High predictability, low false positives with pregnancy-adjusted sepsis screening tools

Key takeaways:

• Pregnancy-adjusted screening tools had high sepsis predictability rates for women with sepsis or chorioamnionitis.
• Pregnancy-adjusted tools also had low false-positive rates for these patient populations.

SAN FRANCISCO — During childbirth hospitalizations, pregnancy-adjusted vs. non-adjusted screening tools had the highest prediction rate and lowest false-positive rates for sepsis, according to results of a case-control study.

“Sepsis in hemorrhage and hypertension is the No. 2 cause of mortality and the third leading cause of severe maternal morbidity after delivery,” Elliot K. Main, MD, clinical professor in the department of obstetrics and gynecology at Stanford University, said during a presentation at the ACOG Annual Clinical & Scientific Meeting. “But if you look at the peripartum periods and the postpartum periods, it is the leading cause of severe maternal morbidity.”

Main and colleagues conducted a case-control study using electronic health record data to identify all delivery admissions of women with sepsis from 59 hospitals to assess the performance characteristics of available screening tools to diagnose sepsis during childbirth hospitalizations.

The first cohort had 647 women with sepsis, of whom 228 had end-organ injury, and the second cohort had 14,591 women with chorioamnionitis-endometritis, of whom 1,049 had sepsis and 238 had end-organ injury. All women with sepsis were matched by gestational age at delivery with women without sepsis, and all women with chorioamnionitis and sepsis were compared with women with chorioamnionitis without sepsis.

Researchers reported sensitivity, false-positive rates and C-statistics for the following widely used sepsis screening tools:

• the California Maternal Quality Care Collaborative (CMQCC);
• systemic inflammatory response syndrome (SIRS);
• the maternal early warning criteria (MEWC);
• the United Kingdom Obstetric Surveillance System (UKOSS); and
• the Maternal Early Warning Trigger (MEWT) tool.

In the first and second cohorts, researchers observed the highest C-statistics for both the CMQCC (0.92 and 0.67, respectively) and UKOSS (0.91 and 0.64, respectively) pregnancy-adjusted criteria. For the first cohort, SIRS (21.3%) and the MEWC (38.3%) had similar sensitivity, but higher false-positive rates compared with the CMQCC (6.9%) and UKOSS (9.6%).

In the second cohort, all screening tools had high false-positive rates. However, the false-positive rates for the CMQCC (60.2%) and UKOSS (67.5%) were substantially lower compared with SIRS (86.6%) and the MEWC (92.3%).

“Careful attention to early warning signs and vital signs during screening and during admission can identify patients with serious infections and sepsis,” Main said. “But a second step is required that includes the bedside evaluation and simple laboratory tests to identify those who need more intensive treatment and support.”

In an accompanying editorial, Andrea D. Shields, MD, MS, associate professor, and Beverly C. Tse, MD, maternal-fetal medicine fellow in the department of obstetrics, gynecology and women’s health at the University of Connecticut Health Center, noted that the findings highlight the importance of using a two-step process involving an initial clinical data screen to alert clinicians of those at higher risk and further clinical and laboratory evaluation to detect and confirm maternal sepsis.

“Although it often seems that detecting every case of maternal sepsis is aspirational and reminiscent of trying to find the proverbial needle in the haystack, this goal remains worthwhile,” Shields and Tse wrote. “Continual refinement of screening and treatment protocols with emerging evidence may help to one day realize the ultimate aspiration ... to eradicate maternal sepsis deaths.”

For more information:

Elliot K. Main, MD, can be reached at emain@stanford.edu.

References:

• Main EK, et al. Obstet Gynecol. 2024;doi:10.1097/AOG.0000000000005477.
• Shields AD, et al. Obstet Gynecol. 2024;doi:10.1097/AOG.0000000000005513.