755 - Shifting Thresholds of Viability: National Trends in Active Care and Outcomes for 21–24 Weeks’ Gestation, 2016–2024

Publication Number: 2736.755

Erwin T. Cabacungan, Medical College of Wisconsin, Children's Wisconsin, Milwaukee, WI, United States; Krishna Acharya, Medical College of Wisconsin, Milwaukee, WI, United States; Ruby Gupta, Children's Hospital of Wisconsin, Milwaukee, WI, United States

Background: Care for infants born at the margins of viability (21-24 weeks' gestation) has shifted towards initiating active perinatal interventions. However, national impact of these evolving practices on birth incidence and survival remains uncertain. Understanding these trends is essential for guiding periviability counseling, resource allocation, and ethical decision-making.

Objective: To examine national trends in active perinatal intervention and evaluate whether increasing Acute Care Prematurity Index (ACPI) is associated with changes in birth and infant death rates among infants born at 21-24 weeks' gestation.

Design/Methods: This IRB-exempt study analyzed U.S. national data from CDC WONDER Birth and Linked Birth/Infant Death Period databases from 2016-2024. The study included only in-hospital births weighing 300-799g, with gestational ages (GA) of 21-24 weeks and no congenital anomalies. We developed an ACPI, defined as receiving one or more of five neonatal interventions: assisted ventilation, ventilation >6 hours, surfactant therapy, antibiotics for suspected neonatal sepsis, or NICU admission. Births and deaths were categorized as no ACPI or ≥ 1 ACPI. Annual birth and infant death rates were calculated per 1,000 live births for each respective GA. Linear regression estimated yearly trends, with β representing the slope (change per year).

Results: Across 21-24 GAs, births with ≥ 1 ACPI increased significantly (slope (β) = 21.4-191.8; p< 0.05)), while births without ACPI declined (β = -21.4 to -192.5; p < 0.05) (Fig.1). Births at 22 GA without ACPI have steadily decreased (β = -21.4; p < 0.001), while those with ≥ 1 ACPI have increased at a comparable rate (β = 21.4; p < 0.001). Infant death rates declined among infants without ACPI at 22 and 23 GAs (β = -0.09 and -0.05; p< 0.05) and among 22 GA with ≥ 1 ACPI (β = -0.06; p=0.001) (Fig. 2). Deaths with ≥1 ACPI significantly declined only among 22 GA who died within 24 hours (β = -0.05; p< 0.001) (Table 1). The national ACPI, derived from the Linked Birth/Infant Death Period database, rose steadily from 2017-2023 across all GAs, particularly among 22 GA. (Fig. 3).

Conclusion(s): From 2016 to 2024, active interventions for births at 21-24 GAs increased significantly, accompanied by declining infant death rates, particularly at 22-23 GAs. Notably, first-day mortality declined only among 22 GA receiving ≥1 ACPI, suggesting that expanded perinatal interventions may have driven recent improvements in early survival. Continued improvement of early perinatal care strategies may further improve outcomes for infants born at the threshold of viability.

Figure 1. Trends in Birth Rates per 1,000 Live Births by Gestational Age and Active Care Index (ACI) Group, 2016 -2024
Annual birth rates per 1,000 live births were calculated for infants born at 21-24 weeks' gestation, stratified by ACPI category: no ACPI (blue squares), and ≥1 ACPI (red circles). Linear regression models estimated the annual change (β) with corresponding p values. Positive β values indicate increasing rates over time, and negative β values indicate decreasing rates. Across gestational ages, births with ≥1 ACPI generally showed increasing trends, whereas births with no ACPI showed decreasing trends.



Figure 2. Trends in Infant Death Rates per 1,000 Live Births by Gestational Age and Active Care Index (ACI) Group, 2017-2023
Annual death rates per 1,000 live births were calculated for infants born at 21-24 weeks' gestation, stratified by ACPI category: no ACPI (blue squares), and ≥1 ACPI (red circles). Linear regression models estimated the annual change (β) with corresponding p-values. Positive β values indicate increasing mortality over time, and negative β values indicate decreasing mortality. Significant declines were observed among 22-23 weeks infants without ACI (β = -0.09 and -0.05; p < 0.05) and among 22 weeks infants with at least 1 ACI (β = -0.06; p = 0.001).



Table 1. Trends of Age of Death with At least 1 ACI, 2017 - 2023 and Figure 3. Temporal Trends in the Active Care Prematurity Index (ACPI) for Infants Born at 21-24 Weeks' Gestation, 2017 -2023 from the Linked Birth/Infant Death Period Database
Table 1 - For each gestational age group, death rates (per 1,000 live births) were categorized by age at death: <24 hours, 1-27 days, and 28-364 days. Annual values represent the corresponding rates for each year. The column labeled β shows the slope from a linear trend model across the study years, and the p-values indicate the statistical significance of these trends. Values in red denote statistically significant changes (p < 0.05). Dots (.) indicate suppressed or unavailable data. A statistically significant decreasing trend was observed only for deaths occurring within <24 hours among infants born at 22 weeks' gestation (β = -0.05, p <0.001).
Figure 3 - The Active Care Index (ACPI) was derived from five Selected Conditions of Newborn indicators-assisted ventilation, ventilation >6 hours, antibiotic use for suspected sepsis, NICU admission, and surfactant therapy. For each gestational week (21-24), condition-specific rates were computed as Yes / (Yes + No), z-standardized, and averaged to form the ACPI (range 0-100, higher = more active care). Temporal trends were assessed by linear regression of ACPI on year, weighted by known denominators. Across all gestational weeks, ACPI increased substantially from 2017 to 2023, reflecting progressive adoption of active perinatal interventions, particularly among infants born at 22 weeks' gestation who died (β= 0.06, p <0.001).