Geographically Targeted Interventions versus Mass Drug Administration to Control Taenia solium Cysticercosis, Peru

Optimal control strategies for Taenia solium taeniasis and cysticercosis have not been determined. We conducted a 2-year cluster randomized trial in Peru by assigning 23 villages to 1 of 3 geographically targeted intervention approaches. For ring screening (RS), participants living near pigs with cysticercosis were screened for taeniasis; identified cases were treated with niclosamide. In ring treatment (RT), participants living near pigs with cysticercosis received presumptive treatment with niclosamide. In mass treatment (MT), participants received niclosamide treatment every 6 months regardless of location. In each approach, half the villages received targeted or mass oxfendazole for pigs (6 total study arms). We noted significant reductions in seroincidence among pigs in all approaches (67.1% decrease in RS, 69.3% in RT, 64.7% in MT; p<0.001), despite a smaller proportion of population treated by targeted approaches (RS 1.4%, RT 19.3%, MT 88.5%). Our findings suggest multiple approaches can achieve rapid control of T. solium transmission.


Sample Size Calculations
We used a 2-sample proportion test in Stata SE12.1 (StataCorp LLC, https://www.stata.com) to calculate sample size requirements for the primary outcome, cumulative seroincidence among pigs, between intervention groups. Assuming a 60% reduction in cumulative incidence from baseline to end in the reference group, ≈800 pigs were needed per group to detect a between-group difference >7 percentage points at a power of 0.8 (2-sided α = 0.05). We allocated villages on the basis of a conservative estimate of the ratio of the human-pig population, to meet the sample size requirements.

Intervention Group Allocation
Among possible villages, 23 met the criteria for population size of 50-500 residents, accessibility by vehicle, and no history of control interventions for taeniasis or cysticercosis. We used Stata SE14.0 (StataCorp, LLC) to allocate each of the 23 villages randomly into 1 of 6 intervention groups. We used the generate rannum = uniform() command, with a specified seed, to assign a random number to each village from a uniform distribution. The villages were sorted numerically in ascending order by their random number, then sequentially allocated to each intervention in blocks of 4 villages; the last block included only 3 villages. We repeated this sequence until the number of residents in each intervention group was within 10% of the total study population divided by 6; groups contained 1,415-1,731 residents each, per a prestudy census of 9,438 total residents in all 23 villages. No other factors were considered in assigning villages to interventions.

Consent Procedures
Prior to the start of the study, we met with leaders in each community to gauge their interest in participating and to request permission to include the village in the study. We also held open community meetings to provide information on the objectives, study design, procedures, anticipated start date, and to solicit comments. At the start of the study, our teams visited each home in the selected villages to provide additional information both verbally and in writing, to answer questions, and to solicit informed consent. During our subsequent visits to the villages to provide study interventions, we also recruited newly arrived residents for enrollment and participation in the study. Adults who chose to participate were required to sign a written consent form with complete information about the study, their rights as participants, and the contact information for study personnel and for members of the oversight board for the study. A parent or legal guardian was required to authorize the participation of any child 2-17 years and complete and sign a separate consent form for this purpose.

Collection and Processing of Pig Serum
Veterinary study teams collected blood samples from pigs by going door-to-door in study villages. We maintained a continuous cold chain from the point of collection all the way through  (2). We excluded GP50 in this study because recent studies showed that this band cross-reacts with T. hydatigena, a cestode that infects pigs and is co-endemic in the region where we conducted our study (3,4). We anticipated this change would reduce the sensitivity of the assay because GP50 is typically the first reactive band to appear after exposure, but we also expected the change to increase the assay's specificity. Because the GP50 band was excluded equally from the results in all study arms, we do not expect this change to affect interpretation of comparative results.

Collection and Processing of Human Stool
Our study teams went door-to-door in all study villages to offer presumptive treatment with niclosamide (NSM) for taeniasis, and to provide each participant with a 500-mL plastic container, soap, and instructions for hygienic collection of a posttreatment whole-stool sample.
We collected posttreatment stool samples to increase the likelihood of detecting parasite material. We returned to each home within 24 hours to collect stool samples from participants and transferred these immediately to a temporary field laboratory. There, we examined samples macroscopically for the presence of Taenia sp. scoleces or proglottids, then placed 10-mL fecal aliquots in 40 mL of 5% formol-phosphate buffered saline at pH 7.2 in sealed propylene tubes. These samples were transported by ground at ambient temperature to the Center for Global Health laboratory (Tumbes, Peru) where they were concentrated by spontaneous sedimentation for 24 hours, then examined by light microscopy for the presence of Taenia sp. eggs (5). We then shipped the fecal samples by air to the CNS Parasitic Diseases Research Unit, Universidad Peruana Cayetano Heredia (Lima, Peru) for further analysis by ELISA to detect Taenia sp. coproantigens, as previously described (6), with the exception that the capture antibody and conjugate used were specific to T. solium (7). This ELISA assay has 96.4% sensitivity and 100% specificity for T. solium (7). We used posttreatment stool samples and multiple evaluation techniques (macroscopy, microscopy, ELISA) to maximize the likelihood of detecting true positive cases.