From March 8, 2004 through January 18, 2008, 134 women enrolled in the study and initiated ART. Twenty-six additional women from the parent cohort were eligible for ART but chose to initiate treatment at a non-research clinic. These women did not differ significantly from those who enrolled in terms of their age, marital status, education, or WHO clinical stage. Pre-ART CD4 counts from outside clinics providing ART to these 26 women were not available for this study. Fewer non-participating women had a documented history of GUD prior to ART initiation (38% vs. 61%, p=0.03).
The enrolled women attended 852 visits from ART initiation through month 6 (median, 7 visits; inter-quartile range [IQR], 6–7 visits). The median CD4 count on ART initiation was 127 (IQR, 79 – 165), and 82 women (61.2%) had a documented history of GUD at any time prior to ART initiation. GUD was reported or observed in 54 women (40.3%) at 85 visits (10.0%). Detection was by examination alone in 32 (37.6%), by report alone in 38 (44.7%), and by both methods for 15 ulcers (17.6%). The presence of GUD by examination was significantly associated with participants’ self report of GUD, both prior to ART initiation (p<0.001) and after starting ART (p<0.001). Syphilis was diagnosed at 5 visits, of which 4 included concurrent examiner identification of GUD. No chancroid was detected by culture of 18 incident atypical ulcers.
GUD prevalence was 9.7% at baseline, increased to 16.7% at month 1, then decreased to 6.4% by month 6. GUD detection was significantly higher among women with a history of GUD (13.9% versus 3.9%, p<0.001), as depicted graphically in . The results of univariate and multivariate logistic regression analysis using GEE are presented in . In univariate analysis, the first month after ART initiation, baseline CD4 count <100 cells/µL, WHO stage 4 disease, and a history of GUD were all significantly associated with GUD detection.
GUD Prevalence by Month from ART Initiation
Predictors of GUD Detected in the 6 Months after ART Initiation
In multivariate analysis, both baseline CD4 count <100 cells/µL and the first month after ART initiation were associated with approximately 2-fold increases in risk of GUD, although the association with baseline CD4 count was reduced to a statistical trend (p=0.06). History of GUD was associated with a 3.8-fold increased risk in GUD after ART initiation. HIV stage was no longer a significant predictor of GUD risk after adjustment for the other variables in the model.
To examine the effect of our GUD definition on the results, we repeated our analyses after restricting our definition of GUD to ulcers detected on examination. In this analysis, point estimates for risk of GUD in months 1–6 after ART initiation were 1.4, 1.2, 0.6, 0.6, 0.4, and 0.4 respectively. This analysis included fewer events, and did not achieve statistical significance for comparisons between different time points. Both CD4 count <100 cells/µL (aOR 2.2, p = 0.04) and history of GUD on examination (aOR 3.5, p = 0.001) remain significant predictors of GUD after ART initiation.