A prospective randomised controlled trial of antimicrobial prophylaxis in hydrocephalus shunt surgery

Despite attempts to reduce their incidence, shunt infections remain a major complication of the treatment of hydrocephalus. Various forms of antimicrobial prophylaxis are in use, but no controlled, statistically valid trial has been conducted to assess their efficacy. Such a trial was therefore carried out and its design is described here. After a 1-year retrospective and prospective study by members of the United Kingdom Hydrocephalus Group to establish feasibility and infection rates, a statistical study showed that at least 712 patients would be required. Six centres were enrolled to fulfil these requirements, and ethical committee approval was obtained at each. The chosen prophylactic regimen was 10 mg vancomycin administered into the ventricular system during surgery. Adults and children undergoing insertion or revision of ventriculoperitoneal shunts were included unless they were receiving therapeutic antimicrobials. Randomisation was by computer-generated numbers. Controls received the antimicrobial regimen, if any, currently used in that centre, the only difference between the two groups being intraventricular vancomycin in the test group. Diagnosis of shunt infection included accepted clinical and microbiological criteria reinforced by measurement of serum C-reactive protein levels. Follow-up was for at least six months. After 2.5 years only 158 patients had been enrolled in the trial, 80 controls and 78 tests. There were 5 preventable infections in the control group and 2 in the test group. In view of the small total the planned statistical analysis was not possible. Therefore, while no problems were encountered with toxicity, the trial failed to enroll enough patients to answer the question of efficacy of antimicrobial prophylaxis in shunt surgery and the reasons for this are discussed.