Coma is a common clinical presentation for critically ill children in sub-Saharan Africa, where major differential diagnoses include cerebral malaria, viral encephalitis, and bacterial and tuberculous meningitis. Lumbar puncture (LP) is a crucial diagnostic test to distinguish between these etiologies and determine optimal treatment. However, clinicians may be hesitant to perform an LP due to concerns of precipitating herniation and death, particularly in environments where pre-procedure neuroimaging is unavailable. We performed a retrospective cohort study of the safety of LP in comatose Malawian pediatric inpatients recruited over consecutive rainy seasons from 1999-2013. Our goal is to assess whether performing an LP changed the mortality within the first 12 hours after the admission and the overall mortality during the hospitalization. Following propensity score matching, all baseline characteristics were balanced between matched pairs of children who did and did not have LPs. After matching, both 12 hour mortality and overall mortality were not significantly different between children who did and did not receive lumbar punctures. Logistic regression, stratification on propensity scores, and inverse probability weighting analyses all showed that the overall mortality has a statistically significant decrease in the group that received an LP. The results of 12-hour mortality from these methods are consistent with the matching method.
Read
