Search

To the Editor—We appreciate the response from Drs Yang, Chang, and Hung to our article “Syphilis treatment response among HIV-discordant couples in Zambia and Rwanda” [1]. It is true that the calculated treatment response rate may be an underestimate or an overestimate based on the limitations of our retrospective analysis, although the size of the human immunodeficiency virus (HIV)–discordant cohort and the duration of extended follow-up adds to previously published literature confirming the utility of single-dose therapy to treat early syphilis in HIV-infected persons. Because we did not have any rapid plasma reagin (RPR) values, symptom information, or documentation of syphilis therapy prior to the period of the study, we decided to dichotomize clients into incident or prevalent disease. Among the group classified as prevalent disease, some had latent infection from >12 months prior and may have benefited from 2 additional weekly doses of benzathine penicillin, although others in this group likely had early-stage infection acquired prior to study onset or had serofast state following previous therapy. These latter 2 groups would likely not have benefited from additional penicillin therapy. Clinical information to allow more precise categorization of these cases would have been helpful but was not available to us. As often encountered in clinical practice, many positive samples had low titer RPR values. To increase testing specificity in our study design, we included only RPR titers ≥1:2 in our subsequent multivariate and Cox proportional hazard analyses. For the subset of samples with treponemal confirmatory testing available, most RPR titers ≥1:2 were confirmed and therefore were unlikely to be biological false-positive tests (see Table 3 [1]). Among those with >1 episode of positive RPR during the study period, the response rate was similar compared to those with a single episode (61% among those with >1 episode and 72% among those with 1 episode). Data presented in the article's Table 4 [1] show separate multivariate modeling for patients who had a single episode and those with >1 episode treated during the study period. Treatment response among coinfected patients remains an important area of research, and we certainly agree with the need for prospective randomized or observational trials in varying populations, including in men who have sex with men.

Background Most new HIV infections in Africa are acquired from cohabiting heterosexual partners. Couples' Voluntary Counselling and Testing (CVCT) is an effective prevention strategy for this group. We present our experience with a community-based program for the promotion of CVCT in Kigali, Rwanda and Lusaka, Zambia. Methods Influence Network Agents (INAs) from the health, religious, non-governmental, and private sectors were trained to invite couples for CVCT. Predictors of successful promotion were identified using a multi-level hierarchical analysis. Results In 4 months, 9,900 invitations were distributed by 61 INAs, with 1,411 (14.3%) couples requesting CVCT. INAs in Rwanda distributed fewer invitations (2,680 vs. 7,220) and had higher response rates (26.9% vs. 9.6%), than INAs in Zambia. Context of the invitation event, including a discreet location such as the INA's home (OR 3.3–3.4), delivery of the invitation to both partners in the couple (OR 1.6–1.7) or to someone known to the INA (OR 1.7–1.8), and use of public endorsement (OR 1.7–1.8) were stronger predictors of success than INA or couple-level characteristics. Conclusion Predictors of successful CVCT promotion included strategies that can be easily implemented in Africa. As new resources become available for Africans with HIV, CVCT should be broadly implemented as a point of entry for prevention, care and support.