Introduction

Primary bilateral macronodular adrenocortical hyperplasia (PBMAH) is a rare cause of ACTH-independent Cushing’s syndrome. It is usually identified by adrenal imaging demonstrating bilateral enlarged adrenal masses (symmetric and asymmetric) [1] composed of multiple bilateral macronodules (>10 mm) with hyperplasia and/or internodular atrophy. Rarely, the PBMAH phenotype is associated with primary aldosteronism and hyperandrogenism. The prevalence of PBMAH is probably underestimated, as bilateral adrenal masses are increasingly identified through imaging performed for reasons unrelated to adrenal disease [2]. Along current guidelines for such incidentalomas assessment of endocrine activity is required [3]. The clinical phenotype of Cushing’s syndrome due to PBMAH is usually mild [4]. The optimal treatment for PBMAH with endocrine activity is the subject of ongoing debate. According to the guidelines of the Endocrine Society, first-line treatment of PBMAH with overt Cushing’s syndrome should be bilateral adrenalectomy with consecutive remission and lifelong requirement for steroid replacement therapy [5]. More recently, several studies indicated the effectiveness of unilateral adrenalectomy in bilateral disease [6,7,8,9,10,11]. Our group has analyzed the long-term outcome after unilateral adrenalectomy for PBMAH and showed that unilateral adrenalectomy leads to clinical remission and a lower incidence of adrenal crisis [12]. It also appeared from this publication that some patients may have persistent hypercortisolism to a clinically relevant degree and potentially higher mortality. However, this has been questioned [13].

The decision of which adrenal to be removed in PBMAH, is generally based on adrenal size. This assumes that a larger adrenal mass causes higher glucocorticoid output. However in endocrinology, morphology does not necessarily predict function. Further methods to evaluate potential lateralization of cortisol production included functional radiotracer imaging to evaluate the adrenal uptake of 131I-methylnorcholesterol [7], [14,15,16,17], but 131I-methylnorcholesterol is no longer available in most countries. Another approach is adrenal venous sampling (AVS) which is commonly used in the diagnostic work-up for primary aldosteronism [18]. AVS might be useful in evaluating the predominant side of cortisol production and may guide the decision for unilateral adrenalectomy in bilateral disease. Four studies with different methods and definitions have evaluated AVS in the context of autonomous cortisol secretion and uni- or bilateral adrenal incidentalomas in small patient groups [17, 19,20,21]. In addition, several case reports addressed this topic [22,23,24,25,26,27,28,29]. In contrast, patients with PBMAH have not been systematically evaluated. Here we report our experience with AVS in 16 patients with PBMAH from three centers. The aim of this study was to define the relevant clinical gain of information through AVS to identify the hormonally dominant side.

Methods

Patients

We performed a retrospective analysis of patients with primary adrenal Cushing’s syndrome and bilateral masses suspicious for PBMAH who underwent AVS to identify the hormonally dominant side of cortisol production. Patients were treated at the endocrine tertiary care centers of the Klinikum der Universität München, Ludwig-Maximilians-Universität Munich, Germany (12 patients), the University hospital of Würzburg, Germany (2 patients) and the University hospital of the University Graz, Austria (2 patients) between 2006 and 2020. Thereby, a total of 16 patients were included. Referral to the centers was triggered by clinical suspicion of Cushing’s syndrome (N = 11) or due to incidental detection of bilateral adrenal masses (N = 5). Patients underwent evaluation for Cushing’s syndrome following current guidelines [5]. The German Cushing’s Registry was approved by the LMU ethics committee (Project number: 152 -10), and all patients gave written informed consent.

Adrenal venous sampling (AVS)

We performed AVS in patients with bilateral adrenal masses based on an individual decision of the treating endocrinologists (until 2012) or based on a multidisciplinary endocrine board decision (since 2012) to evaluate and identify the predominant side of cortisol production. Main criteria to proceed with AVS were radiological asymmetry and patient’s willingness to undergo unilateral adrenalectomy.

Peripheral blood samples were collected simultaneously with each of the selective blood samples. AVS was performed without concomitant infusion of ACTH or dexamethasone suppression.

In addition to cortisol, further adrenocortical and –medullary hormones were measured concomitantly as “reference hormones” to verify catheter placement in the adrenal veins and to adjust for dilution. Throughout the 14 years of observation period different reference hormones were used. The choice was based on the technical capability, the state of science and knowledge about PBMAH at that time, and the center’s experience derived from AVS for subtyping of primary aldosteronism. The following hormones were used alone or in combination: Aldosterone (n = 13), metanephrine and normetanephrine (n = 6), DHEA-S (n = 4), or androstenedione (n = 2).

For further calculations, the result of one basal sample was used, as the results of additional samples, if available, were synonymously. We post hoc defined successful adrenal vein sampling as a selectivity index (SI) ≥ 2.0: absolute cortisol concentration or “reference hormone” concentration in the adrenal veins at least twice as high than the concentration in the simultaneously drawn peripheral sample. Lateralization of cortisol production was evaluated by the absolute cortisol ratios or the ratio of the cortisol to “reference hormone” ratio from both sides:

$${{{\mathrm{Lateralization}}}}\,{{{\mathrm{index}}}}\,\left( {{{{\mathrm{LI}}}}} \right) = \frac{{\left[ {{{{\mathrm{Cortisol}}}}/{{{\mathrm{reference}}}}\,{{{\mathrm{hormone}}}}} \right]{{{\mathrm{right}}}}\,{{{\mathrm{or}}}}\,{{{\mathrm{left}}}}}}{{\left[ {{{{\mathrm{Cortisol}}}}/{{{\mathrm{reference}}}}\,{{{\mathrm{hormone}}}}} \right]{{{\mathrm{left}}}}\,{{{\mathrm{or}}}}\,{{{\mathrm{right}}}}}}$$

As no consensus on the definition of lateralization in this context exists, the AVS was interpreted in analogy to primary aldosteronism. A predominant cortisol production was usually assumed with a LI ≥ 4 [30]. When the concentrations of the measured hormones were below the lower limit of quantification of the assay, half of the lower limit of quantification was used for calculation. For measured concentrations above the upper limit of quantification, the upper limit of quantification was used for calculation.

Biochemical and genetic assessment and assays

Biochemical and genetic assessment and assays are described extensively in the online supplement [31].

Adrenal volume

Adrenal volume was measured three-dimensionally based on the approach of a cuboid by width* depth*height in cm3. These measurements were performed by a single radiologist (L.A.). Asymmetry in adrenal size was defined by a difference of more than 30% in volume. In two patients only axial planes were available. Therefore, in these two patients adrenal size was assessed by the maximum diameter in cm and asymmetry defined by a difference of more than 30%.

Treatment

Recommendation for unilateral or bilateral adrenalectomy was based on the severity of Cushing’s syndrome (clinically and biochemically), the results of radiologic studies and AVS results. Remission after surgery was defined clinically by improvement of catabolic or metabolic features of CS (“clinical remission”) and biochemically by postoperative adrenal insufficiency requiring replacement therapy and/or normalization of all three screening tests for hypercortisolism (“complete biochemical remission”). Partial biochemical remission referred to normalized 24 h urinary free cortisol (UFC), but abnormal 1-mg dexamethasone suppression test (1.8–5 µg/dl) and late night salivary cortisol; biochemical persistence was defined by three abnormal screening tests for hypercortisolism. After surgery the diagnosis of PBMAH was confirmed by an experienced pathologist (T.K.) by the presence of multiple macronodules (>1 cm) and diffuse hyperplasia or atrophy on adrenal specimens.

Statistics

Pearson’s correlation coefficients (R) were calculated for correlations of adrenal size with overall/bilateral biochemical activity measured as 24 h UFC. IBM SPSS Statistics (version 21.0, IBM North America) was used for statistical analysis and calculation of Pearson’s correlation coefficients and p-values. MS Excel (version 2013) was used to create scatter graphs.

Results

Sixteen patients were retrospectively analyzed (15 female, 1 male, mean age at diagnosis: 59 ± 9 years). Fourteen patients had clinically overt Cushing’s syndrome. Two patients had mild Cushing’s syndrome with abnormal screening tests and metabolic features, but no catabolic features (e.g., myopathy, skin atrophy, osteoporosis) of Cushing’s syndrome. The presentation of each patient is summarized in Table 1 in the online supplement [31]. Genetic testing for ARMC5 was performed in 13 patients and revealed single nucleotide variants in 6 patients.

Adrenal size

Mean adrenal volume was 76.5 ml, ranging from 9.5 to 852 ml. 11 of 16 patients had asymmetric adrenal size, equally distributed on the left (n = 6) and right (n = 5) adrenal gland. The mean left adrenal volume was 99 ml (range: 11–852 ml, total: 1383 ml) and the mean right adrenal volume was 51 ml (range: 9–137 ml, total: 719 ml).

Adrenal vein sampling (AVS)

The detailed hormone assessments for each patient are provided in Table 1. Based on the biochemical criteria of absolute cortisol values, all 16 AVS procedures were technically successful. Using the SI calculated on the basis of reference hormones, bilateral catheterization was successful in 13 of 16 (81%) patients, and unilaterally successful in the remaining three patients. The mean SI of all patients on the basis of the absolute cortisol concentrations was 11.3, and that of the reference hormones 17.9.

Table 1 Biochemical results from AVS
Full size table

Cortisol lateralization indices ≥4.0 were present in 7 of 16 patients (reference hormone based LI) and 6 of 16 (based on absolute cortisol levels). In total, in 8 of 16 patients the AVS was defined as lateralized (7 patients based on reference hormone LI plus 1 patient based on absolute cortisol LI). In 4 of these 8 patients LI’s were contrary for reference hormone based LI and absolute cortisol based LI (Patients 2, 7, 10, 13; Table 1). In patients in whom lateralization was present, it occurred toward the larger adrenal in 5 patients, toward the smaller adrenal in no patient, and toward one side in 3 patients with equally enlarged adrenal. In the latter three patients, AVS showed maximum lateralization indices of 11.8, 14.8 and 4.4 (Patients 2, 8 and 11; Table 1), respectively.

Treatment and outcome

Decision on treatment was based on asymmetry in adrenal size and on lateralization in AVS. Table 2 shows for each patient the interpretation of AVS, the applied surgery and postoperative results (pathology and biochemical assessment). Figure 1 shows the clinical and biochemical outcome separated into patients with morphological asymmetry and with adrenals of equal size.

Table 2 Interpretation of AVS, applied surgery, and postoperative results (pathology and biochemical assessment)
Full size table
Fig. 1
figure 1

Decision diagram. Definitions of remission: Clinical remission: Improvement of catabolic or metabolic features of CS. Biochemical remission: Adrenal insufficiency and/or normalization of screening tests for hypercortisolism. AVS: Adrenal venous sampling, Unilat: Unilateral, Bilat: Bilateral, ADX: Adrenalectomy

Full size image

Five patients had both asymmetry in adrenal size and lateralization on AVS (always toward the radiologically larger side). One patient underwent bilateral adrenalectomy with consecutive clinical and biochemical remission (patient 5). Four patients underwent unilateral adrenalectomy: Three patients achieved both clinical and biochemical remission (patients 7, 9, and 13). The remaining patient achieved neither clinical nor biochemical remission and was therefore treated with metyrapone (patient 10).

Six patients had asymmetry in adrenal size, but no lateralization in AVS: One of these patients received medical treatment with ketoconazole (patient 14). Two of these patients underwent bilateral adrenalectomy with consecutive clinical and biochemical remission (patients 4 and 15). The three remaining patients underwent unilateral adrenalectomy of the larger adrenal gland: Two of these patients achieved both clinical and biochemical remission after surgery (patients 1 and 12). The remaining patient (patient 3) achieved clinical, but no biochemical remission.

Three patients had no asymmetry in adrenal size but lateralization in AVS: All three patients underwent unilateral adrenalectomy based on AVS results. Two patients are lost to follow-up (patients 2 and 8) and no postoperative information is available. The other patient had clinical and biochemical remission (patient 11).

Two patients had neither asymmetry in adrenal size nor lateralization in AVS: One of these patients underwent (complete) bilateral adrenalectomy (patient 16) and the other partial bilateral adrenalectomy (left side: complete, right side ½ removal, patient 6). Both patients achieved clinical and biochemical remission.

In summary, outcome of patients with asymmetric enlargement of adrenals, who underwent unilateral adrenalectomy, were similar regardless of lateralization seen in AVS prior to surgery.

Correlation between urinary free cortisol and adrenal volume

Correlation of the 24 h UFC as a marker of overall biochemical activity with the combined bilateral volume from both adrenals from the same patient showed an excellent correlation (Fig. 2: R = 0.88, p < 0.001, N = 13). The relationship was strongly influenced by one case (to the far right in Fig. 2). Without this case, the R coefficient was 0.45 (p = 0.143). This indicates that the volume of adrenals in patients with PBMAH may be one of the main determinants of cortisol production.

Fig. 2
figure 2

Correlation of 24 h urinary free cortisol (UFC) with the volume of both adrenals from the same patient. The X-axis is shown as a logarithmic scale. Triangles represent the patients with ARMC5 variants (N = 3)

Full size image

Discussion

In this study in PBMAH patients with Cushing’s syndrome we provide evidence that AVS is able to detect lateralization of adrenal cortisol production in roughly 50% of patients. In 5 out of 8 patients, lateralization was observed toward the radiologically larger adrenal gland, while the remaining patients had equally enlarged adrenals. In no case lateralization was found toward the smaller gland. In addition, correlation of adrenal size and 24 h UFC (as a marker of overall biochemical activity) indicates that size of the adrenals may be one of the main determinants of cortisol production. Therefore, we conclude that the gain of information through AVS in the context of Cushing’s syndrome due to PBMAH with marked asymmetry of the adrenals is limited, since AVS does not provide additional information on top of radiological imaging. Importantly, the surgical outcomes were similar in patients without AVS gradients compared to those with gradients (Fig. 1). This conclusion concerns only patients with asymmetric enlargement, since our series was too small for patients with symmetric enlargements to draw conclusions.

AVS is mainly used for the diagnostic work-up in primary aldosteronism. In adrenal incidentalomas with autonomous cortisol production four studies have evaluated the usefulness of AVS, and all of them judge AVS as a helpful diagnostic procedure.

In all of these studies dexamethasone was administered prior to AVS to induce full ACTH suppression. As our patients had ACTH-independent symptomatic Cushing’s syndrome (n = 14) or autonomous cortisol production with low or suppressed ACTH levels, dexamethasone was not administered prior to AVS.

One of the main differences in the studies from the literature (Table 3) are the different definitions of correct drainage and lateralization: In some of these studies absolute cortisol concentrations in the adrenal veins were used for interpretation of AVS. At variance, in primary aldosteronism, measurement of a non-target hormone (cortisol) is used in AVS protocols to adjust the target hormone aldosterone, for two reasons: Firstly, hormonal verification of correct drainage is obligatory as drainage of especially the right adrenal vein can be very challenging [32]. Secondly, adjustment for dilution effects of vena cava blood is necessary in order to compare adrenal cortisol concentrations from both sides. In our study 12 of 16 patients had the same interpretation of the AVS comparing absolute cortisol levels with adjusted cortisol lateralization indices. We followed mainly the adjusted cortisol levels. However, the AVS from patient 10 was considered as lateralized because absolute cortisol levels showed almost sevenfold higher levels in the right adrenal vein (compared to a reference hormone based LI of 2.6).

Table 3 Literature review
Full size table

Which hormone to use for defining correct drainage and adjust for dilution is challenging to decide and also the main limitation of our study. Due to the long-lasting time interval of our retrospective analysis and changes in principal investigators during this time the reference hormones have changed over the years. In fact, a potpourri of the adrenal hormones (aldosterone, DHEA-S and androstenedione as androgens and the adrenomedullary hormones metanephrine and normetanephrine) was used for this purpose. During the last years the scientific advances increased significantly, including the discovery of biallelic inactivation of the ARMC5 gene necessary for the growth of adrenal nodules, paracrine ACTH secretion and the autonomous macronodules inside a PBMAH adrenal gland leading to different endocrine phenotypes [33,34,35,36]. Moreover, androgens have a quite long half-time and therefore do not seem to perform superior. The use of DHEA-S as a reference hormone is questionable, as in our study the measured levels of DHEA-S were very low and often close to or below the lower limit of detection. Due to its origin from the medulla and its presumed independency from fluctuations due to stress (in contrast to epinephrine) the use of metanephrine as a reference hormone may be superior. In the setting of primary aldosteronism the measurement of metanephrines has been recommended [37]. With reference to AVS in ACTH-independent Cushing’s syndrome aldosterone itself may also serve as reference hormone, when excess of aldosterone due to PBMAH is excluded.

Most of our patients had clinically overt Cushing’s syndrome and further diagnostic work-up revealed the suspicion of PBMAH. In contrast, in the previous studies mentioned above, AVS was mainly performed in patients with bilateral adrenal incidentaloma and endocrine activity was usually limited and defined as subclinical Cushing’s syndrome or mild autonomous cortisol secretion. A direct comparison is therefore not possible. With respect to PBMAH, lateralization in AVS was mainly to the radiologically larger adrenal gland. We, therefore, argue against routine use of AVS in asymmetric PBMAH. Whether it may be of use in patients with symmetric PBMAH needs to be analyzed in a series of larger size than ours.

Limitations

Co-secretion of aldosterone was not assessed systematically, therefore interpretation of AVS based on aldosterone-corrected cortisol values could be confounding. Main limitation of our study is the retrospective analysis and the long time span. During this time individual decisions for treatment did not follow a predefined protocol as also medical staff changed. This applies also to the diagnostic use of the AVS and the variety of reference hormones used. Unfortunately, of the small number of patients in this study some patients were lost to follow-up. Therefore, evaluation of the postoperative outcome could not be assessed in all patients. Nevertheless, we think that our data is robust enough to derive our main conclusions from it.