The infant feeding schedule:

Why babies want (and need) to be fed “on demand”

© 2008 Gwen Dewar, Ph.D., all rights reserved

What’s the best infant feeding schedule?

For generations, Western “baby experts” have advised parents to feed their babies at regularly-spaced intervals of 3- or 4-hours. Today, official medical recommendations have shifted in favor of feeding on demand.

That’s a good thing, because the timed-interval infant feeding schedule is not dictated by physiology or what’s best for babies. Research suggests that many babies are better off if they are permitted to regulate their own feeding bouts. Unfortunately, parents don’t always get the message. The time-interval infant feeding schedule has become culturally entrenched, and many Western hospitals have been slow to adopt the sorts of practices—-like rooming in—-that promote breastfeeding on demand (e.g., Kersting and Dulon 2001).

This article reviews the evolutionary, cross-cultural, and clinical evidence in favor of frequent, baby-initiated feedings. As noted below, the evidence supports the following conclusions:

• Human infants have evolved to feed on demand

• New mothers are more likely to establish an adequate milk supply if they breastfeed frequently and on demand

• Newborns should be nursed frequently (every two hours or more) and whenever they show signs of hunger—ideally, before they begin to cry

• Breastfeeding sessions should last as long as infants show an interest in feeding

• Demand-style infant feeding schedules permit babies to adjust their intake in response to natural variations in milk quantity and quality

• Infants breastfed on demand may benefit from higher quality breast milk and fewer digestive problems

• Infants (of any age) breastfed on demand may experience less stress and pain

• Infants fed on demand take more naps

Moreover, women who begin breastfeeding on demand are more likely to continue breastfeeding after 1 month.

For these reasons, both the World Health Organization (WHO) and the American Academy of Pediatrics (APP) now recommends that breastfeeding mothers follow a baby-initiated infant feeding schedule.

What about formula feeding? There is growing concern among some researchers that babies on formula may be at increased risk for overfeeding. But more information is needed to resolve this question, and current studies can’t tell us if feeding on demand increases or reduces the risk of overfeeding in formula-fed infants. Meanwhile, we do know that formula-fed infants adjust their appetites in response to the caloric concentration of their formula (see bottom). Given this fact, it seems prudent for parents of formula-fed babies to begin with a demand-style infant feeding schedule, and then consult with their pediatricians about their babies' weight gain over time.

For the parents of children who breastfeed, however, the evidence seems especially clear-cut. So let's review the scientific evidence in favor of a demand-style infant feeding schedule.

The infant feeding schedule in evolutionary perspective

Mammals are named for their milk-secreting mammary glands, and mammal babies everywhere begin life on a diet of milk. But not all milk is alike, and different species follow different infant feeding schedules.

Survey the world of mammalian infant care, and you’ll discover two major feeding strategies: the spaced feeders and the continual feeders.

Spaced feeders “park” their young in nests and leave them there—unattended—for hours at time. Mom uses her time away to forage. The kids must wait for long intervals between feedings. But they don’t go hungry: Spaced feeders have evolved high-fat, high-protein milk. And the babies are designed to suck fast and furious when they get the chance. A good example of a spaced feeder is the rabbit, which produces milk that is 18.3% fat and 13.9% protein (Jenness 1974).

By contrast, continual feeders stay with their babies at all times. Whether they follow their moms around on foot (like cows) or get carried (like monkeys), kids stay in close contact with Mom. They get fed more frequently than spaced feeders do, and they tend to initiate feedings. Consistent with their high-frequency infant feeding schedule, they suck at a more leisurely rate. Their milk is less caloric, more dilute. Cow milk is typically 3.7% fat and 3.4% protein (Jenness 1974).

So what about humans?

In some modern industrial societies, humans act like spaced feeders. Babies are “parked” in cribs or cradles and get fed after intervals of 2-4 hours. But were we designed for this strategy? The answer is no.

Humans belong to the order Primates, a group that includes prosimians, monkeys and apes. A few—mostly nocturnal, prosimian species—are spaced feeders (Teilden and Oftedal 1997). But most—including our closest living relatives, the gorillas, chimpanzees, and bonobos—are continual feeders. And the proof is in our milk: Compared to the milk of spaced feeders, human milk is relatively low in fat (3.8%) and protein (1%), and human infants suck slowly when they feed.

So the composition of our milk gives us away. We’re meant to be continual feeders. (Click here for a more detailed account of the nutrients and calories in human breast milk.) But though human milk is comparatively dilute, this doesn’t mean that human milk is everywhere and always the same. Breast milk composition may vary significantly among individual women, and the same woman may experience significant fluctuations in her milk quality over time.

And—-most interesting for our discussion here—-that fat content of human milk increases when babies feed more frequently.

Human milk quality improves with frequent feedings…and volume of intake

International studies show that the more frequently a woman breastfeeds, the higher the fat content of her milk (Prentice and Prentice 1988; Prentice et al 1981; Jackson et al 1988; Daly et al 1993).

Other factors associated with higher fat content include the volume of intake at previous feed (how much milk the baby ingested at the last feeding) and volume of intake at the current feed (Woolridge 1995).

Volume of intake is important because the highest fat milk comes from soft, apparently empty breasts. When a baby begins feeding from a full breast, the milk she receives (“foremilk”) is relatively low in fat. Fat content increases as the breast empties. For this reason, it’s important not to force babies to switch breasts before they’ve had their fill. Premature switching can rob the baby of the higher-fat hindmilk associated with the latter part of a feed (Woolridge and Fisher 1988). For this reason, mothers should let their babies decide when it’s time to switch—if ever. Some babies may nurse from one breast exclusively during a feeding session. And that’s fine.

These discoveries suggest that women can improve the quality of their breast milk by feeding frequently and on demand.

And quality matters. When milk quality is low, babies need to consume more of it to get the same caloric benefit. Babies vary in their ability to suck quickly and efficiently. If access to the breast is limited—as it is on the typical Western infant feeding schedule—some babies may not get enough to eat (Woolridge 1995).

Low-quality milk can cause other problems as well. As breastfeeding expert Michael Woolridge (MD and PhD) has pointed out, low-fat milk can contribute to colic, vomiting, diarrhea, and flatulence in infants (Woolridge 1995).

So—-from the standpoint of our evolutionary heritage and the quirks of human milk production-—it appears that frequent feedings are best. Humans were designed to follow a demand-style infant feeding schedule.

The hunter-gatherer infant feeding schedule...and beyond

How frequently should babies be nursed? There is probably no single, “right” answer to this question. But the practices of contemporary hunter-gatherers—whose life-ways are probably most similar to those of our ancestors—are strikingly different from those of industrial Westerners.

Among the !Kung San of South Africa, the traditional pattern is for babies to be fed about 4 times an hour. Feeding bouts are brief, lasting only 2 minutes or so (Konner 2006). And nursing continues—on demand—throughout the night. The !Kung might represent the extreme end of feeding frequency, but other hunter-gatherer groups—from South America to the Philippines—follow a similar pattern, nursing at least twice an hour (Konner 2006).

And it’s not just hunter-gatherers. In a recent, cross-cultural survey sampling 48 non-industrial societies—-including nomadic pastoralists and settled agricultural peoples—-demand feeding was the rule in every society for which information about the infant feeding schedule was available (25 out of 25; Severn Nelson et al 2000).

It seems likely that frequent, infant-initiated nursing has been the normal human pattern for over 99% of human history.

But does it matter?

Skeptics might argue that milk fat concentrations and hunter-gatherer practices are irrelevant as long as babies thrive. And it’s true that normal, healthy infants can adapt to a variety of feeding styles.

For instance, in a retrospective study of 1-year-olds who had been fed either “on schedule” or “on demand,” researchers detected no difference in growth rates (Saxon et al 2002). However, most infants in the “on demand” feeding group were fed with formula, not breast milk, so the timing of feedings could have no effect on fat content. Moreover, babies can drain bottles more rapidly than they can drain breasts. These points suggest that a formula-fed baby subjected to an infrequent infant feeding schedule can “make up for lost time” by taking in large quantities of formula at each feed.

But the situation is very different for breast-fed babies, particularly in the first few months. It is also important to recognize that infants need more food during growth spurts. Finally, feeding on demand may have important non-nutritive benefits. Let’s take these points in turn:

Newborns should be fed on demand

Most medical experts recommend that parents feed newborns frequently-—around once every two hours. But this represents an average. Babies vary in their ability to suck efficiently, and individuals have different stomach capacities. In addition, breastfeeding mothers vary in the quality and amount of milk they can deliver at any single feeding. As a result, there is no such thing as a universal infant feeding schedule that benefits all newborns.

Feeding on demand allows newborns to adapt to their particular circumstances. When newborns feed on demand, they regain their birth weight more quickly. And demand feeding helps new mothers establish and maintain an adequate milk supply. For details, see this article on the newborn infant feeding schedule.

Demand feeding is best for growth spurts

Although the timing varies among individuals, babies typically experience growth spurts at around 1-3 weeks, 6-8 weeks, 3 months and 6 months of age. During these times, breastfeeding babies will want to nurse more frequently. More frequent feedings stimulate the breasts to step up milk production and meet their babies increasing energy requirements. When mothers breastfeed on demand, they help ensure that milk production keeps up with their babies’ changing needs.

A rigid infant feeding schedule is ill-suited to deal with growth spurts. If breastfeeding mothers ignore their babies’ cues, they may not produce enough milk.

Other benefits: Less pain, less stress, more naps?

Suckling is known to have a calming effect on babies, even if they are not especially hungry. Newborns cry less and show signs of reduced pain when they receive small amounts of milk, formula, or sucrose (see review by Shaw et al 2007; also Blass 1997a; Blass 1997b; Blass and Watt 1999; Barr et al 1999). The act of suckling is itself an analgesic (Blass and Watt 1999).

And breastfeeding appears to be a powerful painkiller and stress reducer. When newborns randomly assigned to either be breastfed or swaddled during a painful blood collection procedure, the breastfeeding babies experienced far less distress than did controls (Gray et al 2002). Breastfeeding babies cried for only 4% of the total procedure time, as opposed to 43% for controls. Grimacing was also markedly reduced (8% v. 50%), and breastfeeding babies showed less of an increase in heart rate (6 beats per minute v. 29 beats per minute). Interpreting this study is difficult, because breastfeeding babies, but not control babies, received skin-to-skin contact, which is also known to reduce pain. However, the authors of this study believe that breastfeeding is a more effective analgesic than skin-on-skin contact alone. When the authors performed a similar experiment on the effects of skin-on-skin contact, babies grimaced more than did the breastfeeding babies (Gray et al 2000; Gray et al 2002). The authors also note that babies who are held without being fed tend to get frustrated, and require much more time to settle down (Gray et al 2002).

Taken together, these findings suggest that babies fed on demand have more opportunities to soothe pain and reduce distress.

They might also have more opportunities to drowse off.

A retrospective study of 1-year old infants found that babies fed on demand took more frequent naps than those following a more rigid infant feeding schedule (Saxon et al 2002). This was true even though the majority of infants fed on demand (81%) were either primarily or wholly dependent on formula. The results lend support to the possibility that infants fed on demand—-be they breastfed or formula-fed-—feel more satisfied after meals.

But won't babies overfeed?

Some parents worry that if they adopt a baby-led infant feeding schedule they'll end up with an overfed baby. This isn't very likely--not if babies are breastfeeding. At present, there is no evidence that babies who breastfeed are at risk for overfeeding. In fact, breastfed babies tend to gain weight less rapidly than formula-fed infants.

Other evidence supports the idea that infants have an appetite-control mechanism. Studies of both breastfed (Woolridge and Baum 1992) and formula-fed (Fomon et al 1975) infants adjust their intakes in response to changes in the calorie content of their milk or formula.

What about exhaustion?

New moms may find a demand-style infant feeding schedule exhausting, especially during the first few weeks. This may be particularly true in the West, where cultural practices make demand feeding more difficult. For more information about coping with the demands of a baby-led infant feeding schedule, see this anthropological guide to breastfeeding on demand.

The bottom line

Human breastfeeding is meant to operate on a demand basis. When babies determine the timing and duration of their feeds, they are most likely to get what they need. Breast milk production increases or decreases in response to the baby's demand. Breast milk quality may improve. And babies may be better able to cope with pain and stress.

An arbitrary, timed-interval infant feeding schedule is less likely to fit a baby's needs.

References: The infant feeding schedule

Barr RG, Pantel MS, Young SN, Wright JH, Hendricks LA, Gravel R. 1999. The response of crying newborns to sucrose: is it a “sweetness” effect? Physiol. Behav 66: 409-417.

Blass EM. 1997a Milk-induced hypoanalgesia in human newborns. Pediatrics 99: 825-829.

Blass EM. 1997b. Infant formula quiets crying newborns. Journal of Dev Behavioral Pediatrics. 18:162-165.

Daly SE, DiRosso A, Owens RA and Hartmann PE. 1993. Degree of breast emptying explains fat content, but not fatty acid composition, of human milk. Exp Physiol 78: 741-755.

Fomon SJ, Filmer, Jr., JA, Thomas LN, Anderson TA and Nelson SE. 1975. Influence of formula concentration on caloric intake and growth of normal infants. Acta Pediatrica Scandinavica 64: 172-181.

Gray L, Miller LW, Philipp BL, Blass EM. 2002. Breastfeeding is analgesic in healthy newborns. Pediatrics 109: 590-593.

Gray L, Watt L, Blass EM. Skin-to-skin contact is analgesic in healthy newborns. Pediatrics 105(1).

Illingworth RS, Stone DHG, Jowett JH and Scott JF. 1952. Self-demand feeding in a maternity unit. Lancet 1: 683-687.

Jackson DA, Imong SM, Silpraset A, Preunglumpoo Ruckphaopunt S, Williams AF, Woolridge MW, Baum JD, and Amatayakul K. 1988. Circadian variation in fat concentration of breastmilk in rural Northern Thailand. British Journal of Nutrition 59: 365-371.

Jenness 1974. Biosynthesis and composition of milk. Journal of investigative dermatology. 63: 109-118.

Kersting M and Dulon M. 2001. Assessment of breastfeeding promotion in hospitals and follow up survey of mother-infant pairs in Germany: The Su-Se study. Public Health Nutrition 5(4): 547-552.

Konner M. 2005. Hunter-gatherer infancy and childhood: The !Kung and others. In: Hunter-gatherer childhoods: Evolutionary, developmental and cultural perpectives. BS Hewlett and ME Lamb (eds). New Brunswick: Transaction Publishers.

Prentice AM and Prentice A. 1988. Energy costs of lactation. Annual review of nutrition 8: 63-79.

Prentice A, Prentice AM and Whitehead RG. 1981. Breast-milk concentrations of rural African women I. Short-term variations within individuals. British Journal of Nutrition 45: 483-494.

Saxon TF, Gollapalli A, Mitchell MW, and Stanko S. 2002. Demand feeding or schedule feeding: infant growth from birth to 6 months. Journal of reproductive and infant psychology 20(2): 89-99.

Severn Nelson EA, Schiefenhoevel W, and Haimerl F. 2000. Child care practices in nonindustrial societies. Pediatrics 105: 75-79.

Shah PS, Aliwalas L, and Shah V. 2007. Breastfeeding or breast milk to alleviate procedural pain in neonates: a systematic review. Breastfeeding medicine 2:74-82.

Tilden, C.D. & Oftedal, O.T. (1997)Milk composition reflects pattern of maternal care in prosimian primates. American Journal of Primatology 41: 195-211.

Woolridge MW. 1995. Baby-controlled breastfeeding: Biocultural implications. In: Breastfeeding: Biocultural perspectives. P. Stuart-Macadam and KA Dettwyler (eds). New York: Aldine deGruyter.

Woolridge MW and Baum JD. 1992. Infant appetite-control and the regulation of breast milk supply. Children’s hospital quarterly 3:133-119.

Woolridge MW and Fisher C. 1988. Colic, 'Overfeeding,' and Symptoms of Lactose Malabsorption in the Breast-Fed Baby: A Possible Artifact of Feed Management. Lancet 13: 382-384.

For more references pertaining to the infant feeding schedule, see my article on breastfeeding on demand.