PROTECT-Africa: Biomarker Validation for Preeclampsia Triage

Hypertensive disorders of pregnancy account for approximately 16 to 22 percent of maternal deaths in Africa, and the incidence of preeclampsia is estimated to be several times higher in low- and middle-income countries than in high-income settings. Angiogenic biomarkers, particularly the ratio of soluble fms-like tyrosine kinase-1 (sFlt-1) to placental growth factor (PlGF), have demonstrated strong negative predictive value for short-term progression to preeclampsia in European and Asian cohorts, where a ratio cut-off of 38 has been validated. These cohorts did not include African populations. Differences in baseline angiogenic profiles and in the prevalence of comorbidities such as anaemia and infection may affect biomarker performance, so population-specific thresholds are required before these tools can be relied upon locally. The study is organised in two parts using a single enrolled population. Part A establishes and then confirms sFlt-1/PlGF ratio thresholds on the Roche Elecsys reference platform, using a derivation cohort of 553 participants followed by an independent validation cohort of 553 participants. Part B evaluates rapid point-of-care platforms against those confirmed reference thresholds in the same 1,106 participants. The platforms under evaluation include quantitative serum or plasma immunoassays and a qualitative urine-based test that detects misfolded proteins. Stored samples may be used for additional platforms as they become commercially available. Blinding is maintained at all sites: neither participants nor investigators receive biomarker results. Serum samples are batch-processed after collection, and urine tests requiring immediate analysis are performed without results being released to clinical teams. Because the reference and investigational results cannot influence management, the diagnostic accuracy estimates reflect the natural course of disease under routine care. Routine safety laboratory results, including full blood count and liver function tests, are returned to treating clinicians, since these are inconsistently available at the participating sites and support standard care. Participants are grouped for analysis by the timing of diagnosis relative to the baseline visit: within one week, more than one week to four weeks, and beyond four weeks through two weeks postpartum. Those who develop none of the study conditions through four weeks postpartum form the comparator group. A within-trial economic evaluation runs alongside the clinical analysis. It uses a decision-analytic model from a societal perspective, capturing both health system and household costs, with short-term and five-year time horizons. Cost and volume data are collected from between three and seven purposively sampled health workers at participating facilities, who are not part of the trial population.