Incidence of Maternal Sepsis and Sepsis-Related Maternal Deaths in the United States

Maternal sepsis is a leading cause of maternal morbidity and mortality. However, population-based estimates of maternal sepsis occurring after delivery hospitalization have been limited because previous studies have focused on select populations or have not followed up patients longitudinally.^1,2 Thus, the burden of maternal sepsis and sepsis-related deaths may be underestimated. We assessed the nationwide incidence and outcomes of maternal sepsis within 42 days of delivery hospitalization discharge using all-payer data.

We analyzed data from 2013 to 2016 from the National Readmissions Database (NRD), which aggregates all-payer hospital discharges from 27 states, representing 57.8% of the US population, and allows for measurement of readmissions, even across insurance changes. We identified single or multiple live-birth and stillbirth delivery hospitalizations, excluding ectopic pregnancies, molar pregnancies, and abortions.³ We further limited index hospitalizations to those with a January through October discharge to allow for 42-day follow-up in the yearly data sets.

We identified sepsis using diagnosis codes for severe sepsis and septic shock or concurrent codes for sepsis and acute organ dysfunction.² For both delivery hospitalizations and sepsis, we updated International Classification of Diseases, Ninth Revision, Clinical Modification identification methods to International Classification of Diseases, Tenth Revision, Clinical Modification methods. Acute organ dysfunctions were identified using criteria based on the methods of Dombrovskiy et al,⁴ which have been used in previous studies of maternal sepsis.² Site of infection was classified based on the highest-ranking diagnosis indicating an infectious site, and comorbidities were identified using a weighted index from Bateman et al.⁵ Maternal sepsis was defined as sepsis occurring during delivery hospitalization or during a readmission within 42 days of delivery discharge. The 42-day end point was selected in accordance with the World Health Organization definition of maternal sepsis.⁶ Maternal death was defined as in-hospital mortality occurring during index delivery hospitalization or within 42 days of delivery hospitalization discharge. Maternal deaths were considered sepsis-related if they occurred during a hospitalization with sepsis. The NRD does not include data on out-of-hospital deliveries or deaths.

The primary outcomes were the incidence of maternal sepsis and sepsis-related mortality (overall and during and after delivery hospitalization). All-cause maternal deaths were also measured and the proportion that were sepsis-related was determined. Weighted characteristics of maternal sepsis episodes occurring during vs after delivery hospitalization were compared using the F statistic with significance set at a 2-sided P < .05. Nationwide estimates were generated using NRD sampling weights. Results are presented as unweighted numbers and weighted proportions, means, or incidences per 100 000 deliveries. Analyses were performed with Stata/MP15 (StataCorp). This study was deemed exempt by the University of Michigan’s institutional review board.

From 63 665 593 hospitalizations in the NRD (2013-2016), there were 5 957 678 delivery hospitalizations with discharges in January through October, of which 32.5% were cesarean deliveries. There were 2905 deliveries complicated by sepsis (0.038%; 95% CI, 0.037%-0.040%); 49.8% (95% CI, 47.7%-52.0%) were complicated during delivery hospitalization and 50.3% (95% CI, 48.1%-52.4%) after delivery discharge. Sepsis readmissions occurred a mean of 13.6 (SD, 11.2) days after discharge from delivery hospitalization.

Maternal sepsis episodes after delivery discharge, compared with during delivery hospitalization, occurred in women with fewer comorbidities (mean, 1.9 vs 2.6; difference, −0.7 [95% CI, −0.8 to −0.5]; P < .001), were less frequently associated with cesarean deliveries (49.6% vs 63.7%; difference, −15.7% [95% CI, −19.2% to −12.1%]; P < .001), had a higher percentage of acute kidney dysfunction (35.2% vs 23.9%; difference, 10.6% [95% CI, 7.3%-13.9%]; P < .001), and had a lower percentage of genitourinary infection (39.8% vs 49.2%; difference, −7.5% [95% CI, −11.1% to −3.9%]; P < .001) (Table 1).

There were 408 delivery hospitalizations (0.007%; 95% CI, 0.006%-0.007%) that were followed by maternal death; 74.4% (95% CI, 69.6%-78.6%) during delivery hospitalization and 25.6% (95% CI, 21.4%-30.4%) after delivery discharge (Table 2). Overall, 95 (22.6%; 95% CI, 18.7%-27.1%) of the maternal deaths were sepsis-related; 17.3% (95% CI, 13.2%-22.4%) occurred during delivery hospitalization and 38.1% (95% CI, 29.0%-48.1%) occurred after delivery discharge.

Maternal sepsis occurred in 0.04% of deliveries in the United States, and 23% of all maternal deaths were sepsis-related. Substantial percentages of both outcomes occurred after discharge from delivery hospitalization. Study limitations include lack of data on home births and out-of-hospital deaths. The findings highlight the importance of sepsis in maternal mortality, particularly after discharge, and the need for awareness of this condition.

Accepted for Publication: June 17, 2019.

Corresponding Author: Matthew K. Hensley, MD, MPH, Department of Internal Medicine, University of Michigan, 1500 E Medical Center Dr, 3916 Taubman Clinic, Ann Arbor, MI 48109 (hematthe@med.umich.edu).

Author Contributions: Dr Hensley had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Bauer, Admon, Prescott.

Acquisition, analysis, or interpretation of data: Hensley, Prescott.

Drafting of the manuscript: Hensley, Admon.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hensley, Prescott.

Administrative, technical, or material support: Hensley, Prescott.

Supervision: Admon, Prescott.

Conflict of Interest Disclosures: None reported.

Funding/Support: Dr Prescott is supported by grant K08 GM115859 from the National Institutes of Health/National Institute of General Medical Sciences.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views herein do not necessarily reflect the position or policy of the US government or the US Department of Veterans Affairs.