Timo Deutschbein, Giuseppe Reimondo, Guido Di Dalmazi, Irina Bancos, Jekaterina Patrova, Dimitra Argyro Vassiliadi, Anja Barač Nekić, Miguel Debono, Pina Lardo, Filippo Ceccato, Luigi Petramala, Alessandro Prete, Iacopo Chiodini, Miomira Ivović, Kalliopi Pazaitou-Panayiotou, Krystallenia I Alexandraki, Felicia Alexandra Hanzu, Paola Loli, Serkan Yener, Katharina Langton, Ariadni Spyroglou, Tomaz Kocjan, Sabina Zacharieva, Nuria Valdés, Urszula Ambroziak, Mari Suzuki, Mario Detomas, Soraya Puglisi, Lorenzo Tucci, Danae Anastasia Delivanis, Dimitris Margaritopoulos, Tina Dusek, Roberta Maggio, Carla Scaroni, Antonio Concistrè, Cristina Lucia Ronchi, Barbara Altieri, Cristina Mosconi, Aristidis Diamantopoulos, Nicole Marie Iñiguez-Ariza, Valentina Vicennati, Anna Pia, Matthias Kroiss, Gregory Kaltsas, Alexandra Chrisoulidou, Ljiljana V Marina, Valentina Morelli, Wiebke Arlt, Claudio Letizia, Marco Boscaro, Antonio Stigliano, Darko Kastelan, Stylianos Tsagarakis, Shobana Athimulam, Uberto Pagotto, Uwe Maeder, Henrik Falhammar, John Newell-Price, Massimo Terzolo, and Martin Fassnacht
Background\ud \ud The association between cortisol secretion and mortality in patients with adrenal incidentalomas is controversial. We aimed to assess all-cause mortality, prevalence of comorbidities, and occurrence of cardiovascular events in uniformly stratified patients with adrenal incidentalomas and cortisol autonomy (defined as non-suppressible serum cortisol on dexamethasone suppression testing).\ud \ud \ud \ud Methods\ud \ud We conducted an international, retrospective, cohort study (NAPACA Outcome) at 30 centres in 16 countries. Eligible patients were aged 18 years or older with an adrenal incidentaloma (diameter ≥1 cm) detected between Jan 1, 1996, and Dec 31, 2015, and availability of a 1 mg dexamethasone suppression test result from the time of the initial diagnosis. Patients with clinically apparent hormone excess, active malignancy, or follow-up of less than 36 months were excluded. Patients were stratified according to the 0800–0900 h serum cortisol values after an overnight 1 mg dexamethasone suppression test; less than 50 nmol/L was classed as non-functioning adenoma, 50–138 nmol/L as possible autonomous cortisol secretion, and greater than 138 nmol/L as autonomous cortisol secretion. The primary endpoint was all-cause mortality. Secondary endpoints were the prevalence of cardiometabolic comorbidities, cardiovascular events, and cause-specific mortality. The primary and secondary endpoints were assessed in all study participants.\ud \ud \ud \ud Findings\ud \ud Of 4374 potentially eligible patients, 3656 (2089 [57·1%] with non-functioning adenoma, 1320 [36·1%] with possible autonomous cortisol secretion, and 247 [6·8%] with autonomous cortisol secretion) were included in the study cohort for mortality analysis (2350 [64·3%] women and 1306 [35·7%] men; median age 61 years [IQR 53–68]; median follow-up 7·0 years [IQR 4·7–10·2]). During follow-up, 352 (9·6%) patients died. All-cause mortality (adjusted for age, sex, comorbidities, and previous cardiovascular events) was significantly increased in patients with possible autonomous cortisol secretion (HR 1·52, 95% CI 1·19–1·94) and autonomous cortisol secretion (1·77, 1·20–2·62) compared with patients with non-functioning adenoma. In women younger than 65 years, autonomous cortisol secretion was associated with higher all-cause mortality than non-functioning adenoma (HR 4·39, 95% CI 1·93–9·96), although this was not observed in men. Cardiometabolic comorbidities were significantly less frequent with non-functioning adenoma than with possible autonomous cortisol secretion and autonomous cortisol secretion (hypertension occurred in 1186 [58·6%] of 2024 patients with non-functioning adenoma, 944 [74·0%] of 1275 with possible autonomous cortisol secretion, and 179 [75·2%] of 238 with autonomous cortisol secretion; dyslipidaemia occurred in 724 [36·2%] of 1999 patients, 547 [43·8%] of 1250, and 123 [51·9%] of 237; and any diabetes occurred in 365 [18·2%] of 2002, 288 [23·0%] of 1250, and 62 [26·7%] of 232; all p values