BACKGROUND Sacral nerve stimulation is an established treatment for faecal incontinence. Nearly half of all patients experience loss of efficacy at some point. Standard reprogramming restores efficacy for some, but not all, patients. This study aimed to determine whether alternative stimulator settings would increase treatment efficacy. METHODS Patients with sustained loss of efficacy were recruited from two international specialist centres. A number of alternative stimulation parameters were tested using a double-blind randomized crossover study design. Stimulation settings tested were pulse frequencies of 6.9 and 31 Hz, and pulse widths of 90 and 330 µs, compared with one standard setting of 14 Hz/210 µs. Treatment efficacy was evaluated using a disease-specific quality-of-life score (Fecal Incontinence Quality of Life Scale, FIQLS) and a bowel habit diary completed before randomization, during the study period and after 3 months of follow-up with one preferred setting. RESULTS Fifteen patients were analysed. With one preferred setting, three of four subdomains in the FIQLS improved significantly. The mean(s.d.) total number of incontinence episodes dropped from 11.7(10.8) to 4.8(4.5) per 3 weeks (P = 0.011) and improvements were maintained after 3 months of follow-up. Optimal pacemaker settings were individual, but a trend towards highest patient satisfaction and improved treatment outcome was evident for high-frequency stimulation (31 Hz/210 µs), which was preferred by eight of the 15 patients. CONCLUSION Patients experiencing loss of efficacy can experience improvement if alternative pacemaker settings are tested. High-frequency stimulation (31 Hz/210 µs) was preferred by more than half of the patients, and improved treatment outcome was sustained at 3 months.