Meta-regression results
The results of univariate and multivariable meta-regression analyses are summarized in Table 3. We performed two multivariable models, the second multivariable model replacing the categorical variable of “year of data collection” in the first multivariable model with another temporal variable of “year of publication”.
The first multivariable model explained 47.7% of the variation in seroprevalence (Table 3). Compared to general populations, FSWs had the highest HSV-2 seroprevalence with an adjusted risk ratio (ARR) of 1.66 (95% CI: 1.58-1.75), followed by drug users (ARR=1.53, 95% CI: 1.39-1.70), PLWH and HIV-negative individual in HIV discordant couples (ARR=1.22, 95% CI: 1.12-1.32), STI clinic attendees (ARR=1.20, 95%CI: 1.15-1.25), intermediate-risk populations (ARR=1.13, 95% CI: 1.05-1.22) and MSM/MSWs (ARR=1.09, 95% CI: 1.03-1.16).
Compared to those aged <20 years, HSV-2 seropositivity increased with age and was highest in those aged ≥60 years (ARR=1.22, 95% CI: 1.11-1.36). People living in the Northeast have the highest risk of HSV-2 (ARR=1.30, 95% CI: 1.21-1.40) than those living in the North, followed by the Southwest (ARR=1.10, 95% CI: 1.04-1.16) and Central-southern regions (ARR=1.10, 95% CI: 1.04-1.16). Men have a similar HSV-2 seroprevalence as women (ARR=0.96, 95% CI: 0.92-1.01). Studies that used PCR as an assay reported lower HSV-2 seroprevalence than those using ELISA.
Studies with larger sample sizes and non-probability sampling reported lower seroprevalence. Studies with lower response rates reported higher seroprevalence. Compared to data collected before 2000, data collected during 2001-2010 (ARR=0.88, 95% CI: 0.83-0.94) and after 2010 (ARR=0.91, 95% CI: 0.85-0.96) had lower HSV-2 seroprevalence.
The second model explained 45.7% of the variation in HSV-2 seroprevalence, yielding similar results to the first model. The seroprevalence did not vary by year of publication (Table 3 ).