Background. No prospective study has evaluated the long-term effect on mortality of the new acid concentrates added to bicarbonate dialysate. The aim of this pharmacoepidemiological study was to evaluate the association between hydrochloric or citric acid-based dialysate and mortality on haemodialysis (HD). Methods. This study included 117 796 patients with 3 723 887 months on HD recorded in the national French Renal Epidemiology and Information Network registry. Dialysate acid components were retrospectively reconstructed for each facility. All patients on HD were associated each month with an exposure based on that at their facility of treatment. We took each patient's time-varying exposure into account to calculate the monthly mortality rates for each exposure. Incidence rate ratios (IRRs) for mortality were calculated with a Poisson regression, with acetic acid as the reference. Regressions were adjusted for initial clinical characteristics (age, gender, previous cardiovascular events, active malignancy, diabetes, pulmonary disease, mobility), dialysis technique and location (in-centre, outpatient centre, self-care unit) and ESRD vintage, updated monthly. Results. The crude mortality rate per 1000 patient-months with citric acid \11.5 [95% confidence interval (CI) 11.1-12.0] \ was lower than with either acetic acid [12.9 (95% CI 12.8-13.1)] or hydrochloric acid [12.8 (95% CI 12.2-13.5)]. For the 2014-17 period, the IRR for mortality with citric acid [adjusted IRR 0.94 (95% CI 0.90-0.99)] and with hydrochloric acid [adjusted IRR 0.86 (95% CI 0.79-0.94)] were significantly lower than with acetic acid. Conclusion. This post-marketing study of long-term exposure to dialysate acidifiers at the patient level found the use of citric and hydrochloric acid-based dialysates, compared with acetic acid, was associated with lower mortality.