placebo				13 & 15% at 1 & 2 yr
Silva, 2004 (556)	17	0.45	Randomized	CT scan	Fixed arbitrary levels	58 & 73% at 1 & 4 yr
	17	Placebo	Not blinded			40 & 57% at 1 & 4 yr
Nielsen, 2006 (272)	28	0.3	Randomized	US scan	Above 100 Gy	62% at 1 yr
	29	Placebo	Double-blinded		100 Gy	46% at 1 yr
Bonnema, 2007 (70)	14	0.3	Randomized	MRI	Above 100 Gy	53% at 1 yr
	15	Placebo	Double-blinded		100 Gy	34% at 1 yr
Albino, 2010 (565)	8	0.1	Randomized	MRI	Fixed 1110 MBq	37% at 1 yr
	6	0.01	Double-blinded			37% at 1 yr
	8	Placebo				19% at 1 yr

The majority of patients suffered from a compressive nontoxic goiter, but patients with mild toxic multinodular goiter were also included in some of the studies. For further details, please see Ref. 78. For the “equality approach,” the aim is to reduce the amount of radioactivity that equals the rhTSH-induced increase in the thyroid ¹³¹I uptake. The aim of the “superiority approach” is to achieve a greater goiter reduction compared with conventional (without rhTSH stimulation) ¹³¹I therapy. US, Ultrasound.

Most often, a single rhTSH injection, given 2–72 h before ¹³¹I administration, has been used. Comparative trials strongly suggest that a time interval of 24 h before ¹³¹I administration is superior to intervals of 2, 48, or 72 h to maximize the increase in the thyroid RAIU (318, 570). As regards the rhTSH dose, the majority of studies used 0.1 to 0.3 mg, but the doses have ranged from as little as 0.005 mg (564) to 0.9 mg (566). In thyroid cancer patients, the recommended dose is higher (0.9 mg given twice), and it is worth noting that the serum peak TSH does not correlate with the rate of ¹³¹I remnant ablation in these patients (518).

The optimal applied thyroid dose for reduction of nodular goiter is not established, and no prospective dose-response study has been performed. The doses used in previous studies—mostly in the range of 3.7–5.5 MBq/g thyroid tissue—were adapted from ¹³¹I therapy of hyperthyroid diseases. Most rhTSH studies (556, 559, 560, 562–567) have used fixed doses, with the consequence that the goiters were irradiated to a varying degree due to differences in the thyroid RAIU and goiter size. In other studies (70, 272, 561), the thyroid dose was based on thyroid dosimetry, and with rhTSH stimulation the applied thyroid dose reached 150–300 Gy, much more than the targeted 100 Gy in the placebo-stimulated patients. In one study, the amount of radioactivity was restricted to an extent that equaled the increase of the thyroid RAIU, and by this approach the goiter reduction was 40% at 1 yr, comparable to that obtained by conventional ¹³¹I therapy (561). Without rhTSH stimulation, a positive correlation seems to exist between the relative goiter reduction and the retained thyroid dose (63, 70, 272). Such a correlation is less clear, or even absent, with rhTSH stimulation, as indicated by some (70, 272), but not all studies (559). This implies that TSH may amplify the effect of the ¹³¹I therapy beyond that related to the retained thyroid dose. Such an intriguing feature of rhTSH is possibly explained by its ability to render the thyrocyte more radiosensitive. A similar correlation between efficacy (i.e., cure rate) and serum TSH has been observed in hyperthyroid patients treated with ¹³¹I (88, 262, 280, 301), but the mechanisms may be different in that situation. Indeed, we recently demonstrated, in a randomized double-blind study (273), that an equivalent goiter reduction after ¹³¹I therapy can be obtained by a much lower thyroid dose if rhTSH-prestimulation is used (Fig. 10). It strongly supports the view that rhTSH has a preconditioning effect on ¹³¹I therapy; the important implication is that a much lower amount of ¹³¹I may suffice without loss of efficacy. The latter study (273) can be characterized as an “equality approach,” where the aim is to reduce the amount of radioactivity corresponding to the rhTSH-induced increase in the thyroid RAIU. Employing a “superiority approach,” used in most previous rhTSH studies mentioned above, the usual amount of radioactivity is given aiming at a greater goiter reduction due to an increased thyroid irradiation (Table 3).

Patient satisfaction after rhTSH-stimulated ¹³¹I therapy has only been monitored to a limited extent. In our two double-blinded studies, in which the goiter reduction was amplified by rhTSH stimulation, the patient satisfaction measured by a visual analog scale was similar in the placebo and rhTSH groups (70, 272). Neither could a difference in patient satisfaction be demonstrated in the double-blind study by Fast et al. (273), in which the control group and the rhTSH/reduced-irradiation group achieved similar goiter volume reduction after ¹³¹I therapy. However, patient satisfaction is generally high with conventional ¹³¹I therapy, which reduces the possibility of detecting significant differences. Application of a newly developed thyroid disease-specific QoL questionnaire (571) may prove valuable for this purpose. Nevertheless, it can be questioned whether rhTSH stimulation is unequivocally beneficial to the patient, particularly if side effects are more commonly encountered.

The long-term effect and the risk of goiter recurrence is an important issue that has only sparsely been evaluated. A few studies including a limited number of patients (569, 572, 573) have evaluated the effect of rhTSH-stimulated ¹³¹I therapy beyond 1 yr, with conflicting results. In the noncontrolled study by Paz-Filho et al. (572), no further goiter reduction was achieved in the second year (48% goiter reduction during the first year). In contrast, Cardia et al. (573) found that the goiter reduction continued during the additional 3 yr of follow-up, to the same extent as in patients who received conventional ¹³¹I therapy. A few patients in both studies (572, 573) experienced thyroid regrowth during follow-up. In the nonrandomized study by Giusti et al. (569) with a median follow-up of 36 months, patients pretreated with 0.1 mg rhTSH on 2 consecutive days had more goiter reduction and fewer goiter-related complaints compared with non-pretreated patients, but the two groups were not comparable at baseline as regards the thyroid RAIU and the thyroid function. Larger and well-controlled studies are needed before concluding whether rhTSH-stimulated ¹³¹I therapy confers any long-term benefit, compared with conventional ¹³¹I therapy. Strongly supporting a long-term benefit of this therapeutic concept is our recent follow-up study of our previous RCT (574). This study shows that the improved goiter reduction is maintained at a median of 71 months after ¹³¹I therapy, and furthermore the need for additional therapy is reduced, as compared with conventional ¹³¹I therapy (574).

A completely different approach for increasing the thyroid RAIU is to stimulate the secretion of endogenous TSH by methimazole. Albino et al. (575) pretreated patients with nodular goiter with methimazole, aiming at a serum TSH above 6 mU/liter. The thyroid 24-h RAIU increased from 21 to 78%, and 1 yr after ¹³¹I therapy, the goiter reduction was 46%. Whether a marginal hypothyroid state, obtained by ATD, is as effective as rhTSH for increasing the thyroid RAIU and for augmenting the goiter reduction after ¹³¹I therapy remains to be clarified by controlled trials.

5 Potential concerns with rhTSH-stimulated ¹³¹I therapy

Experience with the clinical use of rhTSH is mainly obtained from patients with differentiated thyroid cancer. Generally, rhTSH is well tolerated in these patients treated with repeated doses of 0.9 mg rhTSH (576). There have, however, been reports of serious reactions such as tumor swelling and pain from metastases thought to result from injections of rhTSH (576–582). The rapid onset, the favorable response to glucocorticoids, and radiological findings of peritumoral edema or hemorrhage strongly suggest that the tumor expansion was the result of swelling rather than tumor growth (576). In fact, glucocorticoid premedication, as an adjuvant to rhTSH-stimulated ¹³¹I, is recommended in cancer patients with known or suspected lesions in confined spaces (576). Clinical thyrotoxicosis in a thyroid cancer patient given rhTSH-stimulated therapy has also been described, probably due to massive functional skeletal and soft-tissue metastases (579).

In patients with an intact thyroid gland, the situation is quite different because thyroid swelling and thyrotoxicosis may pose potential problems due to the profound stimulation by rhTSH (78). As discussed earlier, conventional ¹³¹I therapy of large goiters occasionally leads to tracheal compression (64, 89). In addition, rhTSH per se causes a temporary thyroid swelling, but this side effect is unpredictable with a wide interindividual variation (514). This effect of rhTSH is clearly dose-dependent because 0.3 and 0.9 mg increases the thyroid volume by 30–45% (583, 584) with a maximum between d 1 and 4 and subsequent reversion to the initial size before d 7, whereas 0.1 mg rhTSH causes insignificant changes (514). rhTSH in the dose range 0.03–0.3 mg, given to patients with a smaller MNG (median size, 20 ml), does not on average affect the lung function assessed by flow volume loops (585). However, rhTSH and ¹³¹I therapy in combination may lead to a serious thyroid swelling. Indeed, cervical pain and a sense of thyroid growth within the first week are more frequently reported after rhTSH-stimulated ¹³¹I therapy (70, 272, 556, 559, 560). Some of the initial trials in patients with MNG (70, 272, 556, 566) used rhTSH doses of 0.3–0.45 mg, explaining the majority of side effects reported in those studies. Three studies (111, 561, 565) have monitored the goiter volume and the tracheal anatomy in the early phase after rhTSH-stimulated ¹³¹I therapy. Although larger trials are needed, the safety parameters from these studies are reassuring. Thus, using 0.3 mg rhTSH, the goiter volume as well as the smallest cross-sectional tracheal area were unchanged—on average—1 wk after ¹³¹I therapy (70, 111). However, larger deviations from baseline were observed than with conventional ¹³¹I therapy. Similarly, the two other studies (561, 565), using lower doses of rhTSH, did not find acute alterations either. In the study by Bonnema et al. (111), the respiratory function was also monitored. The inspiratory function was unaffected 1 wk after rhTSH-stimulated ¹³¹I therapy, whereas a slight but significant decrease in the forced vital capacity was observed. None of the respiratory parameters were affected among placebo-stimulated patients (111). Glucocorticoids can probably minimize or even prevent a critical thyroid swelling if this is anticipated during rhTSH-stimulated ¹³¹I therapy, in line with the situation in thyroid cancer patients (576)

The effects of rhTSH on thyroid function, both in healthy individuals and in MNG patients, have been evaluated in several studies (514, 551, 557, 570, 583, 584). Within 4–8 h of rhTSH injection, a rise in serum levels of T₄ and T₃ occurs, peaking at 24–48 h, followed by normalization within 3–4 wk. Serum thyroglobulin has a slightly slower rise, peaking at 48 h after rhTSH stimulation, but the kinetics of thyroglobulin is in principle similar to that of the thyroid hormones (514). In parallel with the impact on the thyroid volume, wide individual variations in healthy subjects are seen, but with a clear dose-response relationship. Thus, a more pronounced response in serum levels of T₄, T₃, and thyroglobulin is observed when administering 0.3 or 0.9 mg rhTSH, compared with lower doses (514). A head-to-head comparison of the thyroid function after rhTSH-stimulated (using 0.45 mg) and conventional ¹³¹I therapy, respectively, was done in one study (556). In the rhTSH group, the serum levels of T₃ and T₄ were increased approximately 2.5-fold above the levels in the control group, 24–72 h after therapy (556). In other studies using rhTSH doses of 0.01 and 0.03 mg, most patients maintained thyroid hormone levels within the normal range (561). Taking these studies together and balancing the beneficial effect (long-term goiter reduction and decreased whole-body radiation) with the side effects (thyroid swelling, temporary hyperthyroidism. or permanent hypothyroidism), the rhTSH dose for augmenting ¹³¹I therapy efficacy in patients with MNG is most likely in the range of 0.03–0.1 mg. Patients with mild or overt toxic multinodular goiter are those most at risk of developing a clinically significant rise in serum thyroid hormone levels, even with 0.1 mg rhTSH (273, 567). Moreover, one study has implied that the goiter reduction after rhTSH-stimulated ¹³¹I therapy is most pronounced in euthyroid patients with a normal serum TSH, as compared with patients with a thyrotoxic goiter (567). Knowledge about the acute effects and the relevant dose of rhTSH on the thyroid volume and tracheal compression is still limited (111, 565), and larger series are needed to evaluate more closely whether rhTSH-stimulated ¹³¹I therapy confers a clinically significant risk. To avoid inadvertent thyroid stimulation, a “modified-release rhTSH” has been introduced. This compound has a slightly different serum profile with a delayed serum peak after injection, and a recent phase II trial investigated its effect in patients with MNG undergoing ¹³¹I therapy (586). The goiter reduction after 6 months was 23% in patients prestimulated with either placebo or 0.01 mg modified-release rhTSH, whereas the goiter reduction was 33% in those prestimulated with 0.03 mg. The smallest cross-sectional area of the trachea increased most in the latter group, although it was not significantly different from the other two groups (586). Whether there is any clinically relevant difference between the new and the old formulation of rhTSH is unknown.

Some patients may be reluctant to accept ¹³¹I therapy due to the fear of developing radiation-induced cancer, a concern also shared by some physicians. There are no studies on the risk of malignancy after ¹³¹I therapy of MNG, whether this is given with or without rhTSH prestimulation. The strategy, by which the amount of radioactivity is reduced according to the rhTSH-induced increase in thyroid RAIU (273, 561), leads to a considerably lower whole-body radiation (587). In addition, the possible preconditioning effect of rhTSH allows an even further reduction of the radioactivity (273), thus potentially reducing the theoretical risk of ¹³¹I-induced malignancy.

The most significant long-term adverse effect of rhTSH-stimulated ¹³¹I is an up to 5-fold increased risk of hypothyroidism (up to 65% after 1 yr), as demonstrated in three randomized trials (70, 272, 556). Some patients with hypothyroidism have a reduced QoL (588, 589), but it can be questioned whether thyroid failure after ¹³¹I therapy should be considered a serious side effect. The alternative to ¹³¹I therapy, i.e., surgery, would lead to hypothyroidism in the majority of patients because a total (or near-total) thyroidectomy is performed at an increasing number of centers in case of bilateral thyroid lesions (590). No study has evaluated to which extent QoL is affected if goiter treatment—by any method—leads to permanent hypothyroidism. Along the same line, it is unclear how hypothyroidism, even if treated with thyroid hormone, affects morbidity and mortality (591). It remains to be clarified whether the augmented goiter reduction and the improved upper airway dynamics, seen with rhTSH-stimulated ¹³¹I therapy, justify the increased prevalence of permanent hypothyroidism. QoL is clearly an important parameter that needs to be addressed in future trials.

Theoretically, rhTSH may provoke an autoimmune response due to its profound effect on the thyroid. In fact, a few reports have described a possible association between rhTSH stimulation and the development of TAO in patients with thyroid cancer (592–594). Similar serious side effects have not been reported in patients with benign goiter treated with rhTSH-stimulated ¹³¹I. Thyroid autoantibodies did not appear in healthy subjects undergoing repeated rhTSH stimulation (514), whereas one study reported that TPOAb developed in eight of 15 patients with MNG after rhTSH-stimulated ¹³¹I therapy (595). The concentration was significantly higher at 3 months than in patients treated with conventional ¹³¹I therapy, but at 12 months there was no difference between the two groups (595).

6 Extrathyroidal effects of rhTSH

TSH also has extrathyroidal effects, which have only recently been studied. TSH receptors are present on osteoclasts, osteoblasts, and adipocytes, and acute rhTSH administration to thyroidectomized patients increases the osteoblast activity, reflected by an increase in N-terminal propeptide of type I procollagen and serum receptor activator of nuclear factor-kappaβ ligand in postmenopausal (but not in premenopausal) women (596). In addition, serum leptin increases proportionally to the adipose mass (597), and serum ghrelin is temporary depressed (598). Vascular factors are also affected by TSH. In mice grafted with human thyroid tissue, rhTSH stimulates angiogenesis and local vascular endothelial growth factor (VEGF) expression (599). VEGF—a growth factor elevated in patients with metastases—is associated with disease progression and a poor prognosis (600). It is reassuring, however, that no rise in VEGF is seen in thyroid cancer patients stimulated with rhTSH (600), and some studies have even demonstrated a significant reduction in serum VEGF (601), suggesting that TSH can regulate VEGF production from nonthyroidal tissues.

Results are inconsistent as regards the impact on the circulation. Short-term stimulation with rhTSH transiently increases the endothelial-derived factor nitric oxide 3–6 d after injection (602). Indeed, a slight fall of the arterial blood pressure has been demonstrated after rhTSH administration in some (603) but not in all studies (604). Moreover, a marked and rapid activation of the endothelial-mediated vasodilation after rhTSH administration is seen, with no significant changes in vascular cell adhesion molecule-1, TNF-α, and IL-6 (604, 605). In contrast, other studies found rhTSH stimulation associated with platelet activation (606), impairment of the endothelium-dependent vasodilation, and elevation of blood IL-6 and TNF-α, along with a reduction of nitric oxide availability (607). It is at present unknown whether the conflicting results are caused by as yet unidentified confounding factors. It is also unclarified to what extent the various effects are transient, and whether the effects of rhTSH on the vascular system are dose-dependent. Regardless of the possible impact on the vascular bed, the cardiac function is not clinically affected by rhTSH stimulation, either in thyroid cancer patients (603, 604) or in mildly hyperthyroid MNG patients given ¹³¹I therapy (608).

E Other compounds

The NIS, the primary mediator of iodine uptake by the thyrocyte, is regulated by a number of factors (609). NIS expression, being highly correlated with the ¹³¹I accumulation, may be decreased or even lost in follicular and papillary thyroid carcinomas, and approximately 30% of the cancer cell lines lose their capability of ¹³¹I concentration (609). Such dedifferentiation hinders an effective adjuvant ¹³¹I therapy and leads to a worse prognosis. Failure of iodine uptake might be reversed by redifferentiating agents such as retinoic acid (RA; isotretinoin) and rexinoids (bexarotene). RA, used for many years in the treatment of severe acne vulgaris, mediates a wide spectrum of genetic, proliferative, and differentiation processes by binding to RA receptors and retinoid X receptors in the cellular nucleus. Subsequently, the RA receptor–retinoid X receptor complex modulates the transcription of target genes, including an increased expression of NIS-mRNA. Cis-RA (but not trans-RA) given to rats stimulates NIS activity and results in more hydrogen peroxide production, whereas thyroid peroxidase activity is unaffected (610). RA receptors are to some extent expressed in human thyroid carcinomas (611), and in vitro studies have shown that RA increases the expression of NIS in tumor cells in a dose-dependent manner (612, 613). Indeed, a small number of pilot studies indicate that 6–8 months of RA treatment redifferentiate thyroid cancer cells, leading to an increased RAIU, a higher increase of thyroglobulin (614, 615), and a better response to ¹³¹I therapy (614, 616–618). Unfortunately, other and larger studies (619–622) do not support these initial findings, and a clinically relevant response was found in only a minority of patients. The inconsistent response to RA may rely on the fact that RA and retinoid X receptors are variably expressed in thyroid tumors (623). Most studies used 13-cis-RA, but all-trans-RA and bexarotene have also been evaluated (618, 624). Whether the response differs between these drugs is unknown.

The concept of RA treatment has not been tested in benign thyroid diseases. It might be speculated that RA is able to increase the expression of NIS in, for example, scintigraphically cold nodules or in amiodarone-induced hyperthyroidism, a condition known to be associated with a very low thyroid RAIU. RA treatment for up to 3 months is generally well tolerated. The most common side effects are dry skin and xerostomia. However, a potential hindrance for the use of these agents in patients with an intact thyroid gland is that the thyrotrophs in the pituitary gland are rapidly suppressed, resulting in central hypothyroidism, as shown for bexarotene (625). Even more worrying is the observation that RA-augmented ¹³¹I therapy may provoke the development of TAO (592).

Also, the effects of glitazones on the thyroid have been investigated. In rats, rosiglitazone reduces the levels of plasma thyroid hormones, liver deiodinases, and the number of thyroid hormone receptors in adipose tissue, and these effects in combination with a reduced sympathetic stimulation of adipose tissue contribute to the glitazone-induced reduction of the energy expenditure (626). In vitro studies demonstrated that glitazones are able to reverse the dedifferentiation of some thyroid tumor cell lines, to increase the RAIU, and interestingly, at the same time to promote apoptosis (627–630). Troglitazone seems superior to rosiglitazone and pioglitazone in this setting (628). A few small clinical trials have shown that treatment with rosiglitazone in thyroid cancer may enhance the RAIU in some patients (631–633). However, results have been inconsistent (634), probably due to variations in the expression of peroxisome proliferator-activated receptor-γ (635).

Many other compounds, besides those already discussed, have been investigated for their potential to increase NIS expression in the thyrocyte and/or the intracellular ¹³¹I retention time. These agents include valproic acid, histone deacetylase inhibitors, azacytidine, arsenic trioxide, bortezomib, rapamycin, apicidin, lovastatin, and resveratrol (636–639). As with RA and glitazones, none of these have been evaluated in patients with benign thyroid disorders, and with the advent of rhTSH it is difficult to see a future role for any of these compounds in the context of ¹³¹I therapy. Of potential clinical importance is the recent observation that the expression of genes responsible for the ¹³¹I uptake in thyroid cancer cells can be restored by a range of inhibitors targeting the MAPK and PI3K/Akt/mTOR pathways (640). Telomerase-driven expression of NIS is another novel concept that may have applications for ¹³¹I therapy (641).

The content of selenium within the thyroid gland is high, and the role of this trace element in thyroid diseases has been paid much attention recently. Selenoproteins are involved in the cellular antioxidative defense, and they are probably important in the protection of the thyrocyte from reactive oxygen species (407). It is therefore likely that the selenium status of the patient influences the effect of ¹³¹I therapy, but that possibility has yet to be explored.

VIII Closing Remarks, Remaining Questions, and Directions for Future Research

Despite the accumulated knowledge during nearly 70 yr of study within the field of ¹³¹I therapy, our evidence base is hampered by most studies being inadequately designed and underpowered. Thus, to some surprise, high-level solid evidence is lacking for many aspects of this treatment. It is therefore difficult to provide well-founded recommendations. Evidently, the outcome from ¹³¹I therapy of hyperthyroid diseases is potentially confounded by a range of variables, such as age, goiter size, applied radioactive dose, severity and type of disease, as well as pre- and post-radioiodine use of ATD. Although most of the individual randomized trials did not find these drugs to be radioprotective, a reduction of the cure rate by the use of ATD in conjunction with ¹³¹I therapy is supported by a meta-analysis. Whether ATD have inherent radioprotective properties awaits additional well-powered and well-designed clinical trials and the utility of novel in vitro models. ATD also affect other factors such as serum TSH and the thyroid RAIU, and these may confound the influence of ATD on the cure rate after ¹³¹I therapy. Finally, the temporal relation between the treatment with ATD and the ¹³¹I administration needs to be taken into consideration and studied further.

Much focus has been on finding the most appropriate applied thyroid dose. Because ¹³¹I therapy initiates a destructive process of the thyroid, rather than being a cure of the thyroid disorder, the long-term outcome is permanent thyroid failure in most cases, at least when treating Graves' disease. Therefore, the issue is rather a choice between the use of a high amount of radioactivity to rapidly obtain euthyroidism or hypothyroidism or, alternatively, to restrict the use of radioactivity at the expense of a higher rate of treatment failure. On a group level, a dose-response relationship exists. However, a pronounced individual variation in the effect of the radiation dose is evidenced by numerous studies. Calculation of the applied dose is based on standardized models, but dosimetric factors such as the thyroid follicle size and the metabolic state of the thyrocyte undoubtedly modify the biological effects of the radiation. It is likely that one or more important factors that determine the outcome from ¹³¹I therapy remain to be identified. Research into the cellular responses to radiation and lessons from radiotherapy of cancer patients indicate that the radiosensitivity of the cell relies not only on endogenous and exogenous factors, but also on genetic susceptibility and modification by the interaction with the immune system. The applications of new biotechnological methods to assess the radiosensitivity at the individual level and interventions that may increase the radiosensitivity are future research areas that may open novel avenues for a more accurate and individualized tailoring of ¹³¹I therapy.

Pertinent issues in the management of patients with nontoxic goiter also need clarification. TSH-suppressive therapy with levothyroxine should be discouraged due to an unfavorable risk-benefit ratio, whereas both surgery and ¹³¹I therapy are effective for goiter treatment. Each carries distinct advantages and disadvantages, and a correct selection of patients for therapy, whether surgery or radioiodine, is of pivotal importance. Indeed, patients with limited symptoms may be better off without intervention. Depending on factors like age, goiter size, and comorbidity, some patients are obvious candidates for surgery, whereas ¹³¹I therapy is better suited for others. Future randomized trials, comparing different modalities including observation, should focus on QoL as well as cost-effectiveness. In addition, the long-term consequences of ¹³¹I therapy of patients with nontoxic goiter—in particular the risk of overlooking or causing malignancy—remain to be evaluated in large-scale studies.

Accepting that the concept of rhTSH-stimulated ¹³¹I therapy is beyond a doubt more effective than conventional (without rhTSH) ¹³¹I therapy for goiter reduction, it should be emphasized that rhTSH-stimulated ¹³¹I therapy is still an off-label treatment for MNG. Moreover, the experience with this therapy is limited, and whether QoL is improved remains unclarified. From a patient's point of view, the beneficial effect may be counterbalanced by the higher incidence of permanent hypothyroidism necessitating permanent medication with levothyroxine. Perhaps it is more rational to follow an “equality strategy,” whereby the amount of radioactivity is reduced by a factor that equals the rhTSH-induced increase in the thyroid RAIU. Although very few studies have explored such an approach, the results have been intriguing. If rhTSH has a preconditioning effect, i.e., the ability to increase the radio-susceptibility of the thyroid gland, as suggested by a few studies, this could reduce the need for radioactivity even more and turn out to be very attractive for both patients and physicians. Keeping the dose low, the risk of rhTSH administration to a patient with an intact thyroid gland is very modest, significant adverse effects are rare, and in principle, this allows a higher number of patients to be treated on an outpatient basis.

We are confident that the answers to these and many more questions will not await another 70 yr of study.

Acknowledgments

We thank Søren Fast and Viveque E. Nielsen for their important role in a number of our rhTSH studies. We are indebted to the many patients who have participated in our numerous trials, thereby advancing our knowledge in this field.

S.J.B. was supported by grants from The Agnes and Knut Mørk Foundation, Desiree and Niels Yde's Foundation, and Thyreoidea Landsforeningen. L.H. was supported by grants from The Novo Nordisk Foundation.

It should be emphasized that rhTSH for augmenting the effect of radioiodine therapy of nontoxic goiter is an off-label use. The manufacturer of rhTSH (Genzyme) has had no influence on any aspect of our own studies utilizing rhTSH.

Disclosure Summary: S.J.B. has nothing to declare. L.H. is an advisory board member and has received consultancy fees from Genzyme Corporation, Cambridge, Massachusetts.

Abbreviations

 
• ATD

  Antithyroid drug(s)

 
• 53BP1

  p53-binding protein 1

 
• CT

  computed tomography

 
• DSB

  double-strand breaks

 
• HRR

  homologous recombination repair

 
• LMDS

  locally multiply damaged sites

 
• MIRD

  medical internal radiation dose

 
• MNG

  multinodular nontoxic goiter

 
• MRI

  magnetic resonance imaging

 
• NHEJ

  nonhomologous end-joining

 
• NIS

  Na+/I− symporter

 
• PET

  positron emission tomography

 
• PTU

  propylthiouracil

 
• QoL

  quality of life

 
• RA

  retinoic acid

 
• RAIU

  radioiodine uptake

 
• RCT

  randomized controlled trial

 
• rhTSH

  recombinant human TSH

 
• TAO

  thyroid-associated ophthalmopathy

 
• TcTUs

  technetium thyroid uptake under exogenous or endogenous suppression

 
• TgAb

  thyroglobulin antibodies

 
• TPOAb

  thyroid peroxidase antibodies

 
• TRAb

  TSH-receptor antibodies

 
• TSAb

  thyroid-stimulating autoantibody

 
• VEGF

  vascular endothelial growth factor.