When can babies go through the night without feeding, and what approaches to night weaning does the evidence support?
At some point the 2am feed stops being a nutritional necessity and starts being a habit — for the baby, or for the exhausted parent doing the feeding. When is that point, and is there an evidence-backed way to drop night feeds?
There is no proven age at which all babies stop needing night feeds, and no tested "best method" for night weaning. What the evidence does support: night waking and night feeding are normal for months — at 3–6 months babies need fewer night feeds, some sleep 8 hours in one stretch but many don't, and around 1 in 3 babies still haven't slept through the night by their first birthday [3]. One small 1993 trial found that a package of strategies (including a late-evening "focal feed") helped newborns sleep a five-hour midnight–5am stretch by 8 weeks — but it was 26 babies, a bundled package, and a five-hour definition of "through the night" [1]. For the actual methods parents ask about — gradually shrinking feeds, partner-led resettling, dropping feeds one by one — there are no controlled trials at all. (Overall evidence rating: D.)
"Sleeping through the night" is a slippery phrase. In research it usually means a 5–6 hour unbroken stretch, not 8–10 hours. And it often means the parents didn't hear a waking — video studies show babies keep waking briefly through the night for months; older babies just increasingly resettle without signalling [7]. Babies over 5 months still wake 4–6 times a night as they cycle through sleep stages; the hard part is falling back asleep, not the waking itself [5]. So when someone says their baby "sleeps through", ask what they mean.
Waking and feeding are two different things. This is the single most useful distinction in the whole topic. Babies wake between sleep cycles all night long at every age — adults do too. Night weaning is not about stopping the waking; it's about whether the waking needs a feed to resolve [5][7]. A baby who stirs, grizzles, and resettles without milk is not a baby who "needs" a night feed — and many parents discover, once they pause before offering the breast or bottle, that some wakings resolve on their own [3][5]. Conversely, a baby who wakes hungry, feeds eagerly, and settles is telling you the feed is still doing nutritional work [3][4]. Watch the baby, not the clock.
The developmental timeline: slow, and wildly variable. Newborns sleep and feed around the clock, waking every 1–3 hours to eat [5]. The longest sustained sleep period grows from roughly 3–4 hours at birth to around 6 hours by six months [7]. At 3–6 months babies need fewer night feeds; some manage 8 hours in one stretch, but many don't [3]. Even at one year, around 1 in 3 babies have not slept through the night [3]. UNICEF's guidance for health professionals is blunter: most young babies wake and feed frequently day and night regardless of feeding method, "this is normal behaviour and... not something they should try to prevent" [4]. There is no validated weight or age at which night feeds become unnecessary for every baby — it varies with gestational age, growth, illness, daytime intake, and feeding method [3].
The best trial evidence is one tiny, old, bundled study. Pinilla and Birch (1993) randomised 26 first-time parents of exclusively breastfed newborns [1]. The treatment group got a package: a "focal feed" between 10pm and midnight every night, gradually lengthening intervals before middle-of-the-night feeds using alternative soothing (reswaddling, nappy change, walking), and stronger day/night environmental cues. By 8 weeks, all 13 treatment babies (1,000 in 1,000) versus 3 of 13 controls (231 in 1,000) were sleeping the study's five-hour "night" (midnight to 5am) — and total 24-hour milk intake didn't differ, because treatment babies took more milk in the early morning [1]. What this doesn't tell you: whether the focal feed alone did anything (it was a bundle), whether it works beyond 8 weeks, or whether it applies to formula-fed babies, later-born children, or anyone outside 1993 Illinois. Outcomes were unblinded parent diaries — the parents knew which group they were in.
The other RCT is also a bundle, not a night-weaning trial. The INSIGHT trial randomised 279 first-time mothers to a broad responsive-parenting programme (bedtime routines, earlier bedtimes, sleep location, responses to waking, self-soothing, dream feeding before 4 months, and more) versus a control [2]. Babies in the programme group slept longer at night — by 35 minutes at 8 weeks, 25 at 16 weeks, and 22 at 40 weeks — were less likely to be fed immediately before bed, and more likely to self-soothe; effects had faded by one year [2]. Useful, but it cannot tell you whether dream feeding, dropping night feeds, or any single component worked — and it wasn't testing night weaning.
Dream feeding specifically: weak standalone evidence. The NCT's evidence summary notes that "dream feeding" means different things to different people, that there is no evidence it's suitable for all babies, and that it may not be recommended for underweight babies, those with reflux, or certain medical conditions [3]. As a deliberate night-weaning step (add a late feed, then drop the rest), it is essentially untested in isolation.
One myth worth retiring: formula doesn't reliably buy more sleep. Parents who formula-feed often report better infant sleep — but video studies find that what parents report about their baby's night often doesn't match what actually happened [7]. No randomised trial has tested whether switching to formula improves babies' sleep, and the observational comparisons are inconsistent. Switching to formula to fix nights is not an evidence-backed strategy — and if you value breastfeeding, the trade-off isn't supported by the evidence either.
Here is the uncomfortable truth: nobody has properly studied whether night weaning helps parents. No trial was found that randomises families to night-weaning methods and measures parental sleep, mood, relationship strain, return to work, or finances. Everything below is therefore reasoning from adjacent evidence, not findings — and we won't dress it up as more. (Parental-outcomes evidence rating: D — absent.)
What we can say with confidence is that the broken sleep is real and normal. Babies waking 4–6 times a night through sleep cycles is developmental, not a disorder [5], and frequent night waking and feeding in the early months is normal behaviour, not something to prevent [4]. Knowing it's normal doesn't make 3am easier, but it does mean the baby isn't broken and you aren't failing.
The genuine decision is a trade-off with no evidence-backed winner. Continuing night feeds preserves a feeding pattern the family may value and avoids the engorgement and supply dips that abrupt stopping can cause (see below). Reducing feeds might redistribute the night work — for instance, a partner settling the baby for some wakings, which the NHS suggests as practical help [8] — but whether that actually improves parental sleep or mood has not been tested. Be wary of anyone selling certainty here: the evidence simply doesn't exist.
Partners and the division of night labour: unstudied, but probably decisive. No study was found on whether partner-led night settling works as a night-weaning method or improves anyone's sleep — it is the standard practical suggestion, not an evidence-backed one [5]. What is known from the sleep-location literature is that the night work falls overwhelmingly on mothers in most studied families (see our room-sharing topic), so any weaning plan that doesn't explicitly assign the 2am shift tends to default to the mother. If you have a partner, the most important conversation may not be about the baby's readiness but about who does what, on which nights.
Normalise the range before you judge your nights. Around 1 in 3 babies still haven't slept through by their first birthday [3] — so a 10-month-old waking for a feed is ordinary, not a failure of parenting or of the baby. Much of the distress around night feeding comes from comparing with other families' babies (or with books' promises), not from anything the evidence marks as abnormal [3][5]. If you do decide to night-wean, do it because the current arrangement isn't working for your family — not because a milestone chart says you should have done it months ago.
One caution from the sleep-training literature: behavioural sleep interventions are generally studied from around 6 months, and reviews find little benefit to applying them earlier (see sleep training and attachment). The same caution applies to night weaning — pushing it early has the thinnest evidence of all.
The evidence base is too thin for a proper benefits-and-risks table, so here is what the numbers that do exist say — and what they don't:
None of these are evidence-backed methods — they are the options families actually use, with the evidence status stated honestly: