Maternal Intensive Care Unit Utilization by Mode of Delivery in the United States: A Retrospective Cohort Study

BACKGROUND:

Improved understanding of intensive care unit (ICU) utilization among obstetric patients is essential for optimizing outcomes, managing severe maternal morbidity (SMM), and guiding resource allocation. Despite the clinical importance of ICU care in obstetrics, large-scale data examining utilization patterns and differences by mode of delivery remain limited. To address this gap, our objectives were to determine the proportion of delivery hospitalizations involving ICU admission, to describe the indications for ICU admissions, and to assess factors associated with ICU admission.

METHODS:

We conducted a retrospective cohort study using delivery hospitalizations in the Premier Database (October 1, 2015–December 31, 2020). ICU admission indications were characterized using Centers for Disease Control and Prevention (CDC)-defined SMM events. Mode of delivery was stratified as vaginal birth, intended cesarean birth (cesarean without trial of labor), and intrapartum cesarean birth (cesarean after trial of labor). Multivariable logistic regression was used to evaluate factors associated with ICU admission. We also performed a secondary analysis to determine the rate of ICU admission in non-SMM and SMM patients with and without comorbidities.

RESULTS:

We analyzed 4.7 million delivery hospitalizations; 3261,071/4789,673 (68.1%) were vaginal births, 539,670/4,789,673 (11.3%) intrapartum cesarean births, 988,932/4789,673 (20.6%) intended cesarean births. Overall, 35,837/4789,673 (0.75%) patients required ICU admission, including 16,490/3261,071 (0.51%) vaginal births, 8243/539,670 (1.53%) intrapartum cesarean births, and 11,104/988,932 (1.12%) intended cesarean births. Non-transfusion SMM occurred in 18,552/4789,673 (0.39%) of patients. Among ICU patients, 6574/35,837 (18.4%) experienced non-transfusion SMM events, while 27,329/35,837 (76.3%) patients did not meet criteria for a SMM event. The most common SMM events in the ICU were acute respiratory distress syndrome (2509/35,837; 7.0%), ventilation (2401/35,837; 6.7%), shock (2186/35,837; 6.1%), hysterectomy (2150/35,837; 6.0%), and disseminated intravascular coagulopathy (1433/35,837; 4.0%).

In multivariable analysis, both intrapartum cesarean and intended cesarean births were associated with increased odds of ICU admission compared with vaginal birth—adjusted odds ratio (aOR) 1.87 (95% CI, 1.81–1.94) and 1.46 (1.41–1.52), respectively. Additional factors associated with ICU admission included transfusion of ≥4 units of packed red blood cells (aOR 116.3; 95% CI, 110.3–122.7); cardiomyopathy (18.5; 16.7–20.4); preeclampsia with severe features (6.72; 6.48–6.97); pulmonary hypertension (6.53; 5.17–8.25); arrhythmia (3.99; 3.60–4.43); placenta accreta spectrum (3.90; 3.46–4.39); preterm birth (3.55; 3.45–3.66); and chronic renal disease (2.48; 2.24–2.75).

In a secondary analysis, SMM occurred in 24,358/2,831,900 (0.9%) patients with comorbidities compared with 4128/1957,773 (0.2%) of patients with no comorbidity present. Additionally, patients with comorbidities without SMM had higher rates of ICU admission across modes of delivery when compared to patients without comorbidities.

CONCLUSIONS:

There is approximately one ICU admission per 135 delivery hospitalizations in this US database. Intrapartum and intended cesarean births confer higher risks of ICU admission compared with vaginal birth. Among all investigated factors, significant transfusion requirements were most strongly associated with ICU admission. These findings highlight the importance of early recognition and preparation for potential morbidity associated with intrapartum cesarean birth and underscore the need for proactive obstetric critical care planning.