Background: Equipoise exists regarding the benefits of restricting caloric intake during electrolyte replacement for refeeding syndrome, with half of intensive care specialists choosing to continue normal caloric intake. We aimed to assess whether energy restriction affects the duration of critical illness, and other measures of morbidity, compared with standard care. Methods: We did a randomised, multicentre, single-blind clinical trial in 13 hospital intensive care units (ICUs) in Australia (11 sites) and New Zealand (two sites). Adult critically ill patients who developed refeeding syndrome within 72 h of commencing nutritional support in the ICU were enrolled and allocated to receive continued standard nutritional support or protocolised caloric restriction. 1:1 computer-based randomisation was done in blocks of variable size, stratified by enrolment serum phosphate concentration (>0?32 mmol/L vs =0?32 mmol/L) and body-mass index (BMI; >18 kg/m2 vs =18 kg/m2). The primary outcome was the number of days alive after ICU discharge, with 60 day follow-up, in a modified intention-to-treat population of all randomly allocated patients except those mistakenly enrolled. Days alive after ICU discharge was a composite outcome based on ICU length of stay, overall survival time, and mortality. The Refeeding Syndrome Trial was registered with the Australian and New Zealand Clinical Trials Registry (ANZCTR number 12609001043224). Findings: Between Dec 3, 2010, and Aug 13, 2014, we enrolled 339 adult critically ill patients: 170 were randomly allocated to continued standard nutritional support and 169 to protocolised caloric restriction. During the 60 day follow-up, the mean number of days alive after ICU discharge in 165 assessable patients in the standard care group was 39?9 (95 CI 36?4-43?7) compared with 44?8 (95 CI 40?9-49?1) in 166 assessable patients in the caloric restriction group (difference 4?9 days, 95 CI -2?3 to 13?6, p=0?19). Nevertheless, protocolised caloric restriction improved key individual components of the primary outcome: more patients were alive at day 60 (128 [78 of 163 vs 149 [91 of 164, p=0?002) and overall survival time was increased (48?9 [SD 1?46] days vs 53?65 [0?97] days, log-rank p=0?002). Interpretation: Protocolised caloric restriction is a suitable therapeutic option for critically ill adults who develop refeeding syndrome. We did not identify any safety concerns associated with the use of protocolised caloric restriction. Funding: National Health and Medical Research Council of Australia.