Trends in vitamin K administration among infants

Introduction

Routine newborn intramuscular vitamin K prophylaxis has nearly eliminated vitamin K deficiency bleeding in the US since initiation of universal administration in 1961.¹ Nonetheless, there are reports of increasing parental vitamin K refusal and resultant bleeding.² To mitigate risk of such bleeding, newborns require prophylaxis at birth because of minimal vitamin K placental transfer, poor gastrointestinal absorption, and low concentrations in breast milk. Parents sometimes refuse newborn vitamin K administration because of concerns regarding necessity, injection pain, and potential adverse effects.^1–3 Public skepticism regarding pediatric preventive interventions has grown after the COVID-19 pandemic.⁴ Whether national rates of intramuscular vitamin K administration have changed in this period is unknown. This study aimed to evaluate whether the proportion of newborns not receiving intramuscular vitamin K has increased in recent years and identify factors associated with non-receipt.

Methods

We performed a retrospective cohort study of newborns in Epic Systems’ Cosmos research platform (January 2017- December 2024), a US-based electronic health record database with anonymized, deidentified, patient-level data. Hospitals with at least 10 births per year in every year of the study period and newborns 35–43 weeks’ gestation were included; 403 hospitals across all 50 states and the District of Columbia contributed data. After calculating the proportion of newborns not receiving intramuscular vitamin K, we examined associated maternal and newborn characteristics, as well as changes in characteristics over time, with a standardized difference >0.1 considered significant (Supplement). Race and ethnicity data were obtained from electronic medical records; prior studies have reported differences in vitamin K administration by race.⁵ Adjusted percentages of newborns not receiving vitamin K over time were calculated using a logistic regression model adjusted for covariates associated with vitamin K administration in bivariate analysis (Supplement). The STROBE reporting guideline for cohort studies was followed. The Children’s Hospital of Philadelphia Institutional Review Board deemed this study non-human subjects research per 45 CSF 46.104.

Results

Of 5,096,633 newborns, 199,571 (3.92%) did not receive intramuscular vitamin K, which increased from 2.92% (15,354/510,987) in 2017 to 5.18% (37,187/680,215) in 2024 (unadjusted trend p <0.001, Figure 1) (Table 1). No maternal or infant characteristics changed significantly over time. Race and ethnicity, with highest rates among other/unknown and non-Hispanic White populations, and vaginal birth were associated with not receiving vitamin K (Table 1). Multivariable adjusted percentages of newborns not receiving vitamin K also increased throughout the study period from 2.57% (95% CI: 2.52%, 2.63%) in 2017 to 4.62% (95% CI: 4.55%, 4.70%) in 2024.

Figure 1. Percentage of newborns not receiving prophylactic intramuscular vitamin K administration after birth.

Overall, 199,571/ 5,096,633 (3.92%) did not receive vitamin K. The results demonstrated a linear trend via chi-square test (P<0.001)

Table 1.

Characteristics among infants 35 – 43 weeks of gestation who did and did not receive routine, prophylactic, intramuscular vitamin K in Epic Cosmos, January 2017- December 2024 (n= 5,096,633)

	Vitamin K Given		No Vitamin K
	n=4,897,062(96.1%)		n=199,571(3.92%)
Maternal Characteristics	n	(row %)	n	(row %)	Standardized Difference
Age (years)
<20	152,543	(96.7)	5,137	(3.3)	0.05
20-<35	3,681,004	(96.1)	147,474	(3.9)
35+	1,063,502	(95.8)	46,958	(4.2)
Race and ethnicity					0.14a
American Indian or Alaska Native	50,528	(96.3)	1,936	(3.7)
Asian	300,536	(96.3)	11,469	(3.7)
Hispanic	897,989	(96.7)	30,484	(3.3)
Native Hawaiian or Other Pacific Islander	29,753	(97.4)	799	(2.6)
Non-Hispanic Black	835,719	(96.6)	28,977	(3.4)
Non-Hispanic White	2,641,985	(95.7)	117,604	(4.3)
Other/Unknownb	140,552	(94.4)	8,302	(5.6)
Infant Insurance					0.03
Public	1,940,239	(96.2)	75,762	(3.8)
Otherc	2,956,823	(96.0)	123,809	(4.0)
SVI d					0.04
High Risk	1,617,137	(96.4)	61,238	(3.6)
Low Risk	3,279,925	(96.0)	138,333	(4.0)
Residence e					0.05
Urban	4,246,452	(96.0)	174,652	(4.0)
Rural	577,904	(96.2)	23,046	(3.8)
Mode of birth					0.14a
Vaginal	3,352,096	(95.8)	148,589	(4.2)
Cesarean	1,544,966	(96.8)	50,982	(3.2)
Infant Characteristics
Birth weight (g), mean (SD)	3,312 (492)		3,342 (506)		0.06
Sex					0.04
Female	2,396,209	(95.9)	102,066	(4.1)
Male	2,500,853	(96.2)	97,505	(3.8)

SD, standard deviation;

^aDissimilarities were quantified between groups using standardized differences. A standardized difference of >0.1 indicates substantial imbalance between groups (see Supplement 1 for details).

^bOther includes missing or unspecified race and ethnicity

^cOther includes private, Tricare, and unspecified insurance.

^dSVI, Centers for Disease Control and Prevention Social Vulnerability Index at the census tract level, high and low risk SVI are the top and bottom three quartiles of SVI values across the cohort, respectively;

^eResidence is determined using maternal ZIP code which then is grouped into urban or rural using rural-urban commuting area (RUCA) codes (1–3 considered urban and 4–10 considered rural), which were missing for 0.01% of patients.

Discussion

The proportion of newborns not receiving intramuscular vitamin K increased from 2.92% to 5.18% in a large US-based electronic health record dataset from 2017 to 2024. These rates are consistent with a 2016 North Carolina study of over 18,000 newborns, but higher than in a 2018 study of 102,878 newborns across 34 states (0.6%).^3,5 Changes in vitamin K administration rates may not have resulted solely from COVID-19–related public skepticism of pediatric medical interventions, as the increase in infants not receiving vitamin K subtly began prior to the pandemic’s 2020 onset. While refusal is likely the major contributor to not receiving vitamin K,⁶ this study lacked data from parents or clinicians on the reasons for non-receipt. These findings are only generalizable to infants ≥35 weeks’ gestation. Vitamin K administration in out-of-hospital birth was not captured in this analysis; thus, the observed rates of non-receipt may be an underestimate since vitamin K refusal rates are higher for planned home births.¹ Further study is warranted to evaluate whether these trends are associated with increased risk of major bleeding such as intracranial hemorrhage. A multipronged approach comprised of interventions including public health regulation at the state-level and standardized practice regarding clinician communication with patients regarding vitamin K refusal at the hospital-level is urgently needed to improve rates of highly effective prophylactic intramuscular vitamin K administration to prevent bleeding and its associated morbidity and mortality.