Background: The neutrophil-to-lymphocyte ratio (NLR), a marker of host inflammation, has been associated with poor outcome in several solid tumors. Here, we investigated associations of the derived NLR (dNLR) and duration of initial androgen deprivation therapy (ADT) with survival of men with metastatic castration-resistant prostate cancer (mCRPC) receiving first-line chemotherapy. Patients and methods: Data from the multinational randomized phase III studies VENICE and TAX327 included a total of 2230 men with mCRPC randomized to receive first-line chemotherapy, and were used as training and validation sets, respectively. Associations of dNLR and duration of initial ADT with overall survival (OS) were evaluated by multivariable Cox regression analysis in the training set stratified for performance status and treatment arm. The model was then tested in the validation set. Subsequently, we investigated the treatment effect of docetaxel on OS in subgroups according to dNLR and duration of initial ADT. Results: In the training set, both dNLR >= median (2) and duration of initial ADT <median (15 months) were associated with increased risk of death [hazard ratio (HR) 1.29; 95% confidence interval (CI) 1.11-1.50, P < 0.001 and HR 1.41; 95% CI 1.21-1.64, P < 0.001, respectively] after adjustment for age, alkaline phosphatase, hemoglobin, and pain at baseline. In the validation set, dNLR remained an independent prognostic factor for OS (HR 1.43; 95% CI 1.20-1.70, P < 0.001), whereas duration of initial ADT was not (HR 1.16; 95% CI 0.97-1.37, P = 0.10). In subgroup analyses of the TAX327 study, docetaxel improved OS irrespective of dNLR and duration of initial ADT. Conclusion: The dNLR was prognostic for OS in men with mCRPC receiving first-line chemotherapy in two randomized phase III trials. A high dNLR (= 2) was associated with shorter survival irrespective of the received treatment. This readily available biomarker may serve for risk stratification in future clinical trials and could be incorporated into prognostic nomograms.