Home / Guides

Why the WHO and CDC growth charts differ, and why the switch happens at two

Two charts, two completely different projects

A US child gets plotted on two different charts before their third birthday, and most parents never find out why. The short version: the two charts were built to answer two different questions, and only one of them is a question about how children should grow.

The CDC 2000 Growth Charts are a reference. They describe how a large sample of American children actually did grow, drawn from national survey data, and they were released in 2000 as a revision of the 1977 NCHS charts. A reference is a mirror. It reports the distribution of a population as it was, whatever that population was eating and however it was fed.

The WHO Child Growth Standards are a standard. They come from the WHO Multicentre Growth Reference Study, which ran from 1997 to 2003 across six sites, in Brazil, Ghana, India, Norway, Oman and the United States, following roughly 8,500 children, and were published in 2006. The study deliberately selected children raised under conditions believed to support healthy growth, and every infant in the reference group was breastfed, so the curves describe how children grow when circumstances are favourable rather than how a given national population happens to grow.

That distinction, descriptive reference versus prescriptive standard, is the root of every other difference between the two charts.

Why US practice uses one for infants and the other for toddlers

In September 2010 the CDC published a recommendation in its Morbidity and Mortality Weekly Report (Vol 59, No. RR-9, “Use of World Health Organization and CDC Growth Charts for Children Aged 0 to 59 Months in the United States”) that clinicians in the United States use the 2006 WHO standards for children under 24 months, and continue using the CDC charts for ages 2 through 19.

The stated reasoning is that breastfeeding is the recommended standard for infant feeding, so the appropriate yardstick for an infant is a population of breastfed infants growing under good conditions. Above two, the growth patterns the two charts describe converge enough that the CDC’s own national reference is used, and it carries on continuously all the way to 19.

The same recommendation notes two practical consequences worth knowing. On the WHO charts, the recommended cut-offs for screening are the 2.3rd and 97.7th percentiles, which correspond to plus and minus two standard deviations, rather than the 5th and 95th used with the CDC charts. And the WHO curves show slower weight gain than the CDC curves through much of the first eighteen months, which is the expected pattern for a breastfed reference group.

This site implements exactly that split, at exactly 24.0 months, and never blends the two into one smoothed curve. It also never uses the CDC’s separate birth-to-36-month infant file at any age.

The step you can see at the seam

Because the two charts describe different populations, their curves do not meet at the boundary. On the tables this site loads, at 24 months:

The weight step goes one way and the length step goes the other, and the length step has a second cause on top of the population difference: the measurement itself changes at two years. Under 24 months a child is measured lying down, which is recumbent length. From 24 months a child is measured standing, which is stature, and a standing measurement reads shorter than a lying one because the spine compresses. WHO and the CDC both switch measurement technique at the same age the chart switches, which is not a coincidence.

The practical upshot is that a child’s percentile can shift a little at their second birthday without anything at all having happened to the child. The chart tool marks the seam with a labelled line so you can see it rather than wonder about it.

Head circumference is the exception

The CDC 2000 charts carry no head-circumference reference past 36 months, and this site does not use the CDC infant file, so head circumference stays on the WHO standard across the whole 0 to 60 month range the tool supports. There is no seam on the head-circumference chart, and the tool says so in the caption when you plot one.

Which chart is “right”

Neither, and the question is the wrong shape. A standard and a reference are different instruments. A prescriptive standard answers “how do children grow under good conditions”, and a descriptive reference answers “how did the children in this survey grow”. Both are honest answers to their own question, and US practice uses each where its question is the more useful one.

What follows for a parent reading a number is only this: know which chart your number came from, because the same child can carry two slightly different percentiles on the same day depending on which curve was used. Every result on this site names the reference it used, for exactly that reason.

Percentiles describe position in a reference population, not health; growth patterns are for your pediatrician to read. The full parameter tables behind both charts, with the source file and retrieval date for each, are on the reference tables page.

Keep going: plot a measurement, read the published reference tables, or browse the other guides.