The record is a study published December 8, 2025, in JAMA. Kristan Scott and colleagues at the Children’s Hospital of Philadelphia pulled the delivery records of 5,096,633 newborns across 403 hospitals in all fifty states and found that the share of parents declining the routine vitamin K injection rose from 2.92 percent in 2017 to 5.18 percent in 2024. Adjust for demographics and the curve holds, 2.57 percent to 4.62 percent. The refusal roughly doubled in seven years, climbing from 2019 and accelerating through the pandemic and past it. What those parents are turning down is a single shot that all but erases a bleed which, when it reaches the brain, kills roughly one in five of the infants it strikes.
The shot is phytonadione, given once, into the thigh, within hours of birth. It prevents vitamin K deficiency bleeding, or VKDB, and above all the late-onset form that arrives without warning between roughly two weeks and six months of age. Newborns are born with almost no vitamin K, breast milk carries little of it, and an infant’s gut has not yet learned to make its own, so the clotting system runs on empty until it does. In the late-onset presentation, more than half of the infants who bleed bleed into the brain. By the CDC’s count, as reported by ProPublica, roughly one in five of the babies who develop VKDB die, and an infant who does not get the shot is 81 times more likely to develop the late form than one who does. The injection does not merely reduce that risk. It very nearly erases it.
So the question the documents force is not whether the shot works. It is why a cheap, decades-old intervention with that kind of record is being turned away in a hospital hallway, more every year.
Follow the fear back to its source and it terminates in a single paper. In 1992, the British epidemiologist Jean Golding and colleagues reported in the BMJ a statistical association between intramuscular vitamin K and later childhood cancer, a leukaemia risk they put at roughly 2.65-fold. Intramuscular, not oral. The signal was serious enough to take seriously, and it was taken seriously. It was also tested, repeatedly, by people who went looking for it and did not find it. A population-based case-control study in Lower Saxony found no association. A pooled analysis of six case-control datasets found no convincing evidence linking the injection to childhood leukaemia. The United Kingdom Childhood Cancer Study, the large national effort built in part to settle the matter, concluded there was no convincing evidence that neonatal vitamin K influences a child’s risk of leukaemia or any other cancer, by any route of administration. The 2021 review in BMC Pediatrics records the arc plainly: an early alarm, a body of later work, and no confirmed link. The scare was answered before most of today’s new parents finished school. The answer never traveled the way the alarm did.
The cheapest thing in the nursery
The accountability question splits here, and both sides carry receipts. One of them cuts against the loudest voices in the refusal movement. Search the parent forums and you find the shot described, in one Facebook comment collected by ProPublica, as “poisons from big pharma.” That framing has a problem it never mentions. Phytonadione is generic and off-patent, one of the cheapest things on the labor-and-delivery floor. There is no patent to defend, no marketing budget, no sales force working a drug that cheap. It is among the least profitable objects in the building. When an influencer sells you outrage about a product its maker barely earns a margin on, the incentive worth tracing is not the drugmaker’s. It is the seller’s. Candace Owens told her audience that the recommendation amounts to telling parents their babies “were born wrong,” a line that moves subscriptions and does nothing for a clotting factor. The outrage has a business model. The injection does not.
The other side is harder, and it does not let the institutions off. Parents are not refusing in a vacuum. They are refusing after five years in which public-health authority spent its credibility faster than it has begun to earn back: oversold certainty on COVID, reasonable questions filed under “misinformation” before the questions were answered, guidance that shifted and was presented each time as settled. A parent who watched that does not sort interventions by patent status. They sort by whether the people recommending it have been straight with them before. And the reflexive institutional reply to vitamin K refusal, that it is simply misinformation, is the same move that taught these parents to distrust the messenger in the first place. As the journalist Maryanne Demasi reports, many of these parents are not chanting slogans. They are asking what specific product is going into their child, what the alternatives are, and why national recommendations differ across borders.
Those are answerable questions, and the honest answer persuades where the label does not. Some countries do use an oral vitamin K regimen instead of the shot. The oral route requires multiple doses over weeks, depends on a parent completing the schedule, and is documented to fail more often at preventing late VKDB, particularly in infants with undiagnosed cholestatic liver conditions who cannot absorb it. That is the real reason the single intramuscular dose became the standard: not because a company willed it, but because it does not depend on anyone remembering the third dose. A parent who gets that explanation is being treated as a citizen. A parent who gets “misinformation” is being managed. The institutions keep choosing the second, then act surprised at the refusal rate.
The cost of that choice is not abstract, and it is not evenly borne. It lands on infants who had no vote. ProPublica’s reporting on hospitals watching the trend puts it in flat numbers: at the Mercy system across four states, 1,552 babies went without the shot last year against 536 in 2021; at Idaho’s St. Luke’s, refusal ran from 3.8 percent in 2020 to 9.8 percent in 2025, with one hospital at 20 percent. And it documents where that curve ends for a small number of families: a seven-week-old in Maryland with sudden seizures, an infant in Alabama who stopped breathing, one in Kentucky, one in Texas not yet two weeks old bleeding at the navel. Each is a preventable death from a condition a cheap injection would have stopped.
Nobody is counting
There is, at present, no national system that counts these cases. The United States does not systematically track vitamin K refusals, VKDB cases, or the deaths that follow. In July 2026, Senator Angela Alsobrooks and Representative Kim Schrier wrote to the Acting Director of the CDC asking the agency to begin doing exactly that. As of this writing that letter has produced no tracking program, and the agency that would have to build one has not announced whether it will. The evidence on the shot has been settled for twenty years. What is unsettled is whether the institutions that lost these parents’ trust intend to earn it back with straight answers, or to keep filing the refusals under “misinformation” while the count they will not keep goes up. The receipts on vitamin K are in. The open question is who reads them out loud, to whom, and whether anyone in a position to act is keeping score.
Sources
- Children’s Hospital of Philadelphia, summary of Scott et al., JAMA, “Trends in newborn vitamin K non-administration, 2017–2024” (Dec 8, 2025)
- ProPublica, “Parents Increasingly Reject Vitamin K Shots for Newborns” (hospital refusal figures, CDC mortality data, VKDB deaths, refusal-message sourcing, Owens quote)
- BMC Pediatrics (2021), “Vitamin K prophylaxis in newborns” (VKDB incidence, oral vs. intramuscular regimens, the childhood-cancer concern and its refutation)
- UK Childhood Cancer Study, British Journal of Cancer (no convincing evidence that neonatal vitamin K, by any route, affects childhood cancer risk)
- Maryanne Demasi, “Why are more parents refusing vitamin K for their newborns?” (parents’ stated questions; Alsobrooks–Schrier letter to CDC, July 2026)