September 20, 2010 — The most common charts used by pediatricians to measure head growth in infants and toddlers may be inaccurate, potentially leading to delayed diagnoses, unnecessary tests, and parental worry, according to new research from The Children’s Hospital of Philadelphia.
First study to evaluate validity of head-circumference curves in U.S.
The study, which appears online today in the journal Pediatrics, is the first study to evaluate the validity of these head-circumference curves in the U.S. primary care population.
Head-circumference growth curves are used by pediatricians millions of times annually to identify children who may need extra follow-up or testing because of a large head, also called macrocephaly, or a rapidly growing head. Many children with macrocephaly or a rapidly growing head are healthy, but in some children these findings signify a problem.
“The study was done because primary care doctors were concerned that the curves weren't working well for our patients,” said study leader Carrie Daymont, MD, a pediatric researcher formerly at The Children’s Hospital of Philadelphia.
The researchers used the electronic health records of more than 75,000 infants in the Children’s Hospital primary care network between August 2001 and January 2008. They compared the most widely used charts with actual measurements and found that the most commonly used chart would lead to a diagnosis of macrocephaly, defined as a head circumference above the 95th percentile for age, in 2.5 times more 1-year-olds than expected.
The American Academy of Pediatrics recommends that all children have head-circumference measurements done at each “well child” visit until the age of 2. Doctors in the U.S. plot each measurement on the growth chart from The Centers for Disease Control and Prevention (CDC). The World Health Organization (WHO) made charts in 2006 that are used in many countries.
Neither the CDC nor the WHO charts worked well in the study population, the researchers said. A chart made in 1977 by the National Center for Health Statistics (NCHS) did work well in the study population, classifying 5 percent of children above the 95th percentile. The NCHS growth chart was replaced by the CDC growth chart in 2000.
When using the CDC curves, the percentage of children diagnosed with macrocephaly changes with age. One-fifth of the expected number of one-month olds and 2.5 times the expected number of 1-year-olds were classified as having macrocephaly, or being above the 95th percentile. Many of these older infants above the 95th percentile may have received unnecessary follow-up tests and specialty referrals, and their parents may have been unnecessarily worried, the researchers said.
“We were noticing that we had a larger-than-expected number of children labeled as having large and rapidly growing head circumferences,” Daymont said.
Study results lead to development of new growth curve
Using the CDC curves may cause a delay in diagnosis in babies under two months old with diseases causing large heads, and may cause children six months to 3 years to be labeled as having large or rapidly growing heads, when in fact they are healthy and have a typical head size, Daymont said. Use of the WHO curves would cause an excess of children to be labeled as having large heads at all ages. The researchers at Children’s Hospital used the data from this study to develop a new growth curve.
“More research is also needed to determine how doctors can use head circumference growth curves to identify children with problems without causing unnecessary worry to parents of healthy children,” said David Rubin, MD , MSCE, a pediatric researcher at Children’s Hospital and a senior author on the study. “We are currently working on research related to this question.”
The team at Children’s Hospital is currently evaluating other sources of data to see if the new head-circumference growth curves work in other populations.
Other study authors are Chris Feudtner, MD, MPH, PhD, of Children’s Hospital, and Wei-Ting Hwang, PhD, of the Center for Clinical Epidemiology and Biostatistics at The University of Pennsylvania.
Head-Circumference Distribution in Large Primary Care Network Differs from CDC and WHO Curves, Pediatrics, published online Sept. 20, 2010. Read the abstract.