The epidemiology of herpes simplex virus type 2 in sub-Saharan Africa: systematic review, meta-analyses, and meta-regressions

Background: Herpes simplex virus type 2 (HSV-2) infection is a prevalent sexually transmitted infection with a sizable disease burden that is highest in sub-Saharan Africa. This study aimed to characterize HSV-2 epidemiology in this region. Methods: Cochrane and PRISMA guidelines were followed to systematically review, synthesize, and report HSV-2 related findings. Meta-analyses and meta-regressions were conducted. Findings: From 218 relevant publications, 451 overall outcome measures and 869 stratified measures were extracted. Pooled incidence rates ranged between 2.4-19.4 per 100 person-years across populations. Pooled seroprevalence was lowest at 37.3% (95% confidence interval (CI): 34.9-39.7%) in general populations and high in female sex workers and HIV positive individuals at 62.5% (95% CI: 54.8-70.0%) and 71.3% (95% CI: 66.5-75.9%), respectively. In general populations, pooled seroprevalence increased steadily with age. Compared to women, men had a lower seroprevalence with an adjusted risk ratio (ARR) of 0.61 (95% CI: 0.56-0.67). Seroprevalence decreased in recent decades with an ARR of 0.98 (95% CI: 0.97-0.99) per year. Seroprevalence was highest in Eastern and Southern Africa. Pooled HSV-2 proportion in genital ulcer disease was 50.7% (95% CI: 44.7-56.8%) and in genital herpes it was 97.3% (95% CI: 84.4-100%). Interpretation: Seroprevalence is declining by 2% per year, but a third of the population is infected. Age and geography play profound roles in HSV-2 epidemiology. Temporal declines and geographic distribution of HSV-2 seroprevalence mirror that of HIV prevalence, suggesting sexual risk behavior has been declining for three decades. HSV-2 is the etiological cause of half of GUD and nearly all genital herpes cases.


Introduction
Herpes simplex virus type 2 (HSV-2) infection is a highly prevalent sexually transmitted infection (STI) worldwide.(1) It is a leading cause of genital ulcer disease (GUD) and genital herpes, manifesting in the form of painful, recurrent, and frequent genital lesions.(2-8) Its vertical transmission from mother-to-child can lead to neonatal herpes, a severe and sometimes fatal outcome in newborns. (3,9) HSV-2 is linked to a 3-fold increase in sexual transmission and acquisition of HIV, (10)(11)(12) implying a potential epidemiological synergy between the two viruses. (11,13,14) HSV-2 is typically asymptomatic in most of those who acquire it. (3) This trait coupled with HSV-2's chronic and reactivating nature, as well as its subclinical shedding, (15,16) increases its rate of transmission and leads to high antibody prevalence (seroprevalence) among general and higher-risk populations alike. (11,17,18) Since HSV-2 is more transmissible than HIV and produces long-lasting antibodies, it has been used as an objective biological marker of sexual risk behavior and risk of HIV infection. (19)(20)(21)(22)(23) Analyses using empirical data and mathematical modeling supported the utility of using HSV-2 seroprevalence to predict HIV epidemic potential. (19,20,24) Inadequate understating of HSV-2 epidemiology and its considerable consequences on sexual and reproductive health and the HIV epidemic, (11,13) calls for urgent preventive and control measures to tackle it. The World Health Organization (WHO) outlined a "Global Health Sector Strategy on STIs" (25) that sets goals to eliminate STIs as a main public health concern by 2030 through integration of preventive and control measures. Consequently, WHO, along with their global partners, are spearheading efforts to develop an HSV vaccine as an urgent priority.(26, 27) screened for relevance based on titles and abstracts, followed by full text screening of relevant and potentially relevant records. Additional bibliography screening was performed on both reviews and the relevant articles to identify any missing publications.
Inclusion criteria were met if publications reported primary data on any of the following four outcomes: 1) HSV-2 incidence rate, 2) HSV-2 seroprevalence, 3) proportion of HSV-2 detection in GUD, and 4) proportion of HSV-2 detection in genital herpes. A sample size of ≥10 was required for inclusion for all outcome measures. Exclusion criteria encompassed case reports, series, commentaries, reviews, and qualitative studies. In this review, "publication" refers to a document reporting any outcome measure, whilst a "study" refers to details of a specific outcome measure. Special care was given to ensure that overlapping studies were only included once and not in duplicate.

Data extraction and synthesis
Data extraction and double extraction were performed by MH and FA independently. A list of extracted variables is in Box S2.
Overall outcome measures and their stratified measures were extracted, provided the stratification agreed with a pre-set stratification hierarchy and the subsample in each stratum was ≥10. The pre-set stratification hierarchy sequence for incidence and seroprevalence measures was as follows: population type (see Box S3 for definition), sex, and age. As for proportion of HSV-2 detection in GUD and genital herpes, the sequence was: genital herpes episode status (primary versus recurrent episode), sex, age, and study site (hospital versus outpatient clinic).

Incidence overview and pooled mean estimates for HSV-2 incidence rate
In general populations, HSV-2 incidence rates among women (n=20) ranged between 3.6-21.7 per 100 person-years with a median of 7.5 and a pooled mean of 7.2 (95% confidence interval (CI): 5.5-9.4) per 100 person-years (Table 1). Among men (n=20), incidence rate ranged between 1.5-10.5 per 100 person-years with a median of 5.5 and a pooled mean of 4.1 (95% CI: 3.1-5.3) per 100 person-years. Higher incidence rates were found in higher risk populations. A summary of pooled mean incidence rate by population type and associated forest plots are in Table 1 and Figure S1, respectively.
In higher-risk populations (n=40), almost all studies were conducted among female sex workers (FSWs; n=39) with seroprevalence ranging between 4.3-99.0% with a median of 65.0% (Table   2). High seroprevalence was observed in HIV positive individuals and in individuals in HIV discordant couples, ranging between 42.0-95.2% with a median of 79.7% among women (n=20), and between 44.0-94.2% with a median of 61.4% among men (n=15). Tables 2, 3, and S9 summarize HSV-2 seroprevalence measures for further populations and subpopulations, including by population type, country, subregion, age, sex, and year of publication.  Table 4 shows results of the meta-regression analyses for HSV-2 seroprevalence. Nine variables were eligible for inclusion in the multivariable model (p-value<0.1 in univariable analysis). Two multivariable models were conducted, one including year of publication as a categorical variable and one including it as a linear term.

Pooled mean estimates for HSV-2 seroprevalence
The model including year of publication as a categorical variable explained 57.5% of seroprevalence variation and included population type, age group, sex, African subregion, country's income, assay type, sampling method, response rate, and year of publication category (Table 4). Compared to general populations who had the lowest seroprevalence, seroprevalence Compared to women, men had a 0.61-fold (95% CI: 0.56-0.67) lower seroprevalence.
Studies that had a lower or unknown response rate had a higher seroprevalence. Meanwhile, assay type, study sample size, and study sampling method were not associated with seroprevalence.  (Table S10). Table S11 summarizes extracted proportions of HSV-2 detection in GUD and in genital herpes.

Quality assessment
Quality assessment of diagnostic methods excluded 51 publications due to potential issues in the implemented diagnostic assays. Quality assessment of included seroprevalence measures is summarized in Table S12. Briefly, 288 studies (89.4%) had high precision, 88 studies (27.3%) and 85 studies (26.4%) had low ROB in the sampling method and in the response rate domains, respectively. Twenty-three studies (7.1%) had high ROB in both quality domains.

Discussion
This systematic review presented a detailed assessment of HSV-2 epidemiology in sub-Saharan Africa. Strikingly, the results demonstrate that HSV-2 seroprevalence has been declining by about 2% per year over the last three decades (Table 4). This decline is in line with the observed declines in HIV epidemics in SSA during the same timeframe. (62) Drivers of the decline in HIV prevalence have been subject to debate, with various mechanisms posited including natural epidemic dynamics,(63) increased HIV-associated mortality, (64,65) impact of interventions, (65) heterogeneity in host susceptibility to HIV infection, (66,67) and reductions in sexual risk behavior. (65,(68)(69)(70)(71) Considering that HSV-2 seroprevalence has been shown to provide an objective proxy biomarker of population-level sexual risk behavior, (19)(20)(21)(22)(23)(24) our finding of rapidly declining HSV-2 seroprevalence suggests that sexual risk behavior has been declining, and that this decline has reduced the transmission of both HIV and HSV-2 infections. With evidence of HSV-2 infection increasing risk of HIV acquisition and transmission, (10)(11)(12)(13) it remains to be seen whether the declining HSV-2 incidence may have also contributed to the declines seen in HIV incidence in SSA.
Despite declining HSV-2 transmission in SSA, incidence rate is still high (Table 1), and much higher than that found elsewhere in the world.(1) For instance, the incidence rate in the United States (<1 per 100 person-years) (72) is an order of magnitude lower than that in SSA. The results further demonstrate that age plays a profound effect in HSV-2 epidemiology-age alone accounted for 30% of the seroprevalence variation (Table 4). HSV-2 infection in SSA is All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 26, 2021. ; https://doi.org/10.1101/2021.01.25.21250443 doi: medRxiv preprint typically acquired at a young age not long after sexual debut, especially so for women, manifesting in a rapidly increasing seroprevalence with age before plateauing at high levels by ages 40-50 (Tables 1, 3, 4, and S9).
Population risk classification was also found to play an important role in HSV-2 epidemiology, accounting for 13% of seroprevalence variation (Table 4). Both incidence rate and seroprevalence were found at much higher levels in specific at-risk populations, such as female sex workers (Tables 2 and 4), with both measures displaying the "classical hierarchy" of STI exposure by sexual risk behavior, also seen with other STIs. (73,74) Despite its prominence amongst higher risk populations, HSV-2 infection is also widely disseminated in SSA with high levels of infection even in the lower risk general population where over a quarter of men and nearly half of women are infected (Tables 2 and 4).
Although HSV-2 seroprevalence is high everywhere in SSA, there are still considerable variations by subregion. Infection levels were highest in Eastern Africa followed by Southern Africa, Central Africa, and lowest in Western Africa (Table 4). Incidentally, this pattern is also seen for HIV infection with Eastern Africa and Southern Africa being most affected and Western Africa least affected. (75,76) This further suggests a strong link between HSV-2 and HIV epidemiologies,(24) reflecting a similar mode of transmission and hinting at a biological/epidemiological synergy. (10)(11)(12)(13)(14) The results further demonstrate that women are almost twice as likely to be infected as men (Table 4), reflecting a higher bio-anatomical susceptibility to the infection.(77, 78) Another finding of this study is that HSV-2 infection is the cause of half of GUD cases in SSA (Table 5), confirming that this infection is the main cause of this disease in this part of the world where nearly 60 million individuals are estimated to be affected with HSV-related GUD. (79) All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Although HSV-2 seroprevalence is declining (Table 4), it will likely remain the main cause of GUD in SSA, as other causes such as syphilis have also been declining. (80,81) HSV-2 infection (as opposed to HSV-1 infection) was also found to account for nearly all cases of genital herpes (>97%; Table 5). This finding is presumably due to the nature of HSV-1 being widely acquired in childhood in SSA through oral transmission, (33) and distinguishes this region from other global regions where there is an increasing role for HSV-1 in genital herpes, (33)(34)(35)(36) with some countries already observing HSV-1 as the cause of a large proportion of cases of genital herpes. (33)(34)(35)(36) This study had limitations, principally the unavailability of data for 15 of 45 African countries.
There was also less data for Central and Western Africa than for Eastern and Southern Africa, in addition to data for seroprevalence eclipsing those of GUD and genital herpes. Despite these limitations, a large volume of data was available to sufficiently power an array of analyses.
Included studies exhibited heterogeneity (Tables 2, 3, and S9); however, more than half of this heterogeneity (57%) was subsequently explained through meta-regressions (Table 4). Studies differed by assay type, sample size, sampling method, and response rate, yet none of these study characteristics appeared to affect seroprevalence with the exception of response rate where studies with lower or unknown response rate had a higher seroprevalence (Table 4). Overall, these limitations should not pose a barrier to the critical interpretation of this study's results, or its findings.

Conclusions
In conclusion, HSV-2 seroprevalence is declining rapidly in SSA. Yet, HSV-2 incidence and seroprevalence remain at high levels, with over a third of the population being infected. Age and subregion within SSA play a critical role in HSV-2 epidemiology and explain much of the observed variation in seroprevalence. The geographical distribution of this infection was also found to be similar to that of HIV infection, and the declines in HSV-2 seroprevalence mirrored those for HIV prevalence. These findings suggest that the declines in both infections were driven by reductions in sexual risk behavior following the massive expansion of the HIV epidemic in this continent, and may suggest that some of the declines in HIV incidence could have been attributed to the declines in HSV-2 incidence. HSV-2 infection was found to be the etiological cause of half of GUD cases in this region, and virtually all cases of genital herpes. These findings demonstrate the urgent need for both prophylactic and therapeutic HSV-2 vaccines to tackle the disease burden of this infection, (82) and argue for further acceleration of ongoing efforts for vaccine development. (26,27,83) All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Contributors
MH and FAH conducted the systematic search, data extraction, and data analysis. MH wrote the first draft of the paper. CJ and KJL contributed to the systematic search, data extraction, and interpretation of the results.
LJA conceived the study and led the data extraction and analysis and interpretation of the results. All authors contributed towards drafting and revising the manuscript.

Declaration of interests
MH, FAH, CJ, and LJA declare no competing interests. KL is currently funded by GlaxoSmithKline (GSK) for a gonorrhea vaccine modeling project. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Box S1. List of the 45 countries included in our definition for sub-Saharan Africa by subregion. 2  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 26, 2021.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Supplementary Table 2 Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).
p. 4 Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators. Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating p. 7 5 which were pre-specified.

Results
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.
p. 8; Figure 2 Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.
p. 8 and 9; Tables 1 and  4 Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). p. 10-11; Supplementary Table 3 Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.

Summary of evidence
24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).
p. 11 Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 1. General populations (populations at low risk): these include populations at lower risk of exposure to HSV-2, such as antenatal clinic attendees, blood donors, and pregnant women, among others.
2. Intermediate-risk populations: these include populations who presumably have frequent sexual contacts with populations engaging in high sexual risk behavior, and have therefore a higher risk of exposure to HSV-2 than the general population. These comprise prisoners, people who inject drugs, and truck drivers, among others.
3. Higher-risk populations: these include populations at high risk of exposure to HSV-2 as a consequence of specific sexual risk behaviors such as female sex workers, men who have sex with men, male sex workers, and transgender populations, among others. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.   Figure S1. Forest plot presenting the outcome of the pooled mean herpes simplex virus type 2 (HSV-2) incidence rate by population type in sub-Saharan Africa.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.        3) * I 2 : a measure that assesses the magnitude of between-study variation that is due to actual differences in seroprevalence across studies rather than chance. All meta-analyses had a statistically significant Cochrane's Q-statistic (p<0.001) and a wide prediction interval indicative of strong heterogeneity. † Other countries: Cote d'Ivoire, Eritrea, Gabon, Mali. Abbreviations: CI = Confidence interval, HSV-2 = Herpes simplex virus type 2. Figure S2. Forest plots presenting the outcomes of the pooled mean herpes simplex virus type 2 (HSV-2) seroprevalence among the different at risk populations across the sub-Saharan Africa subregions.
Eastern Africa All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Southern Africa
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Western Africa
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.      All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted January 26, 2021. ; https://doi.org/10.1101/2021.01.25.21250443 doi: medRxiv preprint Table S12. Summary of the precision assessment and risk of bias assessment for the studies reporting HSV-2 seroprevalence in sub-Saharan Africa.