Background: Genital immune activation is suspected to modulate local human immunodeficiency virus (HIV) RNA levels and the risk of sexual HIV transmission.

Methods: A prospective, observational cohort study of 221 HIV-infected men undergoing male circumcision (MC) was conducted in Rakai, Uganda. Penile lavage samples collected from the coronal sulcus at baseline and 4 weekly visits after MC were assayed for pro-inflammatory cytokines and HIV RNA. The main analysis was limited to 175 men with detectable HIV plasma viral load (VL > 400 copies/mL; n = 808 visits). The primary exposures of interest were individual and total cytokine detection at the previous postoperative visit. Adjusted prevalence risk ratios (adjPRR) of detectable HIV shedding (VL > 40 copies/mL) were estimated by Poisson regression models with generalized estimating equations and robust variance estimators and included adjustment for plasma HIV VL.

Findings: Among men with a detectable plasma VL, penile HIV shedding was detected at 136 visits (16.8%). Detectable interleukin (IL)-1β (adjPRR = 2.14; 95% confidence interval (CI) = 1.02-4.48), IL-6 (adjPRR = 2.24; 95% CI = 1.28-3.90), IL-8 (adjPRR = 2.42; 95% CI = 1.15-5.08), IL-10 (adjPRR = 2.51; 95% CI = 1.67-3.80), and IL-13 (adjPRR = 1.87; 95% CI = 1.15-3.03) were associated with penile HIV shedding at the subsequent visit. Men with 2-4 (adjPRR = 2.36; 95% CI = 1.08-5.14) and 5-7 (adjPRR = 3.00; 95% CI = 1.28-7.01) detectable cytokines had a greater likelihood of detectable penile HIV shedding at the subsequent visit, compared to men with ≤ 1 detectable cytokine. The total number of detectable cytokines was also associated with a higher penile log10 HIV VL at the subsequent visit among HIV shedders.

Interpretation: Pro-inflammatory cytokine production had a dose-dependent and temporal association with penile HIV shedding, suggesting that genital immune activation may increase the risk of sexual HIV transmission by driving local HIV replication.