The aim of this study was to review the literature on human studies of drug therapy in cardiac arrest during the last 25 years. In May 2015, a systematic literature search was performed in PubMed, Embase, the Cochrane Library, and CRD databases. Prospective interventional and observational studies evaluating a specified drug therapy in human cardiac arrest reporting a clinical endpoint [i.e. return of spontaneous circulation (ROSC) or survival] and published in English 1990 or later were included, whereas animal studies, case series and reports, studies of drug administration, drug pharmacology, non-specified drug therapies, preventive drug therapy, drug administration after ROSC, studies with primarily physiological endpoints, and studies of traumatic cardiac arrest were excluded. The literature search identified a total of 8936 articles. Eighty-eight articles met our inclusion criteria and were included in the review. We identified no human study in which drug therapy, compared with placebo, improved long-term survival. Regarding adrenaline and amiodarone, the drugs currently recommended in cardiac arrest, two prospective randomized placebo-controlled trials, were identified for adrenaline, and one for amiodarone, but they were all underpowered to detect differences in survival to hospital discharge. Of all reviewed studies, only one recent prospective study demonstrated improved neurological outcome with one therapy over another using a combination of vasopressin, steroids, and adrenaline as the intervention compared with standard adrenaline administration. The evidence base for drug therapy in cardiac arrest is scarce. However, many human studies on drug therapy in cardiac arrest have not been powered to identify differences in important clinical outcomes such as survival to hospital discharge and favourable neurological outcome. Efforts are needed to initiate large multicentre prospective randomized clinical trials to evaluate both currently recommended and future drug therapies.