Long-Term Consequences of Cushing Syndrome: A Systematic Literature Review

Abstract

It is held that the condition of endogenous chronic hypersecretion of cortisol (Cushing syndrome, CS), causes several comorbidities, including cardiovascular and metabolic disorders, musculoskeletal alterations, as well as cognitive and mood impairment. Therefore, CS has an adverse impact on the quality of life and life expectancy of affected patients. What remains unclear is whether disease remission may induce a normalization of the associated comorbid conditions. In order to retrieve updated information on this issue, we conducted a systematic search using the Pubmed and Embase databases to identify scientific papers published from January 1, 2000, to December 31, 2022. The initial search identified 1907 potentially eligible records. Papers were screened for eligibility and a total of 79 were included and classified by the main topic (cardiometabolic risk, thromboembolic disease, bone impairment, muscle damage, mood disturbances and quality of life, cognitive impairment, and mortality).

Although the limited patient numbers in many studies preclude definitive conclusions, most recent evidence supports the persistence of increased morbidity and mortality even after long-term remission. It is conceivable that the degree of normalization of the associated comorbid conditions depends on individual factors and characteristics of the conditions. These findings highlight the need for early recognition and effective management of patients with CS, which should include active treatment of the related comorbid conditions. In addition, it is important to maintain a surveillance strategy in all patients with CS, even many years after disease remission, and to actively pursue specific treatment of comorbid conditions beyond cortisol normalization.

hypercortisolism, comorbidities, hypertension, diabetes, osteoporosis, myopathy, thromboembolism, depression, cognitive impairment, mortality

Cushing syndrome (CS) is a condition that may cause a variety of serious health risks as a result of the chronic hypersecretion of cortisol. The list of comorbidities associated with CS includes cardiovascular (CV) and metabolic disorders, musculoskeletal alterations, as well as cognitive and mood impairment, which adversely impact the quality of life and life expectancy of affected patients.

Given the rarity of CS, with an estimated incidence and prevalence of 0.7 to 2.4 cases/million/year, and 40 cases/million, respectively (1), there is a scarcity of long-term prospective studies, and thus limited information on the long-term consequences of cortisol excess. Despite efforts to normalize cortisol levels as soon as possible in patients with CS, it remains unclear whether disease remission can induce a complete restoration of the comorbidities developed during the active phase of CS, during which the patients are exposed to a chronic cortisol excess. Due to the scarcity of data, it is difficult to provide a definitive answer, since in most studies patients are re-evaluated only after a short (3-6 months) or intermediate (1-4 years) remission period.

The aim of this work was to conduct a systematic search of the evidence in order to retrieve updated information on the long-term consequences of overt CS.

Methods

Search Strategy

In January 2023, a systematic review was conducted to identify scientific papers published from January 1, 2000, to December 31, 2022, using the Pubmed and Embase databases and adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (2); see Fig. 1. The details for the search are reported in the Supplementary Appendix (3). Results were restricted to the English language and studies involving human subjects.

Figure 1.

PRISMA statement flow chart for study screening and selection.

Screening of the Studies

Two independent investigators (A.M.E.P. and C.B.) screened the papers for potential inclusion by reviewing titles and abstracts. Case reports, conference abstracts, and reviews were excluded. Full-text articles were then screened for eligibility. The inclusion criteria were (i) original articles with a cohort of at least 15 adult patients (≥18 years of age at time of diagnosis of CS) with diagnosis of CS of adrenal, pituitary or ectopic origin; and (ii) median follow-up of at least 4 years of disease activity for patients with active disease, or at least 4 years of remission time in the case of patients with disease in remission (Fig. 1). The criteria for establishing either the diagnosis of active CS or remission of CS were not critically appraised, and the specific results were used for this analysis as they were reported in the original studies.

If there was a disagreement, the senior investigators (S.P., F.O., and M.T.) reviewed the full text, and consensus was reached through discussion. We excluded papers on the consequences associated with autonomous cortisol secretion associated with adrenal incidentaloma, a condition previously named as subclinical CS (4), because this evaluation is beyond the scope of our work.

Study Selection

The initial search identified 1907 potentially eligible records. After duplicates were removed, 1847 papers were screened by title and abstract for eligibility, resulting in the exclusion of 1649 papers. Following a review of full-text manuscripts, a total of 79 of the remaining articles on the main long-term consequences of CS (cardiometabolic and thromboembolic diseases; bone and muscle damage; impairment of mood, cognitive functions, and quality of life; mortality) were identified and included in this systematic review (Fig. 1).

Papers were classified by the main topic. Each comorbidity was covered in a sub-section, with a brief description of the condition, the underlying evidence, and our interpretation of the available data.

Results

Cardiometabolic Risk

Comorbidities such as hypertension, hyperglycemia, and dyslipidemia are well-known consequences of chronic cortisol excess and contribute to the increased cardiometabolic risk in CS. However, the reversibility of all these conditions and, consequently, the possibility to normalize the risk of CV events with the remission of CS remains a debated issue.

Impairment of glucose metabolism is frequently found in patients with CS, with diabetes mellitus (DM) diagnosed in 20% to 47% and impaired glucose tolerance in 21% to 64% of cases (5). Cortisol excess leads to a deranged glucose metabolism through several mechanisms: (i) promoting gluconeogenesis through the induction of the hepatic gluconeogenic enzymes and stimulation of lipolysis and proteolysis that provide substrates for gluconeogenesis; (ii) reducing insulin sensitivity of liver and muscle through the impairment of the intracellular signaling pathway of insulin; and (iii) inhibiting insulin secretion by the pancreatic β-cells (5).

Dyslipidemia is also commonly found in patients with CS, with high cholesterol levels reported in 16% to 60% and high triglycerides levels in 7% to 36% of cases. These findings are the result of the cortisol-mediated stimulation of several enzymes involved in lipolysis, lipogenesis, and adipogenesis (5).

Hypertension is reported in up to 95% of patients with CS and there are several mechanisms contributing to its pathogenesis in the context of cortisol excess, including: (i) saturation of the enzyme that physiologically converts cortisol to cortisone (11β-hydroxysteroid dehydrogenase type 2) by the elevated cortisol levels, resulting in an increased mineralocorticoid mimetic action that causes salt retention and volume expansion; (ii) inhibition of the production of vasodilators (prostacyclin, prostaglandins, kallikreins, nitric oxide); (iii) increase in the levels of vasoconstrictor agents; and (iv) increase in the vascular sensitivity to vasoconstrictor agents (angiotensin II, catecholamines) (6).

Studies have reported that cure of hypertension in patients in long-term remission ranged from 36% to 75% (7-10) (Fig. 2A). Interestingly, the frequency of hypertension among patients with CS in remission remains higher than among healthy controls (11, 12), while similar rates have been described between patients with CS in remission and patients affected by nonfunctioning pituitary adenoma, who served as controls (9). After long-term remission, cure of DM/impaired glucose tolerance ranged from 56% to 83% (Fig. 2B), from 23% to 60% for dyslipidemia (Fig. 2C), and from 44% to 60% for obesity (Fig. 2D) (7, 9, 10). In these studies, the median or mean remission time ranged from 4.1 to 11.2 years (Table 1).

Figure 2.

Percentage of patients who normalized blood pressure (A), glucose levels (B), lipids (C) and body mass index (D) after remission of Cushing syndrome. *In panel B, data from reference (9) and (10) concern the category of diabetes only; data from reference (7) concern both diabetes and glucose intolerance. ^#In panel C, data from reference (9) concern low-density lipoprotein (LDL) levels only.

Table 1.

Characteristics of the studies included in the systematic review

Reference	N CS patients	Type of Cushing	Patient management	% Patients in remission	FU (years)	Comorbidities or parameters assessed	Controls	Study period
(7)	50	50 Adrenal	Surgery (UAX)	100%	RT = mean 11.2 (range, 1.1-22)	Cardiometabolic risk (hypertension, diabetes, dyslipidemia, obesity)	NA	Jan 1980 - Dec 2000
(8)	71	50 Pituitary 21 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT = mean 4.1 ± SD 4.4	Cardiometabolic risk (hypertension)	NA	1978 -2003
(9)	75	24 ACS: 17 Pituitary 4 Adrenal 3 EAS 51 RCS: 33 Pituitary 15 Adrenal 3 EAS	Pituitary: Pituitary surgery ± RTX and/or BAX Adrenal: UAX or BAX EAS: Surgery of ACTH-secreting tumor Medical therapy for patients who refused surgery or with occult EAS	68%	RT for RCS = median 4.7 (range, 1-16) FU for ACS = median 2 (range, 1-16.7)	Cardiometabolic risk (hypertension, diabetes, dyslipidemia obesity) Mortality	Patients with NFPA	1991-2010
(10)	118	52 Pituitary 58 Adrenal 8 EAS	Surgery (23.7% BAX)	100%	RT for RCS = median 7.9 (range, 2-38)	Cardiometabolic risk (hypertension, diabetes, dyslipidemia, obesity) Mood disturbances (depression)	NA	2012-2018
(11)	29	21 Pituitary 8 Adrenal	Pituitary surgery ± RTX and/or BAX UAX or BAX	100%	RT for RCS = mean 11 ± SD 6	Cardiometabolic risk (coronary artery disease, hypertension)	Healthy subjects matched for age and sex	Since 1982
(12)	37	27 Pituitary 10 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (unilateral or BAX)	100%	RT = mean 11 ± SD 6 (range, 0.7-22)	Cardiometabolic risk (body fat mass, hypertension)	Healthy subjects matched for age, BMI, and sex	Since 1982
(13)	58	40 Pituitary 18 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT for RCS = mean 13.6 ± SD 8.0	Cardiometabolic risk (body fat mass, inflammatory markers)	Healthy subjects matched for age, BMI, and sex	NA
(14)	50	38 Pituitary 12 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT for RCS = median 13 (IQR 5-19)	Cardiometabolic risk (body fat mass) Bone impairment (BMD)	Healthy subjects matched for age and sex	NA
(15)	51	14 ACS: 12 Pituitary 2 Adrenal 37 RCS: 27 Pituitary 10 Adrenal	NA	72.5%	DA = mean 5.8 ± SD 0.4 RT = mean 11 ± SD 6 (range, 0.7-22)	Cardiometabolic risk (body fat mass, inflammatory markers)	Patients affected by rheumatoid arthritis, treated with exogenous GCs	NA
(16)	41	31 Pituitary 10 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	70.7%	RT for RCS = mean 11 ± SD 6 DA = mean 5.8 (range, 1.1-20.1)	Cardiometabolic risk (coronary calcifications)	Healthy subjects matched for age and sex	Since 1982
(17)	58	40 Pituitary 18 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT for RCS = median 13.6 (range, ± 8.0)	Cardiometabolic risk (vascular endothelial dysfunction)	Healthy subjects matched for age, BMI, and gender	NA
(18)	346	346 Pituitary	Pituitary surgery ± RTX and/or BAX In case of persistent disease, medical therapy	89.4%	FU = mean 6.3 (range, 0.1-30)	Cardiometabolic risk (CV events, including AMI, stroke, and VTE), Mortality	NA	1980-2011
(19)	343	211 Pituitary 132 Adrenal	NA	NA	FU > 1 to 30	Cardiometabolic risk (AMI, stroke) Thromboembolic disease (VTE) Bone impairment (fractures) Mortality	General Denmark population	1980-2010
(20)	502	502 Pituitary	364 Pituitary surgery 129 RTX 102 BAX 31 Medical treatment	83.5%	RT for RCS = median 7 (IQR 3-14)	Cardiometabolic risk (AMI, stroke) Thromboembolic disease (VTE)	Swedish general population	Jan 1990-Dec 2013
(21)	473	360 Pituitary 113 Adrenal	Pituitary: Pituitary surgery ± RTX and/or BAX ± cortisol-lowering medication Adrenal: Surgery (unilateral or BAX) ± cortisol-lowering medication	88.6%	FU = median 6.0 (range, 0.7-6.0)	Thromboembolic disease (VTE)	Patients with NFPA	Jan 1990-Jun 2010
(22)	18	9 ACS: 9 Pituitary 9 RCS: 8 Pituitary 1 Adrenal	NA	50%	DA for ACS = average 2.8 ± SD 0.9 RT for RCS = median 9.7 ± SD 5.9;	Thromboembolic disease (vascular endothelial dysfunction, hypercoagulability)	Healthy subjects matched for age, BMI, and sex	Jan 2012-Jan 2016
(23)	19	16 Pituitary 2 Adrenal 1 EAS	Pituitary: Pituitary surgery Adrenal: UAX EAS: BAX	100%	HT for RCS = median 5 (range, 4-6)	Thromboembolic disease (hypercoagulability)	Healthy subjects matched for age and gender	2011-2014
(24)	30	24 Pituitary 6 Adrenal 2 EAS	16 no treatment 9 surgery ± ketoconazole 5 ketoconazole	0%	DA = mean 4.84 ± SD 4.53	Bone impairment (BMD)	Healthy subjects matched for age, BMI, and sex	NA
(25)	52	38 Pituitary 14 Adrenal	NA	0%	DA = mean 5.25 ± SD 5.68	Bone impairment (Fractures)	NA	Feb 2007-Dec 2012
(26)	33	26 Pituitary 7 Adrenal	Surgery	100%	18 RCS patients had long-term RT (mean 5.9; range, 3.6–9)	Bone impairment (BMD)	Healthy subjects matched for age, BMI, and sex	1994-1999
(27)	231	188 Pituitary 43 Adrenal	Pituitary: Pituitary surgery ± RTX and/or BAX Adrenal: unilateral or BAX 8 patients were not treated and 8 patients were on medical therapy	NA	89 RCS patients had long-term RT (data at 5, 10, 15, 20, and 25 years after treatment)	Bone impairment (BMD, fractures)	NA	1968-2020
(28)	51	14 ACS 13 Pituitary 1 Adrenal 37 RCS: 27 Pituitary 10 Adrenal	Pituitary: Pituitary surgery ± RTX and/or BAX Adrenal: UAX	72.5%	RT for RCS = mean 11 ± SD 6	Bone impairment (BMD)	Healthy subjects matched for age, BMI, and sex	Since 1982
(29)	20	15 Pituitary 3 Adrenal 2 EAS	Pituitary: Pituitary surgery ± BAX Adrenal: UAX EAS: BAX	100%	FU = mean 4.75 (range, 1-10)	Bone impairment (BMD)	NA	1998-2009
(30)	104	76 Pituitary 28 Adrenal	Surgery	85% Pituitary 100% Adrenal	66 RCS patients had long-term RT (>5)	Bone impairment (fractures)	Healthy subjects matched for age and sex	1985-1995
(31)	196	47 ACS: 33 Pituitary 10 Adrenal 4 EAS 139 RCS: 95 Pituitary 38 Adrenal 16 EAS	Surgery	70.9%	RT for RCS = mean 13 (range, 2-53)	Muscle damage	Patients referred with suspicion of CS, which was subsequently ruled out	NA
(32)	88	49 Pituitary 34 Adrenal 5 EAS	Surgery	100%	22 patients had long-term RT (FU at 4 years)	Muscle damage	Healthy subjects matched for age, BMI, and sex	2012-2018
(33)	17	10 Pituitary 7 Adrenal	Pituitary surgery Surgery (UAX)	100%	RT for RCS = median 11.3 (range, 4-28)	Muscle damage	Healthy subjects matched for sex, age, BMI, estrogen status, smoking, ethnicity, physical activity	1985-2009
(34)	36	28 Pituitary 8 Adrenal	Pituitary surgery ± RTX Surgery (UAX)	100%	RT for RCS = mean 13 ± SD 8	Muscle damage	Healthy subjects matched for age, BMI, and sex	NA
(35)	179	179 Pituitary	Pituitary surgery ± RTX	89.9% (116 out of 129 patients with FU at 5 years)	FU of entire cohort = median 7.0 (IQR 4.0-9.8); 76 RCS patients had long-term RT (median 9.3; IQR 8.1-10.4)	Mood disturbances (use of psychotropic drugs)	Healthy controls matched for sex, age, and residential area	1991-2018
(36)	166	99 Pituitary 48 Adrenal 16 Cancer associated 3 Other	Surgery	NA	FU = median 8.1 (range, 3.1-14)	Mortality Quality of life	General Danish population	1985-1995
(37)	343	305 Pituitary 17 Adrenal 21 EAS	NA	100%	RT = mean 11.8 ± SD 4.9	Quality of life	Healthy subjects matched for age and sex	NA
(38)	55	43 Pituitary 12 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT = median 13 (IQR 5-19)	Mood disturbances Cognitive impairment	Healthy controls matched for age, gender, and educational level	NA
(39)	58	58 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 13.4 ± SD 6.7	Quality of life Mood disturbances	Healthy subjects matched for age and sex	1978-2002
(40)	25	25 Pituitary	Pituitary surgery ± BAX	100%	RT for RCS = mean 8.6 ± SEM 1.6	Quality of life Mood disturbances Cognitive impairment	Subjects matched for age, gender, living area, socioeconomic status, and educational level	NA
(41)	123	99 Pituitary 24 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT = mean 13.3 ± SD 3.2	Quality of life Mood disturbances	Healthy subjects matched for age and sex	1967-2007
(42)	15	15 Pituitary	Pituitary surgery ± RTX	100%	RT = mean 7.7 (range, 4.6-10.9)	Quality of life Mood disturbances	Treated patients with NFPA, macroprolactinoma, or acromegaly	NA
(43)	36	25 Pituitary 11 Adrenal	Pituitary surgery ± RTX and/or medical therapy before surgery Surgery (UAX)	100%	RT = Mean 6 ± SD 12.8	Quality of life Mood disturbances	Healthy controls matched for sex, age, and BMI	NA
(44)	51	51 Pituitary	Pituitary surgery ± RTX	100%	RT = mean 11 ± SD 9	Mood disturbances	Healthy controls matched for age, gender, and educational level	NA
(45)	19	14 Pituitary 5 Adrenal	Pituitary surgery Surgery (UAX)	100%	RT = median 7 (IQR 6-10)	Mood disturbances Cognitive impairment	Healthy controls matched for age, gender, and education	NA
(46)	51	39 Pituitary 12 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT = median 12 (IQR 4-18)	Mood disturbances Cognitive impairment	Healthy controls matched for age, gender, and education	NA
(47)	25	25 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 11.2 ± SD 8.2	Mood disturbances Cognitive impairment	Healthy controls matched for age, sex, and years of education	NA
(48)	22	22 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 11.87 ± SD 8.49	Cognitive impairment	Healthy matched controls for age, gender, and education	NA
(49)	21	21 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 10.8 ± SD 7.9	Mood disturbances Cognitive impairment	Healthy controls matched for age, gender, and education	NA
(50)	25	25 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 11.2 ± SD 8.2	Mood disturbances Cognitive impairment	Healthy controls matched for age, gender, and education	NA
(51)	35	NA	8 ACS (4 on medical therapy, 4 without treatment) 7 controlled CS on medical therapy 20 RCS after surgery	57.1%	RT for RCS = mean 6.1 ± SD 6.9	Mood disturbances Cognitive impairment	Healthy controls matched for sex, age, and years of education	NA
(52)	50	50 Pituitary	49 Pituitary surgery 13 RTX 5 Ketoconazole	74%	RT = mean 7.1 ± SD 10.4	Mood disturbances	Age- and gender-matched mentally healthy controls and patients with NFPA	NA
(53)	102	102 Pituitary	Pituitary surgery ± RTX and/or BAX	92%	FU = mean 7.4 (range, 0.4-32.4)	Quality of life Mood disturbances	NA	1980-2012
(54)	141	90 Pituitary 51 Adrenal	Pituitary surgery ± RTX NA	51.1%	RT for 72 RCS = mean 9.69 ± SD 8.05	Quality of life	NA	March 2007-March 2017
(55)	297	206 Pituitary 91 Adrenal	NA	NA	Time from diagnosis = mean 7.8 ± SD 10.1	Quality of life	NA	Sep 2007-Apr 2014
(56)	70	21 EAS vs 59 Pituitary	Surgery of the ACTH-secreting tumor or BAX	100%	RT = median 6.1 (IQR 7.7) in EAS and 8.7 (IQR 15.2) in Pituitary Cushing	Quality of life	NA	From 1990
(57)	269	184 Pituitary 67 Adrenal 12 Other	Surgery ± RTX and/or BAX and/or medical treatment	74.2%	RT for RCS = mean 6 ± SD 6	Quality of life	NA	NA
(58)	34	34 Pituitary	17 BAX vs 17 other treatment (pituitary surgery and/or RTX and/or medical therapy)	100%	RT = mean 8.4 ± SD 5.6	Quality of life	NA	2000-2015
(59)	52	46 Pituitary 6 Adrenal	Pituitary surgery ± RTX and/or adrenalectomy Surgery (adrenalectomy)	100%	RT = mean 16 ± SD 12	Quality of life	Comparison with reference populations	NA
(60)	42	42 Pituitary	Pituitary surgery ± RTX	100%	RT = mean 13 ± SD 10	Mood disturbances	Healthy patients, patients with acromegaly or NFPA	NA
(61)	176	176 Pituitary	Pituitary surgery ± RTX	NA	Time since last surgery = mean 6.8 ± SD 6.66	Quality of life Mood disturbances	NA	2005-2013
(62)	63	63 Pituitary	Pituitary surgery	NA	19 patients had long-term DA (>6)	Cognitive impairment (cerebral atrophy)	Healthy subjects matched for age and sex	Jul 1992-Dec 1995
(63)	33	11 ACS: 7 Pituitary 4 Adrenal 22 RCS: 18 Pituitary 3 Adrenal 1 EAS	11 ACS: 10 on medical treatment (2 failed surgery), 1 not treated 22 RCS: surgery ± RTX	66.7%	DA for ACS = mean 5.5 ± SD 3.7; RT for RCS = mean 7.3 ± SD 2.4	Cognitive impairment	Healthy controls matched for age, sex, and years of education	NA
(64)	36	15 ACS: 10 Pituitary 4 Adrenal 1 EAS 21 RCS: 18 Pituitary 3 Adrenal	15 ACS: 12 on medical treatment 21 RCS: surgery ± RTX	58.3%	DA for ACS = mean 5.18 ± SD 4.92 RT for RCS = mean 5.15 ± SD 2.68	Cognitive impairment (cerebral cortex volume)	Healthy controls matched for age, sex, and years of education	NA
(65)	120	120 Pituitary	Pituitary surgery		DA for 60 ACS = mean 6.56 ± SD 9.07 DA for 28 short-term RCS = mean 4.65 ± SD 5.67 DA for 32 long-term RCS = mean 4.95 ± SD 4.69	Cognitive impairment (gray matter volumes)	Healthy controls matched for age, sex, and education	NA
(66)	15	15 Pituitary	Free of therapy	0%	DA = median 4 (IQR 7)	Cognitive impairment (dendritic density)	Healthy controls matched for age and sex	NA
(67)	18	15 Pituitary 3 Adrenal	Pituitary surgery ± RTX Surgery (UAX)	100%	RT = mean 8.5 ± SD 3.2	Cognitive impairment (hippocampal dysfunction)	Healthy controls matched for age and years of education	NA
(68)	25	25 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	RT = mean 11.2 ± SD 8.2	Cognitive impairment (gray matter volumes)	NA	NA
(69)	87	87 Pituitary	Pituitary surgery	44.8%	DA for 48 ACS = mean 5.4 ± SD 5.9 DA for 39 RCS = mean 4.8 ± SD 4.2	Cognitive impairment (cerebral microbleeds)	Healthy controls matched for age, sex, and education	NA
(70)	32	32 Pituitary	Pituitary surgery	43.7%	DA for 18 ACS = mean 5.5 ± SD 3.57 DA for 14 RCS = mean 4.24 ± SD 4.97	Cognitive impairment (brain activity)	Healthy controls	NA
(71)	24	24 Pituitary	Pituitary surgery	100%	RT = mean 10.92 ± SD 8.36	Mood disturbances Cognitive impairment	Healthy matched controls for age, gender, and education	NA
(72)	19	14 Pituitary 5 Adrenal	Pituitary surgery Surgery (UAX)	100%	RT = median 7 (IQR 6-10)	Cognitive impairment	Age-matched female controls	NA
(73)	16	16 Pituitary	Pituitary surgery ± RTX and/or BAX	NA	DA = mean 14.6 (95% CI 9.4-19.8)	Mood disturbances Cognitive impairment	Patients with NFPA	NA
(74)	74	74 Pituitary	Pituitary surgery ± RTX	100%	RT = mean 13 ± SD 13	Cognitive impairment	Healthy controls matched for age, gender, and education	NA
(75)	53	42 Pituitary 11 Adrenal	Pituitary surgery ± RTX and/or BAX Surgery (UAX)	100%	RT for RCS = median 13 (IQR 5-18)	Cognitive impairment	Healthy controls matched for age, sex, and years of education	NA
(76)	289	289 Pituitary	Pituitary surgery ± RTX and/or BAX	82%	179 patients had long-term FU (median 11.1; range, 0.6-24.1)	Mortality	Age- and sex-matched general population	1975-1988
(77)	253	188 Pituitary 46 Adrenal 19 EAS	Surgery ± RTX and/or BAX	89%	Mean FU 6.4 (range, 0-46)	Mortality	General New Zealand population	1960-2005
(78)	80	80 Pituitary	Pituitary surgery ± RTX and/or BAX	72%	Median FU 10.9 (IQR 4.9-15.6)	Mortality	General population of England and Wales	1988-2009
(79)	418	311 Pituitary 74 Adrenal 33 EAS	Surgery ± RTX and/or BAX; medical therapy for occult EAS	76.5% Pituitary 92% Adrenal 30% EAS	FU: Pituitary = mean 9 (range, 0.1-46) Adrenal = mean 3 (range, 0.1-21) EAS = mean 4 (range, 0.1-18)	Mortality	General population of England and Wales	1962 –2009
(80)	74	74 Pituitary	Pituitary surgery ± RTX and/or BAX	93%	Mean FU = 12.8 ± SD 7.3	Mortality	General Dutch population	1977-2005
(81)	386	240 Pituitary 124 Adrenal (95 benign, 29 malignant) 12 EAS 10 Unproven etiology	Surgery ± RTX and/or BAX and/or medical treatment	79%	Mean FU = 7.1 ± SD 7.8	Mortality	General Bulgarian population	1965-2010
(82)	502	502 Pituitary	Pituitary surgery ± RTX and/or BAX and/or medical treatment	83%	Median FU = 13 (IQR 6-23)	Mortality	General Swedish population	1987-2013
(83)	320	320 Pituitary	Pituitary surgery ± RTX and/or BAX	100%	Median FU = 11.8 (IQR 17-26)	Mortality	General population from UK, Denmark, New Zealand, Netherlands	2009-2014
(84)	371	371 Pituitary	Pituitary surgery ± RTX and/or BAX and/or medical treatment	92% at 5- years FU, 97% at 20-years	Median FU = 10.6 years (range, 0.02-28, IQR 5.7-18)	Mortality	Healthy controls matched for sex, age, and residential area	1991-2018
(85)	172	172 Pituitary	Pituitary surgery ± RTX and/or BAX and/or medical treatment	64% (on 133 cases)	Median FU = 7.5 (IQR 2.4-15)	Mortality	General Mexican population	1979-2018

Abbreviations: ACS, active Cushing syndrome; AMI, acute myocardial infarction; BAX, bilateral adrenalectomy; BMD, bone mineral density; BMI, body mass index; DA, disease activity; EAS, ectopic ACTH syndrome; FU, follow-up; GC, glucocorticoid; HR, hazard ratio; IQR, interquartile range; NA, not available; NFPA, nonfunctioning pituitary adenoma; RCS, Cushing syndrome in remission; RT, remission time; RTX, radiotherapy; SIR, standardized incidence ratios; UAX, unilateral adrenalectomy; UK, United Kingdom; VTE, venous thromboembolism.

Despite the high rates of remission of these cardiometabolic comorbidities, several studies have demonstrated that patients in long-term remission for over 10 years tend to present central adiposity (greater percentage of truncal fat mass and higher waist circumference) and an unfavorable adipokine profile (lower adiponectin concentrations, higher levels of leptin, resistin, soluble tumor necrosis factor [TNF]-α receptor 1, IL-6, apolipoprotein B, and insulin) more frequently than matched controls (12-14) and a higher percentage of trunk and total fat even compared to patients taking exogenous glucocorticoids (GCs) (15). This persistent abdominal adiposity and an accompanying state of low-grade inflammation leads to an increased CV risk, even after several years of remission. Limited data suggest that the increased abdominal fat mass is associated with the polymorphism rs1045642 in the ABCB1 gene and with GC replacement (14).

In 2013, Barahona et al reported a higher prevalence of subclinical coronary artery disease (coronary calcifications and/or noncalcified plaques) in women and young patients (<45 years) in long-term remission (mean ± SD, 11 ± 6 years) than in healthy controls by using a cardiac multidetector computed tomography (MDCT). No patient had a history of ischemic coronary disease at the time of the study, and only a man with CS in remission suffered from acute myocardial infarction that occurred 11 months after MDCT, requiring a triple bypass (11). In 2015, the same authors reported that almost one-third of patients with similar long-term remission had coronary calcifications, as assessed by the Agatston score, and found that plasma soluble TNF-α receptor 1 concentration was the main predictor of coronary lesions (16). However, Wagenmakers et al reported that the effects of cortisol excess on the vessels appear to be reversible. Patients in long-term remission (median 13.6 years, range ± 8.0) who did not have cardiometabolic comorbidities, or had adequately controlled comorbidities, had similar levels of serum biomarkers associated with endothelial dysfunction, values of intima media thickness, values of pulse wave velocity and pulse wave analysis than healthy matched controls (17). These findings highlight the importance of actively managing cardiometabolic comorbidities in conjunction with normalization of cortisol excess.

Interestingly, Lambert et al reported that male sex, depression, DM, and age at diagnosis predicted CV events, including acute myocardial infarction, stroke, and venous thromboembolism (VTE) (18). Schernthaner-Reiter et al found that in patients in long-term remission (median 7.9 years; range, 2-38), age, fasting glucose levels, body mass index (BMI), and the number of comorbidities at diagnosis were positively correlated to the number of long-term cardiometabolic comorbidities, which included hypertension, DM, obesity, and dyslipidemia. In contrast, 24-hour urinary free cortisol levels at diagnosis were inversely correlated to the number of long-term cardiometabolic sequalae, possibly because diagnosis of mild CS is delayed compared to overt CS; therefore, patients are exposed for a longer period to elevated cortisol levels before being diagnosed and treated (10).

Dekkers et al found that the risk of acute myocardial infarction or stroke remained higher than in the general population, even during a long-term follow-up (hazard ratio [HR] 3.6; 95% CI, 2.4-5.5 and HR 1.8; 95% CI, 1.1-3.0, respectively) (19), whereas Papakokkinou et al reported a persisting elevated risk for stroke only (standardized incidence ratio [SIR] 3.1; 95% CI, 1.8-4.9], but not for acute myocardial infarction (SIR 0.6; 95% CI, 0.1-1.8) (20).

Thromboembolic Disease

Endogenous hypercortisolism is known to increase the risk of thrombosis, which at diagnosis of CS has been found to be 10 times greater than in the general populations. The incidence of VTE in the postoperative period is similar to that occurring after major orthopedic surgery (86).

Several studies have shown that this thrombotic diathesis is primarily the result of a marked increase in the von Willebrand factor and factor VIII levels, combined with the inhibition of the fibrinolytic system due to increased levels of fibrinolytic inhibitors, such as the plasminogen activator inhibitor type 1 (87). Other minor alterations have been reported in CS, such as an increase in platelet number and functionality, higher protein C and protein S concentrations, and slight increase in factors II, V, IX, XI, and XII levels. However, some of these are likely expressions of compensatory mechanisms in the balance between prothrombotic and antithrombotic processes (87).

A large multicentric study on 473 patients with CS reported an overall incidence of VTE of 14.6 (95% CI, 10.3-20.1) per 1000 person-years, with most events occurring in the early postoperative period (from 1 week to 2 months after surgery) (21).

Whether the alterations causing the procoagulant imbalance are reversible after long-term remission of CS is still debated. A translational study compared the pro-inflammatory and prothrombotic potential of circulatory factors between a few patients in remission or with active disease and matched controls. The clinical study did not reveal significant differences between groups, while the in vitro study demonstrated that the sera of both groups of patients with active or in remission CS presented a significantly higher expression of the von Willebrand factor and increased platelet adhesion compared to healthy subjects. Therefore, the authors concluded that a greater thrombogenicity persists even after several years of remission (22).

This conclusion is supported by a Swedish nationwide population-based study, reporting an increased incidence of VTE (SIR 4.9; 95% CI, 2.6-8.4) in a large cohort of patients with CS in remission compared to the general population, with no difference between patients receiving GC replacement therapy or not (20). Similarly, Dekkers et al reported an increased risk of VTE at long-term follow-up (HR 1.6; 95% CI, 0.8-3.4), although the highest risk was observed during the first year after diagnosis (HR 20.6; 95% CI, 7.8 53.9), suggesting that surgery plays a key role (19).

However, in CS patients after 5 years of persistent remission, the endogenous thrombin potential (ETP)-ratio (ETP with/without thrombomodulin) was recently found to be significantly lower than before surgery and 6 months after surgery, and it was comparable to the ETP-ratio of the matched controls (23).

Bone Impairment

Endogenous hypercortisolism has a detrimental effect on bone health, with a prevalence of osteoporosis and nontraumatic fractures in CS patients of 28% to 50% and 15% to 50%, respectively (88). Interestingly, significantly higher rates of fractures have been reported in patients with ectopic CS than in patients with CS of pituitary or adrenal origin, and this finding was associated with higher urinary cortisol levels in ectopic CS (88).

The skeletal impairment is the final consequence of different mechanisms. Cortisol excess increases apoptosis of osteocytes and osteoblasts suppressing bone formation. On the other hand, cortisol excess promotes the maturation and survival of osteoclasts inducing bone resorption.

Cortisol excess also indirectly impact bone metabolism, through (i) calcium depletion, which is the result of reduced intestinal absorption and increased renal excretion; (ii) muscle wasting, with loss of the trophic effect on the skeletal system; and (iii) inhibition of gonadal axis, with reduced estrogen levels in females and testosterone concentrations in men, both associated with lower bone mineral density (BMD) (89).

Vertebral fractures, which are frequently asymptomatic, tend to be the most common manifestation of bone damage in CS. This has been historically interpreted as the result of a more deleterious effect of cortisol excess on the trabecular than on the cortical bone, given that the trabecular bone is more preponderant in the lumbar spine. Interestingly, vertebral fractures often occur in patients with CS with normal or minimally altered BMD. This discrepancy between dual-energy X-ray absorptiometry (DEXA) findings and the incidence of osteoporotic fractures suggests that more attention should be paid to the bone microstructure in CS patients. In line with this hypothesis, Dos Santos et al demonstrated that cortisol excess has negative effects on the microarchitecture of cortical bone. In fact, a high-resolution peripheral quantitative computed tomography showed lower cortical area and thickness both at the radius and at the tibia, and lower cortical density at the tibia in patients with long-term active CS than in matched controls (24). Interestingly, Trementino et al demonstrated that disease duration was correlated with the presence of peripheral (rib, metatarsal, wrist, and hip) fractures, while no correlation was found with the incidence of vertebral fractures, although these findings may have been affected by the small sample size of the cohort (25).

Data regarding the complete reversibility of skeletal impairment after several years of cortisol normalization are contradictory. Most studies have demonstrated a progressive improvement of BMD, with values in patients on long-term remission similar to those of the general population (14, 26, 27). Ragnarsson et al reported that BMD was not significantly different at any site between controls and patients, after a median remission time of 13 (interquartile range [IQR] 5-19) years (14). A prospective study, including patients who were evaluated after a mean of 33 months (first follow-up) and 71 months (second follow-up) of remission, reported a progressive increase in Z-scores at all compartments compared with baseline values; however, the full normalization compared with healthy controls occurred only at 72 months (26). A retrospective longitudinal cohort study showed a continuous albeit slow improvement in the Z-score up to 20 years after treatment (27). Conversely, 2 studies have reported a persistence of bone alteration (28, 29). Barahona et al showed that when the comparison with matched controls was made in the context of estrogen sufficiency, women with CS had less whole-body BMD and bone mineral content, and lower lumbar spine BMD and osteocalcin concentrations, despite a mean of 11 years of cortisol normalization. This suggests that the protective effect of estrogens on bones is lacking in CS patients on remission (28). Although there is a general improvement in bone density compared to baseline values, Randazzo et al highlighted that in 8 out of 11 patients, after a median remission of 7 years, the femoral and vertebral T-scores were in the range of osteopenia or osteoporosis (29).

Despite these conflicting results on BMD parameters, data on the real incidence of fractures after a long-term remission are encouraging. In fact, Vestergaard et al did not find an increased fracture rate in 66 patients evaluated with a self-administered questionnaire more than 5 years after successful surgical treatment of CS (30). Dekkers et al reported that the increased risk of fractures characterizing the first year after CS diagnosis (HR 3.8; 95% CI, 1.7-8.7) was markedly reduced with longer follow-up (HR 1.1; 95% CI, 0.8-1.6, at >1 to 30 years from diagnosis) (19). More recently, Van Houten et al confirmed these results, reporting a rapid decrease in the fracture rate within 2 years after treatment of CS (27).

Finally, few studies have reported that replacement therapy with GCs was associated with lower BMD values (14, 28).

Muscle Damage

The prevalence of cortisol-induced myopathy has been reported in 42% to 83% of patients with active CS, mainly affecting the proximal lower limbs. The muscular damage is the consequence of the inhibition of anabolic processes (ie, reduced synthesis of myofibrillar protein) and the promotion of catabolic mechanisms (ie, enhanced proteolysis), resulting in the atrophy of type IIa muscle fibers, which impairs muscle performance (1).

There are only a few studies on long-term residual morbidity in muscle function after cortisol normalization. A cross-sectional study showed reduced hand grip strength in CS patients in remission after a median remission time of 13 years. In addition, their performance at the chair rising test was poorer compared to healthy controls and not significantly different compared to patients with active CS (31). Similarly, an observational longitudinal study demonstrated that, after an initial improvement in the first months after surgery, hand grip and chair rising performance did not improve further at long-term follow-up after remission, with a persistent statistically significant difference compared with control subjects (32).

The reasons why the myopathy does not fully regress after years of cortisol normalization have been investigated by 2 cross-sectional studies. The first study reported a reduced aerobic exercise capacity in CS patients compared with control subjects without differences in muscle fiber, capillarization, and mitochondrial content, therefore suggesting the prevalent role of an impaired cardiac output causing limited blood flow and oxygen supply in the muscle tissue (33). The second study found that the intramuscular fatty infiltration of the thigh in patients with long-term remission was higher than in controls and was associated with impaired performance on functional tests. On the other hand, no difference in the tight muscle volume of the 2 groups was reported, suggesting a persistent alteration of the “quality” rather than “quantity” of the muscle tissue (34).

Mood Disturbances and Quality of Life

Patients with endogenous hypercortisolism are more likely to suffer from psychiatric symptoms, above all anxiety (up to 80%) and depression (up to 70%) (90). They also need more antidepressants, anxiolytics, and hypnotic drugs than matched controls (35). Similarly, the health-related quality of life (HRQoL) is negatively affected in patients with CS, due to the combination of different factors, including a compromised health perception caused by various symptoms and comorbidities, depressive mood, and cognitive impairment with loss of memory as well as disturbances in executive, intellective, and attentive functions (91).

Although a small improvement in HRQoL and mood has been described after surgery, a residual impairment seems to persist even after longer-term remission, and patients in remission have often shown unfavorable scores on HRQoL, psychopathology, and personality scales (36-52). An overview of all the assessment tools used in studies reporting worse mental and functioning outcomes in CS patients than in control groups is provided in Fig. 3.

Figure 3.

Overview of all questionnaires for mood disturbances and HRQoL that were reported to have poorer scores in patients with Cushing syndrome compared with matched controls. Abbreviations and references: Apathy scale (44, 47-50); ASE, Appearance Self-Esteem (41); BAI, Beck Anxiety Inventory (47-50); BDI-II, Beck Depression Inventory-II (43, 51); CES-D, Center for Epidemiological Studies Depression Scale (43); CIS, Checklist Individual Strength Questionnaire (41); CPRS-A, Comprehensive Psychopathological Rating Scale (38, 45); Cushing QoL, Cushing quality of life (40, 41, 43); DAPPs, Dimensional Assessment of Personality Pathology short-form (44); EPQ-RK, Eysenck Personality Questionnaire-RK (52); EuroQoL (43); FACT, Functional Assessment of Cancer Therapy (42); Fear Questionnaire, but only in the social phobia subscale (47-50); FIS, fatigue impact scale (38, 45); GHQ-28, General Health Questionnaire 28 (42); HADS, Hospital Anxiety and Depression Scale (39, 41, 42, 44); IDS, Inventory of Depression Symptomatology (47-50); Irritability scale (44); MASQ-30, Mood and Anxiety Symptoms Questionnaire short-form (44); MADRS, Montgomery-Asberg Depression Rating Scale (40, 47-50); MFI-20, Multidimensional Fatigue Index (39); MFS, mental fatigue scale (46); NHP, Nottingham Health Profile (39, 41); PANAS, Positive Affect Negative Affect Scale (43); PSS, Perceived Stress Scale (43); STAI, State-Trait Anxiety Inventory (43, 51); RAND-36 (41); SF36, Short-Form 36 (36-40); SAS, Social Adjustment Scale−modified (SAS1 and SAS2) (42); TPQ, Cloninger Tridimensional Personality Questionnaire (52); WHOQOL-BREF, World Health Organization Quality of Life Scale−abbreviated version (42).

Despite one study reporting that the normalization rate of depression was 52% at the last follow-up visit (10), patients in remission continued to use antidepressants and hypnotics more frequently than controls, even after years of cortisol normalization (35).

Several predictors of poor HRQoL have been evaluated in patients after remission, but findings are not consistent. The female sex appears to be a negative factor (39, 41, 53, 54), which has only been partially explained with mood disturbances associated with the menstrual cycle and fertility concerns (54) and the causes of this gender difference are still not clear. There are conflicting data on the impact of the etiology of CS. In fact, Wagenmakers et al did not find a difference in patients with CS of adrenal or pituitary etiology that has been treated successfully (41). Other studies showed poorer scores in HRQoL questionnaires in patients with pituitary than those with adrenal origin of hypercortisolism (54, 55). Osswald et al reported more favorable questionnaire scores in patients with ectopic than pituitary CS, particularly among women (56). The effect of age on HRQoL questionnaires is not clear (39, 53), whereas a higher BMI has been associated with a poorer quality of life (53). Some studies reported the negative effect of hypopituitarism (39, 57), but others failed to demonstrate it (53). The duration of exposure to cortisol excess seems to be another significant predictor, as the HRQoL score is associated with both the lag of time from the onset of symptoms to the diagnosis and the duration of remission (40, 41, 43, 56, 57). Only one study in our survey compared the long-term outcome of HRQoL in CS patients that had undergone different treatment approaches, reporting lower scores in HRQoL questionnaires in patients treated with bilateral adrenalectomy over transsphenoidal surgery, radiotherapy, and medical treatment, although the longer duration of cortisol exposure in the first group may have had an impact on these results (58). A recent study conducted on patients in remission found a negative correlation between depression and anxiety scores and serum brain-derived neurotrophic factor (BDNF—a mediator of neuronal differentiation and plasticity expressed in the brain areas, especially in those involved in mood and stress response) and morning salivary cortisone levels (43). Finally, several studies have reported on the relationship between attitude in dealing with the disease and psychosocial impairment, the key role of negative illness perceptions, and the less effective coping strategies in CS patients (53, 59-61).

Cognitive Impairment

Neurocognitive symptoms are common in CS patients during the active phase of the disease, but they may persist even after remission, and seem to involve both structural and functional brain alterations.

Among the structural abnormalities, a higher degree of premature cerebral and cerebellar atrophy has been detected on magnetic resonance imaging (MRI) in patients with long-term active disease than matched controls (62-65). When assessed by neurite orientation dispersion and density imaging (NODDI), CS patients showed a lower dendritic density of gray and white matter than healthy controls (66).

Furthermore, lower volumes of gray matter in brain regions that are important for emotional and cognitive processing (ie, hippocampus, amygdala, anterior and posterior cingulate cortex, cuneus, and precuneus) have been reported even after long-term remission compared with matched controls (47, 50). Similarly, lower fractional anisotropy values have been detected in the white matter of patients after CS remission, especially in the bilateral hippocampal cingulum, bilateral uncinate fasciculus, and corpus callosum (48). Interestingly, the severity of the depression symptoms has been associated negatively with fractional anisotropy values (48, 51).

In 2013, Resmini et al reported lower N-acetylaspartate concentrations as well as higher glutamate and glutamine levels in the hippocampus of patients with long-term remission than in controls, which were interpreted as signs of neuronal dysfunction and glial proliferation, respectively (67).

To explain the persistent structural alterations of the anterior cingulate cortex in long-term remission, a recent study integrated data from brain MRI scans performed in this group of patients with gene expression data derived from the Allen Human Brain Atlas, finding an underrepresentation of deactivated microglia and oligodendrocytes in the anterior cingulate cortex (68). In addition, a higher prevalence of cerebral microbleeds at the quantitative susceptibility mapping images was found in CS patients than in matched controls (69).

Among the functional alterations, altered spontaneous brain activity, which was evaluated with functional MRI and specific software to calculate the amplitude of low-frequency fluctuation and the regional homogeneity, has been found in patients with active disease, particularly in the posterior cingulate cortex/precuneus, occipital lobe/cerebellum, thalamus, right postcentral gyrus and left prefrontal cortex (70). Interestingly, the resting-state functional MRI showed an increased functional connectivity between the limbic network (including the hypothalamus, hippocampus, amygdala, insula, and parts of the nucleus accumbens) and the anterior cingulate cortex (71), but also in the medial temporal lobe and in the prefrontal cortex, in patients with long-term remission compared to healthy controls, with a negative association with the duration of remission (72). In addition, in the prefrontal cortex, patients in long-term remission showed reduced activation when processing emotional faces (49) and decreased functional response in episodic and working memory testing (45).

These structural and functional brain changes likely underpin poor cognitive test performances among CS patients, as has been demonstrated with a series of tests: Cognitive Failure Questionnaire (CFQ) (47, 49, 50); Trail Making Test (TMT) (46, 73), Adult Memory and Information Processing Battery (AMIPB) (73), Test of Everyday Attention (TEA) (73), Wechsler Memory Scale (74), Verbal Learning Test (74), Rey Complex Figure (74), Letter-Digit Substitution Test (74), Digit-Deletion Test (74) and Figure Fluency Test (74), digit symbol coding from the Wechsler Adult Intelligence Scale, third edition (WAIS-III NI) (38), digit span test (38, 75), verbal fluency test (FAS) (38, 75), DLS reading speed test (38, 75), Rey Auditory Verbal Learning Test (63), and the Rey-Osterrieth Complex Figure (63). According to Ragnarsson et al, patients with CS carrying polymorphisms in the GC receptor genes 11βHSD1 (rs11119328, previously associated with increased activity of the hypothalamus-pituitary-adrenal axis), and NR3C1 (Bcl1, previously associated with increased GC sensitivity) performed badly in some of these tests, but evidence on this aspect is limited (75).

In other studies, cognitive test performances of patients with CS in remission were not significantly worse than those of healthy controls, in terms of verbal episodic memory (assessed by the Rey Auditory Verbal Learning Test), verbal speed and fluency (assessed by the Isaac Set Test), visuo-spatial working memory (assessed by the Wechsler Memory Scale), visual memory and visuo-spatial constructional ability (assessed by the Rey-Osterrieth Complex Figure, Object Assembly and Block Design from WAIS-III) (40), and information processing speed (assessed by the Symbol Digit Modalities Test) (51).

Mortality

Although CS patients that do not achieve remission after surgery present a higher mortality risk, the question of whether long-term remission conveys a mortality rate similar to the general population is not clear (92).

Although in the early 2000s there were 2 studies that reported that successful treatment of CS was associated with normal long-term survival (36, 76), an increasing amount of evidence later suggested that an excess in the mortality risk persisted in patients in remission in comparison with the general population (77-79) or with patients harboring a nonfunctioning pituitary macroadenoma (80).

Interestingly, Yaneva et al reported that the mortality risk varies depending on the etiology, with a significantly higher standardized mortality ratio (SMR) for ectopic CS, adrenocortical carcinoma, or CS of unknown origin, compared to CS from pituitary adenoma, adrenal adenoma, or bilateral adrenal hyperplasia (81). However, in this study, the SMR in patients with pituitary CS in remission was not significantly different compared to the general population, whereas patients with active CS had a significantly increased mortality rate (81). Our research group reported a similarly higher mortality rate in active CS than in remitted CS (survival probability at the end of follow-up: 48% vs 75%, P < .0001) (9).

Ragnarsson et al reported that although the SMR in remitted pituitary CS (1.9; 95% CI, 1.5-2.3) was lower than in active pituitary CS (6.9; 95% CI, 4.3-10.4), it remained significantly higher than in the general population (P < .0001) (82). Similarly, a study including only pituitary CS patients with a long-term remission (>10 years) showed an increased overall SMR (1.61; 95% CI, 1.23-2.12), with the exception of patients who had undergone neurosurgery alone, who presented a SMR similar to the general population (0.95; 95% CI, 0.58-1.55) (83).

Finally, a recent study that used a control group matched for age, sex, and residential area, reported that the mortality risk was higher even in patients with pituitary CS with long-term biochemical remission (HR 1.5; 95% CI, 1.02-2.2) (84).

Several predictors of mortality have been identified: older age at diagnosis (18, 77, 79, 82, 84), male sex (81, 84), active disease (77, 81, 84, 85), duration of cortisol exposure (18, 81), higher adrenocorticotropic hormone (ACTH) or cortisol levels at diagnosis (18, 85), and DM (77, 83, 85). Interestingly, Lambert et al reported that depression increased the mortality risk in patients with remitted CS (18). In most of these studies, the main causes of death were CV diseases, followed by infections (81, 82, 84, 85).

To sum up, although most studies included a limited number of patients and are heterogeneous in terms of the definition of remission and report variable therapeutic strategies, evidence supports the persistence of an increased mortality risk even after long-term remission (Fig. 4).

Figure 4.

Forest plot presenting standardized mortality ratio (SMR) for patients with Cushing syndrome in remission. Data from Lindholm et al (1) include patients with pituitary Cushing; data from Lindholm et al (2) include patients with benign adrenal Cushing; data from Bolland et al (1) include patients with pituitary Cushing due to macroadenoma; data from Bolland et al (2) include patients with pituitary Cushing due to microadenoma. *In remission after the first procedure.

Discussion

Despite that long-term remission of CS is generally associated with an improvement in several associated diseases, cure of comorbidities is not universally achieved with the normalization of cortisol excess, requiring lifelong surveillance and appropriate management of the specific conditions.

Studies focusing on the treatment of comorbidities in patients in CS in remission are currently lacking and this precludes strong recommendations; however, several expert opinions suggest reasonable approaches, based on the pathophysiology of the comorbidities in CS.

In patients with CS in remission who remain hypertensive, a strict control of blood pressure (≤130/80 mmHg) is recommended with an intensive drug regimen including angiotensin-converting enzyme inhibitors or angiotensin receptor blockers as a first-line therapy (93). If blood pressure targets are not achieved, the addition of calcium channel-blockers as a second-line therapy should be considered (using caution in case of concomitant peripheral edema), whereas mineralocorticoid receptor antagonists, beta-blockers, and thiazide diuretics should be used as a third-line or fourth-line therapy (using caution in case of concomitant metabolic comorbidities that could be worsened by these drugs) (93).

In patients with CS in remission who have persistent hyperglycemia, insulin sensitizers, especially metformin, should be considered as first-line therapy, given the presence of insulin resistance due to cortisol excess (5, 94). If glycemic targets are not achieved, glucagon-like peptide 1 receptor agonists (GLP1-RA) or sodium glucose co-transporter 2 inhibitors (SGLT2-i) (using caution for the increased risk of genitourinary infections associated with this last drug class) should be considered as a second-line therapy given the positive effect of these classes of drugs on CV disease and mortality (95). On the contrary, sulfonylureas should be avoided due to poor efficacy and risk of hypoglycemia, and thiazolidinediones are not indicated due to the risks of water retention, heart failure and fractures (94). Finally, treatment with insulin analogues should be considered in patients who do not reach the glycemic targets, although insulin treatment may lead to weight gain (5, 94). Interestingly, no studies have been carried out on the use of SGLT2-i or GLP1-RA in patients with CS in remission, despite both classes of drugs demonstrating additional benefits in patients with DM on weight and blood pressure, protection against CV events and reduction of the risk of admission to hospital for heart failure, and reduction of cardiovascular and all-cause mortality (95). We think that the implementation of studies focusing on the benefits of these drugs in patients with residual cardiometabolic comorbidities despite CS remission is an unmet clinical need.

Moreover, it should be considered that all the above-mentioned comorbidities are prothrombotic conditions and this could explain why a higher risk of VTE persists regardless of the reversibility of the procoagulant imbalance associated with normalization of cortisol excess.

Regarding the bone damage, it has been suggested that patients with a presumed mild bone impairment (patients without fractures or young subjects) should be treated only with calcium and vitamin D supplementation. Instead, in patients with presumed severe bone damage (presence of fractures, long-term and/or severe hypercortisolism, age >70 years), antiresorptive drugs should be considered in addition to calcium and vitamin D supplementation (96). However, no clear advantage of anti-osteoporotic treatment after long-term remission of hypercortisolism has been demonstrated yet (27, 29). Caution is required in the choice of the drug because bisphosphonates may further suppress bone turnover and inhibit bone remodeling spontaneously occurring after correction of hypercortisolism. Therefore, teriparatide or denosumab may be more appropriate, although data on their effectiveness and safety in CS patients are lacking and a definitive recommendation cannot be provided. The individual clinical picture should guide the therapeutic choice. Vitamin D and calcium supplementation could be sufficient in CS sustained by benign adrenal adenomas, in which a rapid and definitive correction of hypercortisolism by surgery is usually reached. On the other hand, a more intense pharmacological approach might be more suitable in pituitary CS, which is characterized by a lower probability of remission (96).

Few data are available on the CS-associated myopathy in patients in remission and this condition appears to be only partially reversible after normalization of cortisol levels. Ways to improve the restoration process are being investigated in a prospective randomized trial assessing the effect of physiotherapy in the postoperative period (97).

This review has also highlighted that a poor quality of life perception, mood alterations, and impaired cognitive functions persist in patients with long-term remission, despite successful treatment. As the recovery of the structural and functional alterations is long and often incomplete, patients with CS may still show neuropsychological sequelae even after several years of remission, with a variable degree of impairment for each domain of brain function in different patients. Although appropriate predictors have not yet been fully validated, the inverse association between HRQoL and duration of hypercortisolism further emphasizes the importance of reducing exposure to cortisol excess through early diagnosis and prompt treatment.

The available evidence shows that the risk of death remains elevated after remission of CS, since in our systematic analysis the weighted mean SMR for patients with CS remission was 1.32 (95% CI, 1.12-1.66). However, there is heterogeneity among the relevant studies that could be explained by different methodological approaches. In some studies, disease remission was assessed at the last follow-up visit, while in others after the first surgical approach. Moreover, some studies reported specific SMRs for different etiologies of CS.

In our opinion, only a combined strategy of early recognition and effective treatment of CS and of related comorbid conditions could improve the survival chances in patients with CS in the long-term period of remission.

Conclusions

In summary, recovery from cortisol-induced complications is a lengthy and progressive process, and it is biologically plausible to assume that some patients will not recover entirely. Even after an extended period, the degree of normalization of the associated comorbid conditions depends on different individual factors and intrinsic characteristics of the disease. Therefore, it is important to maintain a proactive surveillance strategy in all patients with CS, even after many years of disease remission, and to actively pursue treatment of comorbid conditions beyond cortisol normalization.

Acknowledgments

We thank Nicoletta Colombi, librarian from the University of Turin, for her precious assistance with the literature search.

Disclosures

M.T. received research grants from HRA Pharma, and advisory board honoraria from HRA Pharma and Corcept Therapeutics; the other authors have stated explicitly that there are no conflicts of interest in connection with this article.

Data Availability

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

Abbreviations

 
• ACTH

  adrenocorticotropic hormone

 
• BMD

  bone mineral density

 
• BMI

  body mass index

 
• CS

  Cushing syndrome

 
• CV

  cardiovascular

 
• DM

  diabetes mellitus

 
• ETP

  endogenous thrombin potential

 
• GC

  glucocorticoid

 
• HR

  hazard ratio

 
• HRQoL

  health-related quality of life

 
• IQR

  interquartile range

 
• MRI

  magnetic resonance imaging

 
• SIR

  standardized incidence ratio

 
• SMR

  standardized mortality ratio

 
• TNF

  tumor necrosis factor

 
• VTE

  venous thromboembolism