Evaluating post-treatment Loa loa microfilarial densities to classify serious adverse events caused by ivermectin: a retrospective analysis.

ABSTRACT

BACKGROUND: The elimination of onchocerciasis requires increasing ivermectin treatment coverage in communities hypoendemic for onchocerciasis. In areas where loiasis is co-endemic, this approach is complicated by the risk of serious adverse events following treatment with ivermectin in individuals with a high Loa loa microfilarial density (MFD). We aimed to evaluate the extent to which the pre-treatment MFD can be inferred from post-treatment MFDs. METHODS: For this retrospective analysis, we used data from seven clinical or community trials (six were used for the main analysis and one for the secondary analysis) conducted in Cameroon, in which MFDs were measured both before and after (within 14 days) receiving a single dose of ivermectin (150-200 µg/kg bodyweight). The primary objective was to establish the receiver operating characteristic curves and the corresponding area under the curve statistics of MFD measured after treatment to classify pre-treatment MFD (MFDD0) according to common risk thresholds of serious adverse events. We assessed the performance of post-treatment MFD to accurately classify MFDD0 according to commonly used thresholds using bootstrap procedures. FINDINGS: 281 individuals with MFD measurements available before and 3-10 days after ivermectin treatment were enrolled. Our results show that an MFD of more than 3500 L loa microfilariae per mL of blood (mf per mL) 3 or 4 days after treatment indicates a 68·6% chance (positive predictive value) of an MFDD0 of more than 20 000 mf per mL. An MFD of more than 3500 mf per mL at day 5-10 corresponds to a 72·2% chance of having an MFDD0 of more than 20 000 mf per mL. Conversely, an MFD of less than 2500 microfilariae per mL at day 3-4 or day 5-10 corresponds to a probability of 92·3% or 92·8% (negative predictive value) of having MFDD0 of less than 20 000 mf per mL. An MFD less than 1500 mf per mL on days 3-4 after treatment was associated with a 78·3% probability of having an MFDD0 less than 8000 mf per mL; this probability increased to 89·6% on days 5-10 after treatment. INTERPRETATION: The MFD threshold of 1000 mf per mL within 1 month of treatment, which is commonly used to attribute the occurrence of a serious adverse event to ivermectin, should be revised. In this study, we present tables that can help to assess this attributability as part of mass or individual treatments. FUNDING: None.