Prophylactic cranial irradiation (PCI) reduces the risk of cranial metastasis in small cell lung cancer (SCLC), but the magnitude and value of this reduction, the risks of radiation morbidity and whether PCI influences survival are unclear. We conducted a randomised trial in patients with limited-stage SCLC who had had a complete response to induction therapy. Initially, patients were randomised equally to (1) PCI 36 Gy in 18 daily fractions, (2) PCI 24 Gy in 12 fractions and (3) no PCI; subsequently, to increase the rate of accrual, randomisation was to clinicians' choice of PCI regimen versus no PCI (at a 3:2 ratio). The endpoints were appearance of brain metastases, survival, cognitive function, and quality of life (QoL). Three hundred and fourteen patients (194 PCI, 120 No PCI) were randomised. In the revised design, the most commonly used PCI regimens were 30 Gy in 10 fractions and 8 Gy in a single dose. With PCI, there was a large and highly significant reduction in brain metastases (HR = 0.44, 95% CI 0.29-0.67), a significant advantage in brain-metastasis-free survival (HR = 0.75, 95% CI 0.58-0.96) and a non-significant overall survival advantage (HR = 0.86, 95% CI 0.66-1.12). In both groups, there was impairment of cognitive function and QoL before PCI and additional impairment at 6 months and 1 year, but no consistent difference between the two groups and thus no evidence over 1 year of major impairment attributable to PCI. PCI can safely reduce the risk of brain metastases. Further research is needed to define optimal dose and fractionation and to clarify the effect on survival. Patients with SCLC achieving a complete response to induction therapy should be offered PCI.