Effectiveness of the Bactiseal Universal Shunt for reducing shunt infection in a sub-Saharan African context: a retrospective cohort study in 160 Ugandan children

Abstract

Antibiotic-impregnated shunts have yet to find widespread use in the developing world, largely due to cost. Given potential differences in the microbial spectrum, their effectiveness in preventing shunt infection for populations in low-income countries may differ and has not been demonstrated. This study is the first to compare the efficacy of a Bactiseal shunt system with a non-antibiotic-impregnated system in a developing country. The Bactiseal Universal Shunt (BUS) was placed in 80 consecutive Ugandan children who required a shunt. In this retrospective cohort study, the outcome for that group was compared with the outcome for the immediately preceding 80 consecutive children in whom a Chhabra shunt had been placed. The primary end points were shunt failure, shunt infection, and death. Shunt survival was analyzed using the Kaplan-Meier method. Significance of differences between groups was tested using the log-rank test, chi-square analysis, Fisher's exact test, and t-test. There was no difference between groups in regard to age, sex, or etiology of hydrocephalus. Mean follow-up for cases of nonfailure was 7.6 months (median 7.8 months, interquartile range 6.5-9.5 months). There was no significant difference between groups for any end point. The BUS group had fewer infections (4 vs 11), but the difference was not significant (p = 0.086, log-rank test). Gram-positive cocci were the most common culturable pathogens in the Chhabra group, while the only positive culture in the BUS group was a gram-negative rod. These results provide equipoise for a randomized controlled trial in the same population and this has been initiated. It is possible that the observed trends may become significant in a larger study. The more complex task will involve determining not only the efficacy, but also the cost-effectiveness of using antibiotic-impregnated shunt components in limited-resource settings.