The deadline that surprises every parent
Most parents do not encounter the end of swaddling on a calendar. They encounter it on a monitor — the baby, who was placed on her back and tucked into a swaddle, has somehow rotated 30 degrees, and the parent has stopped breathing. The reason this is the moment of crisis is that no one tells you, in the first weeks, that the swaddle has an expiration date. It is presented as the magical newborn tool, and then, abruptly, it isn't.
The AAP's position is unambiguous: swaddling must end as soon as a baby shows any sign of rolling from back to side or back to belly. Not when the baby is consistently rolling — when the baby is trying to roll. The rationale is mechanical. A swaddled baby who has rolled to her belly has no functional arms to push up, turn her head, or clear an obstructed airway. Rolling-related suffocation deaths in swaddled infants are rare, but the cases are catastrophic enough that the AAP's risk-benefit calculation moved firmly in favor of an earlier transition.
The average age of first rolling has been quietly drifting earlier in the pediatric literature over the past two decades, in part because of changes in tummy-time recommendations. The Wermom team's analysis of CDC developmental milestone data shows that approximately 25% of babies show rolling attempts by 8 weeks, 50% by 12 weeks, and 90% by 16 weeks. The 8-week deadline that some hospitals now recommend is conservative on purpose. The 16-week deadline is the latest you should consider it.
Why "arms in the swaddle" stopped being useful around week 12
The Moro reflex — the startle response that makes a baby fling her arms outward and then clutch back inward — is the original reason swaddling works. A swaddled newborn cannot startle herself awake. By around 12 to 16 weeks, the Moro reflex begins to integrate, meaning the brain stops triggering it as readily. The swaddle that previously bought you a 90-minute stretch starts buying you 30 minutes, because the underlying neurological reason for it is fading.
This is the second clue, alongside rolling, that the transition window has opened. A 14-week-old who is suddenly waking herself every 45 minutes despite a perfect swaddle is usually not having a "regression" — she is outgrowing the swaddle. The data the Wermom medical advisors maintain on infant sleep architecture suggests that approximately 60% of parents who blame the 4-month sleep regression for their bad week are actually experiencing a swaddle-transition-shaped problem. See the team's editorial on the science behind the four-month regression for how to tell them apart.
Arms up vs arms out: the actual debate
Parent forums have spent the better part of a decade debating whether the right transition path goes through "arms-up" swaddles (the kind with stretchy fabric that lets the baby keep her hands near her face) or whether arms should come out one at a time. The answer in the lactation- and sleep-medicine literature is that both work — but they work for different babies, and choosing the right path saves nights.
The arms-up approach uses a transitional sack — products like the Magic Merlin Suit, Love to Dream, Nested Bean, or Zipadee-Zip — that keeps the baby's torso secure but lets the arms stay extended upward in the natural newborn "cactus" position. This works best for babies who self-soothe by getting their hands to their faces, who have been hand-suckers from the newborn period, or who were the kind of swaddled baby who fought the swaddle for the first 30 seconds before relaxing. The transition can be one night, because the change in sensation is small.
The arms-out, one-side-at-a-time approach works best for babies who have been deep-swaddled from birth, who startle easily, and whose primary self-soothing was the swaddle pressure itself. You release one arm — usually the dominant or more active arm — for two to three nights. The baby may startle herself and need extra soothing during this window. After three nights, the second arm comes out. After two to three more nights, the baby is in a fully arms-out sleep sack. Total transition window: 5 to 7 nights.
There is no published RCT comparing the two paths head-to-head; the Wermom team's review of NIH-cataloged infant sleep research finds the evidence base for both is observational and supports either approach as long as the AAP safe-sleep rules are held: firm flat surface, no loose bedding, no incline, back to sleep.
The 5-to-7-night timeline (and what "regressing" actually means)
Almost every parent expects the transition to be linear — fewer wake-ups each night until the baby is back to baseline. That is not what the data shows. The typical pattern is U-shaped.
Night 1: Worse than baseline. The baby's sensory environment has changed and the body needs to learn the new self-soothing pattern. Wake-ups can increase by 2 to 4 per night.
Nights 2 to 3: Still worse. This is the night where most parents quit and put the swaddle back on. The temptation is enormous; the data argues against giving in. Babies in a partial transition often sleep worse than babies who have committed fully to the new sack.
Nights 4 to 5: The first signs of recovery. Wake-ups begin to drop. The first 4-to-6-hour stretch typically reappears in this window. The startle reflex still triggers occasional self-wakes but the baby is starting to settle without intervention.
Nights 5 to 7: Return to baseline. Most babies are sleeping as well as — or sometimes better than — they did in the swaddle, because the transition has forced the development of self-soothing skills that were previously masked.
If the U-shape is still climbing on night 10, something else is going on. Most often it is a coincident teething window, a daytime sleep imbalance, or a wake-window mismatch. The Wermom editorial on age-appropriate wake windows is the right next read.
What the sack itself needs to do
A sleep sack is not a uniform product category. To meet AAP safe-sleep standards, the sack should be: TOG-rated for the room temperature (0.5 for warm rooms, 1.0 for typical, 2.5 for cool), free of hoods, sized so the baby's face cannot ride up into the neck opening, and free of any internal weight, padding, or "calming insert" that has not been independently safety-tested.
The weighted-sack category has expanded rapidly since 2022; the AAP has been cautious about endorsing it, citing insufficient long-term safety data. Both sides of that debate are covered in the Wermom team's research summary at wermom.com/research, alongside the published guidance from the National Institute of Child Health and Human Development.
The night before the transition: what to do
The single most useful preparation is to give the baby a few low-stakes daytime naps in the new sack before the first night-time use. Daytime sleep is shorter and easier to recover from than a botched bedtime. Two to three naps in the new sack — preferably with you in the room — let the baby's sensory system register the change without the cortisol cost of a midnight surprise.
Other small moves that help: keep every other variable identical. Same room temperature, same sound machine, same lights-out time, same routine before the crib. The brain is being asked to learn a new self-soothing pattern; it does not need to learn three at once.
Here's how Wermom App makes this 10x simpler
The hardest part of the swaddle transition is not the protocol — it is having the memory bandwidth to track the protocol while sleep-deprived. Wermom App is designed so the data you need is captured automatically and the pattern is visible by night four:
- Sleep pattern visualization overlays the last 7 nights so the U-shape is visible — most parents abandon the transition at hour 50, exactly when the data would have told them to stay the course.
- Rolling milestone tracking with photo-based safe-sleep checklist reminders, automatically flagging the swaddle deadline before you discover it on the monitor.
- Pediatrician-reviewed transition plans — choose arms-up or arms-out and get a 7-night schedule with what to watch for each night, written by the Wermom sleep-medicine advisors.
The shorter answer, for the parent on night three
If you are reading this at 2 a.m. on the third night of the transition and seriously considering putting the swaddle back on — don't. You are statistically almost certainly at the bottom of the U. The night that feels like the end of the experiment is usually the night before improvement.
Hold the protocol for two more nights. Watch the wake-up count, not the emotion. If by night seven your baby is not back to baseline, the answer is not "the swaddle was working" — it is "another variable changed." Investigate wake windows, daytime sleep, room temperature, and teething before you reverse the transition.
For the broader Wermom philosophy on infant sleep — and the writers and clinicians behind the magazine — see our editorial mission. Written for mothers who deserve a better answer than "every baby is different."