Nonpregnant volunteers of both sexes, 14–65 years of age, were screened with a brief medical history and physical examination and venipuncture between 10 pm and 2 am for detection of W. bancrofti microfilariae by calibrated thick smear of 60 μL of blood, assessment of W. bancrofti circulating antigen levels by enzyme‐linked immunosorbent assay (TropBio), and hemoglobin level. Patients were excluded from participating in the drug study for the following reasons: a W. bancrofti microfilarial level of 7 beers or other alcohol‐containing drinks per week), temperature >37.5°C, serious medical illness, history of allergy to benzimidazoles or ivermectin, or use of albendazole or ivermectin within the past 6 months.
Study design.Eligible participants were randomly assigned to receive annual treatment with standard‐dose albendazole (400 mg) and ivermectin (150 μg/kg) or twice‐yearly treatment with high‐dose albendazole (800 mg) and ivermectin (400 μg/kg) for 24 months. A brief clinical assessment, including vital signs and pregnancy testing (in women of childbearing age), was performed at each study time point. Study medication was administered under the direct observation of a physician who remained in the study village for 1 week after treatment to assess adverse events. Venipuncture was performed between 10 pm and 2 am at baseline and every 6 months before treatment for quantification of W. bancrofti microfilarial levels (by Nuclepore filtration of 1 mL of blood) and circulating antigen levels, as well as complete blood cell count. All laboratory assessments were performed by trained personnel unaware of the group assignments. Ultrasonography was performed for male and female volunteers at baseline to identify adult worms in the scrotal and breast lymphatics (filarial dance sign). Additional ultrasonography was performed at 12 and 24 months. An experienced radiologist, blinded to the study group assignments, performed all ultrasound studies.
Statistical analysis. The primary end point of the study was the difference in W. bancrofti levels between the 2 groups at 12 months. This was evaluated by examining the W. bancrofti clearance rates at 12 months using 2‐sided Fisher’s exact test and matching confidence intervals [8] and by assessing the difference in the percentage of baseline W. bancrofti microfilarial levels at 12 months by Wilcoxon‐Mann‐Whitney U test. The Fisher’s exact test was performed for all patients with complete data at 12, 18, and 24 months and, in addition, for all randomized patients at 12 months using a conservative pooled imputation sensitivity analysis, where the missing patients in each group are imputed with proportions closest to the total proportion of responders, regardless of treatment group. On the basis of the published data comparing single and multidose regimens for the treatment of lymphatic filariasis and a predicted 10% attrition rate, a sample size of 25 per group was estimated to have 90% power to detect a difference in the 2 groups by Fisher’s exact test with a 2‐sided α level of .05. One patient in the annual group and 2 patients in the twice‐yearly group had no detectable W. bancrofti microfilariae at the baseline visit despite a microfilarial level of >50 microfilariae/mL detected at screening. Those 3 patients have been excluded from the difference in the percentage of baseline W. bancrofti analysis. There were similar dropout rates for both treatment groups, and patients missing responses were assumed to be missing at random and have been excluded from all analyses except the sensitivity analysis, in which conservative imputations were performed. Changes over time in binary variables (eg, presence or absence of eosinophilia) were analyzed using the exact McNemar’s test. Calculations were performed in either Prism software, version 5.0c (GraphPad), or R, version 2.10.1 (R Foundation for Statistical Computing), using the exact2×2 software package, version 1.0.0.