Sex Difference in the Association of Osteoporosis and Osteopenia Prevalence in Patients with Adrenal Adenoma and Different Degrees of Cortisol Excess

Objective Osteoporosis and osteopenia (OS/OP) are frequent in patients with adrenal adenomas associated with cortisol excess (CE). However, the relationship between OS/OP and CE severity considering sex differences is unknown. Design A cross-sectional observational study from January 2006 to December 2015. Patients. 237 patients with adrenal adenoma associated with CE, including Cushing's syndrome and mild autonomous cortisol secretion (MACS), diagnosed in 10 referral centers in Japan. MACS was defined by 1 mg overnight dexamethasone suppression test (DST) cortisol level >1.8 μg/dL. Measurements. Prevalence of fragility fractures, medication for osteoporosis, and bone mineral density. Results In total, 112 of 237 patients, who were predominantly female (P < 0.001) and had lower BMI (P=0.013), had OS/OP. Patients with OS/OP was significantly affected by CE (P < 0.01) than those without. The adjusted odds ratio (OR) for predicting OS/OP was obtained in multivariate logistic regression analysis. Clinical measures of CE, 1 mg DST cortisol levels, were positively associated with OS/OP in total cases (OR 1.124, 95% CI: 1.070–1.181, P < 0.001) and the cases with MACS (OR 1.156, 95%CI: 1.046–1.278, P=0.005). A cutoff value of 1 mg DST cortisol level >5.0 μg/dL was associated with OS/OP differently between men and women. OS/OP risk in men with MACS was significantly affected only by 1 mg DST cortisol levels. However, OS/OP risk in women with MACS was significantly affected by 1 mg DST cortisol levels and age. Conclusions CE severity in adrenal adenoma is positively associated with OS/OP. However, the associated factors of OS/OP in the patients with MACS are different between men and women.


Introduction
Cushing's syndrome (CS) includes several subtypes of disease characterized by cortisol excess (CE). ACTH-independent CS consists of adrenal and iatrogenic CS. Adrenal CS is typically caused by CE originating from adrenocortical neoplasms, and the production of ACTH is suppressed by excess cortisol [1,2]. Suppressed ACTH also decreases the endogenous production of adrenal androgens [3,4]. Iatrogenic CS is induced by the therapeutic administration of synthetic glucocorticoids, causing suppression of endogenous ACTH and adrenal androgens. In contrast, ACTH-dependent CS, including Cushing's disease due to pituitary neoplasm and ectopic ACTH-producing tumors, is caused by excess ACTH. Adrenal androgens are also excessively produced in addition to the CE [3,4].
Osteoporosis is one of the well-known complications of CS. Approximately, 21% of patients with CS experience a fragility fracture before diagnosis [1]. However, the prevalence of osteoporosis and its association with endocrinological severity have previously been discussed mainly in patients with CS including various subtypes [1,5,6].
e clinical background of ACTH-independent and -dependent CS differs with regard to adrenal androgens [3]. Endogenous adrenal androgen is associated with the activation of osteoblasts or bone turnover [7,8].
e osteoporosis seen in adrenal CS has been speculated to be more influenced by CE, especially in women.
Recently, mild autonomous cortisol secretion (MACS), defined by a 1 mg overnight dexamethasone suppression test (1 mg DST) cortisol level >1.8 μg/dL but without the classical symptoms and signs of overt CS, has been shown to be prevalent, affecting up to 45% of patients with adrenal adenomas [9]. e indication for adrenalectomy in MACS is decided based on the severity of CE and coexistent complications, including glucose impairment, hypertension, obesity, and dyslipidemia. In diagnostic guidelines published by the Japan Endocrine Society, a 1 mg DST cortisol level >5.0 μg/dL is an indication for adrenalectomy. Patients with a 1 mg DST cortisol level of 3.0-5.0 μg/dL are recommended adrenalectomy based on the presence of complications [10]. In a crosssectional study with prospective enrollment, MACS was associated with a reduction in osteocyte function and number, with restoration after adrenalectomy [11]. ese data suggest that relatively mild ACTH-independent CE can be a risk factor for osteoporosis. However, how the severity of CE in adrenal adenoma differently affects osteoporosis and osteopenia (OS/OP) by sex has not been demonstrated at all.
In the present study, we conducted a retrospective crosssectional analysis to evaluate the sex differences in association between the severity of CE and OS/OP in patients with adrenal adenoma.

Study Populations.
is retrospective observational study was conducted as a part of the Advancing Care and Pathogenesis of Intractable Adrenal Disease in Japan (ACPA-J) study, which was based on a registration system for a cohort of patients with adrenal tumors. Patients aged 20 to 90 who were diagnosed with MACS and overt CS originated from adrenal adenoma between January 2006 to December 2015 in 10 referral centers, including the National Center for Global Health and Medicine, National Hospital Organization Kyoto Medical Center, St. Marianna University Yokohama City Seibu Hospital, Tottori University Hospital, Tokyo Medical and Dental University, Osaka University Graduate School of Medicine, Kyoto University, Kanazawa University Graduate School of Medical Science, Tenri Hospital, and Saiseikai Yokohama-shi Tobu Hospital, were enrolled. Because all participating centers accept patients considered the indication of adrenalectomy, the proportion of overt CS was higher than previous reports enrolling adrenal incidentaloma [9,12,13]. Patients with ACTH-dependent CS and iatrogenic CS were not included in the study. Patients with MACS and overt CS originated from adrenocortical carcinoma and adrenocortical hyperplasia were excluded from the analysis. e patients' clinical characteristics, biochemical examination results, radiological findings, and pathological diagnoses were collected electronically using a web registry system. e present study was conducted using a dataset validated in April 2019. e ACPA-J study was conducted in accordance with the Declaration of Helsinki and the guidelines for clinical studies published by the Ministry of Health, Labour, and Welfare, Japan.
e study protocol was approved by the Ethics Committee of the National Center for Global Health and Medicine and each institutional ethics committee of the participating centers. is study was registered at UMIN ID 000021437.

Definition of Cortisol Excess.
All participants were diagnosed on the basis of the diagnostic criteria for MACS published by Japan Endocrine Society and Cushingoid features [10], which were defined by the positivity of at least one sign, such as easy bruising, facial plethora, moon face, hirsutism, acne, central obesity, proximal myopathy, and striae. All patients with a 1 mg DST cortisol >1.8 μg/dL were enrolled in this study [14]. Of 292 patients with confirmed adrenal adenoma in each institution, a total of 237 were included in this study. e reasons for excluding patients included a past history of adrenalectomy (n � 18), missing 1 mg DST data (n � 13), and insufficient data for OS/OP (n � 26).

Definition of Osteoporosis and Osteopenia.
Bone mineral density (BMD) at the lumbar spine, femoral neck, and distal radius were measured using dual-energy X-ray absorptiometry using a Discovery system (Hologic Japan, Tokyo, Japan) in three centers, Horizon system (Hologic Japan, Tokyo, Japan) in two centers, QDR4500 system (Hologic Japan, Tokyo, Japan) in one center, PRODIGY system (GE Health Care Japan, Tokyo, Japan) in two centers, and iDXA system (GE Health Care Japan, Tokyo, Japan) in two centers.
BMD was automatically calculated from the bone mineral content (in grams) and bone area (in square centimeters) and expressed as an absolute value in grams per square centimeter. e young adult mean (YAM) is the percentage given from the mean BMD for a healthy young adult (100%) reference population. e diagnosis of OS/OP was based on the criteria for primary osteoporosis published by the Japanese Society for Bone and Mineral Research and Japan Osteoporosis Society Joint Review Committee [15]. Osteoporosis was diagnosed based on a past history of fragility fracture, medication for osteoporosis, and less than 70% of the YAM, which is nearly equal to <−2.5 SD of the T-score in the World Health Organization criteria, obtained from the lumbar spine or femoral neck [15]. Osteopenia was defined as a value of 70%-80% of YAM, which corresponds to −2.5 SD < T-score <−1.7 to −1.8 SD, without a history of fragility fracture and medication for osteoporosis [15].
Plasma ACTH and serum cortisol were evaluated by supine position in the morning, which was between 6 and 10 AM, and night, which was between 9 and 11 PM. 1 mg DST was performed by usual fashion, in which 1 mg of dexamethasone was given at 11 PM and the level of serum cortisol was evaluated between 8 and 9 AM in the following morning.

Statistics.
Statistical analysis was performed using the IBM SPSS software (version 24; IBM Corp., Tokyo, Japan). Continuous variables are expressed as a median and interquartile range. Categorical variables are presented as an actual number and percentages. Continuous variables were compared using the Mann-Whitney U test, and categorical variables were compared using the chi-square test. Based on 10 events per predictor rule, multivariate logistic regression analysis was performed after adjusting for variables including patient background and severity of CE. e association between CE and OS/OP was evaluated separately for as follows: 1 mg DST cortisol results considered as a dichotomous variable (>5.0 µg/dL vs. 1.9-5.0 µg/dL (used as the reference)), 1 mg DST cortisol levels considered as a continuous variable (µg/dL), and the presence of Cushingoid features vs. without Cushingoid features (used as the reference). An odds ratio (OR) > 1 indicated an increased likelihood of OS/OP. Cutoff values of 1 mg DST cortisol levels for predicting OS/OP were also compared between male and female by receiver operator characteristic (ROC) curve analysis. All tests were two-tailed. Values of P < 0.05 were considered statistically significant.

Osteoporosis and Osteopenia in the Patients with Adrenal
Adenoma Associated with Cortisol Excess. A total of 237 patients with adrenal adenoma associated with CE were included in this study. e median age of the patients was 56 years, and 76.8% were female. Cushingoid features were present in 40.5% of participants. Two hundred three of the 237 enrolled patients (85.7%) presented unilateral adenoma.
e median BMD at the femoral neck was at a level indicative of osteopenia. One hundred twelve of the 237 enrolled patients (47.2%) were defined as OS/OP. Eighty-nine of the 112 patients with OS/OP (79.4%) met the definition of osteoporosis (data not shown).
OS/OP was more prevalent in female (54.4%) than in male (23.6%) (P < 0.001). Presence of Cushingoid features was significantly higher in patients with OS/OP (P < 0.001). e body mass index (BMI) in patients with OS/OP was significantly smaller than that in patients without OS/OP (P � 0.013). Size and laterality of adrenal adenoma, serum calcium, renal function, lipid profile, and glucose metabolism were not related to OS/OP status (Table 1).
On endocrinological evaluation, morning, night, and 1 mg DST cortisol levels and 24-hour urinary cortisol levels were significantly higher in patients with OS/OP than in those without (P < 0.01). Morning and night ACTH and DHEA-S levels were significantly lower in patients with OS/ OP (P < 0.005). Positivity of 1 mg DST cortisol >5.0 μg/dl was significantly higher in patients with OS/OP (P < 0.001) ( Table 1). ese results suggest that the severity of CE was related to the presence of OS/OP.
To investigate whether the severity of CE in total cases and the cases with MACS was positively associated with OS/ OP, multivariate logistic regression analysis was used to evaluate the positivity of 1 mg DST cortisol level >5.0 μg/dl (vs. 1 mg DST cortisol level of 1.9-5.0 μg/dL), continuous variables of 1 mg DST levels (µg/dL), and the positivity of Cushingoid features as shown in
ROC analysis for predicting OS/OP by 1 mg DST cortisol levels was compared between male and female.

Discussion
In the present study, we demonstrated that the severity of CE was positively associated with developing OS/OP. Presence of Cushingoid features and biochemical evaluation using 1 mg DST cortisol, morning cortisol, night cortisol, 24-hour urinary cortisol, and DHEA-S levels contributed to the prediction of OS/OP. Multivariate logistic regression analysis demonstrated that positivity of 1 mg DST cortisol level >5.0 μg/dl and 1 mg DST cortisol levels were positively associated with OS/OP in adrenal adenoma associated with CE. A cutoff value for differentiating apparent CE of 1 mg DST cortisol level >5.0 μg/dL [10,14] was consistent with predicting OS/OP in both sexes. However, the influence of CE was smaller in women. e prevalence of OS/OP in women was also affected by age especially in the cases with MACS.
Approximately, 47% of our cohort had OS/OP at time of diagnosis with adrenal adenoma associated with CE. In the previous study, the prevalence of vertebral fractures was 21% in patients with CS including Cushing's disease and ectopic ACTH-producing tumors [1]. However, approximately 46%-83% of patients with adrenal adenoma associated with CE had vertebral fractures although Cushingoid features were absent and they presented with mild CE [16]. ere is a possibility that the severity of OS/OP in patients with adrenal adenoma associated with CE is not only dependent on CE but also other unknown factors.

International Journal of Endocrinology
In a study comparing adrenal CS and Cushing's disease [3], the prevalence of osteoporosis was significantly high in patients with adrenal CS. e authors speculated that adrenal androgen was downregulated by the suppression of ACTH in adrenal CS [17]. Tauchmanova et al. reported that vertebral fractures were common in female patients with CE regardless of the severity [18]. Adrenal androgen plays an important role in osteoblast activation, bone mass, and bone quality, especially in women [4,7,8,19,20]. In the present study, not including patients with ACTH-dependent CS, CE, and female sex were positively associated with developing OS/OP (Table 2). e suppression of ACTH by CE may   contribute to the downregulation of adrenal androgen, which can have a negative effect on bone metabolism, especially in women. OS/OP was diagnosed based on the existence of fragility fractures, medication use, and BMD as assessed using the YAM, which was a percentage calculated in comparison with the mean BMD obtained from young adults, in the present study. BMD negatively correlates with age in general [15]. BMD in postmenopausal women decreases dramatically with age in contrast to in premenopausal women [21][22][23]. Multivariate logistic regression analysis evaluating 1 mg DST cortisol demonstrated that OS/OP was affected by age in female patients with MACS. Female sex was positively associated with OS/OP evaluated using 1 mg DST cortisol in total cases and the cases with MACS ( Table 2).
Our study demonstrated that, in total patients and male patients with MACS, CE was positively associated with OS/ OP (Tables 3 and 4). Chiodini et al. demonstrated that the coexistence of adrenal incidentaloma and MACS increased the incidence of vertebral fractures and low BMD in eugonadal male patients [24]. Male patients can be directly affected by CE. Female patients including overt CS can be affected CE and consequent downregulation of adrenal androgen [18]. However, female patients with MACS can be affected not only by CE. Aging can also be associated with OS/OP as in primary osteoporosis [15,17,21,23,25]. e indications for adrenalectomy for adrenal adenoma associated with CE are based on the diagnosis of overt adrenal CS, 1 mg DST cortisol levels, and severity of complications related to CE in the current guidelines [10,14]. A guideline published by the European Society of Endocrinology and European Network for the Study of Adrenal Tumors (ENS@T), states that autonomous cortisol secretion with a 1 mg DST cortisol level >5.0 μg/dL is an indication of adrenalectomy; a 1 mg DST cortisol level of 1.9-5.0 μg/dL is defined as possible autonomous cortisol secretion. Adrenalectomy is considered after adding other clinical manifestations, including 24-hour urinary cortisol and metabolic disorders to possible autonomous cortisol secretion [14]. Other guidelines including those published by the Japan Endocrine Society provided similar criteria for adrenalectomy [10]. For considering adrenalectomy concerning fractures and decreased BMD, MACS defined by a 1 mg DST cortisol level >3.0 μg/dL was used in several studies [17,18,26]. Morelli et al. reported that a criterion for detecting incidental vertebral fractures on adrenal incidentaloma was a 1 mg DST cortisol level >2.0 μg/dL [27]. Several studies have demonstrated that 1 mg DST cortisol levels >1.8 μg/dL can reduce bone quality [11,[28][29][30].
We calculated the OR among those with 1 mg DST cortisol levels >5.0 μg/dL relative to those with 1 mg DST cortisol levels of 1.9-5.0 μg/dL. e OR for OS/OP was significant for both sexes. However, the value of the OR was remarkably higher in male patients than in female patients. Furthermore, 1 mg DST cortisol levels and age were positively associated with OS/OP in female patients with MACS (Tables 3 and 5). ese results suggested that the indication for adrenalectomy based on OS/OP should be determined by considering CE, sex and age. Compared with male patients, the deleterious effect to OS/OP may be exerted by milder CE in female patients with MACS. e present study had some limitations. First, the evaluation of bone metabolic biomarkers was not performed [21]. e precise mechanism underlying the progression of osteoporosis, including bone quality, was not clearly demonstrated [28]. Second, the method for evaluating endocrinological variables and BMD differed between institutions. BMD was evaluated following Japanese criteria [15]. We enrolled patients with OS/OP, which is equivalent to a T-score <−1.7 to −1.8 SD in the World Health Organization criteria. ird, this was a retrospective and crosssectional study with long enrolling period. e number of overt CS was higher than the previous studies evaluating adrenal incidentaloma [9,12,13]. en, the prevalence of OS/OP can be different from the studies evaluating MACS and nonfunctioning adrenal adenoma. e effect of CE may be weakened by the analysis not comparing with nonfunctioning adrenal adenoma. Further studies with a unified evaluation and prospective design are warranted to confirm our results.
In conclusion, we demonstrated that the severity of cortisol excess associated with adrenal adenoma was independently associated with the coexistence of osteoporosis and osteopenia. Our data suggest that cortisol excess in mild autonomous cortisol secretion had different effects on osteoporosis and osteopenia by sex.

Data Availability
Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.