Background : Dermatologic manifestations are very common among individuals with HIV. An abscess is a pus-filled sack that can occur anywhere in the body, which is usually caused by bacterial infection. Abscesses can also occur in otherwise healthy individuals and among people of all ages. We develop a predictive model of time to abscess among HIV-negative and HIV-positive individuals in Zambia.
Methods: We analyzed data from a longitudinal cohort from the Zambia Emory HIV Research Project (ZEHRP) in Lusaka. Abscess incidence and significant differences between incidence of abscess for each abscess among HIV-positive and HIV-negative individuals was assessed with Mid-p exact tests. We assessed bivariate associations between covariates and time to abscess, stratified by abscess type and HIV status. We performed a repeated outcomes survival analysis to assess multivariate associations between covariates and time to abscess, stratified by abscess type and HIV status. The final model was determined using backward selection.
Results: Since HIV status modified the association of the exposures with the outcomes of interest, the 5038 participants were divided into two groups: HIV positive and HIV negative. There was no significant difference in abscess incidence (by abscess type) between HIV-positive and HIV-negative individuals. Among HIV positive individuals, 66% of deep tissue, furuncle, or skin abscesses were treated by a clinician. Among HIV negative individuals, 74% of abscesses were treated by a clinician. Bivariate and multivariable Cox regression models showed that among HIV positive individuals, a per year increase in age, per unit increase in sedimentation rate, being a male, and HIV stage 2, 3, or 4 (versus HIV stage 1) were associated with increased incidence of abscess. Among HIV negative individuals, being a male was associated with increased incidence of abscess.
Conclusion: Age, sedimentation rate, gender, and HIV stage can be used to predict risk of abscess among HIV-positive individuals. Only gender was predictive among HIV-negative individuals. Further studies examining other risk factors of abscess, such as co-infections, environmental stressors, and other factors that were not measured in this analysis are needed.
Table of Contents
Background 1Methods 5 Results 8 Discussion 11 References 14 Tables 17
About this Master's Thesis
|Committee Chair / Thesis Advisor|
|Incidence of Abscess and Predictors of Time to Abscess among Individuals in HIV Discordant Relationships in Zambia ()||2018-08-28||