P2C375: Documented Vitamin K Administration and Reported Neonatal Bleeding: A Report from the HCA Neonatal Registry

Background: Prophylactic administration of intramuscular Vitamin K (VitK) is recommended by the AAP for all newborns for the prevention of Vitamin K deficiency bleeding (VKDB). Recent experience suggests the rates of VitK refusal are increasing and it is unclear if there is a concurrent rise in VKDB. Thus, we sought to determine in a large community based cohort, the rate of VitK administration and to evaluate the association of VitK administration with bleeding events.

Methods: This is a retrospective, observational cohort study, utilizing our Neonatal Registry curated from 116 HCA Healthcare-affiliated hospitals who deliver newborns at all levels of neonatal care. The cohort consists of term (37 weeks + 0 days to 41 weeks + 6 days) infants inborn between January 1, 2017 and October 14, 2022, excluding those with any Vermont Oxford Network Major Malformation Diagnosis, and those with a DNR ordered within the first 2 hours after birth. Endpoints include VKDB defined by ICD-10 codes P53 and E56.1, procedures indicating VKDB, duration of hospitalization, non-traumatic bleeding and mortality within the first 28 days after birth.

Results: We identified 998,254 neonates, of which, 49% were male with 56% White, 30% Hispanic/Latino and 15% Black/African American mothers. VitK administration was documented in 98.5% with 100% documented as IM injection. No notable differences were identified in the plurality of gestation, gender, birth weight (median 3314 vs 3330 grams), gestational age (median 39 vs 39 weeks) or duration of hospital stay (Median 49 vs 52 hours) between those with and without documented VitK. However, maternal Black/African American race was over represented in the no documented VitK group (15% vs 20%). Importantly, VKDB was underreported at 0.005%. Finally, the rate of procedures indicating VKDB within 28 days after birth were significantly elevated in the no documented VitK group (0.2% vs 1.15%, odds ratio (OR) 5.8, p< 0.0001) as well as death (0.017% vs 0.80%, OR 47.6 p< 0.0001). Of note, under representation of VKDB procedures may exist in the no VitK group due to missingness of VitK administration timestamps, and limiting delineation between VitK treatment and prophylaxis. Additionally, our univariable analysis do not address confounding, so multivariable analysis are likely to reduce the effect size.

Conclusion: Although lack of documented VitK administration was rare (1.5%), it was more common in the Black/African American community and correlated with poor neonatal outcomes including need for VKDB- related procedures and mortality. Our data also suggests that VKDB is substantially under reported. Work remains to improve identification and reporting of VKDB, to identify populations most at risk for VKDB, and identify populations who may benefit from parental education to reduce refusal rates and improve neonatal outcomes in the community setting.