Discordance between adrenal vein sampling and imaging: the role of unilateral cortisol excess

Abstract

Objective

Adrenal vein sampling (AVS) is considered the gold-standard for identifying unilateral primary aldosteronism (PA), but is sensitive to any asymmetry in cortisol production. Subclinical autonomous cortisol production likely contributes to discordance between cross-sectional imaging and AVS.

Design and Methods

Retrospective chart review was performed of patients in Calgary, Alberta who (1) had a diagnosis of PA with clear adrenal mass, (2) had discordant AVS and cross-sectional imaging, and (3) underwent dexamethasone-suppressed NP59-iodocholesterol adrenal scintigraphy (n = 25). Postoperative biochemical and clinical outcomes were evaluated. Surgical pathology was analyzed with immunohistochemical staining for CYP11B1 and CYP11B2.

Results

NP59 scanning demonstrated autonomous steroidogenesis from the same side as the computed tomography (CT)-identified lesion, despite discordant AVS results, in 19/25 cases. Out of the 16 patients who underwent adrenalectomy (guided by NP59), 11 cases had a final diagnosis of cortisol-producing adenoma with bilateral PA, while the final diagnosis in the remaining 5 cases was cortisol-producing adenoma with unilateral PA (defined by a complete biochemical response). All cases preoperatively had low/suppressed adrenocorticotropin hormone that rose postadrenalectomy, in keeping with the resolution of cortisol autonomy. Adrenal vein sampling incorrectly diagnosed the subtype of PA or falsely localized the side of aldosterone excess in 10/16 cases.

Conclusions

Discordant CT and AVS results in patients with PA and obvious adrenal mass are often explained by subclinical asymmetric cortisol excess. Clinicians should be aware of the limitations of AVS in the presence of subtle autonomous cortisol secretion and must be able to counsel patients regarding the possible outcomes from surgery when discordant lateralization is present.

Significance.

In many clinical scenarios, unilateral adrenalectomy may lead to better clinical outcomes than medical therapy in unilateral primary aldosteronism (PA). However, successful surgery depends on the localization of aldosterone excess. Discordance between cross-sectional imaging and adrenal vein sampling (AVS) can be difficult to resolve. We found through functional imaging with NP59 scanning that such discordance could often be explained by asymmetric cortisol excess from a subtle cortisol-producing mass in the setting of unilateral, or more commonly, bilateral PA. Challenging cases like these have previously been excluded from clinical trials or treated as “uncertain”, but we resolved these with functional imaging, corroborated by histopathology and clinical outcomes following NP59-guided adrenalectomy. This study has significant implications for investigation and subsequent surgical decision-making in such complicated cases.

Introduction

Surgery, when possible, may lead to better clinical outcomes compared to medical therapy in patients with primary aldosteronism (PA).^1-3 Definitive cure of aldosterone excess has been shown to improve blood pressure,⁴ reduce LV mass,^5,6 delay the progression of nephropathy,⁷ and improve quality of life.⁸ Recent meta-analysis has shown there is a mortality benefit from surgery for PA compared to medical therapy, with benefit seen as soon as 1 year after treatment.¹

Few individuals who are eligible for PA surgery ever receive it.^9,10 In addition to low screening rates for PA,¹¹ difficulty accessing adrenal vein sampling (AVS) is another major reason few people with unilateral PA are diagnosed or treated.¹⁰ However, 1 major post-AVS diagnostic barrier is the situation of imaging-AVS discordance, which may be seen in up to 30%-40% of PA-AVS series.^12-14 In such cases, a patient with PA may have a visible adrenal mass on cross-sectional imaging, but AVS paradoxically indicates a focal source of aldosterone excess in the adrenal gland opposite to the discrete mass (so-called reverse lateralization), or in other cases, fails to lateralize altogether. With such discrepancies in localization studies, many clinicians and patients may be reluctant to pursue adrenalectomy for fear of removing the incorrect adrenal gland. The net effect may be to further reduce the number of patients getting curative surgery. In these situations of discordance, some investigators have insisted that AVS should be considered the reference standard^9,12,14,15 but there is little outcome data to support that surgery guided by AVS objectively leads to a better outcome, and indeed some studies have failed to show a difference in clinical outcomes for therapy guided by computed tomography (CT) vs AVS.¹⁶

Autonomous adrenal cortisol excess likely coexists within a considerable subset of PA patients^17-21 and this combination may be more deleterious than PA alone.^22-25 Autonomous cortisol excess exists on a spectrum, and while clinically accepted definitions make “mild autonomous cortisol secretion” (MACS) or overt Cushing's syndrome more simple to identify, there likely exists a subset of patients with autonomous cortisol excess not meeting the threshold for even MACS. Our group has previously demonstrated that unsuspected adrenal cortisol excess likely explains many cases of apparent failed adrenal venous catheterization and/or reverse lateralization, which may be detected using an algorithm applied to a patient's AVS data.²⁶ We now provide evidence through dynamic nuclear adrenal NP59 scanning and surgical/pathology outcomes to demonstrate that in many cases of reverse lateralization or unexpected nonlateralization, unsuspected mild cortisol autonomy is a common and likely explanation. Adrenal vein sampling results should be interpreted with caution for surgical decision-making in such cases.

Methods

This retrospective study was performed using the Calgary-AVS database, with approval through the institutional review board (REB18-0199), and was conducted in accordance with the principles of the Declaration of Helsinki. A waiver of individual consent was granted for access to personal identifiable health information consistent with the conditions under section 50, the Health Information Act of Alberta. The Calgary-AVS database has been used for many prior studies in PA. It includes all cases of AVS performed by a single operator from 2005 to 2022 and 2 operators from 2022 onwards. Most PA patients in Calgary who get AVS come via the Endocrine Hypertension Clinic where the approach to diagnosis of PA has long been standardized and published previously.^26,27 The majority of cases consist of patients with resistant hypertension, and/or hypokalemia, and/or adrenal mass, and with elevated aldosterone–renin-ratio (ARR).²⁸

Since 2015, our laboratory has reported plasma aldosterone concentration via immunoassay in pmol/L and direct renin concentration in mIU/L using the Diasorin Liaison XL assay. Based on our prior validation study,²⁹ an ARR > 60 pmol/mIU has been considered compatible with a diagnosis of PA, although additional characteristic clinical factors are used to support the diagnosis.^9,30,31 We have previously demonstrated that in cases with multiple high-probability features for PA, confirmatory testing may be omitted,³¹ as also recognized by current guidelines.⁹

Patients with a diagnosis of PA were offered adrenal imaging with consideration of AVS thereafter in a shared decision-making process. Prior to AVS, patients discontinued all antihypertensive drugs except calcium channel blockers, alpha-blockers, beta-blockers, and direct vasodilators for at least 2-4 weeks; glucocorticoids, spironolactone, eplerenone, and amiloride were always stopped at least 6 weeks prior to AVS. Hypokalemia was corrected with oral potassium replacement. Prior to AVS, an early morning ARR, adrenocorticotropin hormone (ACTH) (Siemens Immulite immunoassay), and DHEA-S (Siemens Immulite immunoassay until December 2022, then Roche Cobas e801 assay) were routinely collected. Simultaneous bilateral cannulation of the adrenal veins was performed with sampling from the left, right, and peripheral sites before and after ACTH-stimulation, as previously described.¹³ Based upon local, outcome-supported validation studies, selectivity index (SI) of >2:1 at baseline or >3:1 postcosyntropin was used to define technical success.^32,33 Unilateral PA has been defined and validated by a lateralization index (LI) of >3:1 at either baseline or postcosyntropin.^34,35

Starting in 2017, we began to offer dexamethasone-suppressed NP59-iodocholesterol adrenal scintigraphy to PA patients with a discrete adrenal nodule, but imaging-AVS discordance. For this, patients had similar medication restrictions as AVS with the exception of calcium channel blockers and statins also being stopped a week in advance. Dexamethasone 1 mg qid was given for a 2-week period prior to and during image acquisition to fully suppress ACTH (except in the case of 3 patients where ACTH was already suppressed at baseline, measured within 3 months of NP59 scan). Patients were injected with 37 MBq (I-131) iodomethyl norcholesterol (NP59). Planar and single photon emission computed tomography (SPECT)/CT imaging of the abdomen was done on Days 3, 4, and 5 postinjection, and interpreted by a dedicated nuclear medicine physician.

Patients were offered surgery if there was concordant, unilateral NP59 uptake indicating functional autonomy. Patients were seen by both the surgery and endocrinology teams regularly postoperatively and clinical and biochemical outcomes were collected in accordance with the Primary Aldosteronism Surgical Outcome (PASO) criteria.³⁶

All hematoxylin and eosin-stained slides were reviewed by the study pathologist (MDH). For each case, the presence or absence of background cortical nodules, micronodules, zona glomerulosa hyperplasia, and cortical atrophy were recorded. Additionally, 1 representative section of the dominant adrenocortical tumor and 1 from the background adrenal were selected for further immunohistochemistry. Four micrometer sections were stained with CYP11B1 (Abcam cat #: ab197908, dilution 1:600, run at 20-X-20) and CYP11B2 (Abcam cat #: ab168388, dilution 1:1200, run at 20-X-20) antibodies with a standard 3,3′-diaminobenzidine secondary detection. All CYP11B1 and CYP11B2-stained slides were reviewed by the study pathologist and classified according to the HISTALDO classification^37,38

The present series includes a retrospective review of all PA-NP59 cases performed at our institution to date. Clinical demographics are presented as median (IQR) along with case-series level comparison of concordance between individual CT, AVS, and NP59 lateralization data. Comparison between paired repeat measures was tested by Wilcoxon matched-pairs signed rank test, with significance set at P < .05 (GraphPad Prism 7.0, LaJolla, CA). Comparison of nodule size between groups was tested using an unpaired t-test. For patients who underwent surgery, PASO-related outcomes were reported including postoperative biochemical and clinical data with appropriate cohort and individual-level comparison to preoperative data.

Results

NP59 vs AVS for lateralization in PA patients with AVS-CT discordance

A search of the Calgary-AVS database yielded 25 cases of CT-AVS discordance where NP59 scanning was performed from 2017 (first availability) to 2023; dexamethasone suppression was used in all but 3 cases (where basal ACTH was already suppressed at levels of <1.1, 1.4, and 1.6, ref 2.0-11.5 pmol/L). All patients had at least 1 clear adrenal mass seen on cross-sectional imaging, and CT-AVS discordance was defined as AVS that either (1) lateralized away from the adrenal mass, (2) failed to lateralize despite unilateral mass, or (3) in the case of 2 adrenal masses failed to lateralize to the larger mass. The median patient age was 57 years, and 64% were female. Clinical and biochemical demographics are displayed in Table 1; patients had hypertension for a median of 9 years and used 1-5 defined daily doses of anti-hypertensive medications; 14/25 (56%) patients had a history of hypokalemia (<3.5 mmol/L), and the highest measured median ARR was greater than 4-times the diagnostic upper limit. Dexamethasone-suppressed morning cortisol (either 1 mg overnight dexamethasone suppression test [DST] or during dexamethasone suppression prior to NP59 scanning) was collected in 14/25 patients. Of these, the median cortisol was 47 nmol/L, with an IQR of 33-138 nmol/L; 7/14 patients had a cortisol post-DST that was >50 nmol/L, meeting the 2023 European Society of Endocrinology Clinical Practice Guideline definition of mild autonomous cortisol secretion (MACS).³⁹ An additional 4 patients who did not have a charted cortisol during dexamethasone suppression had 24-urine cortisol collections within the normal range. A total of 11/25 patients had 24-hr urine cortisol results available in their charts, with varying time between collection and AVS. The median cortisol measurement was 118 nmol/day (IQR 58-164 nmol/day). Basal ACTH and DHEA-S levels were at the very low end of normal or overtly suppressed in the majority of cases. Aside from hypertension and obesity, none of the patients had clinical characteristics to suggest classic cortisol excess.

Table 1.

Baseline clinical characteristics of PA patients with discordant lateralization studies who underwent NP59 scanning.

Clinical variable at presentation	Median (IQR)
BMI, kg/m2	31 (27-35)
Systolic blood pressure, mmHg	140 (130-153)
Diastolic blood pressure, mmHg	86 (80-96)
Years of hypertension	9 (4-15)
Defined daily dose of antihypertensive medication	3 (1-5)
First aldosterone, pmol/L	319 (259-492)
First renin, mIU/L	1.8 (1.0-4.0)
First ARR (ref <60 pmol/mIU)	165 (97-293)
Highest ARR (ref <60 pmol/mIU)	269 (157-433)
Lowest potassium, mmol/L	3.4 (3.2-3.9)
ACTH pre-AVS (ref 2.0-11.5 pmol/L)	2.2 (1.6-3.4)
DHEAS pre-AVS (ref 0.3-6.0 μmol/L)	1.3 (0.4-2.1)
CT adrenal mass sizea, mm	22 (16-25)

First aldosterone/renin/ARR denotes the first available measure at the time of first presentation to Endocrinology, while the highest ARR denotes the highest “clean” ARR on the chart (prior to adrenalectomy in cases that underwent surgery).

ARR, aldosterone–renin-ratio; AVS, adrenal vein sampling; PA, primary aldosteronism.

^aIn the 2 cases with bilateral adrenal masses, the size of the larger mass is recorded here.

According to the standard SI of postcosyntropin measures during AVS, 100% of cases met the criteria for successful catheterization (SI > 3). However, in the precosyntropin measures, there were 13 instances where the SI was <2.0, and in 9/13 of such instances, the apparent SI failure was from the adrenal vein contralateral to the visible adrenal mass (where there was likely a relative suppression of cortisol production).²⁶ The remaining 4/13 instances occurred in 2 patients who had basal SI < 2.0 from both adrenal veins. In all instances (13/13) the corresponding baseline ratio of aldosterone levels (aldosterone index²⁶) was >2.0; this in concert with postcosyntropin SI > 3 suggests successful catheterization and relative suppression of cortisol production in the adrenal contralateral to the mass as the explanation for apparent SI “failure”.

In this population selected for discordance as inclusion criteria, compared to the CT findings, AVS results showed either bilateral localization despite unilateral mass (or bilateral masses with a significant size disparity), lateralization despite bilateral masses, or reverse lateralization away from a unilateral mass. In total, 15/25 (60%) showed lateralization by AVS either at baseline or postcosyntropin, but in all such cases, it was reverse lateralization. Dexamethasone suppressed, low-renin NP59 scans suggested functional nodular adrenocortical autonomy in 92% of cases; 76% of cases showed concordance with the CT-visualized adrenal nodule(s), in stark contrast to the AVS results. Therefore, of surgically fit patients, 83% (20/24) could be offered surgical intervention. A comparison between CT, AVS, and NP59 scanning results is shown in Table 2. A representative example of SPECT–CT images taken after the injection of NP59 is shown in Figure 1.

Table 2.

Agreement between CT, AVS, and NP59.

White background denotes subtype test concordance with mass seen on CT imaging, while grey denotes discordance with CT imaging. R = right, L = left, B = bilateral. AVS stim indicates lateralization results measured post-cosyntropin stimulation.

^aUncertain future surgical date.

^bNP59 performed post non-curative prior left adrenalectomy thus no longer candidate for right adrenalectomy.

Figure 1.

Representative SPECT–CT images were taken 3 days after the injection of radiolabeled cholesterol (NP59), following 2 weeks of dexamethasone administration. Images demonstrate uptake of radiotracer within the left adrenal mass and no activity within the right adrenal gland.

Surgical outcomes after NP59-guided adrenalectomy with AVS discordance

There were 16 cases who subsequently received surgical adrenalectomy as guided by the NP59 results; 3 declined surgery. The clinical pre- and postoperative data for the surgical cases are presented in Tables 3 and 4. Where a DST was performed (n = 9), almost half (4/9) showed normal cortisol suppression. In 15 cases, a post-op day 1 morning cortisol was measured and 3 patients had a level <100 nmol/L (consistent with postoperative adrenal insufficiency); these patients received temporary glucocorticoid replacement which was weaned and discontinued within 2 months postoperatively.

Table 3.

Aggregate clinical and biochemical measures pre- and postoperatively in cohort who underwent unilateral adrenalectomy.

Characteristic, median (IQR)	Pre-op	Post-op	P value
BMI	34 (28-35)
Years of hypertension	9.5 (2.5-18)
Systolic blood pressure, mmHg	141 (130-156)	131 (124-141)	.01
Diastolic blood pressure, mmHg	85 (80-96)	83 (76-89)	NS
Defined daily dose of antihypertensive medication	4.2 (0.5-5.6)	1.3 (0.2-3.0)	.007
Aldosterone, pmol/L	322 (270-489)	362 (289-554)	NS
Renin, mIU/L	1.3 (1-4.4)	2.4 (1.4-6.6)	NS
ARR (ref <60 pmol/mIU)	305 (213-522)	141 (40-281)	.01
Potassium, mmol/L	3.4 (3.2-3.9)	4.2 (3.9-4.4)	<.0001
ACTH (ref 2.0-11.5 pmol/L)	2.3 (1.7-3.5)	8.1 (6.0-11)	<.0001
DHEAS pre-AVS (ref 0.3-6.0 pmol/L)	0.8 (0.4-2.5)
PASO—complete clinical response		1/16
PASO—partial clinical response		8/16
PASO—complete biochemical cure		5/16
PASO—absent biochemical cure		11/16

Data are presented as median with interquartile range, comparison between paired repeat measures was tested by Wilcoxon matched-pairs signed rank test. Pre-op ARR is the highest value measured preoperatively, while pre-op aldosterone and renin are the first values measured (ie, at the first presentation to Endocrinology). Pre-op potassium is the lowest value on the chart prior to adrenalectomy. Note that 2 patients had ARR done at outside laboratory with PRA in ng/L/s rather than DRC in mIU/L, and were excluded from the above statistical analysis. Post-op data are that collected at the furthest time point after surgery.

ARR, aldosterone–renin-ratio; AVS, adrenal vein sampling; DRC, direct renin concentration; PASO, Primary Aldosteronism Surgical Outcome; PRA, plasma renin activity.

Table 4.

Individual pre- and postoperative biochemical data with final diagnosis after unilateral adrenalectomy.

Patient #	Pre-op ARR (ref <60 pmol/mIU)	ACTH baseline (ref 2.0-11.5 pmol/L)	DHEAS baseline (ref 0.3-6.0 pmol/L)	CT lesion, size (mm)	AVS base interpretation (B, RL)	AVS stim interpretation (B, RL)	Dexamethasone-suppressed cortisol (nmol/L)	Post-op cortisol (nmol/L)	Post-op ARR (ref <60 pmol/mIU)	Post-op ACTH (ref 2.0-11.5pmol/L)	Final diagnosis
2	282	1.8		14	B	B		394	318	6.9	CPA with BPA
4	403	3	0.6	16	B	RL	38	162	118	6	CPA with BPA
5	269	1.3		24	RL	RL		19	289	10.7	CPA with BPA
6	691	2.3	2.8	23	B	B	43	533	209	10.5	CPA with BPA
7	4192a	1.8	0.8	17, 25	RL	RL	136	903	6	3.6	CPA with UPA
8	532	4.8	3.5	17	B	B	28	66	465	7.4	CPA with BPA
9	157	1.6	0.4	60	RL	B	33	182	303	8.1	CPA with BPA
11	155	1.1	0.4	25	B	RL		179	33	3.6	CPA with UPA
13	378	2.6	2.1	33	RL	RL	143	453	159	9.4	CPA with BPA
14	518	2.2		18	B	B	242	228	156	8.5	CPA with BPA
16	2400a	3.7		25	B	B			59		CPA with UPA
19	232	4.45	1.35	23	B	B	51	598	257	4.9	CPA with BPA
20	327	6.15	0.4	27	RL	B		343	33	14.6	CPA with UPA
22	576	1.6		16	RL	B	119	75	126	10.4	CPA with BPA
23	97	2.2		22	RL	B		273	19	7.2	CPA with UPA
24	256	2.8		15	RL	B		458	77	12.9	CPA with BPA

B = bilateral AVS (LI < 3); RL = reverse lateralization (LI ≥ 3, lateralized away from largest adrenal mass seen on CT). ARR, aldosterone–renin-ratio; BPA, bilateral primary aldosteronism; CPA, cortisol-producing adenoma; DRC, direct renin concentration; PRA, plasma renin activity; UPA, unilateral primary aldosteronism.

^aARR done at the outside laboratory with PRA in ng/L/s rather than DRC in mIU/L. Pre-op ARR is the highest value measured preoperatively.

Postoperative ACTH (measured a median of 3 weeks postoperatively, IQR 2-6 weeks) was substantially higher in all but 1 case, suggesting resolution of any prior adrenal cortisol autonomy (Figure 2). A total of 5/16 patients achieved a complete biochemical resolution of their PA after adrenalectomy with normalization of ARR and resolution of hypokalemia, in keeping with a unilateral subtype of PA, leading to a final diagnosis of cortisol-producing adenoma with concomitant unilateral PA. The remaining 11/16 cases did not have normalization of ARR after surgery and thus the final diagnosis was of a cortisol-producing adenoma with bilateral PA. Though the postoperative ARR normalized in only a third of cases, there were overall large reductions in ARR, systolic blood pressure, and antihypertensive medication use across the cohort that were statistically significant (Figure 2). The median time between surgery and final ARR used to define biochemical outcomes was 21 weeks (IQR 7.5-47), and the median time to final clinical follow-up was 48 weeks (IQR 25-70).

Figure 2.

Aggregate clinical (A-C) and biochemical data (D-F) for patients who underwent unilateral adrenalectomy guided by NP59 localization, compared pre- and postoperatively. Comparison between paired repeat measures (pre- and postoperatively) was tested by Wilcoxon matched-pairs signed rank test, ns, P > .05, *P < .05, **P < .01, ****P < .0001. Note that 2 patients had aldosterone–renin-ratio (ARR) done at outside laboratory with plasma renin activity in ng/L/s rather than direct renin concentration in mIU/L, and were excluded from the above analysis. Preoperative measurements of ARR and serum potassium used the highest and lowest measures, respectively.

Comparison of the AVS lateralization data and final postoperative classification of PA as either bilateral or unilateral revealed that AVS was inaccurate in 10/16 cases: 6 cases were identified by AVS as unilateral (LI >3 either at baseline or postcosyntropin stimulation) when the final outcomes were consistent with bilateral PA; in 3 cases where the final diagnosis was unilateral PA based on surgical outcome, the AVS paradoxically lateralized to the contralateral adrenal; and in 1 case where there was complete biochemical response after unilateral adrenalectomy, the AVS was nonlateralizing. Of the remaining 6 cases, 5 were correctly identified by AVS as bilateral aldosterone excess, and 1 case was unilateral with the correct side identified by AVS (Patient 7). Similarly, comparison of NP59 results and final outcomes demonstrates that while NP59 found 14/16 cases to have unilateral uptake of tracer suggesting adrenocortical autonomy, 10 of these 14 cases had a final diagnosis of bilateral PA.

Table 5 summarizes the results of immunohistochemical analysis of the excised adrenal glands using stains for CYP11B1 (11-beta-hydroxylase) and CYP11B2 (aldosterone synthase). Figure 3 shows representative images of the various HISTALDO classifications identified within the pathology samples.

Table 5.

Immunohistochemical analysis of surgically removed adrenal glands.

Patient #	Dominant nodule CYP11B1	Dominant nodule CYP11B2	HISTALDO classification
2	+	+	APA + MAPM
4	+	−	MAPM
5	No data	No data	No data
6	+	−	APN + MAPM
7	+	+	MAPNa + MAPM
8	+	−	APN + MAPM
9	+	−	MAPM
11	+	−	APDH + MAPM
13	+	−	MAPM
14	+	−	MAPN + MAPM
16	+	−	MAPM
19	+	−	APDH
20	+	−	MAPM
22	+	−	MAPM
23	+	−	APDH + MAPM
24	+	−	APDH + MAPM

Samples were analyzed per the 2020 HISTALDO consensus guidelines³⁷ using CYP11B2 immunostained sections to identify likely sources of aldosterone production. + denotes presence of staining, and − denotes the absence of staining.

CYP11B1, 11-beta-hydroxylase; CYP11B2, aldosterone synthase; APA, aldosterone-producing adenoma; APN, aldosterone-producing nodule; APM, aldosterone-producing micronodule; MAPN/MAPM, multiple aldosterone-producing nodules/micronodules; APDH, aldosterone-producing diffuse hyperplasia.

^aPer HISTALDO classification guidelines, the designation of APA should be reserved for solitary lesions. Patient 7 had both multiple aldosterone-producing micronodules (MAPM) as well as multiple aldosterone-producing nodules (seen by both H&E and CYP11B2 staining), including the dominant nodule.

Figure 3.

Examples of the HISTALDO patterns identified in the study. (A, D, G) Hematoxylin and eosin; (B, E, H) CYP11B1 immunohistochemistry; (C, F, I) CYP11B2 immunohistochemistry. (A-C) Aldosterone-producing adrenocortical adenoma (Patient 2, magnification 20×). (D-F) Aldosterone-producing diffuse hyperplasia (Patient 23, magnification 20×). (G-I) Multiple aldosterone-producing micronodules (Patient 20, magnification 40×).

In the majority of cases (13/15—1 case was not available for analysis) the dominant nodule did not stain positive for CYP11B2, indicating that the mass seen on cross-sectional imaging was likely not the source of aldosterone excess. While staining for CYP11B1 is nonspecific in that it is an enzyme shared between both the aldosterone and cortisol synthesis pathways, absence of CYP11B2 with CYP11B1 positivity suggests the dominant nodules were all producing cortisol (and likely this was what was seen as autonomous hormone production on NP59 scanning), even in the 4/9 surgical cases (Table 4) with “normal” cortisol following DST (ie, not meeting the clinical guideline definition of MACS³⁹). In 5/15 cases, clear cortical atrophy (defined as cortical thickness <2.5 mm) was seen on H&E histology, but in a number of other cases, the presence of extensive multinodularity precluded clear determination of the thickness of the underlying cortex.

Discussion

Both primary autonomous aldosterone excess and autonomous cortisol excess may lead to significant adverse outcomes, although until recently these were considered 2 largely separate disorders: PA and (sub-)clinical Cushing's syndrome. Increasing reports have now demonstrated that these 2 endocrinopathies coexist more frequently than previously appreciated.^17-19 Our data from a patient population with a clear adrenal mass confirms this and shows that in this population the combined endocrinopathy likely takes on at least 2 different forms: the cortisol-secreting adrenal adenoma, with either unilateral or bilateral PA.

In the setting of an adrenal mass and hypertension, when the ARR is normal, PA is likely readily excluded; excess cortisol production alone may be sought and if present, attribution to the mass is generally reasonable unless bilateral masses are present.⁴⁰ However, when the clinical presentation is dominated by hypertension, hypokalemia, and adrenal mass with markedly elevated ARR, subtle cortisol excess may easily be overlooked because classical symptoms and signs may be absent, and conventional laboratory thresholds used to detect traditionally recognized cortisol excess may appear normal. Indeed half of our patients had a normal DST, and isolated low ACTH is not yet widely recognized as a diagnostic criteria of cortisol excess.^39,41 However, the coexistence of even subtle autonomous cortisol excess may jeopardize the reliability of AVS when using cortisol for interpretation, which can lead to either treatment inertia or worse, inappropriate contralateral adrenalectomy.

For the cases presented here, the reason for the apparent discordant lateralization using AVS was on account of the use of cortisol as a “control” measure in the calculation of A/C ratios. The use of cortisol to interpret AVS for PA depends upon the assumption that adrenal cortisol production is symmetric. However, if primary autonomous cortisol excess is present, the underlying assumption is violated and unexpected results, such as reverse lateralization, may appear. Asymmetric cortisol production could cause the LI from AVS to either point away from a unilateral nodule secreting aldosterone, or could make a bilateral phenotype of PA appear to be unilateral.^20,26 The use of a different control hormone such as epinephrine or metanephrine has been suggested and should perhaps supplant cortisol in these cases.^20,42-45 Asymmetric cosecretion of adrenomedullary catecholamines would be expected to be far rarer than cortisol excess, and diagnostic measurement ratios (ie, lateralization indices based on aldosterone/metanephrine ratios) have been shown to be at least as accurate as aldosterone/cortisol ratios at identifying unilateral disease.⁴⁴ Abnormal DST in keeping with MACS as defined by current clinical guidelines should certainly prompt clinicians to consider collection of metanephrine from the adrenal veins and inferior vena cava during AVS, however, as demonstrated by our data, more subtle cortisol autonomy not meeting criteria for MACS may still interfere with AVS interpretation. Thus, we suggest measurement of basal morning ACTH and/or DHEAS within 2 weeks before AVS and if it is found to be suppressed in conjunction with an adrenal mass, consideration should be given to collection of metanephrine during AVS.

The field of functional nuclear imaging in PA is rapidly progressing. Several new tracers with specificity to CYP11B1/11B2 have undergone initial validation studies. ¹¹C-Metomidate, ¹⁸F-CETO, and ⁶⁸Ga-pentixafor all show promise but are presently only available in highly specialized research centers around the world.^46-50 NP59 iodocholesterol scanning is an older nuclear tracer with several disadvantages compared to newer compounds including lower image resolution, higher radioactivity exposure, requirement for extended dexamethasone suppression, and multiple image acquisition visits. Additionally, this tracer does not differentiate between the synthesis of different steroid-based hormones. For these reasons, although its use in PA has been previously reported,^51,52 it has never been suggested as a suitable replacement for AVS.⁹ And indeed, in a majority of complex, imaging discordant cases presented here, the final subtype of PA was found to be bilateral—in contrast to the finding of unilateral uptake on NP59 scan, which as mentioned above is not specific for synthesis of cortisol vs aldosterone. Thus, while NP59 did not prove to be reliable at subtyping PA in this patient population, the more striking (and clinically relevant) finding is that the present data convincingly shows that AVS cannot be considered a diagnostic gold-standard: false lateralization is not rare. In this case series, AVS incorrectly diagnosed the subtype of PA in 10/16 cases. The use of a functional nuclear adrenal tracer, specific for aldosterone synthase, will likely help resolve apparent discordance between obvious adrenal anatomical abnormalities and unexpected AVS lateralization results.

The goal of any lateralization test in PA is to confirm the surgical candidacy of the patient for unilateral adrenalectomy. The value of surgical intervention for unilateral PA is established, as is the value of surgery for a classical cortisol-producing adrenal adenoma. The clinical benefit of surgery to remove an adrenal mass with subclinical cortisol autonomy (MACS) is controversial, although there is mounting evidence to show that these are linked to multiple adverse health outcomes even without the classically recognizable cortisol excess syndrome.^53-55 The question of adrenalectomy is particularly nuanced in many of the cases presented here, as they would not even meet accepted criteria for MACS; therefore, the possible clinical benefit from the perspective of addressing cortisol excess is even more unclear. It is generally accepted that the clinical syndrome of cortisol excess exists on a spectrum,⁵³ but it remains to be elucidated precisely at what level and to what degree mild cortisol excess affects cardiometabolic and other outcomes. Clinical decision-making should involve consideration of how adrenalectomy might address both cortisol and aldosterone excess, existing comorbidities that may be predicted to be affected by addressing hormonal excess, the patient's fitness for surgery, and patient preferences. Most importantly, our experience demonstrates that the moniker of “cortisol cosecretion” in PA is more complicated than may have been believed; most cases represent a cortisol-producing adenoma with bilateral PA, a few represent a true dual-secreting adenoma.

We had initially hypothesized that in our cohort of patients selected for AVS-imaging discordance in the context of clear adrenal mass, the discordance was explained by the mass cosecreting aldosterone and cortisol. We were surprised to find that for the majority of our patients, the final diagnosis was unilateral cortisol-producing mass in the context of bilateral PA, with PA as the dominant presenting phenotype. Further, detailed immunohistochemical analysis revealed that in fact, in only 2 cases was the dominant nodule secreting aldosterone. Indeed, when classified using the recently updated HISTALDO classification schema, the adrenal pathologies in terms of aldosterone production were all compatible with either MAPN, MAPM, or APDH—classifications that have been shown in case series to be associated with a higher prevalence of absent biochemical response compared to a single aldosterone-producing adenoma.^37,56 These results raise the question of whether longer-term follow-up would reveal recurrence of hyperaldosteronism even among the patients with a complete biochemical response by PASO at postoperative follow-up (median time to biochemical follow-up ∼5 months). Indeed, for one of the patients we studied (Patient 9), an initial postoperative ARR at 3 weeks post-op had normalized, but an ARR repeated almost 10 months later was markedly abnormal. This patient had pathology in keeping with MAPM. Rather than a cosecreting mass confounding AVS to lateralize away from a unilateral problem, it appears that in a significant number of cases, unilateral cortisol excess caused AVS (and NP59 scanning) to make bilateral aldosterone excess look falsely unilateral. While we cannot definitively prove that unilateral cortisol excess arose from the visualized adenoma for the reported patients, this was strongly inferred based on (1) unilateral autonomous steroidogenesis seen by NP59 scan (reported as avidity of the radiotracer to the dominant nodule in the majority of cases), (2) absence of CYP11B2 staining within the adenoma in 13/15 cases excluding aldosterone-producing tissue as the cause of norcholesterol uptake, (3) AVS demonstrating significant cortisol asymmetry (and in a significant number of cases relative suppression of cortisol from the adrenal contralateral to the visible mass compared to the periphery, as evidenced by SI <2), and (4) low/suppressed ACTH at baseline that would otherwise be inexplicable.

Therefore, in cases of subtyping discordance, patients should be counseled that resection of a functional cortisol-producing adrenal mass may be of some benefit (though this remains controversial and is unclear how the existing evidence may apply to patients who do not meet clinical criteria for MACS),⁵⁵ but clinical and biochemical PA cure cannot be assured preoperatively due to the possibility of bilateral aldosterone excess. In a recent retrospective cohort study, Turcu et al.⁵⁷ found that CT-AVS disagreement was an independent risk factor for residual PA following unilateral adrenalectomy in patients with PA that lateralized by AVS (though analysis included cases with normal adrenals or bilateral nodules on imaging, and thus represents a somewhat different patient population than that studied here). A useful debulking effect of unilateral adrenalectomy in bilateral PA has been described^58,59 and is consistent with our data where a considerable number of patients (7/11) had partial or complete clinical responses (without a partial or complete biochemical response) after surgery.

In summary, we have presented convincing data showing that discordant CT and cortisol-based AVS results in patients with PA and obvious adrenal masses are likely explained by unexpected subclinical autonomous asymmetric cortisol excess, frequently not detected by DST. Asymmetric cortisol excess can cause AVS to lateralize to the normal adrenal, or perhaps more commonly, make bilateral aldosterone excess appear unilateral. NP59 scanning is an additional piece of information that further supported the hypothesis of unilateral cortisol excess, however, was not accurate at identifying the source of aldosterone excess, most likely because the tracer does not differentiate between different steroid synthesis pathways. Until adrenocortical nuclear imaging with specific tracers is more widely available, clinicians should be alerted to the possibility of cortisol cosecretion by recognizing the pattern of reverse lateralization with AVS along with biochemical clues from the AVS data itself,²⁶ and consider using metanephrine as an alternative to cortisol in the AVS procedure. Further studies to validate discordant findings in PA lateralization studies are warranted.

Funding

A.A.L. is supported by the Heart and Stroke Foundation of Canada. This work was funded in part by the Canadian Institutes of Health Research (PJT-159533).

Authors’ contributions

Megan Grundy (Conceptualization [supporting], Data curation [lead], Formal analysis [lead], Investigation [equal], Methodology [equal], Visualization [lead], Writing—original draft [lead], Writing—review & editing [lead]), Alexander Leung (Conceptualization [lead], Data curation [supporting], Formal analysis [supporting], Funding acquisition [lead], Investigation [equal], Methodology [supporting], Project administration [equal], Resources [equal], Supervision [supporting], Writing—review & editing [supporting]), Martin Hyrcza (Data curation [supporting], Formal analysis [supporting], Investigation [supporting], Methodology [supporting], Visualization [supporting], Writing—review & editing [supporting]), Janice Pasieka (Conceptualization [supporting], Data curation [supporting], Formal analysis [supporting], Investigation [supporting], Writing—review & editing [supporting]), Denise Chan (Investigation [supporting], Methodology [supporting], Writing—review & editing [supporting]), Stefan Przybojewski (Investigation [supporting], Methodology [supporting], Writing—review & editing [supporting]), Benny So (Data curation [supporting], Investigation [supporting], Methodology [supporting], Writing—review & editing [supporting]), Dennis Orton (Methodology [supporting], Writing—review & editing [supporting]), Adrian Harvey (Data curation [supporting], Investigation [supporting], Writing—review & editing [supporting]), Gregory Kline (Conceptualization [lead], Data curation [supporting], Formal analysis [supporting], Investigation [equal], Methodology [equal], Project administration [lead], Resources [lead], Supervision [lead], Validation [equal], Visualization [supporting], Writing—original draft [equal],), and Writing—review & editing [supporting])