Racial and ethnic disparities in postnatal growth among very low birth weight infants in California

Postnatal growth is a key concern for preterm infants due to its association with brain growth and neurodevelopment. While survival of preterm infants has improved, variability in postnatal growth remains, influenced by neonatal morbidities, nutrition, and small for gestational age (SGA) status. Prior quality improvement efforts have reduced growth failure at discharge, yet disparities by race/ethnicity in fetal growth restraint, postnatal growth, and nutrition practices have been documented. The study aims to investigate racial/ethnic differences in postnatal growth failure (PGF) among VLBW infants in California, with attention to temporal trends and gestational age strata.

Multiple studies report variability in defining and measuring postnatal growth failure among preterm infants. In the 1990s, over 95% of VLBW infants were discharged below the 10th percentile for weight-for-age. Quality improvement initiatives (e.g., Vermont Oxford Network) improved discharge growth; PGF declined substantially from 2000 to 2013, and CPQCC data showed decreasing falls in weight Z-score between 2005 and 2012. CPQCC incorporates growth velocity into the Baby-MONITOR NICU quality composite. Disparities in fetal growth and nutrition practices exist by race/ethnicity: higher SGA risk among Black infants versus white; SGA more likely with Black or Asian parents; some evidence of higher weight gain velocity in Black preterm infants with comparable body composition; decelerated growth associated with Asian maternal race; and elevated low birth weight/SGA among Asian Indian infants compared with white. Breast milk feeding at discharge varies by race/ethnicity, with prior CPQCC work showing higher human milk use among Hispanic infants, particularly with foreign-born mothers. These findings underscore the need for granular, race/ethnicity-specific growth data to inform targeted interventions.

Design and data sources: Retrospective cohort using the California Perinatal Quality Care Collaborative (CPQCC) database from 132 NICUs (covering >95% of California VLBW care) linked with California’s Department of Health Care Access and Information for maternal demographics (race/ethnicity, education, payer). Inclusion: 37,122 infants born 2008–2016 with birth weight 500–1500 g or gestational age 23–34 weeks, discharged alive before 50 weeks postmenstrual age (PMA). Exclusions: severe congenital anomalies; missing birth/discharge weight or sex; implausible birth/discharge weight Z-scores (>5 SD from mean). A predefined subgroup included 20,019 infants with birth weight <1000 g or gestational age 23–28 weeks. Outcomes and definitions: Weight Z-scores at birth and discharge were calculated using Fenton growth charts (standardized to PMA). SGA was defined as weight Z-score < −1.28 (10th percentile) at birth or discharge. Primary outcome PGF was defined as change in weight Z-score from birth to discharge < −1.28 (per Fenton). Exposures and covariates: Maternal race/ethnicity (White, Black, Asian American/Pacific Islander [Asian/PI], Hispanic, Other/Unknown; ethnicity categorized as Hispanic or non-Hispanic; Hispanic treated as its own category), infant sex, gravidity, maternal education, maternal age, insurance, SGA at birth, maternal hypertension, chorioamnionitis, chronic lung disease (oxygen at 36 weeks PMA), necrotizing enterocolitis (stage ≥2), gastrointestinal perforation, late-onset sepsis, severe IVH (grade 3–4), severe ROP (stage ≥3 or requiring treatment), and feeding at discharge (formula only vs any human milk). Statistical analysis: Bivariate associations by race/ethnicity via one-way ANOVA and chi-square tests. Linear regression assessed changes in PGF incidence over time and by gestational age, stratified by race/ethnicity. Multivariable logistic regression with hospital as a random effect estimated adjusted odds of PGF, including prespecified covariates; length of stay was excluded due to collinearity with morbidities. Significance at P < 0.05. Analyses conducted in SAS 9.4.

• Cohort: 37,122 infants (47% Hispanic, 26% white, 13% Black, 11% Asian/PI, 3% other); subgroup <1000 g or 23–28 weeks: 20,019 infants (49% Hispanic, 25% white, 13% Black, 10% Asian/PI, 3% other).
• SGA patterns: SGA at birth and discharge were highest among Asian/PI infants and lower among Hispanic infants.
• PGF incidence and trends: Overall PGF decreased from 27.4% (2008) to 22.8% (2016) (also reported as 27% to 23% in-text); incidence decreased with increasing gestational age from 52% at 23 weeks to 3% at 34 weeks (P < 0.0001). Each additional gestational week was associated with lower PGF risk (OR 0.73; 95% CI 0.72–0.74; P < 0.0001).
• Racial/ethnic disparities: • Hispanic infants had the highest PGF (approximately 30%), higher than white (24%), Black (22%), and Asian/PI (23%). In annual trend models, infants of Hispanic mothers had higher PGF risk than white (OR 1.33; 95% CI 1.25–1.41), Black (OR 1.50; 95% CI 1.39–1.62), and Asian/PI (OR 1.38; 95% CI 1.28–1.49) infants. • In multivariable logistic regression (reference = white): Black vs white had lower odds of PGF (adj OR 0.71; 95% CI 0.60–0.84), Hispanic vs white had higher odds (adj OR 1.13; 95% CI 1.00–1.27), Asian/PI vs white not significantly different (adj OR 0.91; 95% CI 0.76–1.09). Subgroup <1000 g/23–28 weeks showed similar patterns (Black vs white adj OR 0.66; 95% CI 0.54–0.80; Hispanic vs white adj OR 1.08; 95% CI 0.93–1.26; Asian/PI vs white adj OR 1.04; 95% CI 0.83–1.30).
• Additional independent associations with PGF (adj ORs for 23–34 weeks or <1500 g): • Male vs female: 0.72 (0.69–0.84) (lower odds) • Maternal hypertension: 0.56 (0.50–0.63) • Formula feeding at discharge: 0.81 (0.71–0.93) • Respiratory distress syndrome: 1.74 (1.54–1.97) • Necrotizing enterocolitis: 2.57 (2.35–2.80) • Gastrointestinal perforation: 2.62 (2.36–2.91) • Patent ductus arteriosus (PDA): 2.06 (1.86–2.37) • Late-onset sepsis: 1.61 (1.37–1.90) • Severe IVH: 1.31 (1.09–1.63) • SGA at birth: 1.19 (1.03–1.38)
• Temporal improvements were observed across races with notable gains among certain gestational age strata (e.g., around 29 weeks), though racial differences in PGF persisted.
• Subgroup <1000 g/23–28 weeks: similar racial patterns; some attenuation of PGF incidence among Hispanic infants relative to the full cohort.

The study demonstrates that racial/ethnic disparities exist in postnatal growth among VLBW infants, with Hispanic infants exhibiting the highest incidence of PGF despite lower SGA prevalence at birth compared with Asian/PI infants. The findings highlight that growth trajectories depend on both intrauterine growth status (SGA) and postnatal course. PGF decreased over time statewide and with increasing gestational age, indicating improvement in care processes and nutrition practices; however, disparities persisted, suggesting that system-level gains did not uniformly benefit all groups. The multivariable analysis implicates both clinical morbidities (NEC, GI perforation, PDA, sepsis, RDS, severe IVH) and perinatal factors (SGA at birth) as key drivers of PGF risk. Formula feeding at discharge was associated with lower odds of PGF, consistent with higher short-term weight gain relative to human milk feeding, though human milk confers other important benefits. These results underscore the need for targeted nutritional strategies, close growth monitoring, and efforts to reduce morbidities, with particular attention to infants of Hispanic mothers to mitigate inequities in short- and long-term outcomes.

Racial and ethnic disparities in early postnatal growth persist among VLBW infants in California. PGF declined from 2008 to 2016 and decreased with advancing gestational age, yet Hispanic infants remained at highest risk. Interventions focused on optimizing nutrition, preventing morbidities, and monitoring growth—especially for high-risk racial/ethnic groups—are warranted. Future work should evaluate targeted, culturally informed nutrition strategies, track longitudinal anthropometrics beyond discharge (including length and head circumference), and assess how unit-level quality improvement efforts can reduce disparities.

Key limitations include lack of quantitative nutritional intake data; assessment of growth only at birth and discharge without longitudinal in-hospital trajectories or length/head circumference; feeding type captured only at discharge; absence of maternal smoking, obesity, and BMI data; and potential selection/survivor bias due to exclusion of infants who died before discharge or were discharged after 50–51 weeks PMA. Observational design limits causal inference due to potential unmeasured confounding.