Chagas disease vector control campaigns are being conducted in Latin America, but little is known about medium-term or long-term effectiveness of these efforts, especially in urban areas. After analyzing entomologic data for 56,491 households during the treatment phase of a Triatoma infestans bug control campaign in Arequipa, Peru, during 2003-2011, we estimated that 97.1% of residual infestations are attributable to untreated households. Multivariate models for the surveillance phase of the campaign obtained during 2009-2012 confirm that nonparticipation in the initial treatment phase is a major risk factor (odds ratio [OR] 21.5, 95% CI 3.35-138). Infestation during surveillance also increased over time (OR 1.55, 95% CI 1.15-2.09 per year). In addition, we observed a negative interaction between nonparticipation and time (OR 0.73, 95% CI 0.53-0.99), suggesting that recolonization by vectors progressively dilutes risk associated with nonparticipation. Although the treatment phase was effective, recolonization in untreated households threatens the long-term success of vector control.