The Influence of Patient Age on Thyroid Nodule Formation, Multinodularity, and Thyroid Cancer Risk

Introduction:

Although advancing age is known to influence the formation of thyroid nodules, the precise relationship remains unclear. Furthermore, it is uncertain whether age influences the risk that any thyroid nodule may prove cancerous.

Aim:

The aim was to determine the impact of patient age on nodule formation, multinodularity, and risk of thyroid malignancy.

Method:

We conducted a prospective cohort analysis of consecutive adults (ages 20–95 y) who presented for evaluation of nodular disease from 1995 to 2011. A total of 6391 patients underwent ultrasound and fine-needle aspiration of 12 115 nodules ≥1 cm. Patients were divided into six age groups and compared using sonographic, cytological, and histological endpoints.

Result:

The prevalence of thyroid nodular disease increases with advancing age. The mean number of nodules at presentation increased from 1.5 in the youngest cohort (age, 20–30 y) to 2.2 in the oldest cohort (age, >70 y; P < .001), demonstrating a 1.6% annual increased risk for multinodularity (odds ratio, 1.02; P < .001). In contrast, the risk of malignancy in a newly identified nodule declined with advancing age. Thyroid cancer incidence per patient was 22.9% in the youngest cohort, but 12.6% in the oldest cohort (odds ratio, 0.972; P < .001), demonstrating a 2.2% decrease per year in the relative risk of malignancy between ages 20 and 60 years, which stabilized thereafter. Despite a lower likelihood of malignancy, identified cancers in older patients demonstrated higher risk histological phenotypes. Although nearly all malignancies in younger patients were well-differentiated, older patients were more likely to have higher risk papillary thyroid carcinoma variants, poorly differentiated cancer, or anaplastic carcinoma (P < .001).

Conclusion:

With advancing age, the prevalence of clinically relevant thyroid nodules increases, whereas the risk that such nodules are malignant decreases. Nonetheless, when thyroid cancer is detected in older individuals, a higher-risk histological phenotype is more likely. These data provide insight into the clinical paradox that confronts physicians managing this common illness.

Thyroid nodular disease is a common adult illness, with published prevalence ranging between 26 and 67% (1–8). Although most nodules are benign, 8–15% will prove malignant (9, 10). Therefore, accurate differentiation between benign and malignant nodules is important and has long been a focus of investigation. Preoperatively, no single clinical, radiological, or molecular finding has proven perfectly predictive of malignant disease. In lieu of this, clinicians collectively integrate multiple variables to improve cancer risk assessment (11). Although there is increasing reliance upon sonographic and cytological interpretation (9, 12), such interpretations also exhibit significant inter-rater variability (13, 14). Whereas inter-rater variability does not detract from their importance, such findings do support the integration of other secondary risk factors. One such risk factor is patient age, which in contrast is objective and not limited by such variability. Therefore, age has long been considered a potentially attractive addition to any risk stratification scheme.

Although epidemiological analyses have demonstrated a positive association between thyroid nodule formation and advancing patient age (1, 3–5), the specifics of this relationship are poorly defined. Does this risk increase in a linear fashion or demonstrate a threshold effect? And, perhaps most importantly, does age modify the risk that a newly detected thyroid nodule is malignant? This latter question has long been subject to debate. Some argue that age influences thyroid cancer risk in a bimodal fashion, with increasing cancer risk among nodules detected in both younger (<30 y) and older (>70 y) patients (15), but this idea has been challenged by recent investigation (16, 17). However, in these studies, limited follow-up, lack of routine sonographic evaluation, and sampling bias prevent a definitive conclusion.

Additionally, the type and extent of malignancy may differ in younger vs older populations and may influence diagnostic and therapeutic decisions. Although thyroid cancer prevalence may be lower in a given age group, the risk of patient morbidity and mortality may still prove greater if such malignancies are more aggressive. In support of this, data confirm an increased risk of thyroid cancer recurrence and disease-specific mortality in older (>45 y) individuals (18–20). Thus, to determine the impact of patient age on thyroid cancer risk, one needs to examine both cancer prevalence and the aggressivity of the malignancy.

Beginning in 1995, we initiated a prospective, structured data collection of all patients referred to the Brigham & Women's Hospital for thyroid nodule care. All patients were evaluated in a uniform fashion, integrating clinical, radiological, cytological, histopathological, and surgical evaluation as part of a team-based approach. Our database now depicts the presentation and care of over 6000 consecutive patients with more than 12 000 thyroid nodules. Together, this large, unbiased database provides the opportunity to clarify the relationship of patient age to thyroid nodular disease.

Patients and Methods

Data from all euthyroid adult patients (ages, 20–95 y) evaluated at the Brigham & Women's Hospital Thyroid Nodule Clinic between 1995 and 2011 were collected and analyzed as part of a prospective cohort analysis. Patients with a suspected thyroid nodule underwent sonographic evaluation and, when indicated, ultrasound-guided fine-needle aspiration (UG-FNA). Ultrasound examination was performed by a Brigham & Women's Hospital radiologist with specific thyroid expertise. For each nodule, the size (length, width, and depth) and cystic component (solid, 25–75% cystic, >75% cystic) were collected as reported at the time of the initial study. Subjects were classified as having a solitary nodule when only one nodule ≥1 cm was present or having a multinodular gland when two or more nodules each ≥1 cm were detected. For each subject, age at the time of first UG-FNA and the total number of nodules ≥1 cm were documented.

Consistent with clinical recommendations contemporary to the study period, UG-FNA was generally recommended for all noncystic nodules ≥1 cm. UG-FNA was performed with a 25-gauge needle after local anesthesia. Typically, two to four passes from different areas of the nodule constituted a single aspiration. Samples were processed by the ThinPrep 2000 system (Hologic Corp) and reviewed by a Brigham & Women's Hospital cytopathologist. All aspirates were classified following the Bethesda System for Reporting Thyroid Cytopathology (21).

All therapy and follow-up monitoring were standardized depending on cytological result. When FNA cytology was “positive for malignant cells/malignant” or “suspicious for carcinoma,” patients were typically advised to undergo near-total thyroidectomy. Usually, lobectomy was advised for patients with cytology “suspicious for a follicular neoplasm” or “suspicious for a Hurthle cell neoplasm.” If cytology demonstrated “atypia of undetermined significance,” clinical recommendations were for repeat FNA cytology evaluation, lobectomy (especially if the nodule was large or if high-risk sonographic features were noted), or more recently, repeat FNA with molecular testing using the Afirma gene expression classifier (n = 13). Finally, patients with “nondiagnostic” cytology were advised to undergo repeat UG-FNA within 6 months. For those patients with benign nodules, follow-up was generally recommended in 12–24 months, pending clinical and sonographic risk assessment.

For the purposes of investigating thyroid nodule formation and multinodularity among different age groups, we investigated all nodules ≥ 1 cm identified with ultrasound examination in the entire cohort of 6653 patients. To determine the risk of malignancy in different age groups, nodules were classified as malignant after histological interpretation, or rarely (∼3%) when cytology was “positive for malignant cells” but no subsequent surgery occurred because of unique clinical circumstances. Nodules were classified as benign when any of the following were confirmed: 1) benign UG-FNA cytology (68%); 2) benign histopathological interpretation (16%); 3) >75% cystic parenchyma (6%); 4) functionality on thyroid scintigraphy (“hot” nodule; 3%); or 5) if very low-risk sonographic characteristics (eg, spongiform appearance; 5%) or clinical characteristics (eg, >50% reduction in nodule volume during follow-up; 2%) were noted, as supported by recent American Thyroid Association guidelines (22). A very small proportion of patients (<4%) was lost to follow-up before final pathological diagnoses could be made, although initial sonographic data were available. Incidental intrathyroidal microcarcinomas (<1 cm) separate from the primary lesion commonly detected postoperatively upon histopathological review were not considered malignant for this investigation (23). Histopathology was classified using the Tumor Node Metastasis System recommended by the American Joint Committee on Cancer (24).

All subjects were grouped by age as follows: 20–29, 30–39, 40–49, 50–59, 60–69, or ≥70 years. For each age group, we then determined the mean number of thyroid nodules ≥1 cm per patient, the proportion of patients with multinodularity, the proportion of nodules proven cancerous, and the type of malignancy. For the purpose of further analysis, those with a diagnosis of malignancy were further categorized by their histological subtype as having either well-differentiated thyroid carcinoma (papillary or follicular carcinoma) or high-risk thyroid carcinoma (anaplastic, medullary, poorly differentiated, or distant metastatic carcinoma). Logistic regression analyses were used for statistical comparison and corrected by gender. Statistical analysis was performed using SPSS version 22 (SPSS IBM), and P values <.05 were considered significant. Approval from the Brigham & Women's Hospital Office of Human Subjects Research was granted to perform this investigation.

Results

From 1995 to 2011, there were 6653 consecutive euthyroid patients evaluated at the Brigham & Women's Hospital Thyroid Biopsy Clinic. Of these, 262 subjects were excluded due to the absence of a nodule ≥ 1 cm (n = 79) or patient age < 20 years (n = 183). This resulted in a final study population of 6391 patients with 12 115 nodules. As shown in Table 1, the population was predominantly female (85%), with a mean age of 47.4 years. Thyroid nodule size averaged 2.1 cm (±1.1 cm), and the majority of nodules were solid. Less than 1% of our study cohort had previous exposure to ionizing neck irradiation during childhood or harbored a family history of medullary thyroid carcinoma.

Table 1.

Baseline Patient and Nodule Characteristics

Patient characteristics (n = 6391)
Sex
Female	5447 (85)
Male	943 (15)
Age, y
Mean ± SD	47.4 ± 10.8
Range	20–95
Decile cohorts
20–29 y	420 (7)
30–39 y	908 (14)
40–49 y	1475 (23)
50–59 y	1606 (25)
60–70 y	1209 (19)
>70 y	773 (12)
Nodule characteristics (n = 12 115)
Maximal diameter, cm
Mean ± SD	2.1 ± 1.1
Range	1.0–12.0
Cystic content
<25% cystic	8975 (74)
25–75%	2007 (16)
>75%	1070 (9.5)
Unknown	63 (0.5)

Data are expressed as number (percentage) unless stated otherwise.

In this total population referred for initial evaluation of nodular disease, an average of 1.9 clinically relevant thyroid nodules was detected per patient. However, when stratified by age, the extent of nodularity increased in a linear fashion (Figure 1). Patients 20–29 years old had an average of 1.55 nodules ≥1 cm, whereas patients ≥70 years old had an average of 2.21 nodules ≥1 cm (P < .001), representing a 43% increase in the total number of thyroid nodules between these two groups (P < .001).

Figure 1.

Left axis depicts the mean number of thyroid nodules (>1 cm) in each age cohort.

Right axis depicts the proportion of patients with multinodularity (two or more nodules each >1 cm) in each age cohort.

The prevalence of multinodularity similarly increased with advancing age (Figure 1). With each successive age group, the relative risk for multinodularity increased by 1.6% (odds ratio [OR], 1.022; P < .001). The oldest cohort (>70 y) demonstrated a 30% higher absolute risk of multinodularity compared to the youngest cohort (56 vs 26%, respectively; P < .01).

We next investigated the association between patient age at nodule evaluation and malignancy risk. Of the 1018 (15.9%) patients diagnosed with thyroid cancer during this 16-year period, 988 (97%) were confirmed by surgical histology, whereas 30 (3%) were based on positive FNA cytology without thyroidectomy. Notably, the risk that a nodule was cancerous decreased with advancing age (P < .001), as shown in Figure 2A. For patients ages 20–29, 30–39, 40–49, 50–59, 60–69, and >70 years, the cancer prevalence was 22.9, 21.8, 17.1, 13.0, 13.7, and 12.6%, respectively (P < .001). For comparison, when the malignancy rate was analyzed “per-nodule,” the youngest cohort (20–29 y) demonstrated a 14.8% malignant risk per nodule at diagnosis in comparison to 5.6% in the oldest cohort (>70 y; P < .01). Between the ages of 20 and 60 years, each advancing year was associated with a 2.2% reduction in the relative risk that any newly evaluated thyroid nodule was malignant in a patient (OR, 0.972; P < .001). This risk of malignancy stabilized after age 60 years.

Figure 2.

A, Patient age and risk that a thyroid nodule > 1 cm will prove malignant when presenting for initial evaluation. B, Patient age and the proportion of identified malignancies representing high-risk, aggressive disease (anaplastic, medullary, poorly differentiated, or distant metastatic carcinoma).

To determine the type and extent of thyroid cancers in this study, all histopathological diagnoses were reviewed for those nodules that underwent resection (n = 988). We categorized all malignancies as either well-differentiated thyroid carcinoma (papillary or follicular carcinoma) or high-risk thyroid carcinoma (anaplastic, medullary, poorly differentiated, or distant metastatic carcinoma). Table 2 and Figure 2B show the distribution of well-differentiated vs high-risk carcinomas by age deciles. The proportion of thyroid carcinoma with high-risk histology increased with advancing age, ranging from 0% in the youngest cohort to 16% of carcinomas detected in the oldest cohort (P < .001). This increase was most pronounced in patients >40 years old, where a 7.0% increase in relative risk of high-risk carcinoma was detected annually (OR, 1.073; P < .001). Next, we analyzed all patients with papillary thyroid carcinoma (PTC) (n = 851), seeking to compare specific histological PTC variants at diagnosis. In this analysis, diffuse sclerosing PTC was disproportionately detected at higher rates in younger patients, with the majority under age 50 years at diagnosis. In contrast, higher-risk variants of PTC such as tall cell variant were more commonly detected in older patients (Table 3).

Table 3.

Variants of PTC in 851 Patients With Histologically Confirmed PTC

Age Cohort	PTC, n	Classical Variant, n (%)	Follicular Variant, n (%)	Tall Cell Variant, n (%)	Diffuse Sclerosing Variant, n (%)	Other Variants
						n (%)	Specifics
20–29 y	79	28 (35)	47 (59)	0 (0)	3 (4)	1 (1)	1 Cribriform
30–39 y	170	73 (43)	86 (51)	0 (0)	7 (4)	4 (2)	1 Columnar
							1 Macrofollicular
							2 Oncocytic
40–49 y	224	89 (40)	113 (50)	11 (5)	6 (3)	5 (2)	1 Cribriform
							2 Oncocytic
							1 Solid
							1 Warthin Tumor like
50–59 y	178	59 (33)	106 (59)	10 (6)	1 (1)	2 (1)	1 Macrofollicular
							1 Solid
60–69 y	128	39 (30)	71 (55)	8 (6)	0 (0)	10 (8)	1 Columnar
							1 Clear cell
							1 Macrofollicular
							5 Oncocytic
							2 Solid
>70 y	72	16 (22)	45 (63)	4 (6)	1 (1)	6 (8)	2 Macrofollicular
							4 Oncocytic

Table 2.

Thyroid Cancer Among 988 Patients With Histological Confirmations, Stratified by Age Cohort

Age Cohort	No. of Patients	Malignant, n (%)	Type of Malignancy
			Well-Differentiated Thyroid Carcinoma, n (%)a	High-Risk Thyroid Carcinoma, n (%)a
20–29 y	420	93 (22)	93 (100)	0 (0)
			■ 79 PTC	■ 0 Anaplastic/poorly diff
			■ 14 FTC	■ 0 MTC
				■ 0 DMC
30–39 y	908	191 (21)	187 (97.9)	4 (2.1)
			■ 170 PTC	■ 2 Anaplastic/poorly diff
			■ 17 FTC	■ 1 MTC
				■ 1 DMC
40–49 y	1475	246 (17)	240 (97.6)	6 (2.4)
			■ 224 PTC	■ 2 Anaplastic/poorly diff
			■ 16 FTC	■ 3 MTC
				■ 1 DMC
50–59 y	1606	204 (13)	195 (96.0)	9 (4.4)
			■ 178 PTC	■ 2 Anaplastic/poorly diff
			■ 17 FTC	■ 4 MTC
				■ 3 DMC
60–69 y	1209	163 (13)	142 (87.1)	21 (12.9)
			■ 128 PTC	■ 12 Anaplastic/poorly diff
			■ 14 FTC	■ 1 MTC
				■ 8 DMC
>70 y	773	94 (12)	79 (84.0)	15 (15.9)
			■ 72 PTC	■ 9 Anaplastic/poorly diff
			■ 7 FTC	■ 1 MTC
				■ 5 DMC

Abbreviations: PTC, papillary thyroid carcinoma; FTC, follicular thyroid carcinoma; Anaplastic/poorly diff, anaplastic or poorly differentiated thyroid cancer; MTC, medullary thyroid carcinoma; DMC, distant metastatic carcinoma. The proportion of well-differentiated thyroid carcinoma (papillary or follicular carcinoma) vs high-risk thyroid carcinoma (anaplastic, medullary, poorly differentiated, or distant metastatic carcinoma) is shown.

^a

P < .001 for trend.

Because a small minority of nodules were persistently nondiagnostic (<3%) or had indeterminate cytology without definitive histology (∼4%), we sought to determine whether these nodules were equally distributed among all age cohorts, thus avoiding any sampling bias. Such cases were widely distributed among all age cohorts, without a statistically significant association (P = .46).

Discussion

Until now, the influence of patient age on thyroid nodular disease, multinodularity, and thyroid cancer risk has remained poorly defined, with conflicting findings (15–17). Our data depict a large-scale, prospective analysis examining this relationship in unselected consecutive patients presenting for thyroid nodule evaluation in which standardized, high-frequency ultrasound evaluation and UG-FNA of nearly all relevant nodules were performed. Our findings confirm an increased prevalence of thyroid nodules and multinodularity with advancing age but, importantly, document a reduced risk that such nodules will prove cancerous with advancing age. However, thyroid cancer is nonetheless detectable in approximately 12–13% of older individuals and, when detected, is significantly more likely to be of higher risk disease. Anaplastic, poorly differentiated, medullary, and distant metastatic carcinoma were more likely in patients presenting with nodular disease over age 60 years. Even among those with well-differentiated PTC, higher-risk variants such as tall cell PTC were similarly more common in older patients. Together, our data emphasize a clinical paradox facing affected patients and their physicians. Although thyroid nodules are more common and more likely benign in older individuals (leading many to support a conservative strategy of noninvasive evaluation), this study also demonstrates that thyroid cancers in older patients are more likely to be aggressive, and thus early identification may prove critical for optimal outcome.

Our series investigated 6391 patients over a 16-year period in which nearly all aspirated nodules ≥1 cm were classified as benign or malignant. We acknowledge that a small proportion of nodules lacked precise cytological or histological results. In our cohort, a few were persistently nondiagnostic. Such nodules, however, were primarily cystic and have been followed over time without incident, leading us to conclude that there is a very high likelihood of benignity. A separate small cohort was cytologically indeterminate, although lost to follow-up or further intervention. Most such nodules were cytologically atypia of undetermined significance or follicular neoplasm. We cannot refute the argument that a small proportion of these nodules would prove malignant (10). However, subanalysis of this cohort demonstrates a similar distribution throughout all age cohorts, thus making it significantly unlikely to impact our conclusions. In contrast, we believe the broad inclusion of all consecutive thyroid nodules evaluated over 17 years in a real-world setting allows the most accurate translation of these data into daily practice. Our study population appears highly comparable to those previously reported (25–27) and confirms that most patients referred for nodular disease are female and middle-aged and typically present with a solid, approximately 2-cm nodule.

Importantly, these data do not describe the prevalence of thyroid nodularity in a general population because this investigation was not designed as a screening study. All patients in this study were referred to the clinic because of a known or suspected thyroid nodule. Given the observed increase in nodularity with age and the increasing use of cross-sectional studies in older patients, one might expect this cohort to represent the largest proportion of patients. Interestingly, patients over age 70 only represented 12% of the referral cohort. One explanation for this finding is that many older patients with thyroid nodules may simply have been followed conservatively (9) because the risk of any potential therapy may outweigh the benefits. However, if true, this practice would likely increase the proportion of thyroid cancer in the older cohort, suggesting that our findings may be even stronger than observed.

Why more thyroid nodules form but are less likely to be malignant with advancing age remains unclear. Oncogenic mutations have been shown to be causative for most thyroid carcinomas (28, 29). Although unproven, mutations are similarly believed to be responsible for benign nodular formation and growth (30, 31). Understanding which genetic mutations associate with benign vs malignant disease, while also better defining a potential multihit hypothesis or the impact of separate epigenetic factors, will be important areas for future research. It has long been known that older age portends a worse prognosis among those diagnosed with thyroid cancer (9, 18). Our data suggest that some of the attributable risk is due to more advanced thyroid cancer and higher risk histology at the time of diagnosis. Others argue that a poorer response to conventional therapy in older compared to younger patients with similar disease burden may also underlie this difference. Regardless, these findings support a complex genotype-phenotype relationship likely to be better deciphered in the decades ahead.

In summary, these data substantially expand our understanding of thyroid nodular disease in the adult population by showing that advancing age clearly increases the risk of thyroid nodule formation although it decreases the risk such identified nodules will be malignant. Ironically, these associations are inversely related. Given a cancer risk >20% in young adults with substantial longevity ahead, it is arguable that nearly all clinically relevant thyroid nodules (especially if solid) should be evaluated with UG-FNA in this population. In contrast, one potential strategy for older adults may include ultrasound evaluation with sonographic risk stratification, thereafter pursuing UG-FNA for nodules identified as sonographically intermediate or high risk. Importantly, older patients with nodular disease must be informed that higher-risk thyroid cancer is nonetheless possible, and continued monitoring of any nodularity is therefore necessary. As we seek to optimize the individualized care for affected patients, multivariable diagnostic algorithms should include patient age in addition to nodule size, sonographic characteristics, and other historical risk factors known to modify cancer risk.

Acknowledgments

The research was supported by National Institutes of Health Training Grant T32 DK007529.

Disclosure Summary: All authors report no disclosures relevant to this work.

Abbreviations

 
• FNA

  fine-needle aspiration

 
• OR

  odds ratio

 
• PTC

  papillary thyroid carcinoma

 
• UG-FNA

  ultrasound-guided FNA.