Prevalence and risk factors associated with tuberculosis and HIV coinfection in Iran: A multivariate firth logistic regression for rare events.

Abstract

Among the complications caused by HIV infection, tuberculosis (TB) is the most important challenge. The study aimed to determine the predictors of TB and HIV coinfection by using Firth logistic regression analysis. This cross-sectional study was conducted on 32,168 HIV-positive patients diagnosed in Iran. The required information was obtained from the national HIV/AIDS surveillance data. The prevalence of TB in HIV patients was 3.2%; the prevalence of TB in males and females was 944 (91.83%) and 84 (8.17%), respectively. Based on results of multivariate Firth logistic regression, male gender OR male/female = 1.91; injecting drug use (IDU) OR YES/NO = 1.46; illiterate or primary groups OR illiterate or primary /university = 2.23; high school group OR high school/ university = 2.24 all increased the risk of TB and HIV coinfection (p < 0.05). Also, having CD4 > 500 [OR CD4 > 500 / < 200 = 0.39], CD4 351-500 [OR351 / < 200 = 0.65], and CD4 200-350 [OR200-350 / < 200 = 0.64] decreased the risk of TB and HIV coinfection (p < 0.001). According to the results of this study, male gender, low education, injecting drug users, and low CD4 count at the time of diagnosis, were significant predictors of TB and HIV coinfection. Therefore, regular and periodic screening programs and linkage to the care and treatment of HIV patients need special attention.