Abstract
The number of centers with established adrenal venous sampling (AVS) programs for the subtype diagnosis of primary aldosteronism (PA) is limited.
Aim was to develop an algorithm for AVS based on subtype prediction by computed tomography (CT) and serum potassium.
A multi-institutional retrospective cohort study in Japan.
A total of 1591 patients with PA were classified into four groups according to CT findings and potassium status. Subtype diagnosis of PA was determined by AVS.
Prediction value of the combination of CT findings and potassium status for subtype diagnosis.
The percentages of unilateral hyperaldosteronism on AVS were higher in patients with unilateral disease on CT than those with bilateral normal results on CT (50.8% vs 14.6%, P < 0.01), and these percentages were higher in those with hypokalemia than those with normokalemia (58.4% vs 11.5%, P < 0.01). The prevalence and odds ratio for unilateral hyperaldosteronism on AVS were as follows: bilateral normal on CT with normokalemia, 6.2% (reference); unilateral disease on CT with normokalemia, 23.8% and 4.8 [95% confidence interval (CI), 3.1 to 7.2]; bilateral normal on CT with hypokalemia, 38.1% and 9.4 (95% CI, 6.2 to 14.1), and unilateral disease on CT with hypokalemia, 70.6% and 36.4 (95% CI, 24.7 to 53.5).
Patients with PA with bilateral normal results on CT and normokalemia likely have a low prior probability of a lateralized form of AVS and could be treated medically, whereas those with unilateral disease on CT and hypokalemia have a high probability of a lateralized form of AVS.
Primary aldosteronism (PA) is the most common cause of secondary hypertension, accounting for 5% to 10% of hypertensive patients (1–3) and increasing various cardiometabolic complications associated with excess production of aldosterone (4, 5). Adrenal venous sampling (AVS) is the standard procedure for a subtype diagnosis of PA recommended by current guidelines in guiding appropriate treatment (6, 7). AVS presents various clinical issues for its wider application: a requirement of technical expertise, the invasive nature of the procedure, its high cost, and the fact that only a limited number of centers have successful AVS programs (8). Thus, it is desirable to optimize the indication for AVS by first predicting the subtype with an alternative procedure.
Adrenal computed tomography (CT) is widely available in most centers and is less expensive than AVS. The Endocrine Society Clinical Practice Guidelines recommend adrenal CT as the initial investigation for the subtype diagnosis of PA to exclude adrenal carcinomas (7). Although several studies have shown that the diagnostic accuracy of CT for subtype diagnosis is inadequate (9–12), CT findings with various biochemical markers, including serum potassium, plasma aldosterone concentration (PAC), and the results of confirmatory tests (13, 14), have been shown to be useful for predicting the unilateral subtype. However, the reference range of PAC varies greatly between centers, and the type of confirmatory test is not standardized (7, 15). By contrast, hypokalemia is the most characteristic clinical feature of aldosterone unilateral adenomas, and measurement of serum potassium is fully standardized.
The aim of the current study was to investigate the utility of a combination of CT findings and serum potassium, which are the most commonly available clinical investigations for the subtype diagnosis of PA determined by AVS, and to refine the algorithm for the indication for AVS with graded recommendation strength based on the probability of unilateral hyperaldosteronism on AVS determined by CT findings and serum potassium.
Methods
Study design and patients
The study was conducted as a part of the Japan Primary Aldosteronism Study (JPAS) study. The nationwide PA registry in Japan was established in 28 centers, including 15 university hospitals and 13 city hospitals. Patients with PA who were diagnosed and underwent AVS between January 2006 and October 2016 were enrolled (Supplemental Table 1). Patients eligible for JPAS were men and women aged 20 to 90 years. Patients who participated in the preceding WAVES-J (West Adrenal Vein Sampling in Japan) study (16) were included. The clinical characteristics, biochemical findings, results of confirmatory testing, imaging findings, AVS results, treatment, surgical findings, and related follow-up data were electronically collected using the WEB registry system. System construction, data security, and maintenance of the registered data were outsourced to EPS Corporation (Tokyo, Japan).
This retrospective study was analyzed using the data set valid in October 2016. Patients with available data on clinical characteristics, biochemistry, CT findings, and successful adrenocorticotropic hormone (ACTH)–stimulated AVS were included. Patients with suspected autonomous cortisol secretion defined by serum cortisol levels >3 µg/dL after 1 mg dexamethasone were also excluded (17).
Diagnosis of PA
The diagnosis of PA was made in accordance with guidelines from the Japan Endocrine Society (18) and the Japan Society of Hypertension (19). PA was diagnosed by positive case detection with a ratio of PAC (measured in ng/dL) to plasma renin activity (PRA) (measured in ng/mL/h) >20 and at least one positive result of confirmatory testing, including the captopril challenge test, the saline infusion test, the furosemide upright test, and the oral salt-loading test. Antihypertensive medications were usually changed to calcium channel blockers and/or α-adrenergic blockers, as appropriate, until the final diagnosis was made. Hypokalemia was considered present if serum potassium was <3.5 mEq/L before or at the diagnosis of PA or a patient was taking a potassium supplement. Oral potassium was supplemented if hypokalemia was present. Estimated glomerular filtration rate was obtained using the following equation: estimated glomerular filtration rate (mL/min/1.73 m2) = 194 × serum creatinine(−1.094) × age(−0.287) × 0.739 (if female) (20).
Adrenal lesion on CT
All patients underwent thin-slice (1- to 3-mm-thick) adrenal CT. The findings were evaluated by radiologists in each institution and classified into three categories: unilateral disease, bilateral normal, and bilateral disease. A nodule on CT was defined as a nodule >10 mm in diameter. Bilateral normal appearance was defined if the size of the nodule or thickness of adrenal gland was <10 mm on both sides. CT was conducted before AVS in routine clinical practice, and hence radiological evaluation of CT was done with a clinical diagnosis such as PA or suspected PA but in the absence of AVS data.
AVS and subtype diagnosis
The subtype diagnosis of PA was determined by the AVS, the procedures of which have been described (16). Adrenal blood samples were collected sequentially in 24 centers and 4 centers where simultaneous catheterization was conducted. ACTH 1-24 (cosyntropin) was administered by bolus injection or by bolus injection followed by continuous infusion or only continuous infusion throughout the procedure. Adrenal vein cannulation was defined as successful if the selectivity index was >5 (6). The selectivity index was defined as the ratio of cortisol concentration in the adrenal vein to that in the inferior vena cava. Unilateral hyperaldosteronism on AVS was defined if the lateralization index (LI) was >4 (6). The LI was calculated by dividing the aldosterone to cortisol ratio on the dominant side by that on the nondominant side. AVS data after ACTH stimulation were used for analysis of the current study.
Assay methods
PAC and PRA were measured by commercial kits. PAC was determined by radioimmunoassay (SPAC-S Aldosterone kits; Fuji Rebio, Tokyo, Japan) in all centers. The reference range of PAC in the supine position was 3.0 to 15.9 ng/dL. PRA was measured by radioimmunoassay or enzyme immunoassay. The reference range of PRA in the supine position was 0.3 to 2.9 ng/mL/h (PRA-FR RIA kits; Fuji Rebio) in 16 centers, 0.2 to 2.3 ng/mL/h (PRA EIA kits; Yamasa, Choshi, Japan) in 8 centers, and 0.2 to 2.7 ng/mL/h (PRA RIA kits; Yamasa) in 3 centers. Plasma active renin concentration (ARC) was measured by immunoradiometric assay (Renin IRMA-FR; Fuji Rebio) in one center. ARC was converted to PRA and divided by 5 according to Japan Endocrine Society guidelines (18). The reference range of ARC in the supine position was 2.5 to 21.4 pg/mL. ARC value was used for analysis by converting to PRA.
Measurements and statistical analysis
Clinical characteristics were compared among patients with unilateral disease on CT, bilateral normal on CT, and bilateral disease on CT. Concordant rate of CT findings with AVS result was investigated. CT findings (unilateral disease on CT and bilateral normal on CT), serum potassium status (hypokalemia, normokalemia), and their combination were investigated in the subtype diagnosis of PA determined by AVS. For prediction of subtype diagnosis, patients with bilateral disease on CT were excluded from the analysis because of an insufficient number for analysis.
Statistical analyses were performed with EZR statistical software (Saitama Medical Center, Jichi Medical University, Saitama, Japan) (21). Continuous variables were expressed as median and interquartile range or range. Categorical variables were expressed as number and percentage. Continuous variables were analyzed by the Mann-Whitney U test and categorical variables by the Fisher exact test. Analysis of group difference used Kruskal-Wallis and Fisher exact test with post hoc Bonferroni analysis. We did a logistic regression analysis to assess the probability of unilateral hyperaldosteronism on AVS. Patients with bilateral normal results on CT were the reference group, and normokalemia was the reference group for potassium status. An odds ratio >1 was designated to indicate an increased likelihood for unilateral hyperaldosteronism on AVS. Ordinal numbers 1 to 4 were assigned to the categories as follows: (1) unilateral disease on CT and hypokalemia; (2) unilateral disease on CT and normokalemia; (3) bilateral normal on CT and hypokalemia; and (4) bilateral normal on CT and normokalemia. Patients with bilateral normal results on CT and normokalemia were the reference group for combination of CT findings and potassium status. Adjustment was made by age (year, continuous) and sex. All tests were two-tailed, with differences reported as significant when P < 0.05.
The study was conducted according to the clinical studies published by the Ministry of Health and Labor in Japan and approved by the ethics committee of the National Hospital Organization Kyoto Medical Center as the project leader center and by the institutional ethics committees of the participating centers. This observational study was registered at UMIN ID 18756.
Results
Patients
Of 2131 patients with PA who had undergone both CT and AVS, we identified 1591 with successful AVS and data available on clinical characteristics, laboratory findings, and CT findings. Reasons for exclusion of 540 patients from the analysis were as follows: no data on AVS after cosyntropin stimulation (n = 157), incomplete data on AVS (n = 43), unsuccessful AVS (n = 209), lack of baseline clinical data (n = 23), and suspected autonomous cortisol secretion (n = 108).
Clinical characteristics of the patients are shown in Table 1. Median age was 53 years (male to female ratio, 762/829). Prevalence of hypokalemia, including patients on oral potassium supplementation, was 40.3% (637 of 1591 patients). Of the 637 patients, 614 (96.4%) were diagnosed as having hypokalemia based on the decreased potassium concentration (<3.5 mEq/L) before the diagnosis (n = 403) and at the time of diagnosis of PA (n = 211). Twenty-three patients (3.6%) were classified as the hypokalemic subgroup based on the potassium supplementation at the time of diagnosis, although the biochemical data of hypokalemia were not available. Statistically significant differences in several clinical parameters, including the proportion with hypokalemia and PAC, were present between patients with unilateral disease on CT and bilateral normal results on CT. Similar differences were also seen between patients with bilateral normal results on CT and bilateral disease on CT, whereas no significant differences in clinical parameters were seen between those with unilateral disease on CT and those with bilateral disease on CT (Table 1).
Comparison of Clinical Findings in Patients With Unilateral Disease or Bilateral Normal/Disease on CT
| Characteristic | Total (N = 1591) | Unilateral Disease on CT (n = 654) | Bilateral Normal on CT (n = 899) | Bilateral Disease on CT (n = 38) | Overall P Value | Pairwise Comparison (P Values) | ||
|---|---|---|---|---|---|---|---|---|
| Unilateral Disease vs Bilateral Normal | Unilateral Disease vs Bilateral Disease | Bilateral Normal vs Bilateral Disease | ||||||
| Age, y | 53 (45–62) | 54 (44–62) | 53 (45–61) | 58 (53–63) | 0.06 | NA | NA | NA |
| Female, n (%) | 829 (52.1) | 324 (49.5) | 492 (45.3) | 13 (34.2) | 0.01 | 0.14 | 0.21 | 0.06 |
| Body mass index | 24.4 (21.8–27.3) | 24.2 (21.5–27.0) | 24.6 (22.0–27.6) | 24.0 (22.2–27.0) | 0.06 | NA | NA | NA |
| Duration of hypertension | 5 (2–12) | 7 (2–14) | 4 (1–10) | 11 (5–19) | <0.01 | <0.01 | <0.01 | <0.01 |
| Systolic blood pressure, mm Hg | 140 (128–152) | 142 (131–150) | 140 (130–152) | 137 (128–155) | 0.31 | NA | NA | NA |
| Diastolic blood pressure, mm Hg | 86 (77–95) | 86 (76–93) | 87 (78–96) | 81 (77–91) | 0.03 | 0.09 | 0.88 | 0.25 |
| Number of antihypertensive medications | 1 (1–1) | 1 (1–2) | 1 (1–1) | 1 (1–2) | <0.01 | <0.01 | 1.0 | 0.1 |
| Serum potassium, mEq/L | 3.8 (3.4–4.0) | 3.6 (3.2–3.9) | 3.9 (3.6–4.1) | 3.5 (3.0–4.0) | <0.01 | <0.01 | 1.0 | <0.01 |
| Oral potassium supplement, n (%) | 327 (20.6) | 217 (33.2) | 97 (12.7) | 13 (34.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Hypokalemia, n (%) | 637 (40.3) | 377 (57.6) | 236 (10.8) | 24 (63.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Estimated glomerular filtration rate, mL/min/1.73 m2 | 78 (66–91) | 78 (66–91) | 77 (66–91) | 78 (67–98) | 0.40 | NA | NA | NA |
| PAC, ng/dL | 18.0 (12.7–28.0) | 22.1 (14.5–36.1) | 16.1 (12.0–22.5) | 24.5 (12.6–39.2) | <0.01 | <0.01 | 1.0 | 0.03 |
| PRA, ng/mL/h | 0.3 (0.2–0.5) | 0.3 (0.2–0.5) | 0.4 (0.2–0.6) | 0.3 (0.2–0.5) | <0.01 | <0.01 | 1.0 | 0.69 |
| PAC to PRA ratio, ng/dL per ng/mL/h | 52.0 (31.2–102.0) | 71.4 (35.8–157.2) | 43.3 (29.2–75.1) | 60.3 (31.1–171.3) | <0.01 | <0.01 | 1.0 | 0.11 |
| Characteristic | Total (N = 1591) | Unilateral Disease on CT (n = 654) | Bilateral Normal on CT (n = 899) | Bilateral Disease on CT (n = 38) | Overall P Value | Pairwise Comparison (P Values) | ||
|---|---|---|---|---|---|---|---|---|
| Unilateral Disease vs Bilateral Normal | Unilateral Disease vs Bilateral Disease | Bilateral Normal vs Bilateral Disease | ||||||
| Age, y | 53 (45–62) | 54 (44–62) | 53 (45–61) | 58 (53–63) | 0.06 | NA | NA | NA |
| Female, n (%) | 829 (52.1) | 324 (49.5) | 492 (45.3) | 13 (34.2) | 0.01 | 0.14 | 0.21 | 0.06 |
| Body mass index | 24.4 (21.8–27.3) | 24.2 (21.5–27.0) | 24.6 (22.0–27.6) | 24.0 (22.2–27.0) | 0.06 | NA | NA | NA |
| Duration of hypertension | 5 (2–12) | 7 (2–14) | 4 (1–10) | 11 (5–19) | <0.01 | <0.01 | <0.01 | <0.01 |
| Systolic blood pressure, mm Hg | 140 (128–152) | 142 (131–150) | 140 (130–152) | 137 (128–155) | 0.31 | NA | NA | NA |
| Diastolic blood pressure, mm Hg | 86 (77–95) | 86 (76–93) | 87 (78–96) | 81 (77–91) | 0.03 | 0.09 | 0.88 | 0.25 |
| Number of antihypertensive medications | 1 (1–1) | 1 (1–2) | 1 (1–1) | 1 (1–2) | <0.01 | <0.01 | 1.0 | 0.1 |
| Serum potassium, mEq/L | 3.8 (3.4–4.0) | 3.6 (3.2–3.9) | 3.9 (3.6–4.1) | 3.5 (3.0–4.0) | <0.01 | <0.01 | 1.0 | <0.01 |
| Oral potassium supplement, n (%) | 327 (20.6) | 217 (33.2) | 97 (12.7) | 13 (34.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Hypokalemia, n (%) | 637 (40.3) | 377 (57.6) | 236 (10.8) | 24 (63.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Estimated glomerular filtration rate, mL/min/1.73 m2 | 78 (66–91) | 78 (66–91) | 77 (66–91) | 78 (67–98) | 0.40 | NA | NA | NA |
| PAC, ng/dL | 18.0 (12.7–28.0) | 22.1 (14.5–36.1) | 16.1 (12.0–22.5) | 24.5 (12.6–39.2) | <0.01 | <0.01 | 1.0 | 0.03 |
| PRA, ng/mL/h | 0.3 (0.2–0.5) | 0.3 (0.2–0.5) | 0.4 (0.2–0.6) | 0.3 (0.2–0.5) | <0.01 | <0.01 | 1.0 | 0.69 |
| PAC to PRA ratio, ng/dL per ng/mL/h | 52.0 (31.2–102.0) | 71.4 (35.8–157.2) | 43.3 (29.2–75.1) | 60.3 (31.1–171.3) | <0.01 | <0.01 | 1.0 | 0.11 |
Data are presented as median (interquartile range) unless otherwise indicated.
Abbreviation: NA, not assessed.
Comparison of Clinical Findings in Patients With Unilateral Disease or Bilateral Normal/Disease on CT
| Characteristic | Total (N = 1591) | Unilateral Disease on CT (n = 654) | Bilateral Normal on CT (n = 899) | Bilateral Disease on CT (n = 38) | Overall P Value | Pairwise Comparison (P Values) | ||
|---|---|---|---|---|---|---|---|---|
| Unilateral Disease vs Bilateral Normal | Unilateral Disease vs Bilateral Disease | Bilateral Normal vs Bilateral Disease | ||||||
| Age, y | 53 (45–62) | 54 (44–62) | 53 (45–61) | 58 (53–63) | 0.06 | NA | NA | NA |
| Female, n (%) | 829 (52.1) | 324 (49.5) | 492 (45.3) | 13 (34.2) | 0.01 | 0.14 | 0.21 | 0.06 |
| Body mass index | 24.4 (21.8–27.3) | 24.2 (21.5–27.0) | 24.6 (22.0–27.6) | 24.0 (22.2–27.0) | 0.06 | NA | NA | NA |
| Duration of hypertension | 5 (2–12) | 7 (2–14) | 4 (1–10) | 11 (5–19) | <0.01 | <0.01 | <0.01 | <0.01 |
| Systolic blood pressure, mm Hg | 140 (128–152) | 142 (131–150) | 140 (130–152) | 137 (128–155) | 0.31 | NA | NA | NA |
| Diastolic blood pressure, mm Hg | 86 (77–95) | 86 (76–93) | 87 (78–96) | 81 (77–91) | 0.03 | 0.09 | 0.88 | 0.25 |
| Number of antihypertensive medications | 1 (1–1) | 1 (1–2) | 1 (1–1) | 1 (1–2) | <0.01 | <0.01 | 1.0 | 0.1 |
| Serum potassium, mEq/L | 3.8 (3.4–4.0) | 3.6 (3.2–3.9) | 3.9 (3.6–4.1) | 3.5 (3.0–4.0) | <0.01 | <0.01 | 1.0 | <0.01 |
| Oral potassium supplement, n (%) | 327 (20.6) | 217 (33.2) | 97 (12.7) | 13 (34.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Hypokalemia, n (%) | 637 (40.3) | 377 (57.6) | 236 (10.8) | 24 (63.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Estimated glomerular filtration rate, mL/min/1.73 m2 | 78 (66–91) | 78 (66–91) | 77 (66–91) | 78 (67–98) | 0.40 | NA | NA | NA |
| PAC, ng/dL | 18.0 (12.7–28.0) | 22.1 (14.5–36.1) | 16.1 (12.0–22.5) | 24.5 (12.6–39.2) | <0.01 | <0.01 | 1.0 | 0.03 |
| PRA, ng/mL/h | 0.3 (0.2–0.5) | 0.3 (0.2–0.5) | 0.4 (0.2–0.6) | 0.3 (0.2–0.5) | <0.01 | <0.01 | 1.0 | 0.69 |
| PAC to PRA ratio, ng/dL per ng/mL/h | 52.0 (31.2–102.0) | 71.4 (35.8–157.2) | 43.3 (29.2–75.1) | 60.3 (31.1–171.3) | <0.01 | <0.01 | 1.0 | 0.11 |
| Characteristic | Total (N = 1591) | Unilateral Disease on CT (n = 654) | Bilateral Normal on CT (n = 899) | Bilateral Disease on CT (n = 38) | Overall P Value | Pairwise Comparison (P Values) | ||
|---|---|---|---|---|---|---|---|---|
| Unilateral Disease vs Bilateral Normal | Unilateral Disease vs Bilateral Disease | Bilateral Normal vs Bilateral Disease | ||||||
| Age, y | 53 (45–62) | 54 (44–62) | 53 (45–61) | 58 (53–63) | 0.06 | NA | NA | NA |
| Female, n (%) | 829 (52.1) | 324 (49.5) | 492 (45.3) | 13 (34.2) | 0.01 | 0.14 | 0.21 | 0.06 |
| Body mass index | 24.4 (21.8–27.3) | 24.2 (21.5–27.0) | 24.6 (22.0–27.6) | 24.0 (22.2–27.0) | 0.06 | NA | NA | NA |
| Duration of hypertension | 5 (2–12) | 7 (2–14) | 4 (1–10) | 11 (5–19) | <0.01 | <0.01 | <0.01 | <0.01 |
| Systolic blood pressure, mm Hg | 140 (128–152) | 142 (131–150) | 140 (130–152) | 137 (128–155) | 0.31 | NA | NA | NA |
| Diastolic blood pressure, mm Hg | 86 (77–95) | 86 (76–93) | 87 (78–96) | 81 (77–91) | 0.03 | 0.09 | 0.88 | 0.25 |
| Number of antihypertensive medications | 1 (1–1) | 1 (1–2) | 1 (1–1) | 1 (1–2) | <0.01 | <0.01 | 1.0 | 0.1 |
| Serum potassium, mEq/L | 3.8 (3.4–4.0) | 3.6 (3.2–3.9) | 3.9 (3.6–4.1) | 3.5 (3.0–4.0) | <0.01 | <0.01 | 1.0 | <0.01 |
| Oral potassium supplement, n (%) | 327 (20.6) | 217 (33.2) | 97 (12.7) | 13 (34.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Hypokalemia, n (%) | 637 (40.3) | 377 (57.6) | 236 (10.8) | 24 (63.2) | <0.01 | <0.01 | 1.0 | <0.01 |
| Estimated glomerular filtration rate, mL/min/1.73 m2 | 78 (66–91) | 78 (66–91) | 77 (66–91) | 78 (67–98) | 0.40 | NA | NA | NA |
| PAC, ng/dL | 18.0 (12.7–28.0) | 22.1 (14.5–36.1) | 16.1 (12.0–22.5) | 24.5 (12.6–39.2) | <0.01 | <0.01 | 1.0 | 0.03 |
| PRA, ng/mL/h | 0.3 (0.2–0.5) | 0.3 (0.2–0.5) | 0.4 (0.2–0.6) | 0.3 (0.2–0.5) | <0.01 | <0.01 | 1.0 | 0.69 |
| PAC to PRA ratio, ng/dL per ng/mL/h | 52.0 (31.2–102.0) | 71.4 (35.8–157.2) | 43.3 (29.2–75.1) | 60.3 (31.1–171.3) | <0.01 | <0.01 | 1.0 | 0.11 |
Data are presented as median (interquartile range) unless otherwise indicated.
Abbreviation: NA, not assessed.
Diagnostic concordance between CT findings and AVS results
The prevalence of unilateral disease on CT was 41.1% (654/1591 patients) and that of bilateral normal results on CT was 56.5% (899/1591), and the overall prevalence of unilateral hyperaldosteronism on AVS was 30.3% (482/1591). The percentages of unilateral hyperaldosteronism on AVS were higher in patients with unilateral disease on CT than those with bilateral normal results on CT (50.8% vs 14.6%, P < 0.01) (Table 2). Although the CT findings were significantly associated with the subtype diagnosis of PA determined by AVS, the overall diagnostic concordant rate of CT findings was 68.1% (1084/1591) (Table 2). Thirty-five patients with unilateral disease on CT (5.4%, 35/654) showed unilateral hyperaldosteronism on AVS on the contralateral adrenal gland (Table 2).
Diagnostic Concordant Rate Between CT Findings and AVS (N = 1591)
| CT Finding | Subtype Diagnosis by AVS, n | Concordance of CT Findings, % (n/N) | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | ||
|---|---|---|---|---|---|
| Unilateral | Bilateral | ||||
| Right | Left | ||||
| Unilateral disease (n = 654) | |||||
| Right | 112 | 18 | 114 | 45.4 (297/654) | 50.8 (332/654) |
| Left | 17 | 185 | 208 | ||
| Bilateral (n = 937) | |||||
| Normal (n = 899) | 51 | 80 | 768 | 85.4 (768/899) | 14.6 (131/899) |
| Disease (n = 38) | 10 | 9 | 19 | 50.0 (19/38) | 50.0 (19/38) |
| CT Finding | Subtype Diagnosis by AVS, n | Concordance of CT Findings, % (n/N) | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | ||
|---|---|---|---|---|---|
| Unilateral | Bilateral | ||||
| Right | Left | ||||
| Unilateral disease (n = 654) | |||||
| Right | 112 | 18 | 114 | 45.4 (297/654) | 50.8 (332/654) |
| Left | 17 | 185 | 208 | ||
| Bilateral (n = 937) | |||||
| Normal (n = 899) | 51 | 80 | 768 | 85.4 (768/899) | 14.6 (131/899) |
| Disease (n = 38) | 10 | 9 | 19 | 50.0 (19/38) | 50.0 (19/38) |
Diagnostic Concordant Rate Between CT Findings and AVS (N = 1591)
| CT Finding | Subtype Diagnosis by AVS, n | Concordance of CT Findings, % (n/N) | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | ||
|---|---|---|---|---|---|
| Unilateral | Bilateral | ||||
| Right | Left | ||||
| Unilateral disease (n = 654) | |||||
| Right | 112 | 18 | 114 | 45.4 (297/654) | 50.8 (332/654) |
| Left | 17 | 185 | 208 | ||
| Bilateral (n = 937) | |||||
| Normal (n = 899) | 51 | 80 | 768 | 85.4 (768/899) | 14.6 (131/899) |
| Disease (n = 38) | 10 | 9 | 19 | 50.0 (19/38) | 50.0 (19/38) |
| CT Finding | Subtype Diagnosis by AVS, n | Concordance of CT Findings, % (n/N) | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | ||
|---|---|---|---|---|---|
| Unilateral | Bilateral | ||||
| Right | Left | ||||
| Unilateral disease (n = 654) | |||||
| Right | 112 | 18 | 114 | 45.4 (297/654) | 50.8 (332/654) |
| Left | 17 | 185 | 208 | ||
| Bilateral (n = 937) | |||||
| Normal (n = 899) | 51 | 80 | 768 | 85.4 (768/899) | 14.6 (131/899) |
| Disease (n = 38) | 10 | 9 | 19 | 50.0 (19/38) | 50.0 (19/38) |
Potassium status and prevalence of unilateral hyperaldosteronism on AVS
The prevalence of unilateral hyperaldosteronism on AVS was 58.4% in patients with hypokalemia and 11.5% in those with normokalemia. On the contrary, the prevalence of hypokalemia was 77.1% in patients with unilateral hyperaldosteronism on AVS and 23.9% in patients with the bilateral hyperaldosteronism on AVS. Potassium status was associated with the subtype diagnosis of PA determined by AVS (P < 0.01) (Table 3).
Subtype Diagnosis by AVS in Patients With Different Potassium Status (N = 1591)
| Characteristic | Subtype Diagnosis by AVS | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | |
|---|---|---|---|
| Unilateral | Bilateral | ||
| Potassium status | |||
| Hypokalemia | 372 | 265 | 58.4 (372/637) |
| Normokalemia | 110 | 844 | 11.5 (110/954) |
| Prevalence of hypokalemia, % (n/N) | 77.1 (372/482) | 23.9 (265/1109) | |
| Characteristic | Subtype Diagnosis by AVS | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | |
|---|---|---|---|
| Unilateral | Bilateral | ||
| Potassium status | |||
| Hypokalemia | 372 | 265 | 58.4 (372/637) |
| Normokalemia | 110 | 844 | 11.5 (110/954) |
| Prevalence of hypokalemia, % (n/N) | 77.1 (372/482) | 23.9 (265/1109) | |
Subtype Diagnosis by AVS in Patients With Different Potassium Status (N = 1591)
| Characteristic | Subtype Diagnosis by AVS | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | |
|---|---|---|---|
| Unilateral | Bilateral | ||
| Potassium status | |||
| Hypokalemia | 372 | 265 | 58.4 (372/637) |
| Normokalemia | 110 | 844 | 11.5 (110/954) |
| Prevalence of hypokalemia, % (n/N) | 77.1 (372/482) | 23.9 (265/1109) | |
| Characteristic | Subtype Diagnosis by AVS | Prevalence of Unilateral Hyperaldosteronism on AVS, % (n/N) | |
|---|---|---|---|
| Unilateral | Bilateral | ||
| Potassium status | |||
| Hypokalemia | 372 | 265 | 58.4 (372/637) |
| Normokalemia | 110 | 844 | 11.5 (110/954) |
| Prevalence of hypokalemia, % (n/N) | 77.1 (372/482) | 23.9 (265/1109) | |
Prevalence of unilateral hyperaldosteronism on AVS of PA predicted by combination of adrenal CT findings and potassium status
The prevalence of unilateral hyperaldosteronism on AVS of PA was investigated by combining the CT findings and serum potassium status (Fig. 1). Patients with bilateral disease on CT were excluded from the analysis because of the small number (n = 38). The prevalence of unilateral hyperaldosteronism on AVS in patients with unilateral disease on CT and hypokalemia was 70.6% (266/377 patients), and that in patients with normokalemia was 23.8% (66/277). The prevalence of unilateral hyperaldosteronism on AVS in patients with bilateral normal results on CT with hypokalemia was 38.1% (90/236 patients) and that in those with normokalemia was 6.2% (41/663 patients). The odds ratio for unilateral hyperaldosteronism on AVS was evaluated by categorizing the combination of CT findings and serum potassium status, and patients with bilateral normal results on CT and normokalemia were defined as the reference group. An odds ratio for unilateral hyperaldosteronism on AVS was 36.4 [95% confidence interval (CI), 24.7 to 53.5] in patients with unilateral disease on CT and hypokalemia, 9.4 (95% CI, 6.2 to 14.1) in patients with bilateral normal results on CT with hypokalemia, and 4.8 (95% CI, 3.1 to 7.2) in patients with unilateral disease on CT and normokalemia (Fig. 2). This result was consistent after adjustment of age and sex. In those with bilateral normal results on CT and normokalemia, there was a significant difference between those with unilateral disease on AVS and those with bilateral disease, but these values overlapped (Supplemental Table 2), whereas it is noteworthy that the minimum PAC was 8 ng/dL in this subgroup with unilateral disease on AVS.
Prevalence of unilateral hyperaldosteronism on AVS and proposed recommendation strength of AVS indication in the subgroup categorized by CT findings and serum potassium status.
The odds ratio for unilateral hyperaldosteronism on AVS by categorizing the combination of CT findings and serum potassium status.
Discussion
Our results have shown that the combination of two readily available clinical investigations, adrenal CT and serum potassium, could be of great value for predicting the subtype of PA in an uncomplicated way and for adding graded strength to the recommendation for the indication of AVS in patients with PA: strong recommendation in patients with adrenal tumors on CT and hypokalemia and weak recommendation in patients with no tumors on CT and normokalemia. Given that the number of centers with established technical expertise is very limited all over the world and specific treatment of hyperaldosteronism by mineralocorticoid antagonist is available, it can be justified to refine the indication of AVS based on the predicted probability of the subtype.
The current study found that the overall concordant rate between CT findings and AVS was 68.1%. This agrees with the previous result of 62.2% reported in a systematic review (9), in which the concordance between CT/magnetic resonance imaging and AVS was studied. However, there was a difference between the current study and the previous report (9) in the diagnostic concordance for unilateral disease on CT (45.4% vs 66.9%) and bilateral normal results on CT (84% vs 56.6%). This could be attributable to the different definitions of abnormal findings on CT as well as to differences in cutoff and methodology of AVS across the centers (6, 22, 23).
Whatever the diagnostic accuracy of the CT findings, the discordance of subtype diagnosis between the CT findings and the true subtype by AVS merits serious consideration when deciding for or against adrenal surgery. Basing the choice of treatment on the presence of unilateral disease on CT could lead to inappropriate unilateral adrenalectomy in 324 of 654 (49.5%) patients, whereas basing it on bilateral findings on CT would miss the possibility of a cure by adrenalectomy in 150 of 937 (16.0%) patients (Table 2). Overall, the frequency of selecting an inappropriate treatment was 31.8% (1084 of 1591) when based on CT findings. It should be noted that the prevalence of unilateral hyperaldosteronism on AVS, corresponding to an inappropriate choice of treatment, was only 14.6% in patients with bilateral normal results on CT.
Spontaneous hypokalemia is one of the classic features of PA (24). Nevertheless, the prevalence of hypokalemia has been shown to be ∼50% in patients with unilateral hyperaldosteronism and 15% in those with bilateral hyperaldosteronism (2). Although there was a close relationship between potassium status and the subtype diagnosis on AVS (Table 3), potassium status itself did not precisely distinguish the unilateral subtype from the bilateral subtype of PA when determined by AVS.
We then combined the CT findings and serum potassium status to further refine the diagnostic utility of these commonly available clinical investigations in predicting the subtype of PA. The prevalence of unilateral hyperaldosteronism on AVS was highest in patients with unilateral disease on CT and hypokalemia followed by bilateral normal results on CT with hypokalemia and unilateral disease on CT with normokalemia (Fig. 2), and these tendencies were consistent after adjustment of age and sex. The prevalence of unilateral hyperaldosteronism on AVS was <10% in patients with bilateral normal results on CT and normokalemia. Because clinical practice guidelines recommend AVS as the crucial step in subtype diagnosis prior to adrenalectomy, AVS is ideally performed when a patient desires surgery. In support of this, numerous observational studies have shown that treatment based on AVS leads to excellent clinical outcomes (11). The combination of CT findings and potassium status does not replace the essential role of AVS in subtype diagnosis, but it could provide evidence for a graded recommendation for it according to the prevalence of unilateral hyperaldosteronism on AVS (Fig. 1). A graded recommendation is of help in providing objective evidence for indicating AVS in patients with PA with heterogeneous subtypes and in avoiding unnecessary physical and economic burdens to the patients associated with the routine use of AVS. Our simple approach, combining two commonly available and standardized parameters, serum potassium concentration and CT findings, could be generally applicable in most centers involved in the clinical management of PA.
Whether AVS could be spared is one of the important issues for clinical practice in PA. The Endocrine Society Clinical Practice Guideline (7) suggests that young patients with spontaneous hypokalemia, marked aldosterone excess, and unilateral adrenal lesions with radiological features consistent with CT adenoma may not need AVS before proceeding to surgery. In our data set, it was demonstrated that all patients younger than 35 years with unilateral disease on CT and hypokalemia showed unilateral hyperaldosteronism on AVS and/or postoperative cure or improvement of PA, validating the suggestion by the Endocrine Society guideline (7) (data not shown). AVS also could be spared in patients with normal results on CT and normokalemia when PAC is <8 ng/dL because all patients showed bilateral aldosteronism on AVS (Supplemental Table 2).
Limitations
The main limitation of our study is that it was retrospective. A prospective study is needed to validate the present approach. However, our study had a multicenter design and a greater number of patients than any previous study. A second limitation is the objectivity of the CT findings. Diagnosis of an adrenal lesion on CT was based on the reports of the radiologists in each center, which may have resulted in between-center variation of the measurements. In addition, 139 of 899 patients with bilateral normal results on CT showed 8- to 9-mm diameter enlargement on CT. However, we defined an adrenal nodule as a nodule >10 mm in diameter, a size readily detectable. In addition, we tried to confirm interobserver agreement in reading the CTs by using the patients’ data of the first author’s institution (Kyoto Medical Center) by blindly and randomly selecting 50 cases (data not shown). The concordant rate between observers (first author vs radiologist finding) was nearly 90%. We considered that this cutoff (10 mm) was applicable for clinical practice.
Third, subtype diagnosis of PA in the current study was based on the AVS results. Although AVS is the referral standard for subtype diagnosis as recommended by clinical practice guidelines (7), its sensitivity and specificity for diagnosis are not 100%, and false-positive and false-negative results are to be expected. Therefore, the present results should be validated from the postoperative outcome, although it could only be achieved by the next data set with more sufficient follow-up data. In addition, postoperative histology was not available in the data set. A lack of conclusive subtype diagnosis after adrenalectomy is a major limitation of the study. We have used a LI of >4 as the criterion for unilateral hyperaldosteronism on AVS. Although that criterion has been the most commonly used and specific cutoff, lateralization criteria have not been fully standardized (25, 26). In this context, some of the patients classified as having bilateral hyperaldosteronism on AVS in the current study could have unilateral PA. Some of the unilateral aldosterone-producing adenomas have been demonstrated to show LI of <4 (23, 27) and therefore lead to selection of patients with the most florid aldosterone-producing adenoma phenotype, thus neglecting those with a milder phenotype. This could be one of the selection biases in the present study.
Fourth, our patients underwent AVS with cosyntropin stimulation. The use of cosyntropin during AVS has not been established in the subtype diagnosis of PA, as shown in the Adrenal Vein Sampling International Study (28), in which half of (11 of 20) the participating centers used cosyntropin stimulation in AVS. Although it has an advantage in assessing the selective catheterization, it could exert a confounding action such as a loss of lateralization in some cases (29). Thus, our results may not be adopted in patients without cosyntropin stimulation during AVS.
Fifth, median age of studied patients was relatively high, probably because of the previous recommendation of Japan Endocrine Society guidelines to perform screening testing in all hypertensives. On the contrary, a higher prevalence of hypokalemia could be mainly attributed to the selection bias in indicating AVS. The bias may affect the results when adopting real clinical practice. In addition, catheterization of the adrenal vein during AVS was different in each center (sequentially, 24 centers; simultaneously, 4 centers), and slice of CT was different at each center, leading to possibilities of bias. These methodological variations could be a bias affecting the results and a limitation of the retrospective study in adopting evidence in real clinical practice.
Sixth, at least one positive result of the confirmatory tests is now a common requirement for the diagnosis of PA (7, 19). Although the fludrocortisone loading test has been described as the most reliable confirmatory test, it has not been widely used and not included in the guidelines in Japan. In addition, it remains not established which confirmatory test is superior to others in its sensitivity and specificity. Therefore, whenever there was a conflicting result in the confirmatory tests, at least one positive result of any test was considered evidence for the diagnosis of PA. However, the influence of the difference of the confirmatory test could be a bias of our study. Finally, we could not confirm the biochemical evidence of hypokalemia in these 23 patients because of the retrospective nature of the study. However, it is quite uncommon and not approved to prescribe potassium supplements without definite evidence of hypokalemia in Japan. In addition, the prevalence of unilateral hyperaldosteronism did not show a significant difference after deleting those 23 patients from the analysis: 58.4% (n = 637) vs 58.9% (n = 614). It is therefore unlikely that the small number of patients (n = 23) affects the results of the current study.
Conclusions
We investigated the diagnostic utility of a combination of CT findings and serum potassium in the subtype diagnosis on AVS in 1591 patients with PA. Our results indicated that patients with bilateral normal results on CT and normokalemia could be treated medically because of a low probability of unilateral disease on AVS, whereas AVS was strongly indicated in patients with unilateral disease on CT because of a high probability of unilateral disease on AVS.
Abbreviations:
- ACTH
adrenocorticotropic hormone
- ARC
active renin concentration
- AVS
adrenal venous sampling
- CI
confidence interval
- CT
computed tomography
- JPAS
Japan Primary Aldosteronism Study
- LI
lateralization index
- PA
primary aldosteronism
- PAC
plasma aldosterone concentration
- PRA
plasma renin activity.
Acknowledgments
We thank Dr. Shintarou Okamura at Tenri Hospital, Dr. Shozo Miyauchi at Uwajima City Hospital, Dr. Tomikazu Fukuoka at Matsuyama Red Cross Hospital, Izawa Shoichiro at Tottori University, Takashi Yoneda at Misato Kenwa Hospital, Hashimoto Shigeatsu at Fukushima University, Masanobu Yamada at Gunma University, Tatsuya Kai at Saiseikai Tondabayashi Hospital, and Ryuichi Sakamoto at Saiseikai Fukuoka Hospital for collecting clinical data.
Financial Support: This study was supported in part by grants-in-aid for the study of PA in Japan (JPAS) from the Practical Research Project for Rare/Intractable Disease from the Japan Agency for Medical Research and Development (AMED) (15Aek0109122) and a grant from the National Center for Global Health and Medicine, Japan (24-1402).
Disclosure Summary: The authors have nothing to disclose.
