Clinically Relevant Outcomes in Adrenal Incidentalomas
Risk of cardiovascular diseases
During the last 20 years, increasing evidence from cross-sectional studies showed
that patients with adrenal tumors associated with ACS have an impaired cardiovascular
profile, as highlighted by the high prevalence of several cardiovascular risk factors,
mainly hypertension and diabetes. The link between hypertension and mild cortisol
excess has been established in several studies, as shown in a recent review of the
literature [2]. Even though a precise calculation of the prevalence of hypertension among patient
with ACS is cumbersome, due to the low comparability among the different methods of
the studies, hypertension is by far the most frequent cortisol-related co-morbidity
found in patients with ACS. Indeed, the prevalence of hypertension in patients with
ACS is over 60% in 2/3 of the studies published in the last 15 years [2]. Moreover, patients with ACS seems to be more prone to develop resistant hypertension,
as defined by treatment with more than three antihypertensive medications. Studies
investigating the prevalence of ACS among hypertensive patients have shown that excessive
cortisol secretion may be found in a negligible number of subjects with essential
hypertension (1%) [3]. However, a study analyzing only patients with resistant hypertension showed that
8% of hypertensive patients had biochemical evidence of cortisol excess, as defined
by cortisol after DST>1.8 µg/dL and two confirmed records of elevated midnight salivary
cortisol levels [4]. It is worth mentioning that among 423 patients included in the initial cohort,
112 (26.5%) had cortisol values after DST>1.8 µg/dL with or without additional alterations
of the hypothalamic pituitary adrenal axis, a condition that it now recognized as
possible autonomous cortisol secretion, in keeping with the current guidelines on
adrenal incidentalomas [1].
In addition to hypertension, several markers of subclinical atherosclerosis have been
studied in patients with ACS. Among them, higher values of intima-media thickness
(IMT) and lower flow-mediated dilation (FMD) have been reported in cross-sectional
evidence, in patients with ACS, when compared to non-functioning tumors [5]
[6]
[7]. The association between altered IMT and severity of hypercortisolism has been recently
investigated in a meta-analysis of 14 studies that characterized the cardiovascular
risk of patients with Cushing’s syndrome. IMT, FMD, and atherosclerotic plaques were
independently associated with severity and duration of cortisol excess, as shown in
patients with active disease [8]. Moreover, improvement of IMT was correlated with the median time of disease remission
in cured patients, providing indirect evidence on the role of duration of cortisol
excess as an important determinant of the development of subclinical atherosclerosis
[8]. Instead, the association between markers of subclinical atherosclerosis and ACS
is not yet completely understood, as shown by a few studies providing contrasting
results. The lack of agreement among the different studies is mainly due to the different
parameters that were selected as markers of cortisol hypersecretion, like urinary
free cortisol or midnight cortisol, which are frequently within the normal range in
patients with ACS. Nonetheless, the association between ACS and increased IMT seems
to be a relevant issue only in older patients [5]. Several studies shed light on the potential association between altered markers
of subclinical atherosclerotic and non-functioning adrenal incidentalomas (reviewed
in [9]). Even with the limitation due to the cross-sectional design, those studies raise
the question whether steroids other than cortisol may have a role in impairing the
cardio-metabolic profile of patients with adrenal incidentalomas. Indeed, retrospective
evidence showed that patients bearing non-functioning adrenal incidentalomas have
a higher incidence of diabetes in a long-term setting [10].
Thrombotic diathesis is a well-known consequence of Cushing’s syndrome. Thromboembolic
events have been identified as an important causative factor for cardiovascular events,
together with hypertension, in up to 20% of the patients [11]. Even though fewer studies have investigated the hypercoagulable state in patients
with ACS, alteration of several parameters (protein C, free protein S, thrombomodulin,
and alpha-1 antitrypsin) resembling that of Cushing’s syndrome have been reported
in patients with subclinical cortisol hypersecretion [12].
Long-term retrospective studies shed light on the association between cardiovascular
events and excessive cortisol production in patients with adrenal incidentalomas.
Three independent studies published in 2014 showed that the incidence of myocardial
infarction and stroke was indeed higher than that of non-functioning tumors [13]
[14]
[15]. As shown by Cox regression models, the incidence of cardiovascular events was associated
with increasing values of cortisol secretion over time, giving feed to the hypothesis
that the progression of hypercortisolism may be an important contributing factor for
development cardiovascular events [13]. Interestingly, among the potential contributing factors, previous cardiovascular
events, absence of lipid lowering treatment, and hypertension were independent contributors
to the incidence of cardiovascular events [13]. Similarly, all-cause and cardiovascular-related mortality were also increased in
patients with ACS, with respect to non-functioning tumors [13] and to the general population [15]. Therefore, cortisol levels and their progression over time seem to have a relevant
role in determining cardiovascular events in patients with adrenal incidentalomas.
However, the cortisol-related mechanisms (direct effects of cortisol on cardiovascular
system vs indirect actions through the associated co-morbidities) leading to the cardiovascular
damage are still not completely understood. It is important to notice that among cardiovascular
diseases, the myocardial infarction is more prevalent over stroke in patients with
ACS [16]. Recent evidence highlighted that cardiac performance may be altered in patients
with hypercortisolism, adding interesting and novel features to the cardiovascular
profile of patients with ACS. A recent cross-sectional analysis of a large cohort
of patients with incidentalomas revealed that left ventricular mass index, a marker
of cardiac hypertrophy associated with hypertension, and pulse wave velocity, a marker
of arterial stiffness, were significantly higher in patients with ACS [17]. Moreover, concentric and eccentric left ventricular hypertrophy was more prevalent
in patients with ACS. Notably, those alterations occurred independently of the presence
and duration of hypertension, and the number of hypertensive drugs, which were not
different between patients with non-functioning tumors and those with ACS. Those direct
and surrogate markers of altered cardiac performance were independently associated
with cortisol levels after DST [17].
Another interesting aspect of the relationship between cardiovascular diseases and
ACS is related to two distinct features of hypercortisolism, the severity and the
progression of cortisol hypersecretion. As previously mentioned, the occurrence of
cardiovascular diseases is independently predicted by the increasing severity of hypercortisolism
over time [13]. Additionally, all-cause mortality, with cardiovascular diseases as the most frequent
cause, was associated with the mean of cortisol after DST during follow-up, uncovering
the importance of the persistence of hypercortisolism over time [13]. The role of cortisol as a risk factor for mortality in ACS has been confirmed by
the observation that patients belonging to three different classes of increasing severity
of hypercortisolism (cortisol after DST<1.8 µg/dL, between 1.8 and 5 µg/dL and>5 µg/dL,
respectively) are at parallel increased risk of all-cause and cardiovascular-related
mortality [15].
Additionally, the alteration of the circadian cortisol rhythm has been recognized
as a potential contributing factor for the development of cardiovascular events. Evidence
from population-based studies and targeted investigations revealed that flattening
of diurnal cortisol rhythm is associated with increased cardiovascular-related mortality
in the general population and a higher incidence of adverse cardiovascular events
in patients undergoing coronary artery bypass graft surgery [18]
[19]. Less is known about the alterations of the circadian cortisol rhythm in patients
with ACS. This issue has been investigated in a few studies, providing contrasting
results, mainly due to the different definition of ACS, the different sampling conditions
(inpatients for serum cortisol vs outpatients for salivary cortisol), and the different
assays for cortisol (immunometric assays vs mass spectrometry-based measurements).
A recent cross-sectional study by Ceccato et al. provided the results of the analysis
of the circadian salivary cortisol measured by mass spectrometry in 106 patients with
adrenal incidentalomas [20]. Patients with cortisol levels after DST>1.8 µg/dL showed a higher area under the
curve (AUC) for salivary cortisol in morning measurements and throughout the day,
than those with non-functioning tumors. The difference in the AUC was mainly due to
significantly increased levels of cortisol in the morning. No difference was detected
in evening and late-night cortisol levels between the two groups. Even though this
study points toward a higher morning cortisol exposure in patients with cortisol after
DST>1.8 µg/dL, no subgroups analysis for morning ACTH levels nor analysis of the association
between cortisol exposure and clinical correlates have been performed. Apart from
AUC, more informative analyses have been proposed as integrated measurements of circadian
rhythmicity, such as cosinor analysys [21], deconvolution parameters and approximate entropy [22]. However, none of those parameters have been applied to the study of adrenal incidentalomas.
The role of circadian cortisol rhythm disruption as a potential marker of severity
of hypercortisolism in adrenal incidentalomas is still under investigation. It may
be speculated that the altered cortisol rhythm may identify a subgroup of patients
at higher risk of co-morbidities. Indeed, in a recent proof of concept study, Debono
et al. investigated the potential implications of the restoration of a normal circadian
cortisol rhythm in patients with incidentalomas, by using a steroidogenesis inhibitor
[23]. Among six patients with increased evening cortisol levels, administration of Metyrapone
500 mg at 6 P.M. and 250 mg at 10 P.M. was able to re-assess the cortisol rhythm.
This effect was associated with a parallel normalization of IL-6 levels, a known inflammatory
marker associated with cardiovascular risk and endothelial disfunction. Larger prospective
studies are strongly needed to investigate the clinical impact of such a targeted
treatment in a long-term run.
An important, still unresolved, issue is the impact of co-secretion of aldosterone
and steroid precursors in patients with ACS and the implication for cardiovascular
diseases. A recent study investigating the 24-h urinary output of steroid metabolites
by mass spectrometry pointed toward a prevalent glucocorticoid excess in patients
with primary aldosteronism in a non-negligible proportion of patients [24]. The co-secretion was highlighted biochemically, by the increased levels of several
glucocorticoid precursors in urine samples, and clinically, by the onset of adrenal
insufficiency after surgery. In a recent case series, the immunohistochemical analysis
of five tumors associated with ACS and concomitant primary aldosteronism revealed
that those tumors were mainly composed by zona fasciculata-like cells, with a heterogeneous
expression of CYP11B1 and CYP11B2
[25]. A recent study investigating the serum steroid profiling by mass spectrometry in
patients with different subtypes of Cushing’s syndrome showed increased levels of
11-deoxycortisol and 11-deoxycorticosterone in adrenal-dependent hypercortisolism,
when compared to controls [26]. Those alterations were not associated with a parallel increase in aldosterone,
giving feed to the hypothesis that those steroid precursors may be produced mainly
by fasciculata cells. Additionally, an earlier pilot study published in 2015, showed
a hyperresponsiveness of several steroid precursors under 1-24 ACTH stimulation in
patients with unilateral adrenal adenomas associated with ACS [27]. The potential pejorative impact of the co-secretion of other steroids in association
with cortisol on the cardiovascular system of patients with ACS is an open topic and
deserves further investigation.
Risk of bone fractures
Despite the well-known association between steroid excess and bone fracture risk in
patients taking exogenous glucocorticoids, a few studies have been performed in patients
with endogenous hypercortisolism. Indeed, endogenous and exogenous glucocorticoid-induced
osteoporosis represent two different models of diseases, in terms of severity and
duration of cortisol exposure. The mechanisms of glucocorticoid-induced osteoporosis
are beyond the scopes of this review and are summarized in a recent paper [28].
The appropriateness of the assessment of bone fractures risk in patients with hypercortisolism
by evaluation of bone mass density (BMD) has been largely questioned. Indeed, cortisol-related
bone fractures have mainly been associated with altered bone quality and occur independently
of the BMD [29]. The occurrence of osteoporosis has been described in 28–50% of patients with overt
Cushing’s syndrome [30], whereas vertebral fractures have been reported in a variable proportion of patients
with hypercortisolism (21–76%) [31]
[32]. Such a wide range is mainly due to the study design and the different methods of
assessment of vertebral fractures (retrospective evaluations vs targeted investigation
by radiological images). Notably, the prevalence of asymptomatic vertebral fractures
among patients with endogenous hypercortisolism was 48% [31]. Fewer studies have investigated the association between ACS and bone fractures.
According to a recent metanalysis, the prevalence of vertebral fractures has been
estimated in 64% (95% confidence interval – CI 56-71%) [29]. Similarly, the incidence of vertebral fractures during follow-up has been recognized
as a major issue, occurring in 48% of the patients [29]. Among other known risk factors, the increased incidence of new vertebral fractures
was associated with the presence of ACS (odd ratio 12, 95% CI 4-37), defined by the
presence of at least two alterations among cortisol after DST>3 µg/dL, increased urinary
free cortisol levels, and ACTH levels<10 pg/ml [33].
Considering the limitations of using the BMD as a predictor of vertebral fractures
in patients with Cushing’s syndrome, and that the major problem of cortisol-induced
osteoporosis is qualitative rather than quantitative, several techniques have been
studied to investigate the bone quality in patients with hypercortisolism. Spinal
deformity index (SDI) and trabecular bone score (TBS) have been used frequently, because
of their feasibility and relative low cost. SDI is a semiquantitative method that
integrates number and severity of bone fractures as a surrogate marker of bone quality
[34]. TBS is a grey-level texture measurement calculated from the 2-dimensional images
of the dual energy X-ray absorptiometry (DEXA) scans at the lumbar level, which gives
information about the microarchitecture of the bone [35]. Patients with ACS showed higher SDI than non-functioning adrenal tumors, in cross-sectional
and prospective settings [33]
[36]. The application of TBS to the study of patients with adrenal incidentalomas revealed
that ACS is characterized by lower TBS values than those of patients with non-functioning
adrenal tumors and control subjects. Moreover, TBS was associated with occurrence
of vertebral fractures and cortisol levels, independently of other potential contributing
factors [37]. Interestingly, TBS was also associated with the incidence of new vertebral fractures
over a 2-year period, independently of the BMD [37]. A recent large cross-sectional study has confirmed the superiority of TBS over
BMD in the evaluation of bone fragility in patients with various degree of hypercortisolism,
including ACS and Cushing’s syndrome of different etiologies. Patients with ACS showed
lower TBS values (1.3±0.1), when compared to patients with non-functioning adrenal
tumors (1.4±0.1) (P<0.04), regardless of BMD [38].
Bone fractures in patients with hypercortisolism have been associated with either
exposure and duration of cortisol hypersecretion [39]. Even though the excessive cortisol production undoubtedly plays a pivotal role
in this context [28], the involvement of additional factors associated with hypercortisolism in increasing
the risk of fractures, like sarcopenia, as a potential risk of fall, is still under
investigation. Indeed, patients with hypercortisolism are characterized by severely
altered muscle mass and performance, which is correlated with severity of hypercortisolism,
as shown by radiological and clinical evidence [40]
[41]. Even though the differences were clearer for overt Cushing’s syndrome, a reduced
muscle mass was also detected in patients with ACS, who do not have clinical stigmata
of sarcopenia. A thorough evaluation of muscle performance in patients with ACS is
still lacking to date. Recent evidence highlighted that some degrees of altered protein
metabolism may feature the biochemical phenotype of patients with ACS, with several
disturbances of amino acid metabolism, including reduction of histidine, proline,
and kynurenine levels [42]. Those alterations have been found in patients with overt Cushing’s syndrome and
in patients with ACS and were associated with the severity of hypercortisolism [42]. Considering the role of histidine and proline as component of collagen matrix,
it may be speculated that their alteration may have a potential impact on muscle strength
and sarcopenia. To further support the muscular implication of ACS, a recent report
on a small cohort of patients showed that women affected by ACS had a significant
reduction in muscle mass than those with non-functioning tumors, as assessed by bioelectrical
impedance analysis [43]. Additionally, the role of reduced androgens has been identified as an important
determinant of reduced muscle mass, which is a well-known cortisol-related clinical
consequence [44]. Several reports have highlighted that patients with either Cushing’s syndrome or
ACS have lower levels of DHEA and androstenedione than subjects bearing non-functioning
tumors and controls, independently of age [26]
[27]. The relationship between sarcopenia, androgens, and risk of fall has been highlighted
in a very recent large population-based studies, showing that lower DHEA and DHEA-sulphate
were associated with increased incidence of falls [45].
A thorough evaluation of muscle performance and its relationship with reduced androgens
in patients with ACS deserves further evaluation.
Risk of infectious diseases
Infectious diseases are a known complication of hypercortisolism. A recent population-based
cohort study showed that patients with overt Cushing’s syndrome are at increased risk
of mortality due to infections, with hazard ratio of 4.9 (95% CI 3.7-6.4) [46]. The increased mortality rate was associated with cortisol secretion independently
of the specific subtypes of Cushing’s syndrome and persisted elevated for several
years after treatment [46]. The cortisol-related mechanisms underlying immune system disruption are still not
completely understood. Direct and indirect effects of cortisol, through immunosuppression,
vascular damage, and hyperglycemia are among the most frequent causes of susceptibility
to infections of patients with Cushing’s syndrome. A summary on the molecular mechanisms
driving alterations in immune response in hypercortisolism has been recently published
elsewhere [11]. Apart from the severity of hypercortisolism, altered circadian cortisol rhythm
seems to play a significant role in the alteration of the immune response, mainly
by disruption of the circadian lymphocyte rhythm. This concept has been highlighted
in a recent randomized controlled trial, showing that restoration of a pseudo-physiological
circadian rhythmicity of cortisol improved the immune cell profile of patients with
adrenal insufficiency [47]. Specifically, patients receiving modified-release hydrocortisone tabs showed a
reduction in the number of pro-inflammatory monocytes CD14(+) CD16(-), lowering their
susceptibility to infections [47].
The relevance of infectious complications in patients with ACS has been underestimated
and rarely investigated until recently. A long-term retrospective study on the natural
history of ACS, previously mentioned, revealed that the infectious complication was
the second cause of mortality in patients with ACS, after cardiovascular diseases
[15]. Moreover, the prevalence of infections was higher than that of the general population.
This was the first study highlighting the clinical impact of prolonged, although mild,
hypercortisolism on the immune system in the long-term run. Moreover, the reduction
in IL6 levels after restoration of cortisol rhythm by Metyrapone pointed toward a
possible link between altered circadian rhythmicity and infectious complications [23]. Further confirmation is awaited from targeted studies.
The Identification of High-Risk Patients
A summary of the risk factors for clinical outcomes in patients with ACS is depicted
in [Fig. 1]. Considering the clinical picture of ACS, it is feasible to speculate about the
potential phenotype of patients at risk of developing clinically-relevant outcomes.
Cortisol levels after DST>1.8 µg/dL may be considered as a reliable cut-off to identify
a subgroup of patients with adrenal incidentalomas who are at risk of developing cardiovascular
events, provided that the measurement is confirmed over time. Moreover, increasing
levels of cortisol after DST during subsequent controls should be interpreted as an
additional risk factor for cardiovascular diseases. The cortisol-related risk seems
to be an independent contributor to cardiovascular events, mainly in patients with
altered cardiovascular profile. The main implication of this hypothesis is that patients
with adrenal incidentalomas presenting with hypertension, dyslipidemia, and previous
cardiovascular events should be considered for cortisol-lowering treatments, if affected
by ACS, or addressed to a careful hormonal follow-up to rule out progression of cortisol
secretion, if bearing a non-secreting tumor. Whether co-secretion of additional steroids
and altered circadian cortisol rhythm are pejorative factors is unknown to date.
Fig. 1 Summary of the potential predictive factors for cardiovascular morbidity, bone fractures,
and infectious complications associated with autonomous cortisol secretion. Continuous
arrows indicate putative mechanisms with evidence in the literature. Dashed arrows
indicate potential mechanisms under investigation. HPAA: hypothalamic-pituitary-adrenal
axis. DST: dexamethasone suppression test.
Conversely, the risk of incident bone fractures seems to be relevant in patients with
cortisol levels after DST at a higher cut point (i. e.>3 µg/dL) and additional hormonal
alterations, such as reduced ACTH levels and elevated urinary cortisol. It is feasible
that a more severe hormonal profile may be necessary for the development of bone fragility.
Even though additional factors like decreased androgens and sarcopenia seem to contribute
to the risk of fractures, stronger evidence is needed. Overall, the occurrence of
asymptomatic vertebral fractures has been recorded in half of the patients with ACS.
Therefore, a fracture risk assessment should be recommended in all patients with ACS
by morphometric evaluation of the spine and assessment of surrogate markers of bone
fragility (e. g. SDI or TBS).
The phenotypic characterization of patients at risk of infectious diseases is cumbersome.
Given that patients with disrupted circadian rhythm seem to be prone to develop infectious
complications, a careful examination of daily cortisol profiles may be useful in all
patients with ACS. However, it is important mentioning that the impact of infectious
diseases on the well-being of patients with ACS is still under investigation.
