Background: Acute kidney injury occurs in one in four children admitted to an intensive care unit (ICU) and its severity is independently associated with increased patient morbidity and mortality. Early prediction of acute kidney injury has the potential to improve outcomes. In smaller, single-centre trial populations, we have previously derived and validated the performance of a renal angina index, a context-driven risk stratification system, to predict severe acute kidney injury in children and adolescents. Here, we tested the predictive accuracy of this index for severe acute kidney injury in a large heterogeneous population. Methods: We did a prospective, observational study (AWARE) that recruited patients in the ICUs of 32 hospitals in nine countries across Asia, Australia, Europe, and North America. All patients aged between 3 months and 25 years who were admitted to an ICU at least 48 h previously were eligible. Exclusion criteria were a history of stage 5 chronic kidney disease (ie, estimated glomerular filtration rate <15 mL/min per 1·73m2 or on maintenance dialysis) or kidney transplantation in the preceding 90 days. Patients' medical records were reviewed to collect data up to 3 months before (serum creatinine only), daily during the first 7 days, and on day 28 after ICU admission. For the assessment of the renal angina index, we included patients from the AWARE study who had full data from the day of ICU admission, day 3, and day 28, including serum creatinine concentrations and urine output measurements. The primary outcome was the presence of severe acute kidney injury (stage 2–3 acute kidney injury, according to Kidney Disease Improving Global Outcomes [KDIGO] guidelines) on the third day after ICU admission. We compared the performance of the renal angina index with changes in serum creatinine relative to baseline for prediction of the primary outcome. A score of eight points or more on the renal angina index defined fulfilment of renal angina; serum creatinine concentration relative to baseline was calculated using maximum serum creatinine concentration in the first 12 h of ICU admission). This trial is registered with ClinicalTrials.gov, number NCT01987921. Findings: Between Jan 1 and Dec 31, 2014, we obtained data for 1590 patients. 286 patients (18%) had fulfilment of renal angina. At day 3, severe acute kidney injury occurred in 121 (42%) patients positive for renal angina and 247 (19%) patients negative for renal angina (relative risk [RR] 2·23, 95% CI 1·87–2.66, p<0·0001). Of 368 (23%) patients with severe acute kidney injury, more had increased use of renal replacement and increased mortality than of the 1222 (77%) patients without severe acute kidney injury (40 [11%] vs 18 [2%], p<0.0001; and 28 [8%] vs 53 [4%], p=0·01). Fulfilment of renal angina showed better prediction for severe acute kidney injury than serum creatinine greater than baseline (RR 1.61, 95% CI 1·33–1·93; p<0·0001), which was maintained on multivariate regression (independent odds ratio for fulfilment of renal angina 3·21, 95% CI 2·20–4·67 vs serum creatinine greater than baseline 0·68, 0·49–4·94). Interpretation: Earlier and better prediction of severe acute kidney injury has the potential to improve patient outcomes associated with acute kidney injury. Compared with isolated, context-free changes in serum creatinine, renal angina risk assessment improved accuracy for prediction of severe acute kidney injury in critically ill children and young people. Funding: US National Institutes of Health.