Evaluating the short child (part 1 of 2)

By David Wyatt, MD

Growth is a process that reflects the vigor of the organism. It is a faithful proxy for the health and nutritional status of the child. Accurate monitoring of a child's growth is an essential part of well childcare and therefore should be a part of every well child visit. A robust growth record, properly interpreted, is reassuring. An abnormal growth pattern, recognized, may be an early warning of a clinically significant but as yet undiagnosed condition.

Developing a functional growth record requires proper equipment, good technique and consistent record keeping. A wall-mounted stadiometer or a supine length device, when used correctly, can provide the accuracy needed for proper growth assessment. Weight, height (or length) and BMI should be obtained and plotted at least once yearly on growth charts based upon the National Center for Health Statistics surveys (http://www.cdc.gov/growthcharts).

Once a child has been measured and plotted, data can be compared to national standards – the child's peers. Height also should be interpreted in light of the genetic potential – the child's parents. But most importantly, the height data should be evaluated as a pattern. We will consider each of these comparisons below.

Peers

Comparison to a national U.S. cohort permits a larger view of the child's growth; that is, how current height corresponds to many other children of similar background. The database consists of a broad statistical sampling of U.S. children representing common medical and nutritional influences. The less a particular child fits the U.S. profile, the less useful the comparison to these standards. For the great majority of children, however, these standards will allow a meaningful "first look."

The "normal range" is a statistical concept with many theoretical boundaries. Commonly used height-for-age charts show centile bands extending from the 5th to the 95th centile for the population. This implies that children outside of these centiles may be abnormal. Of course, 1 in 20 children will be below the 5th centile, and the great majority of these children will be healthy and "normal." Newer growth charts add negative standard deviation lines below the 5th centile. These lines allow a more precise assessment of the shortest children. Any child more than three standard deviations below the mean will fall in the lowest 1 percent and is more likely to have an abnormal medical or nutritional condition. Children referred to pediatric endocrinologists for short stature have a mean height standard deviation of -2.6. Even within this population, however, most of the children are healthy.

A single height measurement, then, provides some estimate of the child's health. Much more information can be added if the genetic potential of the child can be assessed.

Parents

Tall parents tend to have tall children, and short parents tend to have short children. Indeed, "tall marries tall and short marries short." Nevertheless, the extremes of height within the population do not increase because of a phenomenon termed regression to the mean. That is, tall parents tend to have tall children who are somewhat shorter than the parents and short parents tend to have short children who are somewhat taller than the parents. Furthermore, within any family, there is a large biologic variability in height of the children. For these reasons, estimation of the true genetic target range for a set of parents, the range within which, as an example, 90 percent of children would be expected to fall, is complex. In general, a calculation of the mid-parental target height is based upon the parental heights and corrected for the sex of the child. Then a confidence interval describing the target range is calculated around that target height. This range must allow for the normal standard deviation of the population and the regression to the mean within a family. In the Endocrine Clinic at Children's Hospital of Wis-consin, the calculations are as follows:

target height (cm) for males =

(father's height + mother's height +   13 cm)/2

target height (cm) for females = (father's height - 13 cm + mother's height)/2

genetic target range =

target height ± 9 cm

If a child's height falls below the lower end of this target range, it is much more likely to represent abnormal growth, no matter the actual height centile. For example, a child who is at the 15 percent is well within the normal population range and compares well with his peers. But if he has tall parents and his genetic target range extends from the 30 to 90 percent, he does not compare well with his genetic potential. In this way, assessing the genetic target range adds significantly to the evaluation. Comparing the child to his peers and to his genetic potential yields an increasingly accurate estimate of normalcy. The most revealing comparison, however, is the current height to the child's own prior growth pattern.

Pattern

Except for the beginning (first two years) and end (adolescence), growth on an annual basis is remarkably steady. There are seasonal variations (faster growth in spring and summer), but from year to year, a child should stay close to the same isopleth (height centile line) between age 3 and adolescence. Acceleration above or deceleration below that channel often indicates abnormality. Note that acceleration can be as worrisome as a decline, since it may indicate precocious puberty, hyperthyroidism or other abnormal conditions that stimulate growth.

An abnormal fall (or rise) across height centiles cannot simply be quantified as "across two lines," since the commonly used lines are not evenly spaced. For example, a two-line decline from the top centile (95th to 75th) represents a decline of 20 centiles; a fall between the next two lines (90th to 50th) represents 40 centiles; and a fall between the next two lines (75th to 25th) represents 50 centiles. Assuming accurate measurements, any decline of more than 20 centiles over a full year should raise concern. This applies even if the current height centile is well within the "normal" range.

In the first 2 years, about one in three healthy infants will show a 10 to 30-centile decline. This represents the normal shift from the intrauterine environment to the true genetic inheritance, and makes it much more difficult to detect abnormal growth patterns. In general, if an infant is healthy and well-nourished, a moderate decline may simply be monitored. After the first 2 or 3 years, however, all children should have "found their channel" and should show no further decline.

All children show a deceleration prior to the pubertal growth spurt. Those who mature later show a greater deceleration and will have a greater dip on their growth chart. The great majority of these children are normal and will reach a normal adult height once puberty begins. If they are in good general health and well-nourished, they may be reassured with a few simple screening tests (i.e., bone age X-ray, free T4 and TSH). Some may benefit from a brief course of androgens, but most do well with reassurance.

Having sufficient data to perform these three comparison – the child's current height to his peers, his parents and his pattern – is often more valuable than any laboratory test. The comparison to the prior pattern of growth is the most important of the three and only can be done with a well-maintained growth record.

Part 2 of this article, which will examine various sample growth  patterns, will appear in the Spring 2006 issue of Pediatric Rounds.

David Wyatt, MD, is program director of Endocrine at Children's Hospital of Wisconsin. He also is a professor of Pediatrics (Endocrinology) at the Medical College of Wisconsin.

For more information Endocrine Clinic (414) 266-6750

To make an appointment Central Scheduling (414) 607-5280 local or (877) 607-5280 toll-free

To earn CME credit, click here to open the CME quiz.
You may need Adobe Acrobat Reader to view these documents. Click here to download the reader for free.