Syphilis prevalence among people living with and without HIV in sub-Saharan Africa: A systematic review and meta-analysis

Abstract

Background:

Syphilis is a curable sexually transmitted infection which, untreated, is associated with significant morbidity and mortality. In people living with HIV (PLWH), syphilis carries greater risks of disease progression. We estimated syphilis prevalence among PLWH in the general population in sub-Saharan Africa and compared prevalence among PLWH and without HIV.

Methods:

We searched for studies published 1^st January 2011 to 28^th March 2022 reporting syphilis prevalence among PLWH in sub-Saharan Africa (PROSPERO#: CRD42020167328). We excluded studies in high-risk subpopulations. We estimated pooled syphilis prevalence among PLWH using random effects modelling and compared prevalence to people without HIV when included in the same study. We examined influences of region, study setting, and test type in subgroup analyses.

Results:

We identified 926 studies; 53 were included in the meta-analysis. Pooled syphilis prevalence among PLWH was 7.3% (95% CI: 6.3–8.5%). Prevalence differed by region: 3.1% (95% CI: 2.2–4.0%) in Southern, 5.5% (95% CI: 2.3–9.3%) in West/Central, and 10.5% (95% CI: 8.0–13.1%) in Eastern Africa. Prevalence also differed by study setting: 13.8% (95% CI 5.7–23.0%) in SRH/STI care, 8.7% (95% CI 5.0–12.8%) in HIV care, 7.1% (95% CI 5.8–8.5%) in antenatal care and 3.8% (95% CI 2.0–5.8%) in household/community-based settings. Syphilis prevalence was higher among PLWH than without HIV (RR: 3.5, 95% CI: 2.8–4.5).

Conclusions:

Syphilis is highly prevalent among PLWH in sub-Saharan Africa and is more common among PLWH than without HIV. Integration of syphilis screening and management into HIV care may reduce complications of HIV-syphilis co-infection among PLWH in sub-Saharan Africa.

Summary

Syphilis prevalence among PLWH in sub-Saharan Africa is high and is more prevalent among people with than without HIV. This study provides evidence to support integration of syphilis screening, monitoring and management into existing care for PLWH in sub-Saharan Africa.

Introduction

Syphilis is a curable sexually transmitted infection (STI) caused by the bacterium Treponema pallidum subspecies pallidum.¹ If untreated, syphilis is associated with significant morbidity and mortality and can lead to severe neurologic and cardiovascular complications, particularly in people living with HIV (PLWH).¹ Failure to treat syphilis in pregnancy results in adverse birth outcomes including miscarriage, stillbirth, neonatal death, low birth weight, and congenital infection in over 50% of cases.²

The World Health Organization (WHO) estimates that in 2020, there were 7.1 million new cases of syphilis worldwide.³ WHO estimates from 2016 indicate a global syphilis prevalence of 0.5% (95% confidence interval [CI]: 0.4–0.6%) in both men and women and that the majority (61%) of congenital syphilis cases occur in Africa.^4,5 WHO priority actions include reducing Treponema pallidum incidence by 90% globally by 2030, and reducing congenital syphilis to ≤50 cases per 100,000 live births in 80% of countries.^6,7

In PLWH, syphilis has additional clinical implications. First, syphilis can increase the risk of acquiring and transmitting HIV, including via vertical transmission.^2,8,9 A previous study reported risk of HIV diagnosis was 4% in the first year after syphilis diagnosis and reached 18% ten years after syphilis diagnosis.¹⁰ Second, syphilis can worsen the clinical course of HIV infection, increasing HIV viral load and decreasing CD4 counts.^11–13 Likewise, HIV infection may negatively impact the course of syphilis infection with more frequent chancres;¹⁴ larger and deeper lesions;^15,16 and higher rates of treatment failure for primary and secondary syphilis in PLWH.¹⁷ HIV may also accelerate progression to neurosyphilis.¹⁸

Given the increased risks associated with HIV-syphilis co-infection, understanding the prevalence of syphilis among PLWH can help inform development of interventions and guide decision-making around resource allocation to improve health outcomes in PLWH. Recognizing and treating HIV-syphilis co-infection is particularly important in sub-Saharan Africa, as the region bears 40% of the global burden of STIs and 70% of the global HIV burden.^6,19,20 Currently, data on syphilis prevalence among PLWH outside of the US and Europe is limited given the lack of comprehensive STI surveillance, particularly outside antenatal care settings. In light of this, meta-analysis is one of the few options for understanding the epidemiology of syphilis in sub-Saharan Africa.

Within this context, we aimed to estimate the prevalence of syphilis among PLWH in sub-Saharan Africa and to analyze the relative risk of syphilis among PLWH vs. people without HIV, using data from the available literature.

Materials and Methods

The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO#: CRD42020167328) and the systematic review and meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.²¹

Search strategy

A medical librarian performed a literature search of journal articles and conference abstracts published between 1^st January 2011 and 28^th March 2022, using PubMed, Embase, African Index Medicus via Global Index Medicus, African Wide Information via EBSCO and Web of Science Core Collection (Supplemental Table S1). Keywords and controlled vocabulary related to HIV/AIDS, syphilis, and sub-Saharan Africa were used for database searching. A manual search of country-level reports (e.g., Demographic and Health Surveys, antenatal surveillance data, etc.) from the same period was also conducted. Reference lists of eligible studies and the 2018 WHO Report on Global Sexually Transmitted Infection Surveillance were manually searched for further references. Data from PEPFAR Population-Based HIV Impact Assessment (PHIA) surveys were included where the data were presented within published manuscripts. Citations were uploaded to Covidence Systematic Review Software (Melbourne, Australia) for screening.

Inclusion/exclusion criteria

Observational studies and clinical trials reporting syphilis prevalence among PLWH in sub-Saharan Africa were included if participants were aged ≥13 at time of syphilis testing, serologic or molecular testing for Treponema pallidum was used, and the study was published in one of the three languages spoken by the study team (i.e., English, French or Spanish). Studies were included regardless of whether HIV status was clinically confirmed or self-reported. Only studies reporting data from 2011 onwards were included to reflect current-day syphilis prevalence. Studies were excluded if they: included participants aged <13 with no disaggregation of syphilis prevalence by age; reported only incident HIV infection; enrolled <10 participants with HIV; enrolled only populations known to be at higher risk for syphilis, including female sex workers, men who have sex with men, transgender persons, people who inject drugs, prisoners, clients of female sex workers, victims of sexual assault, military personnel, truck drivers and fishermen; reported only prevalence among blood donors; were only published as abstracts; presented syphilis prevalence not disaggregated by HIV status; or included only participants with complications of syphilis e.g., uveitis or neurosyphilis.

Screening

In the first phase of screening, titles and abstracts were independently screened by two reviewers to determine eligibility. Studies not reporting syphilis prevalence among PLWH in sub-Saharan Africa were marked as irrelevant. In the second phase, full texts were reviewed by two independent reviewers. Conflicts were resolved through consensus among the review team and consultation of a third reviewer, where needed.

Data extraction and management

Following screening and removal of duplicates, data were independently extracted by two reviewers and entered into a REDCap (Vanderbilt University, USA) form. Extracted data included study design, data collection period, location(s), study setting (general care, antenatal care, sexual and reproductive health (SRH)/STI care, HIV care or household/community), clinical characteristics of study population (sex and pregnancy status), syphilis test type (treponemal [e.g., T. pallidum passive particle agglutination [TP-PA] assay], non-treponemal [e.g., Venereal Disease Research Laboratory [VDRL] or rapid plasma reagin [RPR] test], combined, polymerase chain reaction [PCR] or other), sample size, and number of people testing positive for syphilis.

Methodological quality assessment

We used a risk of bias tool for prevalence studies, adapted from Hoy et al., to assess external and internal validity and thus determine overall study quality (Supplemental Table S2).²² Overall risk of bias scores ranged from 0–10, with a score of 0–3 classified as low risk, 4–6 moderate risk and 7–10 high risk.

Statistical analysis

We estimated pooled prevalence with 95% CIs among PLWH and compared prevalence to people without HIV (when included in the same study) using random effects modelling. We assessed heterogeneity between studies using the I² statistic; a value >50% indicated substantial heterogeneity. To explore causes of heterogeneity, we conducted subgroup analyses evaluating the influence of geographic region, study setting, data collection period (using midpoint year of data collection), test type, and study quality. We then performed univariable and multivariable meta-regression using a backwards step-wise approach; variables were retained in the final model if P-values <0.10.

Additionally, we performed comparative analyses to determine the relative risk (RR) of syphilis among PLWH versus without HIV, females versus males living with HIV, and prevalence of syphilis by test type (combined versus treponemal-only and combined versus non-treponemal-only), when these comparative groups were included in the same study. Prevalence and 95% CIs were reported for subgroup analyses and RRs and 95% CIs were reported for comparative analyses. The pooled, sub-group, and comparative analyses were performed and presented using MetaXL version 5.3,²³ while the meta-regression was performed in Stata version 17 (StataCorp; College Station, TX).

Results

We identified 926 unique records through database searches, of which 689 were irrelevant and 188 were excluded after full-text review (Figure 1). Four country reports that met inclusion criteria were identified through a search engine and Ministry of Health websites. A total of 53 studies were included in the pooled prevalence analysis among PLWH while 38 studies were included in the comparative analysis between PLWH and without HIV. Studies included were conducted in 23 different countries in sub-Saharan Africa and included 211,976 PLWH who were tested for syphilis. A list of studies included in the meta-analysis, grouped by region are shown in Supplemental Table S3.

Figure 1.

PRISMA flow chart. PLWH, people living with HIV.

Among the 53 studies included in the meta-analysis, 34 (64.2%) were conducted in Eastern Africa, nine (17.0%) in Southern Africa, nine (17.0%) in West/Central Africa and one (1.9%) in both Eastern and West/Central Africa (Table 1). Most studies were conducted in antenatal care settings (n=28, 52.8%) while 10 studies (18.9%) took place in HIV care settings, eight (15.1%) in household/community-based settings and seven (13.2%) in SRH/STI care settings. Thirty (56.6%) studies included pregnant women. The majority of studies did not report on the presence or absence of STI symptoms among participants. Among studies that reported on symptoms, one (1.9%) enrolled participants asymptomatic for STIs while three (5.7%) enrolled only symptomatic participants and eight (15.1%) enrolled participants with and without STI symptoms. Syphilis test types also varied substantially between included studies with 10 (18.9%) studies using treponemal testing only, 11 (20.8%) using non-treponemal testing only, and 27 (50.9%) using a combined treponemal and non-treponemal testing algorithm, whilst five (9.4%) did not report on the syphilis test used.

Table 1.

Summary characteristics of included studies with pooled prevalence subgroup analyses and meta-regression results.

	Number of studies (%)	Pooled prevalence, % (95% CI)	I2, %	P-value, univariable meta-regression	P-value, multivariable meta-regression
Overall estimate	53 (100)	7.3 (6.3–8.5)	95	n/a	-
Region *
Eastern Africa	35 (66.0)	10.5 (8.0–13.1)	95	ref	ref
Southern Africa	9 (17.0)	3.1 (2.2–4.0)	94	0.001	<0.001
West/Central Africa	10 (18.9)	5.5 (2.3–9.3)	89	0.134	0.076
Study setting
HIV care	10 (18.9)	8.7 (5.0–12.8)	97	0.230	0.072
Antenatal care	28 (52.8)	7.1 (5.8–8.5)	93	0.054	0.011
SRH/STI care	7 (13.2)	13.8 (5.7–23.0)	91	0.019	0.008
Household/community	8 (15.1)	3.8 (2.0–5.8)	93	ref	ref
Data collection period
2011–2015	26 (49.1)	8.3 (6.0–10.9)	93	ref	-
2016 or later	27 (50.9)	6.8 (5.5–8.2)	96	0.571	-
Test type †				n/a**
Treponemal	10 (18.9)	9.6 (4.5–15.5)	93	-	-
Non-treponemal	11 (20.8)	11.0 (5.9–16.7)	97	-	-
Combined	27 (50.9)	6.6 (5.0–8.4)	92	-	-
Not specified	5 (9.4)	-	-	-	-
Risk of bias score
Low (0–3)	26 (49.1)	5.6 (4.6–6.7)	94	ref	-
Moderate (4–6)	27 (50.9)	9.7 (6.8–12.9)	95	0.310	-
High (7–10)	0 (0)	-	-	n/a	-

SRH sexual and reproductive health; STI sexually transmitted infection.

^*One study by Gilbert et al (2021) was conducted in both the Eastern and West/Central African regions.

^†Each study is represented once; if studies used multiple test types, classification was prioritised as follows: combined treponemal + non-treponemal > treponemal only > non-treponemal only.

^**Since different assay types were used on the same population within a given study in many instances, we could not conduct a meta-regression on these nested populations; rather, we evaluated prevalence by assay within a single study in the comparative analyses (see Table 2).

Using the risk of bias tool adapted from Hoy et al.,²² 26 (49.1%) studies were determined to have low risk of bias, 27 (50.9%) studies were moderate risk and zero studies were high risk. The most common sources of bias were poor representation of the target population and sampling frame, less rigorous sampling methods such as convenience sampling, participation rate <75% or not reported, and lack of specified numerators and denominators for prevalence estimates.

In meta-analysis of the included studies, the pooled prevalence of syphilis among PLWH was 7.3% (95% CI: 6.3–8.5%) (Figure 2). There was a high degree of heterogeneity between studies with an I² of 95%.

Figure 2.

Forest plot of pooled prevalence estimates for syphilis among people living with HIV in sub-Saharan Africa.

Pooled prevalence estimates for subgroup analyses are shown in Table 1 and Supplemental Figures S1a–f. Prevalence among PLWH differed by region; in subgroup analyses, pooled prevalence was 3.1% (95% CI 2.2–4.0%) in Southern, 5.5% (95% CI 2.3–9.3%) in West/Central, and 10.5% (95% CI 8.0–13.1%) in Eastern Africa. In univariable meta-regression, pooled prevalence in Southern Africa was significantly lower compared to Eastern Africa (P=0.001). In subgroup analyses, syphilis prevalence among PLWH also differed by study setting with the highest prevalence among studies conducted in SRH/STI care settings (13.8%, 95% CI 5.7–23.0%), followed by HIV care (8.7%, 95% CI 5.0–12.8%), antenatal care (7.1%, 95% CI 5.8–8.5%), and household/community-based settings (3.8%, 95% CI 2.0–5.8%). In univariable meta-regression, syphilis prevalence in SRH/STI care settings was significantly different compared to household/community (P=0.019) settings. The prevalence in antenatal and HIV care settings were not significantly different compared to household/community settings (P=0.054 and P=0.230, respectively). However, in the multivariable meta-regression, prevalence in both SRH/STI and antenatal care settings were significantly different compared to household/community setting (P=0.008 and P=0.011, respectively). When limiting studies to the combined treponemal and non-treponemal approach, pooled prevalence among PLWH did not change substantially (6.6%; 95% CI 5.0–8.4%). Syphilis prevalence among PLWH did not differ significantly by data collection period: 8.3% (95% CI 6.0–10.9%) for studies conducted between 2011 and 2015 and 6.8% (95% CI 5.5–8.2%) for studies conducted in 2016 or later (P=0.571). Pooled prevalence among PLWH was higher in studies classified as moderate risk (9.7%, 95% CI 6.8–12.9%) compared to studies classified as low risk (5.6%, 95% CI 4.6–6.7%), but this was not significant in univariable meta-regression (P=0.310). Limiting the analysis to pregnant women with HIV also did not alter the pooled prevalence (7.0%; 95% CI 5.8–8.4%) substantially (Supplemental Figure S2).

Multivariable meta-regression revealed that much of the heterogeneity observed in the pooled prevalence analysis can be explained by region and setting; both remained statistically significant in the multivariable model with P<0.01.

In comparative analysis between people living with and without HIV, syphilis prevalence was higher among PLWH (RR: 3.5, 95% CI 2.8–4.5) (Figure 3, Table 2). Estimates were not substantially different with exclusion of studies in which HIV status was self-reported. In subgroup analyses, the increased risk of syphilis in PLWH compared to people without HIV differed between Eastern (RR=4.1; 95% CI 3.1–5.4), West/Central (RR=2.0; 95% CI 1.2–3.1) and Southern Africa (RR=2.9; 95% CI 1.8–4.6) (Supplemental Figure 3a); the difference between Eastern and West/Central Africa was significant in univariable meta-regression (P=0.015) while the difference between Eastern and Southern Africa was not significant (P=0.339). The increased risk of syphilis in PLWH compared to people without HIV was higher among females (RR=4.7; 95% CI 3.3–6.8) than males (RR=3.7; 95% CI 2.3–5.9) (Supplemental Figure 3b). When limiting studies to only those that had data for both males and females living with HIV, there was no substantial difference in syphilis prevalence between males and females living with HIV (RR=0.9; 95% CI 0.8–1.2) (Supplemental Figure 3c). When limiting the comparative analysis to studies with both combined and either treponemal-only or non-treponemal-only tests within the same study, pooled prevalence was lower when using a combined test (combined test vs treponemal-only test: RR=0.3, 95% CI 0.2–0.5); combined vs non-treponemal-only test: RR=0.7, 95% CI 0.6–0.8) (Supplemental Figure 3d).

Figure 3.

Forest plot for comparative analysis of syphilis prevalence between people living with and without HIV.

Table 2.

Comparative analysis by HIV status, sex, and test type, limited to studies that included both groups included in the analysis.

	Number of studies	Number of people with syphilis/number tested (%)		Relative risk (95% CI)	I2, %
		HIV+	HIV−
HIV status (living with HIV versus without HIV)	38	687/17394 (3.95)	3280/157445 (2.08)	3.5 (2.8–4.5)	83
Region
Eastern Africa	27	471/10373 (4.54)	2103/114036 (1.84)	4.1 (3.1–5.4)	81
Southern Africa	5	132/5775 (2.29)	79/11993 (0.66)	2.9 (1.8–4.6)	49
West/Central Africa	6	84/1246 (6.74)	1098/31416 (3.50)	2.0 (1.2–3.1)	65
Sex
Males	5	68/2448 (2.78)	216/28554 (0.76)	3.7 (2.3–5.9)	47
Females	5	143/4957 (2.88)	256/35595 (0.72)	4.7 (3.3–6.8)	49
Among people living with HIV			-
Sex (female vs male)	10	47/1267 (3.71)	-	0.9 (0.8–1.2)	31
Test type (combined vs treponemal-only)	6	68/1831 (3.71)	-	0.3 (0.2–0.5)	80
Test type (combined vs non-treponemal-only)	5	72/993 (7.25)	-	0.7 (0.6–0.8)	0

Discussion

In this systematic review and meta-analysis, we have addressed a gap in epidemiological data on syphilis and HIV coinfection by estimating the prevalence of syphilis among PLWH. We found a high prevalence of syphilis among PLWH in sub-Saharan Africa, with a pooled prevalence of 7.3% (95% CI: 6.3–8.5%). We also found that the prevalence of syphilis was higher among PLWH than people without HIV. These findings highlight the need to strengthen syphilis screening programmes as part of routine HIV care in sub-Saharan Africa.

Our pooled syphilis prevalence among PLWH (7.3%) was higher than the WHO 2020 prevalence estimate of 1–2% among adults aged 15–49 years in Africa,²⁴ and higher than estimates from a 2018 meta-analysis which reported a pooled syphilis prevalence of 3.04% (95% CI 2.84–3.24%) among adults in the WHO African Region general population from 1990–2016.²⁵ Our pooled syphilis prevalence among PLWH was also higher than estimates reported by Rowley et al., (2019) among the general population in Africa (1.6%, 95% CI: 1.2–2.0) and globally (0.5%, 95% CI 0.4–0.5%).²⁶ Due to lack of available globally/regionally representative estimates of syphilis prevalence among PLWH, we could not make direct comparisons with prior literature. However, the higher prevalence of syphilis among PLWH in our study compared with the above-mentioned population-level estimates aligns with the findings of our comparative analysis which showed a higher prevalence among people with than without HIV. Possible reasons for the higher syphilis prevalence among PLWH include the increased susceptibility to infection due to decreased CD4 cells and the well-established increase in risk of HIV among people with syphilis.^8,27 Prevalence data on syphilis in sub-Saharan Africa is lacking in general and the majority of prevalence estimates come from studies conducted in antenatal care settings. Our findings of a high syphilis prevalence among PLWH highlights the need to invest in syphilis screening programmes beyond the antenatal context, with prioritization of populations affected by HIV.

We found no significant difference in syphilis prevalence between males and females living with HIV in sub-Saharan Africa, echoing the findings of a global pooled meta-analysis among the general population by Smolak et al., (2018).²⁵ This result is, however, in contrast to findings from other settings where a higher syphilis prevalence has been reported among men versus women living with HIV.^28–30 Our pooled syphilis prevalence among pregnant women with HIV (7.0%; 95% CI 5.8–8.4%) was higher than in two systematic reviews and meta-analyses conducted among the general population of pregnant women in sub-Saharan Africa, both of which found a pooled syphilis prevalence of 2.9%.^31s,32s These findings of a higher syphilis prevalence among pregnant women with HIV support global efforts to integrate combined HIV/syphilis testing in antenatal care as part of dual elimination efforts.

We found that syphilis prevalence among PLWH differed by region, with the lowest prevalence in Southern Africa (3.1%) and highest in Eastern Africa (10.5%). This finding is in contrast to a systematic review and meta-analysis conducted between 2015–2020 among pregnant women (with or without HIV), which found that Central Africa had the highest pooled prevalence of syphilis compared to other African regions.^32s The reasons underlying these regional differences in syphilis prevalence are unknown, however, they may be attributed to variations in age at first sex and marriage rates between regions.^33s STI prevalence estimates in our study also differed between study settings; syphilis prevalence was highest in studies conducted in SRH/STI and HIV care settings compared to antenatal care and household/community-based settings. While a higher syphilis prevalence in SRH/STI care is anticipated, the comparably high syphilis prevalence amongst people seeking HIV care illustrates the importance of integrating syphilis screening into routine HIV care. Similarly, offering syphilis screening for people seeking SRH/STI care could have a significant impact in reducing prevalence of syphilis in sub-Saharan Africa.

Improving syphilis surveillance is key to implementing effective syphilis screening and treatment programmes. The majority of studies in the current analysis used a combined treponemal and non-treponemal testing algorithm. In a comparative analysis limited to studies which reported prevalence using both combined and treponemal/non-treponemal-only tests, pooled prevalence was lower when using a combined test versus a treponemal or non-treponemal-only test. This finding is in line with meta-regression results from Smolak et al., which demonstrated that studies using a treponemal-only test (Treponema pallidum haemagglutination; TPHA) had a two- to three-fold higher odds of reported syphilis positivity compared to studies using a combined treponemal and non-treponemal (RPR) approach.²⁵ Due to limitations associated with both non-treponemal tests (low sensitivity to detect primary and late latent syphilis, false-negative results due to prozone reactions or false positive results) and treponemal tests (inability to differentiate between past or present infection),^34s,35s a combined treponemal and non-treponemal testing algorithm is generally considered the more accurate and preferred approach.^36s However, while prevalence studies are more likely to have access to both types of tests, national programmes typically have one or the other which may influence study-based prevalence data compared to national prevalence data. There is a need to improve syphilis surveillance and prioritise combined testing mechanisms, particularly those that can be used at point-of-care. For pregnant women in settings with a high prevalence of syphilis (5% or greater), the WHO recommends an on-site rapid test (treponemal) and if the test is positive, provision of a first dose of treatment to prevent adverse birth and pregnancy outcomes, followed by an RPR test and treatment according to duration of infection if RPR is positive.^37s Development of affordable, point-of-care, combined treponemal and non-treponemal tests could substantially aid in both syphilis surveillance and implementation of rapid syphilis treatment in high-burden settings.^38s

There are some limitations to this systematic review and meta-analysis. First, there were almost four times more studies conducted in Eastern Africa compared to Southern and West/Central Africa. Additional high-quality studies from the latter regions are required to better understand the burden of syphilis in sub-Saharan Africa. Second, the majority of studies included in this analysis did not report on absence or presence of STI symptoms. Syphilis prevalence in studies that recruited from symptomatic participants without reporting such may be higher than in studies that recruited from asymptomatic or mixed symptomatic and asymptomatic populations. Similarly, while we aimed to aggregate syphilis prevalence estimates from studies reflective of the general population of PLWH, it is possible that members of high-risk groups were over-represented in these studies, which may skew estimates higher than national surveillance studies would suggest. We also included attendees of STI clinics based on the rationale that in settings where STI care is not integrated with HIV management, PLWH experiencing STI symptoms may seek care at STI clinics and excluding STI care sites could bias our estimates. However, it is possible that STI clinic attendees, particularly those who are symptomatic, could have a higher risk of syphilis than some other high-risk populations which were excluded from this meta-analysis. Finally, there was a lack of harmonisation in test type between studies, with some studies using a combined treponemal and non-treponemal testing approach and others using one test or the other. Studies collecting syphilis prevalence data should be encouraged to use a combined testing approach as estimates may vary considerably when using treponemal or non-treponemal tests alone.

Despite these limitations, the findings of this study may be helpful to inform regional and country-level surveillance and interventions as countries strive towards achieving the goals for 2030 set out in the WHO Global Health Sector Strategy on Sexually Transmitted Infections.^6,7 In most countries in sub-Saharan Africa, the syndromic management approach is used for the management of STIs, and syphilis screening is only routinely offered during antenatal care.^39s In PLWH, untreated syphilis infection carries greater risks of disease progression, including neurologic complications.^40s Our findings of a high prevalence of syphilis among PLWH in sub-Saharan Africa underlines the need to scale up implementation of routine syphilis screening beyond antenatal care for this population.

Conclusion

Syphilis is a curable sexually transmitted infection but the burden remains high in sub-Saharan Africa, particularly among PLWH. Integration of syphilis screening, monitoring and management into HIV care could have a substantial impact on reducing complications of HIV-syphilis co-infection in sub-Saharan Africa.

Data availability statement

Data collection forms; data extracted from included studies; and analytic code available on request.