Early Warning Systems for Identifying Severe Maternal Outcomes in the Who Global Maternal Sepsis Study (2024)

See all articles by Yamikani Chimwaza

Yamikani Chimwaza

University of Liverpool

Alexandra Hunt

University of Liverpool

Livia Oliveira-Ciabati

Hospital Israelita Albert Einstein

Laura Bonnett

University of Liverpool

Edgardo Abalos

Centro Rosarino de Estudios Perinatales (CREP)

Cristina Cuesta

Centro Rosarino de Estudios Perinatales (CREP)

João Paulo Souza

University of São Paulo (USP) - Department of Social Medicine; University of São Paulo (USP) - Ribeirão Preto Medical School

Mercedes Bonet

World Health Organization (WHO) - UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction

Vanessa Brizuela

World Health Organization (WHO) - UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction

David Lissauer

University of Liverpool; University of Malawi - Malawi-Liverpool-Wellcome Trust Clinical Research Programme

Abstract

Background: Infections and sepsis are major killers of women during pregnancy and the post-pregnancy period. Using data from the WHO Global Maternal Sepsis Study, we explored the use of early warning systems (EWS) in women at risk of sepsis-related severe maternal outcomes.

Methods: We searched the literature for EWS in clinical use or research in obstetric populations. We calculated the proportion of women for whom each existing EWS identified them as at risk for developing maternal sepsis by infection severity (complications and severe maternal outcomes). Sensitivity, specificity, positive and negative likelihood ratios, diagnostic odds ratios, and J statistics were calculated to assess EWS performance. Machine learning was used to test the diagnostic potential of routine maternal sepsis markers.

Findings: 19 EWS were assessed in 2560 women from 46 countries with suspected or confirmed infections. The UK Sepsis Trust screening tool, Early Maternal Infection Prompts, and maternity Systemic Inflammatory Response Syndrome scores had high sensitivity (88.1–97.5%) for identifying sepsis-related severe maternal outcomes. The quick Sequential Organ Failure Assessment in Pregnancy score and Obstetrically modified SOFA had high specificity (90.4 - 100%) for identifying women with sepsis-related severe maternal outcomes. Furthermore, combinations of sepsis markers had very low sensitivity and high specificity using machine learning.