Baby-Led Weaning: What the Trial Actually Found
The two fears about baby-led weaning are iron and choking. A randomised controlled trial tested both, and the results are more specific than either camp claims.
- Writer, standards and the recall record
- Feb 9, 2026
- Updated Jul 9, 2026
- 8 min read
In this guide
Baby-led weaning is the approach where an infant self-feeds appropriately prepared whole foods from around six months, instead of being spoon-fed purées. It has two persistent objections: that babies who feed themselves will not get enough iron, and that they will choke more.
Both objections are testable, and both have been tested. A randomised controlled trial ran a modified version of the approach against conventional spoon-feeding, and the results are considerably more specific than either the advocates or the sceptics tend to report. Here is what it found.
The trial
The Baby-Led Introduction to SolidS study, usually shortened to BLISS, was a randomised controlled trial of 206 participants. 101 were assigned to a control group following conventional advice, and 105 to the BLISS group, who received eight additional visits from before birth through nine months providing education and support.
The critical detail is in the name. BLISS is baby-led weaning modified to address exactly the two concerns above. Parents were specifically advised to offer a high-iron food at every meal, and to avoid foods that present a high choking risk. It was not a test of the approach as commonly practised. It was a test of the approach with the two known weak points deliberately reinforced.
Reading the results without that qualification is the most common mistake made about this trial, in both directions.
What it found on iron
Nothing significant, and that is the finding.
Median dietary iron intake did not differ significantly between the groups at seven months, with a difference of 0.6 mg per day and a confidence interval running from -1.0 to 2.3, or at twelve months, where the difference was -0.1 mg per day with an interval of -1.6 to 1.4. At twelve months there were also no significant differences in plasma ferritin concentration, body iron, or the prevalence of depleted iron stores, early functional iron deficiency or iron deficiency anaemia.
The authors' conclusion is worth quoting in shape rather than paraphrasing loosely: a baby-led approach does not appear to increase the risk of iron deficiency when parents are given advice to offer high-iron foods with each meal.
That conditional clause is the whole finding. The trial did not show that self-feeding is inherently fine for iron. It showed that self-feeding with deliberate high-iron food provision at every meal produces iron status indistinguishable from conventional weaning. Whether unmodified baby-led weaning, without that advice, achieves the same thing is a different question, and this trial did not answer it.
Why iron is the concern at all: infants are born with iron stores that deplete over the first half of the first year, and from around six months dietary iron becomes necessary. Iron-fortified cereals are a convenient vehicle in spoon-feeding, and a self-feeding infant is not eating those, so the intake has to come from elsewhere.
What it found on choking
Less than people claim, and the pilot work is where the useful number sits.
The pilot study that preceded the main trial found that the BLISS group were offered high-choking-risk foods considerably less often than a conventional baby-led weaning group: 0.17 servings per day against 3.24, a difference that reached significance.
Read that carefully, because it is frequently reported backwards. It is not a finding that baby-led weaning causes less choking. It is a finding that the education worked. Parents who were told which foods are high-risk offered far fewer of them. That is a result about an intervention on parental behaviour, not about the safety of an approach.
Which points at the honest conclusion. The risk in baby-led weaning is not the self-feeding, it is the food list, and the food list is entirely within a parent's control.
The foods the education is about
The high-risk list is consistent across guidance and it is short enough to memorise. Whole grapes and cherry tomatoes, which should be quartered lengthwise. Whole nuts. Raw hard vegetables like carrot and apple, which need cooking soft or grating. Chunks of meat and cheese cut into cubes rather than strips. Popcorn. Hard sweets. Marshmallows. Anything round, firm and roughly the diameter of an infant airway.
The preparation rules matter as much as the list. Cook hard vegetables until they yield to gentle pressure between finger and thumb. Cut long and thin rather than round and small, so a piece cannot seat itself crosswise. Offer pieces large enough to be held with a fist protruding, since palmar grasp arrives before pincer grip.
And the one that gets skipped: the baby sits fully upright, supported, at a table, with an adult present for the whole meal. Not reclined, not in a car seat, not in a moving stroller.
Gagging is not choking
This is the distinction that decides whether a parent finds the first month tolerable.
Gagging is loud, involves coughing and retching and a red face, and is a protective reflex working correctly to move food forward. It is common, it is noisy, and the infant is moving air throughout.
Choking is quiet. The airway is obstructed, so there is little or no sound, and colour changes towards blue rather than red.
Silence is the signal. A paediatric first aid course is the single most useful preparation for weaning by any method, and it is worth doing before you start rather than after something frightening happens.
What this means for what you buy
Not much, which is itself useful.
Self-feeding needs a chair that holds a baby fully upright with foot support, a surface that can be wiped, and something to catch what lands on the floor. A suction plate reduces launched dinners and does not eliminate them. Soft-tipped preloaded spoons are useful in either approach, because most families run a mixture rather than a doctrine.
The equipment that does not earn its place is anything promising to make self-feeding safe. Mesh feeders, food-shaped teethers marketed for weaning and specialist cutting gadgets all address a risk that food preparation already addresses better and for nothing.
Our feeding roundup covers the gear that does hold up, and the footprint question that decides which high chair works in a real kitchen is in the high chair that fits a small kitchen. If bottles are the harder problem in your house, that is a different subject and it is in bottle refusal. And the rule that a baby who falls asleep after a meal is moved to a flat firm surface applies exactly as it does anywhere else, which is the argument in the cribs and bassinets guide.
Questions people actually ask
Does baby-led weaning cause iron deficiency?
The BLISS trial found no significant difference in iron intake or iron status at twelve months against conventional weaning. That result came from a version of the approach where parents were specifically advised to offer a high-iron food at every meal, and the authors' conclusion is conditional on that advice.
Is choking more likely with self-feeding?
The trial's pilot work found that education dramatically reduced how often high-choking-risk foods were offered, from 3.24 to 0.17 servings a day. The risk tracks the food list and its preparation rather than the method, and both are under a parent's control.
Can I do a mixture of purées and finger foods?
Most families do, and nothing in the evidence argues against it. The trial compared structured approaches because that is what a trial has to do. In practice the useful parts are the iron provision and the choking-risk food list, and both apply whichever way the food arrives.
What are the best high-iron first foods?
Well-cooked meat cut into strips, iron-fortified cereals mixed thick enough to load onto a spoon a baby can hold, lentils and beans mashed or in soft patties, and dark leafy greens cooked soft. Offering a vitamin C source alongside plant iron improves absorption, which is worth knowing if your household does not eat meat.
When should we start?
Around six months and when the developmental signs are there: sitting upright with support, good head control, and reaching for food. That is a question for your pediatrician for your specific baby, and readiness varies more than the calendar suggests.
Sources
As an Amazon Associate we earn from qualifying purchases. We earn a commission on some links, at no extra cost to you. It never changes which products we recommend or the order they appear in. Full disclosure
Who wrote this
Writer, standards and the recall record
Dan Whitfield writes the standards side of Little Nest: what a certification actually covers, what the CPSC recall database says about a brand, and what the one-star reviews agree on six months in. He is not a clinician and does not write as one. His job is to read the standard rather than the marketing page, quote the real number, and name the limit the packaging skipped.
More from DanThe standard
How this guide was built
- Manufacturer specification sheets and warranty terms, read in full rather than summarised from a listing.
- The federal safety standard the product is certified to, plus anything NHTSA, the AAP or the CPSC has published on it.
- Owner reports at volume, including the one-star reviews, which is where failure modes live.
- Every price checked against the live listing on the date printed beside it.
