AimThe intrathecal baclofen pump is an effective treatment for spasticity. However, long‐term results have reported patients’ dissatisfaction and perception of disability. Potential causes include a frequent need for baclofen pump refill and risks of complications. The aim of the present study was to evaluate the long‐term maintenance, complications and clinical outcome of intrathecal baclofen pumps.Patients and MethodsWe conducted a 16‐year retrospective cohort study of patients with spasticity treated with an intrathecal baclofen pump at a university hospital from 2000 to 2016. The primary outcome was the rate of infection per puncture for baclofen pump refill. Secondary outcomes included the incidence of other complications, such as running out of baclofen causing symptomatic withdrawal symptoms, pump mechanical failure, pump battery end of life and the need for pump replacement. The clinical outcome was assessed by the Modified Ashworth Scale (mAS).ResultsIn total, 340 follow‐up episodes with pump refill procedures were recorded. The average interval between each pump refill was 57.3 days (±15.4 days). The average duration of admission for each pump refill was 4 h and 49 min (from 2 h 23 min to 10 h). There were two events with established infection after puncture for the refill, giving rise to an infection rate per puncture of 0.6 percent (2/340).For the long‐term clinical outcome, at an average follow‐up period of 7.6 years, the postoperative mAS for spasticity was 2.0 ± 0.756, which was significantly better than the preoperative mAS at 3.75 ± 0.462 (P = 0.001).ConclusionLong‐term aftercare with baclofen pump refill was safe, with an infection rate of 0.6 per cent per puncture for each refill. Long‐term intrathecal baclofen pump was effective in the treatment of spasticity with persistent significant improvement in the spasticity scale.