Timing, not method, is what the evidence on starting solids settles

Breastmilk runs out of iron before the baby does

A full-term baby is born with four to six months' worth of iron banked in the liver, and breastmilk was never built to refill it. Human milk carries iron the infant absorbs well, but the quantity is small, and past four to six months it no longer covers a growing infant's needs.¹ The recommended intake rises to 11 milligrams a day between seven and twelve months, a figure an exclusively breastfed baby cannot reach from milk alone.¹ This is the real clock behind the advice to start solids around six months. It is not a birthday. It is a supply running down, which is why the American Academy of Pediatrics recommends iron drops from four months for breastfed infants until iron-rich foods take over.¹

A household at this stage faces two questions, and the research treats them very differently. The first is when to begin. The second is how: baby-led weaning, where a baby self-feeds graspable pieces of family food, or spoon-fed purées. The evidence is firm on the first question and thin on the second, and it is thinnest in exactly the place the marketing is loudest.

Six months is a public-health number, not a biological edge

The World Health Organization advises introducing complementary foods at six months while breastfeeding continues.² The WHO grades its own advice as a strong recommendation resting on low-certainty evidence.² That pairing is deliberate. The WHO is confident enough to advise six months for the whole world while conceding the underlying trials are weak. Part of the six-month logic is reducing an infant's exposure to contaminated food and water, the real dangers in many places. It is a global-health calculation as much as a statement about one baby's metabolism.

Europe's pediatric gastroenterology body reads the same biology into a window rather than a line. Its position is that complementary foods should not start before four months and should not be delayed beyond six.³ By roughly four months the gut and kidneys can handle food that is not milk, and between four and six months most infants gain the motor control to manage it.³ The AAP's own guidance for parents lands at approximately six months and then does the more useful thing: it names signs a family can watch for. Sitting with steady head control. Reaching for food and opening the mouth for it. The fading of the tongue-thrust reflex, so a spoonful goes back and down rather than straight out.⁴ Doubling of birth weight, around four months, is another rough marker.⁴

These sources look like they disagree, and they mostly do not. The WHO is naming a population optimum; Europe and the AAP are naming an individual readiness window. A family that reads "six months" as a hard biological edge is over-reading it, and the WHO's own low-certainty grade concedes as much. The honest synthesis is unglamorous. Somewhere in the four-to-six-month window, when the readiness signs have arrived and iron is becoming the constraint, is the time to start. The exact day is not a finding. It is a judgment call the evidence leaves open.

One trial program is most of what we know

Baby-led weaning skips purées and hands the baby soft, graspable food from the family plate. The claims that travel with it are specific and appealing: better appetite regulation, less fussiness, leaner children, and, on the anxious side, a fear of more choking. Almost all of the careful evidence for any of these traces to a single research program. The Baby-Led Introduction to SolidS study, run in Dunedin, New Zealand, randomized 206 infants to a baby-led approach or to usual care and followed them to age two.⁵

That trial was not the version most families do off a blog. It modified baby-led weaning in two deliberate ways. Families were coached, across eight visits, to offer an iron-rich food at every meal.⁶ They were also taught to avoid the food shapes most likely to lodge in an airway.⁵ And it is, with its companion analyses, close to the whole rigorous literature. A 2026 systematic review searching for randomized trials of baby-led weaning found only four, three of them from this same New Zealand group.⁷ Nearly everything trustworthy that can be said about the method rests on one team's well-run, intentionally modified intervention. That is the first caveat, and it shades all the others: what held up in this trial held up for iron-coached, choking-coached baby-led weaning, not the do-it-yourself kind.

Choking, growth, and iron came out the same either way

Across both arms of the trial, 35% of infants choked at least once between six and eight months, with no significant difference between the baby-led and spoon-fed groups.⁵ The reassuring headline is earned: a structured baby-led approach did not raise choking. Gagging did differ, in a way worth understanding. Baby-led infants gagged more at six months and less by eight, a reversal from a relative risk of 1.56 to 0.60.⁵ Gagging is the protective reflex that keeps food forward of the airway. It is noisy and self-correcting, and it is not choking, which is silent and blocks the airway. The pattern reads like babies meeting textures a little earlier and learning the skill sooner, not choking more often.

The theory that self-feeding trains appetite and holds off overweight did not survive the trial. Body-mass index was no different at one or two years, and if anything overweight was numerically higher in the baby-led group, 10.3% against 6.4%, a gap well within chance.⁸ Energy self-regulation was no better either.⁸ The one difference that held favored baby-led on attitude, not health: parents reported less food fussiness.⁸

Iron told a similar story with a sharper footnote. The worry that self-feeders would fall short on iron did not appear here: iron intake and iron status were no worse than with spoon-feeding.⁶ But the trial's own authors framed that result as beating the low iron reported for unmodified baby-led weaning, not as proof that unguided self-feeding is iron-safe.⁶ The coaching to serve iron at every meal is why the number came out even. One finding did cut against baby-led. At seven months its infants took in 35% more sodium than the spoon-fed group.⁹ That is the predictable price of eating adult food, and a reason to keep the family's own salt down while a baby shares the plate.

What goes on the plate, and the signs that need a doctor

Near six months, once the readiness signs are there, start, and lead with iron-rich foods: meat, iron-fortified cereal, well-cooked legumes, whichever method you use.¹⁴ Keep the forms choking-safe, soft and graspable, with no whole nuts, whole grapes, or hard raw chunks, whether the food arrives on a spoon or in a fist.⁵ Keep the household salt low while the baby shares meals.⁹ Beyond that, the method is a preference, and the data are content to leave it to you.

A few lines are not a preference, and they are where an article stops and a pediatrician starts. A baby born preterm should be timed by corrected age and has different iron needs; that is a conversation with the clinician, not a chart.¹ Signs that can point to iron deficiency, unusual pallor, poor feeding, low energy, or a stall in development, are worth a same-week call.¹ A true choking event, where a baby cannot cough, cry, or breathe, is an emergency, and gagging that never eases or feeding that falters over weeks deserves a visit.⁴ Allergen introduction, the separate and now well-tested question of giving peanut and egg early, is its own decision with its own trial evidence and is not this article's subject.

The shape of the evidence is clearer than the noise around it. It sets a window and a nutrient, and hands a family the rest. That is not a gap in the science so much as its result. Past iron and readiness, feeding a baby is a set of ordinary choices, and the research is comfortable leaving them to the people at the table.¹⁰