Abstract / Summary
Abstract Background Pre-term birth (PTB) remains a leading cause of neonatal morbidity and mortality, requiring scalable approaches to earlier risk identification in routine, real-world settings. Heart rate variability (HRV) is emerging as a noninvasive biomarker of autonomic nervous system regulation with potential relevance to maternal physiologic adaptation and pregnancy risk, particularly given its increasing feasibility for continuous, remote assessment through widely used wearable sensing technologies. Objectives To understand the association between maternal HRV and PTB by synthesizing the existing evidence. Methods A search strategy was applied to PubMed, CINAHL, and Scopus, in collaboration with a medical librarian from inception through September 2025 and updated in October 2025. This systematic review followed a prospectively registered protocol on the Open Science Framework (OSF; Center for Open Science). Two reviewers independently screened titles/abstracts and full texts, with disagreements resolved by a third reviewer. Methods were guided by the Cochrane Handbook and reported according to PRISMA. Data were extracted using a PECO framework, and methodological quality (risk of bias) was assessed using the revised Joanna Briggs Institute (JBI) Critical Appraisal Tool for Cohort Studies. Results Five studies met inclusion criteria; three prospective cohorts and two retrospective analyses. Findings were heterogeneous and largely focused on longitudinal HRV trajectories often relying on a single metric such as RMSSD captured across pregnancy using consumer wearables including WHOOP, Oura, Samsung. Rather than consistently reporting explicit PTB versus non-PTB contrasts, studies emphasized within-pregnancy patterning and trajectory features. Outcome definitions and comparator frameworks also varied including outcome-based PTB comparisons, history-based risk cohorts, and delivery-timing groupings, which limited cross-study comparability. Conclusion HRV shows clear longitudinal patterning across pregnancy, yet RMSSD demonstrates inconsistent discrimination of PTB versus term outcomes and may be insufficient as a standalone marker given metric constraints and PPG-related measurement error. Late-pregnancy HRV trajectories may reflect proximity to delivery more than gestational age, suggesting that within-person trajectory features could outperform GA-aligned cross-sectional comparisons. Predictive performance improves when HRV is integrated with complementary wearable signals including temperature, sleep, activity, supporting a multi-signal approach. Interpretation, however, is limited by variable confounding control, incomplete attrition reporting, and potential selection bias in live-birth-only cohorts. Also, clinically, pairing HRV tracking with pathway-relevant clinical data may better support targeted, timely interventions. Future studies should harmonize outcome definitions, strengthen clinical metadata and confounder adjustment, handle missing data transparently, and apply pathway-aware or time-to-delivery models to establish actionable risk stratification. Systematic Review registration Open Science Framework (OSF) Registries; https://doi.org/10.17605/OSF.IO/E9NU4 .