Association of Adrenal Venous Sampling With Outcomes in Primary Aldosteronism for Unilateral Adenomas

Key Points

Question  Are outcomes improved for patients with primary aldosteronism who undergo adrenal venous sampling prior to adrenalectomy in the setting of a clearly visualized unilateral adenoma and a normal contralateral gland?

Findings  In this multi-institutional study of 125 patients, there was no difference in complete biochemical success (75.9% vs 81.3%) or complete clinical success (43.6% vs 42.2%) based on whether adrenal venous sampling was performed preoperatively, nor was there a significant difference in partial biochemical or clinical success.

Meaning  In the appropriate clinical setting, adrenal venous sampling may not be required for preoperative confirmation of diagnosis of aldosterone-producing adenoma.

Importance  Adrenal venous sampling is recommended prior to adrenalectomy for all patients with hyperaldosteronism; however, cross-sectional imaging resolution continues to improve, while the procedure remains invasive and technically difficult. Therefore, certain patients may benefit from advancing straight to surgery.

Objective  To determine whether clinical and biochemical resolution varied for patients with primary aldosteronism with unilateral adenomas who underwent adrenal venous sampling vs those who proceeded to surgery based on imaging alone.

Design, Setting, and Participants  Retrospective, international cohort study of patients treated at 3 tertiary medical centers from 2004 to 2019, with a median follow-up of approximately 6 months. A total of 217 patients were consecutively enrolled. Exclusion criteria consisted of unknown postoperative serum aldosterone level and imaging inconsistent with unilateral adenoma with a normal contralateral gland. A total of 125 patients were included in the analysis. Data were analyzed between October 2019 and July 2020.

Exposures  Adrenal venous sampling performed preoperatively.

Main Outcomes and Measures  The primary outcome measurements were the clinical and biochemical success rates of surgery for the cure of hyperaldosteronism secondary to aldosterone-producing adenoma.

Results  A total of 125 patients were included (45 cross-sectional imaging with adrenal venous sampling and 80 imaging only). The mean (SD) age of the study participants was 50.2 (10.6) years and the cohort was 42.4% female (n = 53). Of those patients for whom race or ethnicity were reported (n = 80), most were White (72.5%). Adrenal venous sampling failure rate was 16.7%, and the imaging concordance rate was 100%. Relevant preoperative variables were similar between groups, except ambulatory systolic blood pressure, which was higher in the imaging-only group (150 mm Hg; interquartile range [IQR], 140-172 mm Hg vs 143 mm Hg, IQR, 130-158 mm Hg; P = .03). Resolution of autonomous aldosterone secretion was attained in 98.8% of imaging-only patients and 95.6% of adrenal venous sampling patients (P = .26). There was no difference in complete clinical success (43.6% [n = 34] vs 42.2% [n = 19]) or partial clinical success (47.4% [n = 37] vs 51.1% [n = 23]; P = .87) between groups. Complete biochemical resolution was similar as well (75.9% [n = 41] vs 84.4% [n = 27]; P = .35). There was no difference in clinical or biochemical cure rates when stratified by age, although complete clinical success rates downtrended in the older cohorts, and sample sizes were small.

Conclusions and Relevance  Given the improved sensitivity of cross-sectional imaging in detection of adrenal tumors, adrenal venous sampling may be selectively performed in appropriate patients with clearly visualized unilateral adenomas without affecting outcomes. This may facilitate increased access to surgical cure for aldosterone-producing adenomas and will decrease the incidence of morbidities associated with the procedure.

Primary aldosteronism (PA) is characterized by hypertension secondary to the overproduction of aldosterone and is considered to be the most common cause of secondary hypertension.^1,2 Studies estimate the prevalence of primary aldosteronism to be more than 5% of all patients with hypertension, approximately 27% of whom will eventually be diagnosed as having aldosterone-producing adenoma (APA).^2,3 Bilateral adrenal hyperplasia (BAH) and APA are 2 subtypes of PA, between which it is important to differentiate because therapeutic management varies considerably for the 2 conditions. Patients diagnosed as having APA are recommended to undergo laparoscopic adrenalectomy, while those with BAH are treated medically with mineralocorticoid receptor antagonists.⁴ Proper treatment is essential because PA has been found to impair renal function by depressing the glomerular filtration rate.⁵ Arterial hypertension associated with chronic aldosterone excess damages the vasculature to a greater extent than comparable hypertension with normal aldosterone levels, the consequence of which is higher cardiovascular risk in this patient population.⁶ Lastly, PA is associated with left ventricular hypertrophy that is out of proportion to the elevation in blood pressure.⁷ Because many of these complications have been shown to be at least partially reversible with timely management,^5,7 identification of disease subtype and appropriate therapeutic intervention are paramount.

It is recommended by the Endocrine Society that when surgical treatment is feasible, adrenal venous sampling (AVS) should be performed to definitively distinguish between BAH and APA.⁴ This guideline largely rests on a systematic review published in 2009⁸ that asserted a discordance rate of 37.8% when computed tomography (CT)/magnetic resonance imaging (MRI) findings were compared with AVS results. The only exception in the guidelines is for young patients, defined as younger than 35 years, with spontaneous hypokalemia, marked aldosterone excess, and unilateral adrenal lesions consistent with adenoma on adrenal protocol CT scan.⁴ However, AVS is invasive, technically difficult, requires access to a specialty center, and success rates vary widely. Complications occur in approximately 2.5% of cases and can include significantly morbid conditions such as adrenal vein dissection and adrenal hemorrhage.^9,10 Laparoscopic adrenalectomies have been shown to have the same operative risk as laparoscopic cholecystectomies.¹¹ In light of these data, other experts have suggested a more practical approach of selective AVS that accounts for patient preference, age, comorbidities, and the probability of finding an APA based on clinical presentation.¹²

The benefits of AVS need to be further evaluated to support routine performance. Given the high prevalence of PA, the potentially severe consequences of prolonged hyperaldosteronism, and the morbidity of AVS in patients referred for possible cure via adrenalectomy, this study aimed to assess whether outcomes differed between patients with APA who underwent adrenalectomies at 3 tertiary academic centers based on cross-sectional imaging only and those who underwent AVS prior to surgery. Biochemical cure rates as well as clinical hypertension resolution rates were analyzed.

Adult patients with adrenalectomy with APAs treated between October 2004 and September 2019 at 1 of 3 tertiary academic centers, Weill Cornell Medical College, New York, New York; Nancy University Hospital, Nancy, France; or Nantes University Hospital, Nantes, France, were enrolled in a clinical database that was retrospectively reviewed. Primary aldosteronism was defined as a history of persistent, difficult-to-control hypertension, with biochemical evidence of aldosteronism, with or without hypokalemia. Biochemical characteristics of PA included a plasma aldosterone-to-renin ratio greater than 20, with suppressed plasma renin activity less than 1 ng/mL/h, and a plasma aldosterone concentration of at least 15 ng/dL (to convert to picomoles per liter, multiply by 27.74). The diagnosis of APA vs BAH was based on the presence of a unilateral adrenal adenoma with a normal contralateral gland on CT scan or MRI, with or without confirmatory AVS. Computed tomography of the abdomen with and without contrast per institutional adrenal protocol with 2.5-mm slice thickness, or MRI abdomen with and without contrast was performed for each patient. All images were reviewed and diagnoses were confirmed by the operating surgeon. Only patients with completely normal contralateral glands on preoperative cross-sectional imaging and postoperative biochemical markers were included in the analysis. This excluded 92 patients from the study.

The institutional review boards at all institutions approved this study. All patients provided written informed consent for medical record review. Nancy and Nantes University Hospital patients participate in the clinical trial NCT03410394, which is the Registry of Endocrine Tumors (thyroid, parathyroid, adrenal, endocrine pancreas, and endocrine digestive tube). A patient data-sharing agreement was approved between institutions. Protected health information confidentiality was strictly preserved in accordance with hospital policy and the Health Insurance Portability and Accountability Act.

Pertinent preoperative and postoperative clinical, demographic, imaging, and laboratory variables were collected. Selectivity and lateralization indices were calculated for patients who underwent AVS. Adrenal venous sampling was considered successful when the cortisol in the adrenal sample was at least twice that in the inferior vena cava sample bilaterally, as per established guidelines. Three patients were excluded from the calculation of the AVS failure rate because the cortisol and aldosterone values obtained during the procedure were not explicitly documented in the medical record. Small tumors were defined as less than 10 mm in diameter. In accordance with the PASO international consensus definitions, hypokalemia was defined as serum potassium levels less than 3.6 mEq/L (to convert to millimoles per liter, multiply by 1), normalized postoperative aldosterone-to-renin ratio (ARR) of 30 or less, normal ambulatory systolic blood pressure (SBP) as less than 140 mm Hg, and normal ambulatory diastolic blood pressure (DBP) as less than 90 mm Hg.¹³ Additionally, complete clinical success was defined as normal blood pressure (<140/90 mm Hg) without the aid of antihypertensive medication, partial clinical success was defined as unchanged blood pressure with less antihypertensive medication or improved blood pressure with the same amount or less antihypertensive medication, and absent clinical success as unchanged or increased blood pressure with unchanged or increased amount of medication.¹³ The amount of medication is defined as the absolute number of antihypertensive medications and not by defined daily dose (DDD). Complete biochemical success was defined as normokalemia and normalization of the ARR, while partial success is normokalemia with a raised ARR but with an at least 50% decrease in baseline plasma aldosterone concentration.¹³

The cohort was divided into 2 groups for analysis: cross-sectional imaging only for diagnosis (imaging only) and cross-sectional imaging with AVS for confirmation. Descriptive statistics were used to analyze and summarize the 2 subgroups. Pearson χ² test was performed to compare categorical variables. The t test was applied when interval or continuous data was normally distributed, whereas Wilcoxon rank sum test was used for nonnormally distributed data. Statistical significance was evaluated at the .05 α level. Hypothesis tests were 2-sided. All statistical analyses were performed using Stata software, version 15.1 (StataCorp).

A total of 125 patients were included. There were 80 patients who underwent cross-sectional imaging only for diagnosis of APA (imaging only), while 45 patients underwent cross-sectional imaging and AVS for confirmation. Relevant preoperative clinical and demographic variables were similar between groups (Table 1). Ambulatory SBP was higher in the imaging-only group (150 mm Hg; interquartile range [IQR], 140-172 mm Hg vs 143 mm Hg, IQR, 130-158 mm Hg; P = .03), and Asian individuals were more likely to be referred for AVS (0 vs 9; P<.001). Median plasma aldosterone level was also higher in the AVS group (30; IQR, 18-46 vs 41; IQR, 22-78; P = .02), but there was no difference in median ARR (131; IQR, 65-204 vs 93; IQR, 37-244; P = .49) (Table 1). There were 6 patients with small tumors (<10 mm) in each group. Laterality of the adenoma did not differ between groups, with 63.8% of imaging-only patients (n = 51) and 61.4% of AVS patients (n = 27) undergoing left adrenalectomy (P = .79). The AVS failure rate was 16.7%. The median lateralization index was 21.4 (IQR, 7.3-36.2). The imaging concordance rate with AVS was 100%.

The clinical and biochemical outcomes after adrenalectomy did not vary between the imaging-only and the AVS groups (Table 2). Median follow-up overall was 166.5 days (IQR, 52-627) or 5.5 months. Forty-nine percent of patients had at least 6-month follow-up data. With respect to postoperative ARR, 86.2% of patients (n = 50) diagnosed as having APA based on imaging only normalized their ratios, and 87.5% of those with AVS for confirmation of APA (n = 28) had a normal postoperative ARR (P = .86). Resolution of autonomous aldosterone secretion was achieved in 98.8% of imaging-only patients (n = 79) and 95.6% of AVS patients (n = 43) (P = .26). Complete biochemical resolution was attained in 75.9% of imaging-only patients (n = 41), which was not different from the 84.4% of AVS patients (n = 27) (P = .35). The difference between the rate of normalization of ARR and complete biochemical success was a consequence of persistent postoperative hypokalemia, which was seen in 10% of the imaging-only patients (n = 8) and 11.1% of the AVS patients (n = 5) (P = .85). However, the patients with persistent hypokalemia had significantly shorter follow-up times compared with the rest of the cohort (P = .02), with a median follow-up time of 56 days (IQR, 49-95). Regarding clinical resolution of hypertension after adrenalectomy, 91% of imaging-only patients (n = 71) had at least partial success, which was similar to the AVS group at 93.3% (n = 42) (P = .87).

The cohort (n = 125) was then stratified into 3 groups to determine whether hypertension resolution rates after adrenalectomy for APA varied with age: younger than 35 years, between ages 35 and 65 years, and older than 65 years. As patient age increased, the percentage of patients with complete clinical success after adrenalectomy trended down (Figure 1A). However, most patients experienced at least partial clinical success even in the older than 65 years cohort (imaging only 90% [n = 9] vs AVS 80% [n = 4]; P = .27). There was no significant benefit to undergoing AVS with respect to clinical resolution rates for any age group (Figure 1A). Complete biochemical success after adrenalectomy was noted for at least 50% of patients regardless of AVS performance or age at surgery (Figure 1B). An even higher proportion of patients in each cohort, similarly irrespective of AVS confirmation of diagnosis, had normalization of the ARR postoperatively (Figure 2). It should be noted that the sample size for the 65 years and older cohort was small (n = 9).

A subgroup for patients with tumor size less than 10 mm (n = 12) was analyzed. There was no difference in outcomes in this subset of patients based on AVS for confirmation of diagnosis. Partial or complete clinical resolution occurred in 67% of patients with small tumors diagnosed with imaging only (n = 4), and 100% of patients with small tumors who underwent AVS (n = 6) (P = .22). There were 2 patients with small tumors in the imaging-only group without clinical benefit from adrenalectomy. There were 3 patients, 2 of whom underwent AVS, who did not normalize their ARRs in the small tumor cohort (P = .51).

This study analyzed the experience of a group of experienced endocrine surgeons at 3 tertiary care medical centers caring for patients with PA over the past 15 years. Much literature has been published supporting the need for AVS in almost all patients with PA who are possible surgical candidates based on a reported high discordance rate between cross-sectional imaging results and lateralization on AVS. One study⁹ reported that on the basis of CT findings alone overall, 42 patients (21.7%) would have been incorrectly excluded as candidates for adrenalectomy, and 48 patients (24.7%) may have undergone an inappropriate adrenalectomy. However, many of these patients had bilaterally normal or abnormal adrenal glands. They also suggested that patients younger than 40 years with a unilateral macronodule (>10 mm) may proceed directly to adrenalectomy⁹; this is a higher age cutoff than that espoused by the Endocrine Society.⁴ These cutoffs are generally based on the fact that the incidence of nonfunctioning adrenal cortical nodules increases with age, with 10% or more of individuals 70 years or older harboring an adrenal mass detectable on imaging or autopsy.¹⁴ There is a paucity of literature on the outcomes of patients with PA older than 35 years who present with a clear unilateral nodule with a normal contralateral gland on an adrenal protocol imaging study who proceed straight to adrenalectomy. Thus, this study excluded any patient who did not have these characteristics, regardless of age. In this context, the imaging concordance rate with AVS was 100%. It should be noted that the specificity¹⁵ and sensitivity of adrenal CT for adrenal adenoma are 98% and 88%, respectively, while the sensitivity of MRI for lipid-rich adenomas approaches 100%, although that for lipid-poor adenomas (up to 15%-30% of adenomas) is significantly lower (13%-75%).¹⁶

Consequently, the Subtyping Primary Aldosteronism: a Randomized Trial Comparing Adrenal Vein Sampling and Computed Tomography Scan (SPARTACUS) trial¹⁷ attempted to definitively address whether adrenal CT was equivalent to AVS for the discrimination of APA from BAH.¹⁷ Patients were randomly assigned to receive CT-based treatment or AVS-based treatment. The authors found no difference in outcomes between the 2 groups.¹⁷ This study has been heavily criticized for several reasons, including its primary end point of DDD of antihypertensive medication at 1 year instead of the more pertinent finding of biochemical normalization, a sample size too small to prove noninferiority, and a cohort not representative of patients with PA.^18,19 Therefore, whether clinicians may rely on imaging for diagnosis without confirmatory AVS for well-selected patients remains an unresolved question in the PA literature.

The Primary Aldosteronism Surgical Outcome (PASO) study¹³ developed consensus criteria for outcomes of adrenalectomy for unilateral primary aldosteronism. The definitions for clinical and biochemical success in this study were as prescribed by the PASO group, with the exception of amount of antihypertensive medication, which was described by absolute numbers instead of DDDs, a limitation of the database. They reported a complete clinical success rate of 37% and a partial clinical success rate of 47% overall.¹³ This patient cohort experienced similar clinical success rates, independent of whether AVS was performed, with complete clinical success rates of 43.6% (imaging only) vs 42.2% (AVS) and partial clinical success rates of 47.4% (imaging only) and 51.1% (AVS). The complete biochemical cure rate reported by the PASO group (94%) was slightly higher than that observed in this study; however, the aldosterone normalization rate was 98.8% in imaging-only patients and 95.6% in AVS patients, while 10% of imaging-only patients and 11.1% of AVS patients remained hypokalemic postoperatively, accounting for the discrepancy. The relatively high rates of persistent hypokalemia seen postoperatively may be explained by the significantly shorter follow-up time in this subset of patients. Lastly, the PASO group found that younger patients had a higher likelihood of complete clinical success and clinical benefit.¹³ Our study supports this finding, although this was also unrelated to the performance of AVS. Nonetheless, resolution of aldosterone oversecretion and normalization of ARR remained high regardless of age. As previously stated, aldosterone excess on its own confers higher morbidity independent of degree of hypertension.^5-7

It is important to emphasize that patients with small adenomas (<10 mm), bilaterally abnormal glands, or bilaterally normal glands with biochemical evidence of primary aldosteronism should generally undergo AVS prior to adrenalectomy because laterality and disease subtype are not easily characterized based on imaging for these patients. Yet, because many patients with PA have a unilateral tumor with a normal contralateral gland, a large proportion may be able to proceed directly to surgery. A 2018 study²⁰ found that of 1376 patients with benign adrenal tumors found on imaging, including adenoma, cortical thickening, and hyperplasia, only 14% had bilateral adrenal abnormalities. Of those patients with unilateral adenomas, only 23.8% had tumors less than 10 mm.²⁰ At our institution, of 93 patients who underwent adrenalectomy for PA, 14 did not have a clear-cut unilateral adenoma, and an additional 7 had a tumor less than 10 mm. This suggests that, for our population, approximately 77% of patients with PA under consideration for adrenalectomy could potentially forego AVS preoperatively. Although our study found no significant difference in clinical or biochemical resolution between the 6 patients with small tumors who proceeded to surgery based on cross-sectional imaging alone and the 6 patients with small tumors who underwent AVS preoperatively, the sample size was too small to assert that AVS might not always be necessary in this patient subgroup.

The instances of AVS failure that occurred did not appear to ultimately affect outcomes. The calculated AVS failure rate of 16.7% for this study was well in line with previously reported numbers; success rates are reported to range from 30% to 96% and are largely based on operator experience and case volume.^12,21,22 Regarding the 7 patients with PA for whom AVS failed who then proceeded with surgery based on the cross-sectional imaging results, all experienced resolution of their hyperaldosteronism postoperatively. Notably, this is in the context of treatment by expert assessment teams not universally available to patients. One might argue that only clinicians who are specialists in the surgical management of primary aldosteronism should consider proceeding to adrenalectomy without the results of successful AVS.

There were limitations to this study. Approximately half of the patients did not have 6-month follow-up data, although the median follow-up was 5.5 months. As discussed, the database used for this study only contained data regarding the number of antihypertensive medications; therefore, DDDs could not be assigned. In future studies, it would be useful to compare DDDs instead of absolute numbers of medications prescribed. Another limitation was the relatively small sample size. The rigorous exclusion of any patient who did not have a unilateral adenoma with a normal contralateral gland diminished the size of the cohort substantially. Lastly, this was a retrospective study, subject to all the biases intrinsic to that type of review, although the tri-institutional nature of the database may assuage some of the selection bias.

Cross-sectional imaging is sensitive for adrenal adenoma and highly specific. Adrenal venous sampling, while accurate for definitive diagnosis of APA when performed successfully, is invasive and prone to failure even when executed by an expert interventionalist. Adrenalectomy is a safe and low-risk operation in practiced hands. In the setting of the appropriate clinical presentation and a definite unilateral adenoma with a normal contralateral gland, considering patient preference and circumstances, an experienced surgeon may reasonably offer to proceed to adrenalectomy directly or to undergo AVS during the informed consent process without adversely affecting patient outcomes.

Corresponding Author: Rasa Zarnegar, MD, Division of Endocrine and Minimally Invasive Surgery, Department of Surgery, Weill Cornell Medicine, 525 E 68th St, K-836, New York, NY 10065 (raz2002@med.cornell.edu).

Accepted for Publication: August 11, 2020.

Published Online: November 4, 2020. doi:10.1001/jamasurg.2020.5011

Author Contributions: Dr Zarnegar had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Thiesmeyer, Ullmann, Beninato, Finnerty, Fahey, Zarnegar.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Thiesmeyer, Stamatiou, Beninato, Zarnegar.

Critical revision of the manuscript for important intellectual content: Thiesmeyer, Ullmann, Limberg, Stefanova, Beninato, Finnerty, Vignaud, Leclerc, Fahey, Brunaud, Mirallié, Zarnegar.

Statistical analysis: Thiesmeyer, Ullmann, Stamatiou, Limberg, Stefanova, Beninato, Zarnegar.

Administrative, technical, or material support: Ullmann, Stefanova, Finnerty, Leclerc, Fahey, Brunaud.

Supervision: Beninato, Finnerty, Fahey, Mirallié, Zarnegar.

Conflict of Interest Disclosures: Dr Zarnegar reported personal fees from Beckton Dickenson outside the submitted work. No other disclosures were reported.

Funding/Support: The Department of Surgery at Weill Cornell Medical College provided the financial and material support for this work.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.