Hypercortisolism enhances cardiovascular risk factors such as glucose intolerance, central obesity, hypertension, and dyslipidemia. All are linked to an increased incidence of atherosclerosis and coronary disease, and impact on morbidity, cardiovascular disease being the leading cause of death in patients with Cushing’s syndrome (CS). However, this cardiovascular risk profile is not completely explained by conventional cardiovascular risk factors; other still inadequately defined disease-specific factors, partially related to the hypercoagulable and inflammatory state with an unfavorable adipokine profile, have also been observed [1]. Although most of the risk factors improve, cardiovascular risk is clearly increased in CS patients even years after remission (Fig. 2) [2].

Cortisol excess induces a procoagulative phenotype (activation of coagulation cascades and impaired fibrinolysis), so that patients with CS have a greater predisposition to thromboembolic events, especially in the perioperative period. This hypercoagulable state in CS is explained by higher levels of procoagulant factors, mainly factors VIII, IX, and von Willebrand factor, as well as an impaired fibrinolytic capacity, due to increase synthesis of the plasminogen activator inhibitor type 1 (the main inhibitor of the fibrinolytic system) [15]. Consequently, there is a shortening of activated partial thromboplastin time and increased thrombin generation [50, 51]. Moreover, both a rise in platelets and endothelial dysfunction observed in patients with CS predispose to increased cardiovascular risk and play a role in the pathogenesis of the prothrombotic state in patients with CS [52, 53]. The incidence of venous thromboembolism (VTE) in CS is higher than in the general population (2.5–3.1 vs. 1.0–2.0 per 1000 persons/year, respectively) [51, 54]. Patients who undergo transsphenoidal surgery for CD have greater risk of thromboembolism than those for a nonfunctional pituitary adenoma, suggesting a role of cortisol (or ACTH) inducing changes in hemostatic factors [54]. Hemostatic and fibrinolytic parameters did not normalize 80 days after biochemical remission with medical therapy [55]. In the same line, in a systematic review the authors observed that even after remission of hypercortisolism , v Willebrand Factor, VII, and IX factors remained high [51]. An improvement in hemostatic parameters after one year of successful surgery has been described, but complete normalization of hemostasis does not occur [56].

In a recent study, an increase risk for VTE (Hazard Ratio, HR 2.6) in patients with CS was found to be already present 3 years before diagnosis, being highest the first year after diagnosis (HR 20.6) and still remained elevated from 1 to 30 years after diagnosis, although most of the cases occurred during persistent hypercortisolism [28].

Although it is still a matter of debate whether systematic antithrombotic prophylaxis in CS should be used, it seems that thromboprophylaxis could be recommended in patients with CS undergoing surgery. However, there is no consensus on the dose or duration of use of prophylactic anticoagulant therapy. Prospective placebo-controlled trials to evaluate the effects of thromboprophylaxis in patients with CD are still lacking.

Remission rates after pituitary surgery can be achieved for 65–100 % of patients. These percentages are lower in patients with a non-visible adenoma, microadenoma with unfavorable localization or macroadenomas and recurrence rates can reach 5–36 % [4]. Secondary hypothyroidism and hypogonadism are common in patients with CS, due to the functional suppression of thyrotropin and gonadotropin secretion by GC excess. After normalization of cortisol secretion, these endocrine abnormalities usually recover, as well as normal menstrual cycles and sexual activity. However, due to structural damage of the residual pituitary gland after surgical removal of the tumor or prolonged inhibition of the hypothalamic–pituitary–adrenal axis, permanent hormone deficiency may occur (hypopituitarism or adrenal insufficiency) [57].

After surgery the most common pituitary insufficiency observed is GH deficiency (which is not always evaluated), followed by thyrotropin and gonadotropin deficiencies [58]. Some patients require life-long replacement with exogenous GC.

The prevalence of bone disease, mainly osteoporosis, is high and often underestimated in patients with CS, since not all patients undergo DXA scans, and asymptomatic vertebral and rib fractures can remain undiagnosed. Approximately 30–50 % of CS patients present with fractures, particularly vertebral fractures [3]. Additionally, osteoarthritis and osteonecrosis have been reported mainly in patients with iatrogenic CS, but rarely in patients with endogenous hypercortisolism .

GCs have direct and indirect effects on bone, including decrease osteoblastic and increased osteoclastic activity, reduced intestinal calcium absorption, and increased urinary calcium excretion which induces in both cases a modest increase in parathyroid hormone levels [63]. Deleterious effects on bone, especially on cortical bone microstructure, have been observed using a high-resolution peripheral quantitative computed tomography in patients with active CS [64]. Furthermore, secondary hypogonadism and/or decreases in GH or IGF1 levels induced by excessive amounts of cortisol contribute to the loss of bone mineral density (BMD). The pathophysiology of bone disease in CS is detailed in (Fig. 3) [65].

Studies evaluating bone status after biochemical control of hypercortisolism , however, are often conflicting. While some observed a reversal of GC-induced osteoporosis, others showed an incomplete recovery of BMD and quality of bone after remission. Reversal of GC-induced osteoporosis after long-term remission of CS (mean 72 months) has been described in parallel with increased osteocalcin levels [66]. In the same line, after remission of hypercortisolism, bone mass changes were reversible, probably due to the fact that prior exposure time to endogenous hypercortisolism was shorter than in other studies [67]. The mechanisms causing BMD recovery are speculative. They could be attributed to an increase in osteocalcin levels and to the preservation of trabecular architecture despite the thinning induced by GCs, so osteoblasts may continue synthesizing new bone [68].

A partial recovery of BMD and bone quality after treatment for CS has been reported in most studies (in adolescents and adult patients), although the series are small and median follow-up is relatively short (less than 2 years) [69]. In the series with longer follow-up after remission of hypercortisolism (mean 11 years), decreased BMD values were seen in estrogen-sufficient women as compared to age-, sex- and BMI-matched controls, but not in women with estrogen deficiency, suggesting that the protective effect of estrogens on bone mass is lost with hypercortisolism . Prior exposure time to excess endogenous cortisol and the duration of postoperative GC replacement therapy were predictors of low BMD in these patients [70].

In a group of 50 cured CS, with a median remission time of 13 years, BMD was not significantly different at any site between patients and age- and gender-matched controls. The authors observed that the NR3C1 Bcl1 polymorphism of the GC receptor was associated with reduced total and femoral neck BMD, and patients with ongoing GC replacement presented worse skeletal health (reduced total and lumbar spine BMD) [10].

In summary, BMD recovery appears to be only partial in most patients with “cured” CS .

Around 60–80 % of CS patients present with proximal muscle atrophy and weakness, more frequently in males [3]. GC-induced changes in muscle are evident after a few days of GC exposure or administration, with a more prominent effect on proximal muscles [71]. In aging subjects without CS, muscle mass loss were not associated to circulating or urinary cortisol, but muscle strength correlated with quadriceps expression of 11ß-HSD1, supporting the importance of tissue-specific cortisol metabolism and conversion, rather than overall circulating levels in determining negative effects of GCs [72]. Muscle damage can persist both short- and long-term after cure; it has been related to protein synthesis inhibition and increased rate of protein degradation of myofibrillar and extracellular matrix proteins. Indeed, reduced arm muscle area showed no relevant improvement 6 months after successful treatment, despite a reduction of body fat mass [11]. Reduced lean mass due to muscle loss of limb muscle was observed in CS compared to obese controls with same total body fat mass [73]. In a long-term follow-up, patients with CS had reduced limb skeletal muscle mass, but similar lean body mass compared to age- and gender-matched controls [10]. MRI body composition assessment of CD patients 20 months after remission showed that total and limb skeletal muscle is actually reduced compared to active disease, probably due to the GC replacement therapy after cure [9]. Moreover, postmenopausal women in remission presented with similar muscle mass as active disease patients, suggesting a role of estrogen deprivation in muscle mass as well [9]. Creatinin kinase, plasma myoglobin, and muscle fiber conduction velocity were reduced in the active phase of the disease compared to healthy age-, sex-, and BMI-matched controls and correlated with disease duration [74]. It has been suggested that aerobic and resistance exercises could probably be effective in attenuating GC-induced muscle atrophy [75].

Nephrolithiasis has been reported in 50 % of active CD and 30 % of cured CD patients compared to 6.5 % in age- and gender-matched controls [4, 76]. The pathogenesis of nephrolithiasis in CS is not fully elucidated. There is probably a synergistic effect of different metabolic changes (hypercalciuria, hypocitraturia, hyperuricosuria, and hyperoxalaturia) together with hemodynamic changes caused by hypercortisolism. In fact, obesity, hypertension, and diabetes, common features of CS, have been seen more frequently in patients with kidney stones. It seems that normalization of cortisol levels can restore the amino acid profile in urine. In a large series investigating the role of different lithogenic factors in CS, high blood pressure and excessive excretion of uric acid were found to be independent risk factors for the recurrence of nephrolithiasis [76, 77].

Chronic hypercortisolism has been related to changes in memory, behavior, verbal learning, neuronal activity, and other processes of the central nervous system. Psychiatric disturbances have been reported in 54–81 % in different series, major depression and irritability being the most common psychiatric disorders. Emotional lability, mania, paranoia, acute psychosis, anxiety, and panic attacks may also occur in CS [78, 79]. Few reports assess psychopathology after effective surgery; although most of these symptoms and changes improve one year after remission, many persist and do not appear to be fully reversible in the long-term follow-up. An increased prevalence of psychopathology and maladaptive personality traits compared to patients with nonfunctioning pituitary adenomas (NFPAs) and matched controls have been found, indicating that cortisol excess has irreversible effects on the central nervous system, rather than any effect of the pituitary tumor itself [80]. Recently, a retrospective study in a group of patients with CD who underwent bilateral adrenalectomy, with a median follow-up of 11 years, observed improvements in almost all Cushing-specific comorbidities, except for psychiatric morbidities (which included self-reported anxiety, depression, panic attacks, and psychosis ) [81].

On the other hand, the hippocampus, amygdala, and cerebral cortex are important structures involved in cognitive and emotional functions. These structures are rich in GC receptors and, therefore, particularly vulnerable to hypercortisolism. Moreover, 11ß-HSD2 (which inactivates cortisol to cortisone) is not expressed in the hippocampus or limbic structures, which allows the sustained activation of mineralocorticoid receptors by GCs. Since common genetic polymorphism variants in the GC receptor and the 11ß-HSD1 have been recently associated with long-term cognitive impairments in CS in remission (for a median time of 13 years), these results indicate that GC sensitivity and prereceptor regulation of GC action may play a role in the etiology of cognitive dysfunction in these patients [82].

After successful biochemical treatment of CD, psychiatric symptoms may decrease, but patients still show cognitive impairment, decreased quality of life, and a higher prevalence of affective disorders and apathy compared to healthy controls [83–86]. Long-lasting impairments have been reported in several domains of cognitive (attention, visuospatial orienting, alerting, working, verbal and visual memory, verbal fluency, reading speed) and executive functions [83–85]. Higher prevalence of “maladaptive” personality traits in CD, even after long-term cure, has been described [80]. Impaired decision-making together with decreased cortical thickness in selective frontal areas irrespective of the activity of disease has also been observed in CS patients compared to healthy controls, suggesting that chronic hypercortisolemia promotes brain changes which are not reversible after endocrine remission [86]. In the same line, mental fatigue, characterized by mental exhaustion and long recovery time following mentally strenuous tasks, is more common in patients with CS in remission compared to healthy education-, age-, gender-matched controls, according to a very recent study [87].

Decreased hippocampal volume (HV) assessed by 3-T cerebral MRI was seen in CS patients with severe memory impairments compared to controls [83]. Both brain atrophy and reduction in total and cortical grey matter volumes have also been observed in CS compared to controls, but subcortical gray matter reduction has only been seen in those patients with severe memory impairments in parallel to the findings of reduced HV. The negative effects of GC excess on memory and HV seem to be not totally reversible after biochemical cured, since no differences, either in HV or in memory performance between active and cured CS, were found [83]. Brain volumes and cognitive functions have been associated with cardiovascular risk in CS patients in remission [31]. Furthermore, using a proton magnetic resonance spectroscopy, lower N-acetyl-aspartate in the hippocampus (suggesting neuronal dysfunction/loss) and higher levels of glutamate (suggesting glial proliferation as a repair mechanism after neuronal dysfunction) have been observed in cured CS patients compared to matched controls [88]. The authors suggest that these persistently abnormal metabolites could be early markers of GC neurotoxicity, preceding HV reduction [88]. In major depressive disorder patients , similar patterns of reduced HV and reversibility of hippocampal atrophy after treatment have been observed [89]. Moreover, widespread reductions of white matter integrity (reflecting a structural abnormality of white brain matter, like demyelination or loss of axonal integrity) in CD patients with long-term remission (mean 11.9 years) compared with matched controls have also been observed, together with abnormalities in the integrity of the uncinate fasciculus being related to the severity of depressive symptoms [90]. Similarly, structural abnormalities of the brain white matter have been identified with diffusion tensor imaging (DTI) in the brains of CS patients, again suggesting a widespread loss of axonal integrity and demyelination compared to controls [91]. Once present, these alterations seem to be independent of concomitant hypercortisolism, persisting after remission, since a more of these white matter lesions have also been found in CS in remission compared to healthy controls [31, 91]. Moreover, reduced anterior cingulate cortex grey matter volumes and greater volume of the left posterior lobe of the cerebellum in patients with long-term remission of CD (mean 11.2 years) have been observed compared to matched controls [92]. However, another study observed a smaller bilateral cerebellar cortex volume in active CS compared to matched controls [93]. Recently, aberrant resting-state functional connectivity of the brain with the limbic network (responsible for emotional processing and regulation, as well as encoding of memories) and executive control network has been observed in CD patients with long-term remission, suggesting that hypercortisolism may lead to persistent changes in brain functional connectivity (involving episodic memories, semantic knowledge, prospective memory, attention demands, working memory, and cognitive control) [94]. In the same line, altered neural processing of emotional faces after long-term remission of hypercortisolism has been recently reported in CD compared to matched healthy controls [95].