Management of Adrenal Cortical Adenomas: Assessment of Bone Status in Patients with (Non-Functioning) Adrenal Incidentalomas

Our aim is to analyse the bone profile in adults with (non-functioning) adrenal incidentalomas (AIs), specifically addressing the impact of autonomous cortisol secretion (ACS). This narrative review, based on a PubMed search from inception to February 2023 (case reports, non-ACS, and other secondary causes of osteoporosis were excluded), included 40 original studies, a total of 3046 patients with female prevalence (female:male ratio of 1921:1125), aged between 20.5 and 95.5 years old. This three decade-based analysis showed that 37 studies provided dual-energy X-ray absorptiometry (DXA) information; another five studies reports results on bone micro-architecture, including trabecular bone score (TBS), spinal deformity index, and high-resolution peripheral quantitative computed tomography; 20 cohorts included data on bone turnover markers (BTMs), while four longitudinal studies followed subjects between 1 and 10.5 years old (surgical versus non-adrenalectomy arms). Post-dexamethasone suppression test (DST) cortisol was inversely associated with bone mineral density (BMD). TBS predicted incidental vertebral fractures (VFx) regardless of BMD, being associated with post-DST cortisol independently of age and BMD. Low BTMs were identified in ACS, but not all studies agreed. An increased prevalence of ACS-related osteoporosis was confirmed in most studies (highest prevalence of 87.5%), as well as of VFx, including in pre-menopause (42.5%), post-menopause (78.6%), and male patients (72.7%) depending on the study, with a 10-fold increased incidental VFx risk up to a 12-fold increased risk after a 2-year follow-up. No specific medication against osteoporosis is indicated in ACS, but adrenalectomy (according to four studies) should be part of the long-term strategy. This bone profile case sample-based study (to our knowledge, one of the largest of its kind) showed that AIs, including the subgroup designated as having ACS, embraces a large panel of osseous complications. The level of evidence remains far from generous; there are still no homogenous results defining ACS and identifying skeletal involvement, which might be a consequence of different investigation clusters underling adrenal and bone assessments over time. However, bone status evaluations and associated therapy decisions remain an essential element of the management of adults with AIs-ACS.


AI and Potential Cortisol Excess
While the radiological point of view concerns AI as a strictly incidental finding, the endocrine point of view typically relates to a negative hormonal profile and a low expected rate of growth (generally that associated with a cortical adenoma). The overall prevalence of AI is 2%, (between 1% and 8.7% depending on the study criteria) [4][5][6]. AI prevalence increases with age, being very rare in children and adolescents (around 0.2%) and reaching up to 7-10% in patients older than 70 years old (with female predominance) [7][8][9][10].
ACS affects between 5% and 30% of all individuals diagnosed with AI [11,12], subtle cortisol secretion being first described by Beierwaltes in 1974 [13]. The proportion of individuals experiencing cortisol excess in small amounts or intermittent patterns varies due to a heterogeneous diagnostic criteria and cut-off points. Currently, the term ACS is preferred to SCS, but either term involves an associated risk of some comorbidities and alterations of hypothalamic-pituitary-adrenal (HPA) axis regulation due to adrenal cortex-associated hormonal autonomy in the absence of the classical signs or symptoms of overt hypercortisolism (for instance, striae rubrae, proximal muscle weakness, facial plethora, easy bruising, purple striae, etc.) [14]. The standard dynamic test for defining ACS is the 1-mg dexamethasone suppression test (DST), but cut-offs regarding the second-day plasma cortisol vary between 1.8 and 5 µg/dL [15][16][17][18][19]. These values are associated with different risks of metabolic, cardiovascular, and skeletal complications [20].

Cortisol Overproduction Targeting Bone Status
Glucocorticoid excess has a damaging effect on bone mass and quality, being the most common cause of secondary osteoporosis [21]. The condition, underlying increased bone resorption and decreased bone formation, is associated with a reduction in bone formation through suppression of osteoblasts activity mediated by upregulation of peroxisome proliferator-activated receptor (PPAR)-γ and inhibition of the wingless (wnt)/β-catenin signalling pathway [22][23][24]. Additionally, sclerostin is produced by osteocytes and has been recognized as a key negative regulator of bone formation, while chronic glucocorticoid exposure can induce autophagy in osteocytes and consequently decreased sclerostin concentrations [25,26]. Cortisol excess stimulates bone resorption through an alteration of the receptor activator of nuclear factor kappa-B ligand (RANKL)/osteoprotegerin ratio [27,28]. RANKL is a regulator and activator of osteoclasts, while osteoprogerin acts as a decoy receptor for RANKL, preventing its interaction with RANK and causing the inhibition of osteoblastogenesis [29][30][31]. The severity of hypercortisolism-associated skeletal effects also depends on one's sensitivity to hormones, as, for instance, has been shown by polymorphism studies of glucocorticoid receptor [32]. In the general population, BclI and N363S polymorphisms are associated with an increased sensitivity to glucocorticoids and a low bone mineral density (BMD) [33,34], while ER22/23EK is correlated with reduced sensitivity to glucocorticoids [35]. Flowchart diagram of included papers regarding patients with adrenal tumours of AI (non-functioning AI or adrenal cortex adenomas) type (and/or ACS or SCS) in whom bone status was analysed according to our methodology .

Assessing ACS (or SCS) in Patients with Adrenal Tumours
The studies that addressed bone status in patients with adrenal tumours with apparent non-functioning profiles started from 1992 (the only studies that we included Figure 1. Flowchart diagram of included papers regarding patients with adrenal tumours of AI (non-functioning AI or adrenal cortex adenomas) type (and/or ACS or SCS) in whom bone status was analysed according to our methodology .

Assessing ACS (or SCS) in Patients with Adrenal Tumours
The studies that addressed bone status in patients with adrenal tumours with apparent non-functioning profiles started from 1992 (the only studies that we included with overt CS were those with a studied subgroup of patients diagnosed with SCS or ACS). We report them below from a timeline perspective with respect to various criteria for defining ACS, as mentioned (Table 1). Table 1. Studies with AI/non-functioning AI and bone assessments: the analysis of criteria defining ACS (or SCS)-the studies are displayed in the order of publication . Studied population . The mentioned studies used various endocrine criteria in order to define the cortisol excess as seen in ACS (or equivalent) . (Table 2).     According to the aforementioned original studies, ACS, SCS, SH, or preclinical CS was confirmed based on hormonal tests that varied over time, including DST with different doses and cut offs, sometimes in combination with low ACTH (adrenocorticotropic hormone), high UFC (urinary free cortisol), or increased midnight salivary (or serum) cortisol, in addition to lack of clinical phenotype suggestive for CS [43,44,46,47,49,72,73,[75][76][77][78]81]. Another type of analysis enrolled patients with AIs as a general category that sub-included individuals with or without SH, not subjects with AI versus SH (or ACS), as most of the cohorts [65].

First Author
Older studies introduced the term "preclinical" CS. For instance, Reincke et al. [43] investigated the prevalence of ACS in 68 patients with AI (female-to-male ratio of 44:24; aged between 25 and 90 years old), and 12% of them had positive hormonal activity consistent with the diagnosis of preclinical CS. Subjects with non-functioning AIs and preclinical CS did not have osteoporosis; only 37.5% of the individuals with overt CS did. Preclinical CS was defined as the following: no clinical signs of CS and lack of suppression with regard to serum cortisol levels (<3 µg/dL) after 1-mg and 8-mg DSTs. The rate and severity of associated arterial hypertension, obesity, and diabetes mellitus markedly improved after adrenalectomy in individuals with preclinical CS (after a mean follow-up of 28 months), in addition to achieving the normal cortisol suppression after DST [43].
The same prevalence of preclinical CS (12%) was reported by Ambrosi et al. [44]. Thirty-two patients with incidentally discovered adrenal tumours (female-to-male ratio of 23:9; aged between 28 and 74 years old) were tested for cortisol levels that were not adequately suppressed (<140 nmol/L) after 1-mg DST and loperamide testing (16 mg). Despite identifying the sub-group with preclinical CS among subjects with AIs, the entire AI group had a statistically significantly reduced BTM profile in terms of osteocalcin, carboxyterminal cross-linked telopeptide of type I collagen (ICTP), and amino-terminal propeptide of type III procollagen (PIIINP) versus age-matched controls (osteocalcin: 3.9 ± 0.6 versus 5.4 ± 0.15 µg/L; ICTP of 2.4 ± 0.1 versus 4.1 ± 0.1 µg/L; PIIINP of 2.2 ± 0.15 versus 3.3 ± 0.2 µg/L; p < 0.01), thus proving that underlying cortisol activity might be involved regardless of the specific assays that are used [44].

BMD Analysis in Patients with ACS
DXA assessment of patients with AI (or non-functioning AI) provided the prevalence of osteoporosis/osteopenia or the BMD reports versus controls, time-dependent BMD changes, or post-adrenalectomy bone effects, including on DXA exams. As mentioned, ACS (or SCS or SH) involves a distinct type of tumour with positive (yet mild, not overt) cortisol (persistent) excess, and consecutive DXA-BMD evaluations depend on the criteria for defining this hormonal activity, on the age and sex of the studied population (including the menopausal status of women), and on the sample size of the cohort. Except for three, 37 cohorts provided data in terms of DXA and/or VFs [43,, and a confirmation of a negative impact of bone status was generally confirmed.
The first type of mentioned analysis showed a higher rate of osteoporosis in ACS versus non-functioning AI. For example, Ueland et al. [77] found rates of 18.1% versus 8.5%, respectively (in this study, ACS was defined as serum cortisol greater than 50 nmol/L after 1-mg DST) [77]. Podbregar et al. [79] followed 67 patients with non-functioning AIs (female-to-male ratio of 47:20, mean age of 57.9 years old), and 22% of them progressed to mild ACS (MACS) (diagnosed based on serum cortisol ≥ 50 nmol/L after 1-mg DST), which was reflected by an increased rate of osteoporosis (from 17.9% at baseline to 26.9% after follow-up, p = 0.031) [79].
Tauchmanova et al. [51] found anomalies of the finger amplitude-dependent speed of sound (Ad-SoS, which has capacity to detect the structural characteristics of bone changes) as measured in SCS and overt CS subgroups and compared them with healthy matched controls (p < 0.001, all). BMD and fracture prevalence in SCS were similar to those in overt CS and not to the controls [51].
While most studies have confirmed impairment of bone status in terms of low BMD or a higher prevalence of osteoporosis in ACS, other studies have revealed equivocal results. For example, Rossi et al. [47] showed the absence of significant bone loss in patients with AIs, even associated with SCS (as defined by an abnormal response to at least two standard tests exploring the HPA axis: 2 mg DST-cortisol > 3 µg/dL, UFC > +2 SD of the control group, mean daily cortisol > +2 SD of the control group, and reduced ACTH levels), compared with a sex-and age-matched healthy population [47]. Similarly, Osella et al. [49] revealed no BMD difference between subjects with AI (N = 27) and controls, regardless of the presence of SCS in a small subgroup (as established by the confirmation of two abnormal results with respect to the lack of cortisol suppression after DST (>5 µg/dL), elevated UFC (>216 µg/24 h), low ACTH, and elevated night-to-day cortisol ratio [49].
A study conducted by Ahn et al. [74] in 109 patients with SH [as diagnosed by post-1 mg DST cortisol > 138 nmol/L or >61 nmol/L plus ACTH < 2.2 pmol/L or DHEA-S (dehydroepiandrosterone sulphate) < 2.17 µmol/L in men or <0.95 µmol/L in women] versus 686 subjects with non-functioning AIs found that premenopausal women with SH had significantly lower BMD at lumbar spine (by 9.1%, p = 0.008), and femoral neck (by 9.5%, p = 0.012) versus premenopausal subjects with non-functioning AIs, while postmenopausal women with SH had statistically significant lower BMD only at the lumbar level (by 7.1%, p = 0.016). The prevalence of VFs was similar in premenopausal women (0% versus 1.7%, p = 0.578), postmenopausal women (0.0% versus 4.0%, p = 0.209), and men (0.0% versus 0.7%, p = 0.999). As collateral observations, DHEA-S was positively correlated with lumbar BMD in postmenopausal participants (β = 0.096, p = 0.001) and men (β = 0.029, p = 0.038). An additional analysis compared the baseline characteristics of the subjects with SH to the Caucasian population (N = 85) from a previous study; overall, the prevalence of osteoporotic fractures was higher in the Caucasian population than the Asian population (p < 0.001) [74] (Table 3). Table 3. DXA results in terms of BMD assessment and/or osteoporosis/osteopenia prevalence in included studies (to identify each sample size/subgroup, please check Table 1); the display starts from the oldest data to the most recent [43,.

TBS Anomalies Due to ACS-AIs
Mild cortisol overproduction in patients with ACS (or SCS) might impair the skeleton's qualitative features, as reflected by bone microarchitecture analysis. Overall, we identified five such studies, particularly addressing TBS but also, spinal deformity index (SDI) and high-resolution peripheral quantitative computed tomography (HR-pQCT) [63,72,73,81]. The potential cortisol over-production from the tumour correlates with a negative impact at the level of the micro-architecture, while the level of statistical evidence is less convincing than that seen with DXA-BMD analysis.
Cross-sectional analysis of Vinolas et al. [72] proved that TBS is more useful than BMD in individuals associated with various degrees of persistent hypercortisolism (N = 110). Persons with MACS (N = 39, mean age of 57.8 ± 9.3 years) had lower TBS than patients with non-functioning AIs (N = 18, average age of 59.2 ± 9.1 years): 1.30 ± 0.09 versus 1.37 ± 0.12 (p < 0.04) but similar BMD, as shown by the values of 1.06 ± 0.20 versus 1.11 ± 0.18 SD (p = 0.34). During the mid-term evaluation at 15.5 ± 4.8 months following remission of CS (N = 53), the TBS increase was greater than the BMD increase (10% versus 3%, p < 0.02). In patients with overt CS and MACS, no difference was observed regarding TBS and BMD between hypogonadal and eugonadal subjects (1.275 ± 0.10 versus 1.298 ± 0.13, and 1.04 ± 0.18 versus 1.11 ± 0.16 SD, respectively), thus confirming that hypogonadism might not be an essential contributor in such cases [72].
Alternatively, HR-pQCT was used to reflect trabecular bone micro-architectural derangement at the distal radius in one study of 45 subjects with non-functioning AIs (as pointed out by the levels of serum cortisol ≤ 1.8 µg/dL after 1-mg DST) versus 30 individuals with ACS (having cortisol levels between 1.9 and 5.0 µg/dL after 1-mg DST) with a median age of 59 or 60 years old, respectively. Lumbar aBMD was lower in ACS than AIs and was similar at the femoral neck and third distal radius, as were HR-pQC-based trabecular vBMD (p = 0.03), inner zone of the trabecular region (p = 0.01), bone volume-to-tissue volume ratio (p = 0.03), and trabecular thickness (p = 0.04). However, the prevalence of osteoporosis was similar between the groups (75% versus 65%, p = 0.55), as was the rate of fragility fractures (73.7% versus 55.6%, p = 0.24) [75] (Table 4). Table 4. TBS analysis in patients diagnosed with ACS/AIs [63,72,73,81].
Osella et al. [45] showed that patients with AI had a statistically significant osteocalcin reduction and a mild increase in ICTP (marker of bone resorption) compared to controls (6.6 versus 7.8 ng/mL, p < 0.05 and 4.2 versus 3.1 µg/L, p < 0.01); other BTMs, including bone alkaline phosphatase, did not reach statistical significance, suggesting that osteocalcin is more sensitive than bone alkaline phosphatase in to reflecting the actions of glucocorticoids on bone in AIs [45].
A small-sample size study conducted by Torlontano et al. [48] confirmed altered osteoblastic activity, as reflected by lower osteocalcin levels, in the subgroup with SH (N = 8) compared to controls (N = 64): 6.8 ± 3.5 versus 8.8 ± 3.2 ng/mL (p < 0.005), as was BMD at each central site (p < 0.05). PTH was higher (p < 0.05) in individuals with SH than in those who were SH-free (N = 24) and in both groups compared to controls (57.1 ± 13.6, 46.0 ± 14.8, and 37.2 ± 10.9 pg/mL, respectively) [48]. Francucci et al. [53] studied a cohort with CS (N = 15), AI (N = 23) and controls (N = 20) and found a significant reduction in lumbar and femoral neck BMD Z-scores (p < 0.05) in CS subjects versus AI subjects and controls, as well as in osteocalcin and serum phosphorus levels in CS and AI subjects versus controls (p < 0.05) [53].
Another study focused on urinary BTM, namely urinary N-terminal crosslinking telopeptide of type I collagen, as a bone resorption marker. Fifty-five individuals with MACS (mean age of 61.5 ± 10.1 years old) were compared to 12 cases of non-functioning AIs (average age of 66.0 ± 8.9 years old), and there were higher values of this marker (50.6 ± 25.6 versus 26.9 ± 16.6 nmoL BCE/mmolCr, p = 0.017) [78].

Bone Assessment in Patients with Unilateral Versus Bilateral AIs
Approximately, one out of 10 cases with AI has a bilateral tumour, but a direct relationship with a more damaged bone profile still represents an unresolved issue, mostly due to the lack of large cohorts specifically referring to bilateral, rather than unilateral, adenomas with respect to skeleton status. We found two studies of bilateral lesions (please see Tables 1 and 2). Morelli et al. [65] showed a similar SH prevalence (SH was diagnosed in the presence of at least two elements: serum cortisol levels after 1-mg DST of >83 nmol/L, UFC of >193 nmol/24 h; and ACTH levels of <2.2 pmol/L) among subjects confirmed with bilateral AI and unilateral AIs (26.3% versus 23.4 %, p = 0.680); bilateral AIs were correlated with lower BMD and a higher prevalence of VFx than unilateral AIs (52.6% versus 29.7%, p = 0.005); and the diagnosis of VFx was associated with bilateral AIs after adjusting for SH (OR = 1.77, 95% CI 0.85-3.7, p = 0.12). The prevalence of VFx tended to be higher in persons with bilateral AIs and SH than in bilateral AIs without SH (70% versus 46.4%, p = 0.07) [65].
Ognjanovic et al. [71] evaluated 152 patients (105 with unilateral AIs and 47 with bilateral AIs) and identified a prevalence of SH higher in bilateral than unilateral AIs (29.8% versus 16.3%, p = 0.058) according to the serum cortisol levels after 1 mg DST or after LDDST of >50 nmol/L with at least one of the following parameters: midnight serum cortisol of >208 nmol/L, UFC of >245 nmol/24 h or ACTH of <10 ng/L. Participants with bilateral AIs had lower lumbar BMD when compared to unilateral AIs (p = 0.002) and an increased prevalence of osteoporosis (37.1% versus 15.9%, p = 0.011) [71].

Prevalent Fractures in Individuals with AI and ACS
Several studies addressed the issue of prevalent VFx (rarely all types of fragility fractures, all referring to non-vertebral) (Tables 1 and 2). Additional factors, such as hypogonadism (such as menopause), might contribute to VFx, although not all studies agree [52,55,59].
One study revealed that the prevalence of VFx was similar between SCS and overt CS (69% versus 57%, p = 0.56), including clinical VFx (28% versus 11.4%, p = 0.22) and multiple VFx (36% versus 31%, p = 0.92). VFx was independently associated with cortisolto-DHEAS ratio. Lumbar BMD and cortisol-to-DHEAS ratio were the best predictors of VFx (p < 0.01) [56]. Another study reported a VFx prevalence of 35.1% among patients with AIs. The patients with SH experienced a higher rate of VFx (81.5%) after a 2-year follow-up compared to baseline (55.6%, p = 0.04) and deterioration of SDI (2.11 ± 1.85 versus 1.11 ± 1.47, p = 0.032), indicating the importance of periodic check-ups. Additionally, the incidental VFx rate in subjects with SH was increased versus non-SH (48% versus 13%, p = 0.001). The risk of developing new VFx was independently associated with the presence of SH in apparently non-functional adrenal tumours (OR = 12.3, 95% CI 4.1-36.5, p = 0.001). In this study, SH was defined by the presence of at least two alterations among cortisol levels after 1-mg DST > 3 µg/dL, increased UFC levels > 70 µg/24 h, and low ACTH levels < 10 pg/mL [62]. Similarly, Lasco et al. [67] showed that SH was associated with a higher prevalence of VFx independently of BMD; while subjects with SH also had reduced lumbar BMD versus non-SH subjects (p < 0.01) [67].

Longitudinal Studies following Bone Status with Regard to ACS and AIs
The optimal management of patients with ACS has been associated with a dynamic approach over the years due to continuous changes in definition criteria and specific indications, which vary from an individual approach to guideline recommendations. The importance of the topic is related to close follow-up of medical treatment for associated morbidities being needed, especially if surgical treatment is not chosen. Four studies followed patients from 1 year to more than a decade [57,62,79,80] (Table 6).
Regarding an interventional approach specifically addressing medication for osteoporosis in ACS, a limited number of publications were identified. Tauchmanova et al. [57] evaluated the effects of 1-year clodronate (100 mg every week) on lumbar and femoral neck BMD and BTMs in premenopausal women with SCS and osteopenia/osteoporosis (N = 46). Patients were randomized to receive clodronate plus calcium (500 mg daily) and vitamin D3 (800 mg daily) supplements or only calcium plus vitamin D. As expected, clodronate administration increased lumbar BMD (p = 0.04) and conserved BMD values at the femoral neck, while incidental fractures occurred only in the non-clodronate group [57]. Currently, the drug is no longer used or available in daily practice.
Alternatively, removal of the tumour might help the bone status, but not all studies agree. A prospective, randomized study of small size compared laparoscopic adrenalectomy (N = 23) to conservative management (N = 22) in SCS (defined by serum cortisol higher than 2 µg/dL after 1-mg DST). Adrenalectomy improved the cardio-metabolic profile underlying hypertension and diabetes mellitus but not bone parameters [60]. A similar result was reported by Iacobone et al. [64], concluding that adrenalectomy is better than medical therapy in improving the values of high blood pressure hypertension but not the values of DXA-based T-scores [64].
A more complex decision involving surgery is needed in cases with bilateral tumours. Perogamvros et al. [68] published a retrospective study of 33 patients with bilateral AIs; the surgical group included 14 patients who underwent unilateral adrenalectomy of the largest adenoma (all women, mean age of 54.9 ± 6.7 years old), and the non-surgical group enrolled 19 patients (14/19 women, average age of 59.0 ± 8.7 years old). An improvement of comorbidities was registered only in the first group, while two of the three subjects with osteoporosis were associated with an increase in post-operative BMD (p = 0.03) [68]. Salcuni et al. [70] identified a reduced probability of incidental VFx in patients with SH who underwent adrenalectomy versus those who were surgery-free (N = 55 individuals with SH, of whom 32 people underwent adrenalectomy, and 23 received a conservative approach). Adrenalectomy in subjects with SH was associated with a 30% VFx risk reduction (OR = 0.7, 95% CI 0.01-0.05, p = 0.008) regardless of age, gender, follow-up duration, degree of hypercortisolism, lumbar BMD, or prevalent VFx at baseline. In the surgical group, only 9.4% of subjects experienced a new VFx, which was statistically significantly lower than in the non-adrenalectomy group (52.2%, p < 0.0001) [70].
Overall, four studies specifically addressed the outcome of adrenalectomy in patients with AIs-ACs (between 2009 and 2016). While the total number of patients per study remains low (33, 35, 45, and 55), all had a control arm (surgery versus conservative approach); the period of follow-up varied from a mean of 54 ± 34 and 53.9 ± 21.3 months in two cohorts (after adrenalectomy), respectively, and a mean of 56 ± 37 or 51.8 ± 20.1 months to a median of 7.7 (2-17) years in one study (non-surgical group) ( Table 7).

Prevalence of ACS in Patients with Osteoporosis and Fracture
Additionally, two studies were added from the opposite perspective: the assessment of adrenal (cortisol) status in patients diagnosed with osteoporosis and fractures (N = 320, female-to-male ratio of 275:45, mean age of 62.17 years old). This analysis was distinct from the previously described studies. We examined ACS profiles in patients already diagnosed with osteoporosis and with associated fragility fractures, and we identified two additional (transversal) studies (published in 2021 and 2007) to add to the initial panel of 40 original papers. The hypothesis of a negative effect of ACS on skeletal status was supported by Chiodini et al. [83], showing that SH is more common in adults with osteoporosis than in those with abnormal bone profiles. They found a prevalence of 4.8% for SH (between 1.32% and 8.20%) among patients with osteoporosis, while no subjects without osteoporosis had SH. The authors included 219 participants without clinically overt CS or other obvious secondary causes of bone loss (200 women and 19 men). SH was diagnosed according to incomplete cortisol suppression (>50.0 nmol/L) after LDDST, UFC of >165.6 nmol/L (normal ranges between 22.2 and 165.6 nmol/L), and/or midnight cortisol level of >207 nmol/L (normal ranges between 0.0 and 138.5 nmol/L). Patients with osteoporosis had lower T-scores (BMD) at the lumbar spine of −2.88 ± 1.09 SD versus 1.36 ± 1.09 SD, at the total neck of −1.96 ± 0.88 SD versus −0.99 ± 0.97 SD, and at the femoral neck of −2.02 ± 0.87 SD versus −1.12 ± 0.91 SD compared to individuals without osteoporosis (p = 0.001) [83].
The other cohort enrolled 101 subjects with prevalent fragility fractures (75 women and 26 men) with a mean age of 65 ± 10.3 years old. Five of 101 (3 women and 2 men) were diagnosed with "less severe hypercortisolism" (a prevalence of 5%), which was established by unsuppressed serum cortisol less than1.8 µg/dL after LDDST. Lumbar and femoral neck BMDs (Z-score) were similar in individuals with or without this level of cortisol excess as described by Pugliese et al. [84] (Table 8).

Controversies in Defining Subtle (Non-Overt) Cortisol Excess Due to Adrenal Cortex Adenomas
We believe that the most difficult aspect of studying bone profiles in AIs/ACS is the definition of the entities themselves (ACS or SCS) since various criteria and terms have been used according to our 30-year analysis. Pre-CS or preclinical CS historically referred to asymptomatic, but biochemically active hypercortisolism [85,86]. Today, the shift from using the term "SCS" to "ACS" has occurred in daily practice; the key assay remains the serum cortisol level after 1-mg DST [16,17,87,88]. In addition, a value of cortisol between 1.8 µg/dL and 5 µg/dL, low morning plasma ACTH less than 10 pg/mL, and/or increased UFC levels improve the accuracy of ACS diagnosis [89]. Morelli et al. reported that the presence of combined criteria characterized by at least two biochemical anomalies among the values after 1-mg DST cortisol level greater than 3 µg/dL (83 nmol/L), high UFC, and ACTH less than 10 pg/mL ("combined DST-UFC-ACTH criterion") was associate with sensitivity of 61.9% and specificity of 77.1% and thus accuracy of 75.8% in predicting complications [61]. Similarly, a cortisol value higher than 2 µg/dL (55 nmol/L) after 1-mg DST predicts VFx in patients with AIs with sensitivity of 80% and specificity of 68.8% [69]. In most studies, a combination of at least two abnormal HPA tests, especially DST and another assay (such as baseline assays of UFC or ACTH), were used [46,47,49,50,52,55,[57][58][59][61][62][63][64][65][66]68,70,71]. Moreover, two studies included DHEAS assessments of less than 2.17 µmol/L in men and less than 0.95 µmol/L in women or alternatively ACTH values less than the cut off of 2.2 pmol/L in association with 1-mg DST to confirm ACS [73,74].

Bone Status and ACS
The difference in hormonal diagnosis (not only in terminology) represents one major contributor to the heterogeneous bone profiles that we found, in addition to other bias elements, such as different ages (we only identified adult studies), comorbidities (such as menopausal status and, generally hypogonadism, and diabetes mellitus), previous or concurrent medication for osteoporosis, and fracture risk reduction; further, with respect to the tumour itself, bilateral lesions are more prone than unilateral AIs to developing bone profile-associated abnormalities   (Figure 2).
Moreover, the heterogeneity of the mineral metabolism and fracture risk evaluations that we found among the 40 original studies should have limited the data of a systematic review, which is why we included a broader spectrum under the umbrella of a narrative review. Based on our research, this three decade-based analysis of published sample case studies enrolled more than 3000 patients with confirmed AIs/ACS; among the 40 studies, 37 provided DXA-based information, five reported results on bone micro-architecture, including TBS, SDI, and HR-pQCT, and 20 cohorts included data regarding BTMs, while four longitudinal studies (surgical versus non-surgical arms) followed subjects between 1 and 10.5 years old   (Figure 3).
Moreover, the heterogeneity of the mineral metabolism and fracture risk evaluations that we found among the 40 original studies should have limited the data of a systematic review, which is why we included a broader spectrum under the umbrella of a narrative review. Based on our research, this three decade-based analysis of published sample case studies enrolled more than 3000 patients with confirmed AIs/ACS; among the 40 studies, 37 provided DXA-based information, five reported results on bone micro-architecture, including TBS, SDI, and HR-pQCT, and 20 cohorts included data regarding BTMs, while four longitudinal studies (surgical versus non-surgical arms) followed subjects between 1 and 10.5 years old   (Figure 3).
Bone micro-architecture is affected by chronic cortisol excess, as directly shown by histomorphometry/bone biopsy-based studies [93]. However, SDI, but mostly TBS (during last decade), is applicable in daily practice [94]. As mentioned, four TBS-based studies proved its value as a complementary, easy-to-access tool in adults with AIs/ACS [63,72,73,81]. TBS predicts incidental VFx regardless of BMD, being associated with post-DST cortisol levels independently of the patients' age and DXA-BMD [63]. Of course, there are still open questions regarding these subjects, especially in menopausal women diagnosed with type 2 diabetes mellitus, including the exact contribution to TBS damage due to associated glucose profile anomalies. Moreover, SDI is a semi-quantitative method that integrates the number and severity of VFx [95]; an increased SDI has been found in subjects with SH [58,62,63].  . Abbreviations: CC = case-control study; Cs = cross-sectional study; P = prospective study; R = retrospective study.
Bone micro-architecture is affected by chronic cortisol excess, as directly shown by histomorphometry/bone biopsy-based studies [93]. However, SDI, but mostly TBS (during last decade), is applicable in daily practice [94]. As mentioned, four TBS-based studies proved its value as a complementary, easy-to-access tool in adults with AIs/ACS [63,72,73,81]. TBS predicts incidental VFx regardless of BMD, being associated with post-DST cortisol levels independently of the patients age and DXA-BMD [63]. Of course, there are still open questions regarding these subjects, especially in menopausal women diagnosed with type 2 diabetes mellitus, including the exact contribution to TBS damage due to associated glucose profile anomalies. Moreover, SDI is a semi-quantitative method that integrates the number and severity of VFx [95]; an increased SDI has been found in subjects with SH [58,62,63].
Decreased bone formation, as revealed by low osteocalcin, was found in some of the aforementioned studies; of course, this condition is also an effect of the co-presence of diabetes mellitus in certain subgroups [44][45][46]48,53,54,56,76]. Other cohorts did not have mild cortisol overproduction associated with serum osteocalcin anomalies [50,52,78]. Less valuable data are provided for bone alkaline phosphatase, also a bone formation marker, which can be decreased [96] or similar in AIs/ACS and controls [45,48,50]. Additionally, hypercortisolism induces bone resorption, leading to a decrease in sclerostin (secreted by osteocytes). One study found that the severity of cortisol excess was inversely correlated with its levels [76]. Serum C-telopeptide of type I collagen ICTP assays had various results: they were increased [45,56], reduced [44,46], or similar in patients with ACS versus non-ACS [48,76]. Urinary deoxypyridinoline does not seem relevant in ACS/SCS [48,50,52,53]. Urinary N-terminal crosslinking telopeptide of type I collagen, another bone resorption marker, was statistically significant higher in MACS than non-functioning AIs according to one study [78].
Decreased bone formation, as revealed by low osteocalcin, was found in some of the aforementioned studies; of course, this condition is also an effect of the co-presence of diabetes mellitus in certain subgroups [44][45][46]48,53,54,56,76]. Other cohorts did not have mild cortisol overproduction associated with serum osteocalcin anomalies [50,52,78]. Less valuable data are provided for bone alkaline phosphatase, also a bone formation marker, which can be decreased [96] or similar in AIs/ACS and controls [45,48,50]. Additionally, hypercortisolism induces bone resorption, leading to a decrease in sclerostin (secreted by osteocytes). One study found that the severity of cortisol excess was inversely correlated with its levels [76]. Serum C-telopeptide of type I collagen ICTP assays had various results: they were increased [45,56], reduced [44,46], or similar in patients with ACS versus non-ACS [48,76]. Urinary deoxypyridinoline does not seem relevant in ACS/SCS [48,50,52,53]. Urinary N-terminal crosslinking telopeptide of type I collagen, another bone resorption marker, was statistically significant higher in MACS than non-functioning AIs according to one study [78].
Moreover, the relationship between anomalies of DHEA-S in AIs/ACS and bone status is still inconclusive [77,102,103]. One study identified that up to 40% of subjects had AI-associate reduced DHEA-S levels [44], while DHEA-S was positively correlated with lumbar BMD [56], but not all studies showed statistically significant data [77].
Overall, the prevalence of osteoporosis complicated or not with fragility fractures, especially at the vertebral level, is higher in individuals with SH versus non-SH, for example, 87.5% versus 27.8%, 68.8% versus 18.5% [66], 75% versus 65%, and 74% versus 55.6% [75]. Others found a similar prevalence in ACS and non-functioning AIs [74,78]. A greater number of VFx than controls was observed in premenopausal women, also consistent with cortisol rather than oestrogens status (VFx prevalence of 42.9% in premenopause and of 78.6% in post-menopause) [55]. Another study suggested that subjects with SH had a 10-fold increased risk of new VFx, independent of age, gender, and BMD [69], as well as a 12-fold increase after a 2-year follow-up with a VFx prevalence of 81.5% [62]. A similar rate of 72.7% was reported in adult men (most VFx being asymptomatic), not only in women [59]. We identified only one study applying FRAX in patients older than 40 years old, and its values were similar between subjects with ACS and those with non-functional AIs [75].

Interventional Considerations
The therapeutic approach for patients diagnosed with ACS/SCS is based on nonrandomized trials. Weekly clodronate might prevent bone loss and VFx in women with SH [57], but current applications are limited. Adequate calcium and vitamin D are required, while the best treatment strategy against osteoporosis is supported by the general recommendations for adult (primary and/or glucocorticoid) osteoporosis.
Notably, adrenalectomy should be considered when deciding on an anti-osteoporosis strategy (including in cases of unilateral adrenal removal for patients with bilateral AIs) since improvement of bone status is expected after surgery [68,70]; however, some authors showed no remarkable post-operative benefits regarding BMD improvement or fracture prevention [60,64]. Transient post-adrenalectomy adrenal insufficiency, an indirect piece of evidence for prior tumour-related cortisol hypersecretion, might highlight further bone advantages [43,44,104,105].
Cortisol excess due to adrenal tumour-related overproduction might cause glucocorticoid osteoporosis. In patients with ACS/SH, primary (menopausal or age-related osteoporosis) might overlap with type 2 diabetes mellitus-associated bone damage. Despite not having an overt clinical picture of CS, hypercortisolaemia might impair bone status through well-known pathogenic mechanisms of glucocorticoid osteoporosis. The degree of bone influence is related to the duration and level of hormone overproduction, as similarly seen in iatrogenic CS. Glucocorticoids, while acting on osteoblasts, reduce their proliferation and differentiation, also causing elevated apoptosis of osteoblasts and osteocytes and thus reduced bone formation (as mentioned, not all studies identified reduced levels of osteocalcin and bone alkaline phosphatase in subjects diagnosed with ACS/SH). Moreover, cortisol excess induces increased osteoclastogenesis with a greater number of osteoclasts being recruited in association with reduced apoptosis, which is prone to bone resorption. Finally, hormone overproduction impairs the overall bone turnover, favouring a reduced BMD (as pointed out above in subjects with ACS as well). A good level of evidence shows that anti-resorptive medications, such as bisphosphonates and denosumab, as well as bone-forming agents, such as teriparatide, counteract these pathogenic elements with good clinical results; thus, subjects with ACS/SH should be no exception according to the level of evidence that we have so far [131][132][133][134][135].

Further Considerations
Overall, the topic of recognizing mild cortisol excess in these adrenal tumours is mandatory to address the outcome, to decide whether and when surgery is needed, or to select candidates for a specific medication against osteoporosis. As mentioned, the limits of the current topic derive from the lack of standardization both in defining the subgroups with non-overt cortisol excess among patients with apparently non-functional adrenal cortex adenomas and in recommending a standard intervention. We acknowledge that our current work represents a narrative (not a systematic) review, and some studies might be biased because only PubMed was used for the literature search.
Anomalies of bone profiles (including the diagnosis of osteoporosis and fragility fractures) represent an important element in overall decision making; thus, we need not only cross-sectional but also longitudinal data. Future perspectives should address three main issues: the panel of bone-related investigations and its timing during lifelong surveillance to assess the fracture risk; randomized trials concerning medication against osteoporosis and fracture risk reduction in this specific group of patients; and a focus on the role of bone profiles in protocols for deciding on adrenalectomy (including in individuals with bilateral AIs).

Conclusions
According to our bone profile study of published data that is, to our knowledge, one of the largest on its kind, tumours considered AIs or non-functioning AIs, including the subgroup designated as having ACS or SCS (or SH), includes a large panel of osseous complications from lower BMD and TBS to blunt BTMs and a higher prevalence of osteoporosis and/or VFs compared to non-AIs or non-SCS. The level of evidence remains far from generous; there are still no homogenous results, which might be a consequence of different investigation clusters with respect to adrenal and bone evaluations that have been used over time. However, awareness is mandatory, and bone assessments should be an essential element of the management in these adults.