Undetected sexually transmitted infections during pregnancy carry serious consequences — preterm birth, neonatal conjunctivitis, and vertical transmission — yet most low-resource settings still rely on syndromic management protocols that miss the majority of cases. A validated clinical risk algorithm may offer a dramatically better solution without requiring expensive laboratory infrastructure.
In a cohort of 87 pregnant women seeking STI services in Kigali, Rwanda between 2017 and 2020, the overall burden of reproductive tract infections was striking: 79% had at least one infection, with chlamydia or gonorrhea (CT/NG) present in 28% of participants. Four independent predictors of CT/NG emerged from adjusted analysis — age 25 or younger (aPOR 4.92), inconsistent condom use (aPOR 4.86), absence of concurrent candida infection (aPOR 4.23), and clinical signs of endocervical inflammation or discharge (aPOR 4.91). When the researchers' previously developed risk algorithm was applied to this pregnant subgroup, it achieved 92% sensitivity for CT/NG detection, compared with only 46% for Rwanda's 2019 national guidelines and a substantially lower 35% for the updated 2024 guidelines — a troubling regression in guideline performance.
This finding sits within a longstanding tension in global reproductive health: PCR-based diagnostics are the gold standard but remain prohibitively expensive across sub-Saharan Africa, while syndromic protocols — treating based on symptoms alone — are cheap but inaccurate. The 35–46% sensitivity of national guidelines means the majority of CT/NG cases in pregnant women go untreated, a public health failure with measurable neonatal consequences. The algorithm's 92% sensitivity represents a genuinely meaningful improvement using clinical variables that require no additional cost. That said, the cohort is small (n=87) and drawn from a single urban Rwandan site, limiting generalizability across rural or higher-burden contexts. Male partner risk factors were noted as absent from current algorithms, a gap the authors flag for future refinement. This work is incremental in method but potentially high-impact in application if validated at scale.