OBJECTIVE A short vaccination protocol against hepatitis B was compared to the standard approach in patients under haemodialysis who were primarily non-responsive to the vaccine. METHODS This randomized, controlled open trial included 51 chronic haemodialysis subjects previously vaccinated against hepatitis B and with anti-HBs levels of less than 10 IU/mol/L. Twenty-six patients received 20 microg i.m. once a week for 8 weeks (short protocol) and 25 subjects three doses of 40 microg i.m. at months 0, 1 and 6 (standard protocol). Clinical and laboratory data were compared between responders and non-responders. A logistic regression model included selected parameters to assess risk factors for non-seroconversion. RESULTS Seroconversion rates to vaccine at 2 months were 80% and 78% in the short and standard protocol groups, respectively (P = 0.99). Median of anti-HBs levels were similar up to 6 months of follow up, but patients in the standard protocol showed a trend to higher anti-HBs in month 3 and a more steady decline in antibody titres. Non-responders were older, had longer duration of dialysis and a higher prevalence of a prior renal transplant and hepatitis C. In multivariate analysis, only advanced age and hepatitis C remained independently associated with non-responsiveness to vaccination. CONCLUSIONS In haemodialysis patients, a short vaccination protocol against hepatitis B did not provide any benefit compared to the standard approach with respect to peak anti-HBs titres or a higher rate of seroprotection at the end of follow up. Other strategies to increase seroconversion rates should be explored, especially in the elderly and in patients with hepatitis C.