Why newborns are given a vitamin K shot, and what the evidence shows

The vitamin K injection given to newborns prevents vitamin K deficiency bleeding (VKDB), a disorder that is rare but frequently disabling or fatal when it strikes, often as bleeding into the brain in an otherwise healthy-looking baby. Pooled surveillance data put the burden of late-onset VKDB at a median of 35 per 100,000 live births in infants who received no prophylaxis, and a single intramuscular dose at birth reduces that risk to near zero [s2]. It is one of the oldest routine measures in newborn care: the American Academy of Pediatrics has recommended intramuscular vitamin K since 1961 [s1].

This piece reports the evidence and the guidance; it is not medical advice, and because it concerns dosing and route of administration its risk tier is held. Questions about a specific infant belong with a clinician.

Why newborns are uniquely at risk

Babies are born with very little vitamin K, which the body needs to make several clotting factors. Little crosses the placenta, breast milk contains modest amounts, and the newborn gut has not yet built the bacterial population that contributes to vitamin K status. That combination leaves an infant transiently unable to clot normally — the gap that prophylaxis fills [s1].

VKDB is usually divided by timing. The most feared form is late VKDB, appearing roughly between two weeks and six months, because it presents disproportionately as intracranial haemorrhage. The systematic review pooling surveillance studies found the median burden of late VKDB in unprotected infants was 35 per 100,000 live births, with a wide interquartile range of 10.5 to 80 [s2].

A burden concentrated by geography

That average hides a stark split. In high-income countries the median late-VKDB rate in unprotected infants was 8.8 per 100,000 (interquartile range 5.8 to 17.8); in low- and middle-income countries it was 80 per 100,000 (interquartile range 72 to 80) — nearly an order of magnitude higher [s2]. The disparity reflects both underlying nutrition and access to prophylaxis, and it is a reminder that the intervention's value is largest exactly where it is hardest to deliver.

What the shot does, and how routes compare

The direct evidence that prophylaxis works comes mostly from surveillance rather than randomised trials, which the review is explicit about. Across four surveillance studies, intramuscular or subcutaneous vitamin K cut the risk of late VKDB dramatically compared with no prophylaxis, with a pooled relative risk of 0.02 (95% CI, 0.00 to 0.10) [s2]. For the earlier "classical" form, two randomised trials of intramuscular prophylaxis showed reductions in bleeding — in one, any bleeding was reduced (RR 0.73; 95% CI, 0.56 to 0.96) and moderate-to-severe bleeding fell by far more (RR 0.19; 95% CI, 0.08 to 0.46) [s2].

Route matters, and this is where the guidance is firm. A single oral dose is markedly inferior: compared with the intramuscular route, one oral dose was associated with a 24.5-fold higher risk of VKDB (RR 24.5; 95% CI, 7.4 to 81.0) [s2]. Multiple oral doses narrowed but did not clearly close the gap (RR 3.64; 95% CI, 0.82 to 16.3, not statistically significant) [s2]. On that basis the review concluded that routine intramuscular administration of 1 mg at birth is the appropriate default for all neonates [s2].

Why the AAP restated it

The AAP would not normally reissue guidance on a 60-year-old practice, but its 2022 clinical report exists for a specific reason: the incidence of VKDB "appears to be on the rise," and the report attributes that to two things — parental refusal of the injection, and the lower efficacy of the oral alternatives some families request instead [s1]. In other words, the resurgence is not a failure of the drug but a failure to give it, or to give it by the route that works.

That places vitamin K in the same category as other high-value newborn measures whose main threat is now uptake rather than efficacy, a pattern also visible in routine newborn screening and in the safe-sleep guidance whose death toll has stalled despite clear recommendations. The evidence base has limits worth stating plainly — much of it is observational, and the absolute risk to any single unprotected infant is low. But the direction and size of the effect are consistent, and the review's bottom line is that the balance strongly favours prophylaxis [s2].