A couple of recent papers on vitamin K have generated a flurry of media coverage.
Much of that coverage has focused on dramatic headlines suggesting that babies are being put at risk when parents decline vitamin K.
But, as is so often the case, the nuance of the underlying evidence has not always travelled well.
In this blog post, I will explain what this research does — and does not — tell us about vitamin K at birth.
The Sousa et al (2026) vitamin K paper
This ‘paper’ was an abstract, presented at the American Academy of Neurology annual meeting, titled “Rising Parental Refusal of Newborn Vitamin K: Implications for Neurology and Child Development.”
It is likely that, in time, this will be published in a medical journal, but the abstract has been shared as part of a press release.
Many of the media pieces also include quotes from practitioners, almost exclusively focusing on those who think that universal intramuscular (IM) vitamin K is the only valid approach, and that parents who decline this are irresponsible.
What does the Sousa et al research say about vitamin K at birth?
The abstract reports on a systematic review of 25 studies, carried out between 2005 and 2025.
The researchers looked at rates of parental refusal of intramuscular (IM) vitamin K, and outcomes associated with vitamin K deficiency bleeding (VKDB).
They have also discussed parental reasoning, and links between this decision and wider healthcare hesitancy. I want to note, however, that what some clinicians call ‘healthcare hesitancy’, could also be described as, ‘making decisions about one’s own health, which might sometimes involve declining an intervention offered by a clinician or recommended by population-level guidelines.’
Understanding risk: relative vs absolute risk
One of the most striking claims highlighted in both the abstract and media coverage is that babies who do not receive intramuscular vitamin K are “81 times more likely” to develop late VKDB.
That sounds alarming.
But it is a relative risk, and giving this figure without sharing data on absolute risk is very misleading.
That’s because VKDB is rare.
How rare? Well, the most reliable estimate we have for late-onset VKDB is derived from a systematic review for the World Health Organization. Sankar et al (2016) found that the chance of a baby experiencing late-onset VKDB in a high-income country is 1 in 11,363.
There’s more on this in my book, where I discuss other studies, related issues, and also explain why some estimates of this are misleading.
Limitations of the evidence on vitamin K refusal
But that’s not the only issue with the information being share in this abstract and the related articles.
As with all research, it is important to look at the limitations as well as the findings.
This is a review of existing data, not new primary research.
The abstract synthesises existing studies rather than presenting new data. Its conclusions depend on the quality of the studies included. That isn’t a problem in itself, but it’s an important fact. The data aren’t new. And, when researchers don’t have new data to share, they may be encouraged to promote their work using tactics that will attract media coverage.
With or without input from researchers, the media has a tendency to over-emphasise risks, sensationalise issues, judge those making non-standard decisions, and present viewpoints which will upset and anger some people. This is how they attract the attention, comments, and arguments on social media which allow them to generate advertising revenue and thus monetise content.
The studies are very different from each other
The research spans multiple countries, healthcare systems, and birth settings. These differences matter, and combining them can obscure important context. For instance, I was clear to specify that the figure I shared above was from data in high-income countries. That’s because the risks are higher for babies in low- and middle-income countries. Lumping data together can be problematic for everyone, as it will skew the figures and can make the chance of a problem look greater or less than it actually is.
Some data are based on perception
Some of the reports which state that refusal of vitamin K is increasing are based on clinicians’ perceptions rather than robust longitudinal data.
As I have discussed elsewhere, some of the scarier-sounding data that has been shared on this topic was gathered by doctors working in a high-tech neonatal unit.
But, again, they did not put the data into the context of what happens in the wider population.
It’s inevitable that those working in high-tech units are going to see the babies with problems. Those units are where they go for care. What we need are accurate, evidence-based data showing the rates of VKDB in relation to the wider population. Again, context is vital.
There is a lack of discussion of alternatives
This paper (and many like it) focuses on the binary decision of IM vitamin K (via the locally recommended protocol) versus no vitamin K. But this doesn’t reflect reality. Some parents opt for oral vitamin K, ask for IM vitamin K via a different dosing regime, or want to consider other alternatives. These parents would consider that they gave their babies vitamin K, but the researchers who ran the studies in this review might not. As is so often the case, we need more nuance, and consideration of the grey areas.
Outcomes are often drawn from case series
It is common in papers like this one to focus on discussing death and neurological disability when VKDB occurs but, again, this data comes from individual reports and is often not put into context. It’s important to know that VKBD can be severe or fatal, yes, but when researchers and clinicians over-egg the likelihood of this and parents later find out the real figures, this is a key source of distrust in clinicians, as I will discuss further in the next point.
The framing focuses on ‘refusal’ and ‘risk’
Some parents and clinicians find that papers like this one are disrespectful, because they make it clear that, in their opinion, there is only one ‘reasonable’ or ‘sensible’ option — doing what the doctors tell you — and show little understanding of or respect for parents who might have concerns or want to make different decisions.
Concerns are dismissed as being ill-informed or ignorant and, while it is true that there is plenty of misinformation about vitamin K out there, many of the parents who make alternative decisions on this are well informed. And, in a world where misinformation, populism, and polarisation are rife and weaponised, people deserve to be supported rather than shamed, no matter what their position or decision.
There is a lack of information and insight
Papers on this topic, including this one, often contain little exploration on how parents are given information, how decisions are made, or how consent is sought. These are important pieces of the wider picture. Some of the parents who decline vitamin K do so as a result of feeling so patronised, coerced, or bullied by clinicians who, as above, are convinced that their viewpoint is the only valid one.
Ironically, in some cases, these clinicians are causing or exacerbating the very problem that they are seeking to solve.
The Simatou et al (2026) vitamin K paper
In July 2026, a new study on vitamin K hit the headlines, with journalists citing how increased rates of refusal have led to more bleeding.
While it is inevitable that cases will increase alongside any increase in parents declining vitamin K, because vitamin K is a very effective form of prevention, parents need to know more than they are being told.
As I explained above, we need to see not just the relative risk — as in the scary headline that babies are two or three times more likely to bleed — but also the absolute chance of a healthy baby having a bleed.
Again, with vitamin K, this figure is very low. Unfortunately, this fact isn’t mentioned in the headlines and that data aren’t discussed. Even more worryingly in this paper, the measures that have been used to gather the figures are also questionable.
Looking beyond the headlines
Here’s what a closer look at this paper reveals
The absolute risk is tiny, in both groups
The study followed over 2 million Swedish infants. Among those with no record of intramuscular vitamin K, the rate of any bleeding diagnosis in the first six months was 636 per 100,000 (roughly 1 in 157). Among those who received it, the rate was 380 per 100,000 (roughly 1 in 264).
Framed as a headline, that’s “more than 50% higher risk.”
Framed honestly, it’s a difference of about a quarter of one percent.
For intracranial bleeding specifically, the figures are 220 per 100,000 (about 1 in 454) versus 63 per 100,000 (about 1 in 1,577); an NNT of around 637.
None of these numbers appear in the abstract, the plain-language “Key Points” box, or any of the news coverage that I saw.
But don’t get too attached to those numbers, because there are other issues which mean these figures aren’t accurate.
“Bleeding” in this study is not the same thing as VKDB
There is a significant misclassification issue in this paper, and it’s worth spelling out properly.
The primary outcome (or what the researchers were measuring) wasn’t confirmed vitamin K deficiency bleeding (VKDB).
It was any of nine different ICD-10 diagnostic codes. Of those nine, only two (P53, “haemorrhagic disease of newborn,” and E56.1, “vitamin K deficiency”) are specifically about vitamin K status.
The remaining seven codes cover things like unspecified gastrointestinal bleeding, bleeding from the respiratory tract, soft tissue bruising, and non-traumatic intracranial haemorrhage. The latter is a code that can be used for a difficult or instrumental birth as well as a clotting problem.
As I have also discussed in several of my books, medical records also aren’t as accurate as we might like.
The authors acknowledge this themselves, right at the end of the paper, where they admit that their outcome “captures all atraumatic bleedings regardless of aetiology” and was never validated against actual clinical VKDB criteria.
This is very likely why their “1 in 454” figure for intracranial bleeding is so much higher than the figures for confirmed VKDB reported elsewhere (around 1 in 11,000, as above). They are measuring a much broader and less specific category of event, not the condition vitamin K prevents.
That’s why, even if we calculate the absolute rather than relative risk from this paper, the numbers don’t tell us anything helpful.
The authors pre-empt this criticism, but don’t really answer it
To their credit, the authors acknowledge that, “absolute rates are not directly comparable with previous studies,” but argue that “the relative differences between groups remain valid for evaluating the protective effect of intramuscular vitamin K prophylaxis.”
That statement about validity may be true when we are measuring something like an effect size in a research paper.
However, I would argue that it is a different question from what parents actually need in order to weigh a real-world decision for their own baby.
Especially when it is headline-seeking journalists, and not research experts, who are being given just the abstract, and left to communicate the findings in the real world.
The two groups weren’t really comparable
Infants with no recorded vitamin K also had dramatically higher rates of missing data across the board. For instance, Apgar scores were missing 20 times more often (8% vs 0.4%), and mode of delivery and birth weight data were similarly lacking.
This suggests that “no record of vitamin K” is at least partly a marker of babies born outside standard hospital pathways, rather than a group of otherwise-identical babies whose parents simply declined an injection. A lack of documentation often occurs where babies are born not after a planned home birth, but in a situation where labour is fast and everyone is scrambling to help.
Tellingly, 64% of home-born infants in the same dataset had no documented vitamin K. The authors admit they couldn’t draw conclusions about bleeding risk in that group due to small numbers, and yet home birth is clearly overrepresented in the very statistic that is making headlines. Some journalists have specifically mentioned home birth as a risk factor, and yet the authors themselves acknowledge there aren’t enough data to support this.
The “no vitamin K” group also had more emergency caesareans and, where recorded, lower Apgar scores. This is important, because both of these are plausible contributors to intracranial bleeding via a harder birth, independent of whether or not the baby was given vitamin K.
(Please note: this is not support for the ‘traumatic birth increases the need for vitamin K” myth which i debunk here).
Other issues with the paper
These aren’t the only issues with this paper.
There are also questions about the timeframe in which data are gathered, for example. I have focused on writing something quickly about the key issues, to respond to those who have been asking about it.
Pip Wynn Owen has also looked at this paper, and you can find her analysis here.
The paper’s own conclusions lean into a “compliance” framing
I mentioned above that the tone of these papers can be problematic.
In this one, the discussion moves fairly quickly from the data itself to recommending “continued education” of parents and outreach to “address parental refusal”.
Such language frames informed parental decision-making primarily as a problem to be corrected, rather than acknowledging that families weighing this decision are entitled to a fuller, more honest picture which discusses absolute risk and the reasons that a study might not be the best source of data on this.
As above, that’s not respectful, and it’s not aligned with the idea of informed decision making.
Vitamin K, VKDB, and what parents need to know
None of this is to suggest that VKDB is unimportant.
It is a very real condition, and when it occurs it can have serious consequences.
Parents deserve clear, accurate and balanced information about what VKDB is, how vitamin K can reduce the chance of a baby experiencing this, and the pros and cons of the different approaches.
This includes explanation of both the relative and absolute risk, as well as the uncertainties that remain.
Providing this kind of information supports informed decision making.
Relying on fear-based messaging does not.
Supporting informed choice rather than fear
Parents make decisions within their own values, beliefs and circumstances. Trust plays a central role in this process.
When clinicians respond to decisions they do not agree with by using alarmist language, implying irresponsibility, and attempting to pressure compliance, that trust can be damaged.
Without trust, meaningful communication becomes much more difficult, and the problem becomes worse.
What if parents decide to decline vitamin K?
Some parents, having considered the available information, will decide to decline IM vitamin K for their baby. At that point, the question becomes: how do we best support safety?
Rather than using bullying or shaming tactics, I suggest that it would be better if clinicians focused on ensuring that parents have all of the information the need, and then share information on the additional information needed by the parents of babies who have not received vitamin K.
I have a separate blog post on this:
More resources on vitamin K at birth
I have written several other blog posts on this topic, and I have a resources page which lists these. I also wrote a popular book, Vitamin K and the Newborn, which helps parents understand the issues and the evidence, so that they can make the decisions that are right for them.
About the Author: Dr Sara Wickham is an author, speaker, and researcher specialising in pregnancy, birth and maternity care. Her work focuses on evidence-based, woman-centred information and informed decision-making, drawing on more than 30 years of midwifery knowledge and experience.
Want to stay up to date and get regular updates on birth-related research and thinking sent free to your inbox?
Plus news of Sara’s books, projects, events, and courses?
Sign up to our email newsletter list here
We’ll send you regular, friendly updates to help you keep up-to-date and stay informed.
Books to help you become informed and make the decisions that are right for you…
If this blog post or any of Sara’s work has helped you or those you care for, please buy her a coffee and help us continue to provide independent birth information online as well as in her books.
Thank you.