Adrenal adenomata displaying mild autonomous cortisol secretion: a service evaluation of cardiometabolic profile routinely screened patients

Abstract

Introduction

A proportion of adrenal adenomata exhibit autonomous cortisol secretion, now termed mild autonomous cortisol secretion (MACS), as defined by post-1mg overnight dexamethasone suppression test (ONDST) cortisol 51–137 nmol/l. Here, we characterized the cardiometabolic profile of MACS.

Methods

Clinical records of 98 individuals with adrenal adenomata were examined. Subcategorization into MACS1 (ONDST cortisol of 50–137 nmol/l) and MACS2 (ONDST cortisol of >137 nmol/l was created to take account of individuals with ONDST cortisol of more than 137 nmol/l and no diagnosis of Cushing’s syndrome.

Results

Diagnosis of MACS1 associated with a higher diagnosis rate of cardiovascular disease (CVD) (17.7% MACS1) vs. non-MACS = nonfunctioning adenoma (NFA) (3.7%) (P = 0.009) and higher rates of prescription of lipid-lowering agents (51.6%) vs. (29.6%) (P = 0.01). ONDST cortisol levels in MACS1 patients correlated with a more adverse lipid profile (for higher low-density lipoprotein cholesterol, r² = 0.404, P = 0.007; high-density lipoprotein cholesterol r² = –0.346, P = 0.023; for higher serum triglycerides r² = 0.282, P = 0.02) in spite of higher rates of statin prescribing. There was a gradient of increasing numbers of antihypertensives prescribed, going from non-MACS NFA to MACS1 to MACS2. Dunn’s post hoc analysis indicated an overall more adverse lipid profile in MACS1.

Conclusion

The positive direction of associations between serum cortisol and lipid measures highlights that MACS carries a metabolically adverse lipid profile. Diagnosis of MACS was associated with a higher diagnosis rate of CVD and appropriately higher rates of prescription of lipid-lowering agents and a greater number of antihypertensive agents prescribed. The question remains about whether a specific directed treatment of MACS should be offered beyond risk-factor-mitigating management.

Introduction

The increasing use of high-resolution imaging has led to a rising wave of detection of adrenal incidentalomas, with a proportion of these exhibiting autonomous cortisol secretion [1]. This condition, now termed mild autonomous cortisol secretion (MACS), is defined biochemically by a post-1 mg overnight dexamethasone suppression test (ONDST) cortisol level 50 nmol/l or more (there is variation by assay platform), in the absence of the overt physical stigmata of classic Cushing’s syndrome [2].

Once considered a subclinical finding, MACS is now recognised as a clinically significant state that profoundly impacts cardiometabolic health [3]. A substantial body of evidence has demonstrated that the chronic, low-grade hypercortisolism inherent to MACS is a direct driver of multiple adverse outcomes. Compared with patients with nonfunctioning adrenal tumours, those with MACS have a significantly higher prevalence of hypertension, type 2 diabetes (T2D) and dyslipidaemia [4,5]. The risk of developing T2D can be nearly double, while meta-analyses have shown that people with MACS are twice as likely to experience a new major cardiovascular event [6]. The formal recognition of this heightened risk by major endocrine societies globally has underscored the need for active clinical surveillance of this patient population [2].

While national and international data provide a robust framework for understanding these risks, there is a need to quantify the local burden of disease to ensure that clinical services are aligned with patient needs. This must be seen in the context that socioeconomic background and situation, weight gain and coagulative disorders influence cardiovascular risk.

This evaluation aimed retrospectively to analyse all patients who had undergone an ONDST at one centre over a 3-year period. The primary objective was to identify the group of individuals with adrenal adenomata and associated MACS, and comprehensively to characterise their cardiometabolic profile.

By quantifying the local prevalence of MACS-associated hypertension, T2D and dyslipidaemia, this evaluation aimed to provide essential data to inform and optimise local management, guide surveillance pathways and that ensure resources are appropriately allocated to manage this patient group.

Methods

We undertook a service evaluation of the cardiometabolic profile of adrenal adenomata presenting over a period of 3 years at a single centre. All had a 9-a.m. serum cortisol, postmidnight dexamethasone of greater than 50 nmol/l (in an ONDST) which is deemed a ‘fail’ on basis of serum cortisol level on our assay.

All individuals had been seen in a specialist endocrinology clinic and had a radiologically proven adrenal adenoma. As this was a service evaluation, no patients with an adrenal adenoma were excluded.

The clinical records of the patients with an adrenal adenoma who had undergone an ONDST between 6 May 2022 and 22 May 2025 were examined by clinicians from the usual care team. Women taking oestrogen containing medication (combined oral contraceptive pill or oestrogen-containing hormone replacement therapy) were asked to omit this medication for at least 1 month before the test.

Serum cortisol was measured on the Siemens Atellica autoanalyzer (Manchester, UK). The definition of a ‘pass’ on this assay is a 9-a.m. postmidnight dexamethasone cortisol level of less than 50 nmol/l [7].

Data on current health status regarding current and previous diagnoses and up-to-date blood tests of relevance to cardiometabolic profile were recorded as was current weigh, BMI and blood pressure (BP).

Individuals taking glucocorticoids and drugs affecting CYP3A4 metabolism were excluded from the analysis.

Ethical permission was not sought, as this was a service evaluation. All data were obtained from the hospital electronic patient record and were deidentified prior to statistical analysis.

Statistics

All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 29.0 (IBM Corp., Armonk, NY, USA). Continuous variables were inspected for normality using the Shapiro–Wilk test and visual assessment of histograms. Normally distributed variables are presented as mean ± SD, and nonnormally distributed variables as median (interquartile range). Categorical variables are summarised as frequencies and percentages.

For comparisons between the three diagnostic groups [nonfunctioning adenoma (NFA), MACS1 and MACS2], a one-way analysis of variance was used for normally distributed variables, and the Kruskal–Wallis test for nonnormally distributed variables. Post hoc pairwise comparisons were performed using Bonferroni correction for analysis of variance and Dunn’s test for Kruskal–Wallis, with adjustment for multiple testing.

Categorical variables were compared across the three ONDST groups using Pearson’s chi-square test of independence. Where expected cell counts were less than five, Fisher’s exact test was applied. Effect sizes were assessed using Cramer’s V, and statistical significance was set at a P value less than 0.05.

Correlations between ONDST cortisol concentrations and continuous metabolic parameters were assessed using Pearson’s correlation coefficient (r) for normally distributed variables and Spearman’s rank correlation coefficient (ρ) for nonnormally distributed variables. A two-tailed P value less than 0.05 was considered statistically significant.

Results

The clinical records of 98 individuals with adrenal adenomata were examined. Size varied from 2–20 mm. Non-contrast Hounsfield units varied from 2 to 10 Hounsfield units. Cushing’s syndrome was excluded. All other potential adrenal-related disorders were included.

Of the 98 individuals with an adrenal adenoma, 27 (28%) suppressed serum cortisol on the ONDST (NFA), 62 (63%) had an ONDST cortisol of 50–137 nmol/l in keeping with MACS and 9 (9%) had an ONDST cortisol of greater than 137 nmol/l. A subcategorization into MACS1 (ONDST cortisol of 50–137 nmol/l) and MACS2 (ONDST cortisol of >137 nmol/l) was created to take account of those individuals with ONDST cortisol of greater than 137 nmol/l with no evidence of Cushing’s syndrome on subsequent investigation.

Median age was similar across those with a normal ONDST cortisol, MACS1 patients and MACS2 patients, mean age of the MACS1 patients being 64.0 years old. 62.9% of MACS1 patients were women. Mean age was similar across the diagnostic groups.

Median ONDST cortisol was 30 nmol/l for those with a normal ONDST result (NFA) and 66 nmol/l for those designated as MACS1 (Table 1).

Table 1.

Clinical characteristics: values are median (interquartile range)

Variable	NFA (n = 27)	MACS1 (n = 62)	MACS2 (n = 9)
ONDST cortisol (nmol/l)	30 (21.8–41.5)	66 (57–82.8)	201 (186.5–284.0)
Age (years)	61 (56–67)	64 (58–73)	58 (43–69)
BMI (kg/m2)	29.5 (27.0–34.4)	29.0 (26.3–33.5)	28.7 (21.2–32.2)
HbA1c (mmol/mol)	42 (38–50)	41 (37.5–50.5)	41 (35–73)
Systolic BP (mmHg)	128 (120–140)	135 (122.8–153.3)	133 (113.5–141.5)
Diastolic BP (mmHg)	78 (70–83)	80 (70.8–89)	80 (67–96.5)
TC (mmol/l)	5.20 (4.70–6.00)	4.40 (3.60–5.00)	5.10 (3.90–5.50)
HDL (mmol/l)	1.41 (1.20–1.64)	1.46 (1.11–1.80)	1.47 (1.02–1.72)
LDL cholesterol (mmol/l)	3.20 (2.70–3.80)	2.60 (2.00–3.30)	3.40 (2.80–3.80)
Cholesterol/HDL ratio	3.48 (3.03–4.90)	2.97 (2.50–3.50)	3.55 (3.23–4.82)
Non-HDL cholesterol (mmol/l)	3.67 (3.31–4.32)	3.00 (2.30–3.50)	3.70 (3.00–4.40)
Triglycerides (mmol/l)	1.80 (1.40–3.00)	1.40 (0.92–1.90)	1.70 (1.40–2.60)
Sex, female (%)	15 (55.6)	39 (62.9)	7 (77.8)
CVD (%)	1 (3.7)	11 (17.7)	2 (22.2)
Hypertension (%)	16 (59.3)	35 (56.5)	4 (44.4)
T2D (%)	11 (40.7)	23 (37.1)	2 (22.2)
Treatment: surgical (%)	1 (3.7)	7 (11.3)	1 (11.1)
On lipid-lowering therapy (%)	8 (29.6)	32 (51.6)	2 (22.2)
On antiplatelet therapy (%)	3 (11.1)	10 (16.1)	0 (0)
On antihypertensive therapy (%)	15 (55.6)	40 (64.5)	4 (44.4)

Number of antihypertensives (median, interquartile range): 1.1 (1–1.3); 1.8 (1.5–2.1); 2.2 (1.9–2.4).

Agents when prescribed.

BP, blood pressure; CVD, cardiovascular disease; HbA1c, glycated haemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MACS1, mild autonomous cortisol secretion with overnight dexamethasone suppression test cortisol 51–137 nmol/l; MACS2, mild autonomous cortisol secretion with overnight dexamethasone suppression test cortisol >137 nmol/l; NFA, nonfunctioning adenoma; ONDST, overnight dexamethasone suppression test; TC, total cholesterol; T2D, type 2 diabetes.

Although the MACS1 and MACS2 groups demonstrated numerically higher rates of cardiovascular disease (CVD) and a greater use of lipid-lowering therapy and antihypertensive treatment, compared with the ‘normal’ group, these differences did not reach statistical significance, likely reflecting subgroup sizes, particularly in MACS2.

BMI and BP (both systolic and diastolic) were similar across the groups. Median BP for patients with MACS1 was 135/80 mmHg. There was a higher recorded diagnosis of CVD in MACS1 patients (17.7%; P = 0.009) and MACS2 patients (22.2%; P = 0.007) than those with an adrenal adenoma and normal ONDST result at 3.7%, with the diagnosis of T2D made in 37.1% of MACS1 versus 40.1% of adrenal adenomata with a normal ONDST cortisol result. Median glycated haemoglobin was similar in all three groups (Table 1).

Prescription of lipid-lowering treatment (92% were statins) was higher in the MACS1 group than the other two groups at 51.6% for MACS1 versus 29.6% for adrenal adenomata with normal ONDST results. In spite of this, MACS2 patients had a higher cholesterol/high-density lipoprotein (HDL) cholesterol ratio, low-density lipoprotein (LDL) cholesterol and serum triglyceride level than MACS1 patients (Table 1). There was a gradient of increasing numbers of antihypertensives prescribed going from non-MACS NFA to MACS1 to MACS2.

A multivariable logistic regression was performed. The model included age, sex and BMI as covariates. Hypertension, diabetes and lipid-lowering therapy were not included as confounders, as these variables lie on the causal pathway between cortisol and CVD. ONDST cortisol concentration was positively associated with CVD, although this did not quite reach statistical significance [odds ratio (OR): 1.08 per 10 nmol/l increase, 95% confidence interval (CI): 0.99–1.18, P = 0.10]. Age was the only significant predictor, with each additional year increasing the odds of CVD by 10.6% (OR: 1.11, 95% CI: 1.03–1.19, P = 0.005). Female sex was associated with lower odds of CVD (OR: 0.30, 95% CI: 0.07–1.17, P = 0.08), but this was not statistically significant. Model fit was acceptable (Hosmer–Lemeshow P = 0.67), and the overall model was statistically significant (χ² = 13.9, P = 0.008). These results support the trend observed in the categorical analyses, in which higher cortisol levels are associated with increased CVD risk.

Of those patients with a diagnosis of MACS1, 7/62 (11%) went on to unilateral adrenalectomy. Following adrenalectomy, LDL cholesterol was lower in the surgically treated MACS1 versus conservatively managed MACS group, as was non-HDL cholesterol (Table 2).

Table 2.

Lipid profile in patients with mild autonomous cortisol secretion with overnight dexamethasone suppression test cortisol of 51–137 nmol/l by treatment type

Variable	Conservative (n = 52)	Medicala (n = 3)	Surgical (n = 7)	P value
TC (mmol/l)	4.40 (3.60–5.00)	4.10 (3.80–4.50)	4.30 (3.90–4.70)	0.098
LDL cholesterol (mmol/l)	2.60 (2.00–3.30)	2.40 (1.80–2.80)	1.90 (1.70–2.40)b	0.047
Non-HDL cholesterol (mmol/l)	3.00 (2.30–3.50)	2.90 (2.40–3.10)	2.40 (2.00–2.80)b	0.059
HDL cholesterol (mmol/l)	1.46 (1.11–1.80)	1.40 (1.20–1.50)	1.50 (1.30–1.70)	0.346
Triglycerides (mmol/l)	1.40 (0.92–1.90)	1.20 (0.80–1.50)	1.30 (0.90–1.70)	0.714

HDL, high-density lipoprotein; LDL, low-density lipoprotein; TC, total cholesterol.

^aRefers to block/replacement with metyrapone and hydrocortisone.

^bSignificantly different from conservative group, Dunn’s post hoc P < 0.05.

Regarding the relation between ONDST cortisol and lipid parameters, a higher ONDST cortisol in MACS patients correlated in univariate analysis with: lower HDL cholesterol (r = 0.404, P = 0.007), higher cholesterol/HDL cholesterol ratio (r = 0.346, P = 0.023), higher serum triglycerides (r = 0.282, P = 0.067) and prescription of fewer antihypertensive agents and glucose-lowering drugs (r = –0.270, P = 0.034 and r = –0.270, P = 0.034, respectively).

Although some lipid parameters, including LDL cholesterol, showed lower medians in MACS1 compared with the normal ONDST cortisol group, Dunn’s post hoc analysis indicated significantly higher values in MACS1 (Table 3). This suggests that the distributional shifts captured by the rank-based tests may better reflect the underlying impact of cortisol excess on lipid metabolism.

Table 3.

Significant post hoc pairwise comparisons

Variable	Groups	Adjusted P value	Direction
TC	MACS1 vs. normal	0.004	MACS1 higher
LDL cholesterol	MACS1 vs. normal	0.011	MACS1 higher
Cholesterol/HDL ratio	MACS1 vs. normal	0.017	MACS1 higher
Non-HDL cholesterol	MACS1 vs. normal	0.002	MACS1 higher
Triglycerides	MACS1 vs. normal	0.026	MACS1 higher

HDL, high-density lipoprotein; LDL, low-density lipoprotein; MACS1, mild autonomous cortisol secretion with overnight dexamethasone suppression test cortisol 51–137 nmol/l; TC, total cholesterol.

The MACS1 group displayed a wider distribution of lipid values with a greater proportion of patients with elevated lipid values, even though the central tendency (median) was slightly lower. The rank-based approach (Kruskal–Wallis and Dunn’s procedures) is more able to captured the distribution/skewness.

Discussion

The diagnosis of MACS1 and MACS2 (as defined) was associated with a higher diagnosis rate of CVD and in the case of MACS1 with higher rates of prescription of lipid-lowering agents. ONDST cortisol levels in MACS patients (both MACS1 and MACS 2) correlated with a more adverse lipid profile, in keeping with the metabolic impact of hypercortisolaemia. Taken together with the positive direction of the correlations observed between serum cortisol and lipid measures, in spite of prescription of lipid-lowering treatment, these findings highlight that the MACS1 group carries a metabolically adverse lipid profile despite median values appearing lower. Rates of T2D and hypertension were similar in MACS1 and NFA individuals with a gradient of increasing numbers of antihypertensives prescribed going from non-MACS NFA to MACS1 to MACS2.

MACS is characterised by adrenocorticotropic hormone-independent autonomous cortisol secretion, with MACS presenting without the clinical features of full-blown Cushing’s syndrome [2].

Some individuals with a nonfunctioning adrenal adenoma can also have a subtle degree of autonomous cortisol secretion and can progress to MACS over time. Progression from MACS to Cushing’s syndrome is exceptionally rare, except in the context of germline mutations causing primary bilateral macronodular adrenal hyperplasia, where the gradual increase in the size and number of adrenal nodules result in higher autonomous cortisol secretion [8].

By considering the entire distribution, the nonparametric tests used in the analysis more accurately highlighted the adverse lipid burden associated with MACS1. This is consistent with the correlation analysis between ONDST cortisol and lipid values as well as with the higher prevalence of CVD and a greater use of lipid-lowering therapies in the MACS1 group. This explains why statistical reliance purely on median values can underestimate the effect of cortisol excess on lipid metabolism.

Recent European Consensus Guidance [2] has recommended that all patients with MACS should be screened for potential cortisol-related comorbidities that are potentially attributable to cortisol such as hypertension and T2D to ensure these are appropriately treated [9]. It was also recommended that patients with MACS who also have relevant comorbidities surgical treatment should be considered in an individualised approach [10], with any decision about surgical intervention guided by the likelihood of malignancy, the presence and degree of hormone excess, age, general health and patient preference.

Here, we did not report follow-up imaging. However, in a previous study, with a median follow-up of 65 (range: 7–288) months, adrenal adenomata demonstrated minimal enlargement (median growth rate: 2 mm/year) [10]. In the same study on hormonal evaluation, 10% individuals (5/50 tested) had an abnormal ONDST.

The question as to what follow-up should be put in place for people with MACS was addressed in the recent European Consensus Guidance document, which recommended that all patients with adrenal incidentaloma undergo a 1-mg ONDST to rule out cortisol excess [2], with annual reassessment of comorbidities potentially attributable to cortisol for those individuals not undergoing adrenalectomy. A follow-up ONDST may be indicated but the recent innovation of measurement of salivary cortisone instead of serum cortisol on the ONDST morning samples makes this potentially more straightforward [11,12].

Regarding adrenalectomy, an individualised approach was advised to take account of age, degree of cortisol excess, general health and comorbidities. It was recommended that in all patients considered for surgery, adrenocorticotrophic hormone independency of cortisol excess should be confirmed.

Also on the basis of our data, there is a clear cortisol-dependent effect on adverse lipid profile. We speculate that early and proactive medical/surgical control of hypercortisolaemia may be key to improving lipid and metabolic outcomes, ultimately reducing future CVD risk. We accept that isolated posttreatment cholesterol measurements need to be considered alongside prescription/deprescription behaviours. However, only a small number of patients in our cohort received medical or surgical treatment, and the duration of posttreatment follow-up was relatively short. Under these circumstances, deprescribing was not encountered as clinicians typically reconsider lipid therapy only after sustained biochemical improvement and re-evaluation of long-term cardiovascular risk.

Moreover, the primary goal of treating cortisol excess is not necessarily immediate deprescription of anti-lipid medication, as the long-term impact of metabolic memory and prior hypercortisolism on cardiovascular risk remains uncertain. Ultimately, demonstration of benefit will require long-term prospective follow-up to determine whether treatment of cortisol autonomy leads to meaningful reductions in major adverse cardiovascular events.

Hypercortisolism may represent a continuum of hormonal and metabolic abnormalities with varying severity [13]. Cardiovascular risk calculators may underestimate the true risk of a cardiovascular event in people with MACS [14], with the implication that consideration should be given to initiation of statin theory in anyone determined to have MACS [14,15].

The likely next step in terms of management of the MACS adenomata, particularly if there is a level of post dexamethasone 9-a.m. cortisol above 80 nmol/l will be the administration of osilodrostat [16]. This agent is still currently undergoing a clinical trial and, therefore, cannot be formally recommended. Nevertheless, the preliminary data are promising for this indication.

Limitations/strengths

We accept that this is a relatively small case series and that numbers of MACS2 patients were too low to draw any specific conclusions. However, we are able to access the health records of all identified individuals who were consecutive presentations at our clinic, following identification of an adrenal adenoma in consecutively presenting individuals.

We accept that we have used the ONDST 9-a.m. cortisol as our measure here to determine possible adrenal overactivity. However, this is still a usual practice in the UK and most other countries. Our study represents the experience of a single-centre and predominantly Caucasian population which may limit the generalizability of our observations to a broader population. A larger comprehensive NW England Audit is underway.

In conclusion, the diagnosis of MACS was associated with a higher diagnosis rate of CVD and appropriately higher rates of prescription of lipid-lowering agents and a greater number of antihypertensive agents prescribed.

A recent systematic review and meta-analysis emphasised the relevance of MACS. Both cardiometabolic morbidities and mortality were increased in patients with MACS compared with patients with nonfunctioning adrenal incidentaloma [17]. ONDST cortisol levels in MACS patients correlated with a more adverse lipid profile, in keeping with the dysmetabolic impact of hypercortisolaemia. This empathises the importance of vigilance and regular cardiometabolic monitoring in people with MACS.

Acknowledgements

C.S. and S.J. collected the data for analysis. C.S., S.J., Ad.S., A.H. wrote the manuscript with assistance from B.K., R.M., W.M., R.K., F.H., D.M., A.A.F. and I.L. Data analysis was conducted by Ad.S. F.H. provided senior review as did B.K., Ak.S. and A.A.F. who provided context from a laboratory science point of view.

The data that supports the findings of the study are available on reasonable request in fully anonymised form.

Conflicts of interest

There are no conflicts of interest.