Search

Abstract Introduction Of 37.7 million people living with HIV in 2020, 6.1 million still do not know their HIV status. We synthesize evidence on concurrent HIV testing among people who tested for other sexually transmitted infections (STIs). Methods We conducted a systematic review using five databases, HIV conferences and clinical trial registries. We included publications between 2010 and May 2021 that reported primary data on concurrent HIV/STI testing. We conducted a random‐effects meta‐analysis and meta‐regression of the pooled proportion for concurrent HIV/STI testing. Results We identified 96 eligible studies. Among those, 49 studies had relevant data for a meta‐analysis. The remaining studies provided data on the acceptability, feasibility, barriers, facilitators, economic evaluation and social harms of concurrent HIV/STI testing. The pooled proportion of people tested for HIV among those attending an STI service (n = 18 studies) was 71.0% (95% confidence intervals: 61.0–80.1, I2 = 99.9%), people tested for HIV among those who were tested for STIs (n = 15) was 61.3% (53.9–68.4, I2 = 99.9%), people tested for HIV among those who were diagnosed with an STI (n = 13) was 35.3% (27.1–43.9, I2 = 99.9%) and people tested for HIV among those presenting with STI symptoms (n = 3) was 27.1% (20.5–34.3, I2 = 92.0%). The meta‐regression analysis found that heterogeneity was driven mainly by identity as a sexual and gender minority, the latest year of study, country‐income level and region of the world. Discussion This review found poor concurrent HIV/STI testing among those already diagnosed with an STI (35.3%) or who had symptoms with STIs (27.1%). Additionally, concurrent HIV/STI testing among those tested for STIs varied significantly according to the testing location, country income level and region of the world. A few potential reasons for these observations include differences in national STI‐related policies, lack of standard operation procedures, clinician‐level factors, poor awareness and adherence to HIV indicator condition‐guided HIV testing and stigma associated with HIV compared to other curable STIs. Conclusions Not testing for HIV among people using STI services presents a significant missed opportunity, particularly among those diagnosed with an STI. Stronger integration of HIV and STI services is urgently needed to improve prevention, early diagnosis and linkage to care services.

Objectives: Genital Ulcer Disease (GUD) carries a significant disease burden globally. With limited access to diagnostics, the 2001 World Health Organization (WHO) sexually transmitted illnesses (STI) guidelines proposed a syndromic management algorithm that required a clinical decision to determine the management of GUD. We assessed the diagnostic accuracy of this algorithm.Methods: We conducted a systematic review (Prospero: CRD42020153294) using eight databases for publications between 1995 and January 2021 that reported primary data on the diagnostic accuracy of clinical diagnosis to identify aetiological agents of GUD. Titles and abstracts were independently assessed for eligibility, and data were extracted from full texts for sensitivity/specificity. A hierarchical logistic regression model was used to derive pooled sensitivity and specificity. We used GRADE to evaluate the certainty of evidence.Results: Of 24,857 articles, 151 full texts were examined and 29 included in the analysis. The majority were from middle-income countries [(14/29 (48%) lower middle, 10/29 (34%) upper middle)]. We pooled studies where molecular testing was using to confirm the aetiology of GUD: 9 studies (12 estimates) for herpes, 4 studies (7 estimates) for syphilis, and 7 studies (10 estimates) for chancroid. The pooled sensitivity and specificity of GUD for the detection of herpes was 43.5% [95% confidence interval (CI): 26.2–62.4], and 88.0% (95% CI: 67.0–96.3), respectively (high certainty evidence); and for syphilis were 52.8% (95% CI: 23.0–80.7), and 72.1% (95% CI: 28.0–94.5) (moderate certainty evidence); and for chancroid were 71.9% (95% CI: 45.9–88.5) and 53.1% (95% CI: 36.6–68.9) (moderate certainty evidence), respectively.Conclusion: Algorithms requiring a clinical diagnosis to determine and treat the aetiology of GUD have poor sensitivities for syphilis and herpes simplex virus, resulting in significant numbers of missed cases. There is an urgent need to improve access to affordable and efficient diagnostics (e.g., point-of-care tests) to be incorporated into GUD algorithms to better guide appropriate management.Systematic Review Registration: PROSPERO, identifier: CRD42020153294.