<jats:title>Abstract</jats:title><jats:sec><jats:title>Objective</jats:title><jats:p>Neuroleptic malignant syndrome (NMS) is a potentially fatal, idiosyncratic reaction to antipsychotics. Due to low incidence of NMS, research on risk factors of mortality associated with NMS is limited.</jats:p></jats:sec><jats:sec><jats:title>Methods</jats:title><jats:p>Two authors independently searched Medline/Embase/Cochrane/CINAHL/PsychINFO databases for case reports with author‐defined NMS published in English until 05/30/2020. Demographic, clinical, treatment, and outcome data were independently extracted following PRISMA guidelines. NMS severity was rated using the Francis‐Yacoub scale. Mortality risk factors were identified using a multivariable regression analysis including all characteristics that were significantly different between NMS cases resulting vs. not resulting in death.</jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p>683 cases with NMS were analyzed (median age = 36 years, males = 62.1%). In a multivariable model, independent predictors of NMS mortality were lack of antipsychotic discontinuation (odds ratio (OR) = 4.39 95% confidence interval (CI) = 2.14–8.99; <jats:italic>p</jats:italic> < 0.0001), respiratory problems (OR = 3.54 95%CI = 1.71–7.32; <jats:italic>p</jats:italic> = 0.0004), severity of hyperthermia (Unit‐OR = 1.30, 95%CI = 1.16–1.46; <jats:italic>p</jats:italic> < 0.0001), and older age (Unit‐OR = 1.05, 95%CI = 1.02–1.07; <jats:italic>p</jats:italic> = 0.0014). Even in univariate, patient‐level analyses, antipsychotic formulation was not related to death (oral antipsychotic (OAP): <jats:italic>n</jats:italic> = 39/554 (7.0%) vs. long‐acting injectable (LAI): <jats:italic>n</jats:italic> = 13/129 (10.1%); <jats:italic>p</jats:italic> = 0.2413). Similarly, death with NMS was not related to antipsychotic class (first‐generation antipsychotic: <jats:italic>n</jats:italic> = 38/433 (8.8%) vs. second‐generation antipsychotic: <jats:italic>n</jats:italic> = 8/180 (4.4%); <jats:italic>p</jats:italic> = 0.0638). Non‐antipsychotic co‐treatments were not associated with NMS mortality.</jats:p></jats:sec><jats:sec><jats:title>Conclusion</jats:title><jats:p>Despite reliance on case reports, these findings indicate that presence of respiratory alterations, severity of hyperthermia, and older age should alert clinicians to a higher NMS mortality risk, and that antipsychotics should be stopped to reduce mortality, yet when NMS arises on LAIs, mortality is not increased vs. OAPs.</jats:p></jats:sec>