Riluzole is to date the only treatment that prolongs amyotrophic lateral sclerosis (ALS) survival. However, results on the efficacy of riluzole in observational population-based studies with a longer follow-up are conflicting and it is still unclear if the effect of the drug is limited to an early stage of the disease and to some specific subgroups of patients. The objective is: (i) to evaluate the effect of riluzole on ALS survival in a cohort of incident cases; (ii) to examine whether bulbar-ALS benefits from the medication to a greater extent and (iii) to assess the efficacy of the drug in elderly patients. Source of the study was a prospective population-based registry of ALS established in Puglia, Southern Italy. We examined survival of 126/130 incident ALS cases diagnosed during the period 1998-1999. Seventy-three patients were prescribed riluzole and the remaining 53 were not. Riluzole therapy increased survival rates at 12 months by approximately 10% and prolonged survival by 6 months (18.2 months vs. 12.4; peto-test: 2.78; P = 0.09). This beneficial effect was present amongst bulbar-onset ALS (peto-test: 4.11; P = 0.042), but not in subjects with limb-onset (peto-test: 0.48; P = 0.4). In patients aged >70 years riluzole treatment was associated with an 8 months longer median survival time [15.4 months vs. 7.1] and a reduction in mortality rate at 12 months by 27%, regardless of site of symptoms onset. In multivariate analysis, riluzole use was an independent predictor of survival at 12 months from the diagnosis with borderline significance (P = 0.06). Riluzole was effective amongst cases with bulbar-onset ALS (P = 0.04), whereas in subjects with limb-onset there was no effect on survival at 12 months (P = 0.5). In each model riluzole did not influence survival at 24 months. Conversely, riluzole use was associated with an improvement in survival amongst elderly patients both at 12 (P = 0.07), at 24 months (P = 0.03) and in the entire follow-up period (P < 0.04). In this population-based series, we found that riluzole therapy improves ALS survival. The efficacy of the drug was present amongst bulbar-onset ALS and older patients, but not in subjects with limb-onset. The favourable effect of the drug was transient, as it was lost in prolonged follow-up. Our observations support the use of riluzole at an early stage of ALS in bulbar and elderly patients. However, the appropriate duration of riluzole treatment remains to be established.