Procedures for Accurate Measurement

Accurate measurement is a critical component of growth assessment. Weight, in pounds or kilograms, must be determined using an accurate scale. For infants and toddlers, weight, length, and head circumference are obtained. These measures should be performed with the infant naked, and ideally, repeated measures should be performed on the same equipment. Head circumference is determined using a flexible tape measure run from the supraorbital ridge to the occiput in the path that leads to the largest possible measurement. Length is most accurately measured by two examiners (one to position the child), with the child supine on a measuring board. For older children, the measure is stature or height, taken without shoes, using a stadiometer. Measurements obtained in alternative manners, such as marking examination paper at the foot and head of a supine infant, or using a simple wall growth chart with a book or ruler on the head can lead to inaccuracy that may render the measurement useless. It is essential to compare measurements with previous growth trends, repeat any that are inconsistent, and plot results longitudinally.

Derivation and Interpretation of Growth Charts

In 2000, the Centers for Disease Control and Prevention (CDC) published new growth charts, replacing the 1977 version. Modifications since then have not changed the data points. Set 1 includes the 5th to 95th percentiles; set 2, the 3rd to 97th percentiles. These charts contain data from national surveys conducted by the National Center for Health Statistics between 1963 and 1994. Data are representative of the U.S. population, both demographically and in terms of breast-feeding prevalence. Methodological steps have assured that the increase in the prevalence of obesity has not unduly raised the upper limits of normal. Several deficiencies of the older charts have been corrected, such as the over-representation of bottle-fed infants and the reliance on a local data set for the infant charts. The disjunction between length and height, when moving from the infant curves to those for older children, no longer exists. The charts include curves for plotting BMI for ages 2-20 yr rather than weight for height, facilitating identification of obesity.

The data are presented in 5 standard gender-specific charts: (1) weight for age; (2) height (length and stature) for age; (3) head circumference for age; (4) weight for height (length and stature) for infants; and (5) BMI for age for children over 2 yr of age (Fig. 13-1; also see Figs. 9-1 and 9-2). The charts are available at www.cdc.gov/growthcharts/.

Each chart is composed of percentile curves, representing the cross-sectional distribution of weight, length, stature, head circumference, weight for length, or BMI at each age. The percentile curve indicates the percentage of children at a given age on the x-axis whose measured value falls below the corresponding value on the y-axis. On the weight chart for boys 0-36 mo of age (see Fig. 9-1A), the 9 mo age line intersects the 25th percentile curve at 8.6 kg, indicating that 25% of the 9 mo old boys in the National Center for Health Statistics sample weigh less than 8.6 kg (75% weigh more). Similarly, a 9 mo old boy weighing more than 11.2 kg is heavier than 95% of his peers. The median or 50th percentile is also termed the standard value, in the sense that the standard height for a 7 mo old girl is 67 cm (see Fig. 9-1B). The weight-for-height charts (see Fig. 9-2) are constructed in an analogous fashion, with length or stature in place of age on the x-axis; the median or standard weight for a girl measuring 110 cm is 18.6 kg.

For infants, the revised charts represent observed but not necessarily optimal growth because they still incorporate data from many bottle-fed infants. Rates of initiation of breast-feeding in the USA have more than doubled from 26% in 1970 to 74% in 2005, but only 12% of infants are exclusively breast-fed for 6 mo and only 21% of infants receive breast milk for a yr. Compared with current standards, an exclusively breast-fed infant would be expected to plot higher for weight in the first 6 mo, but relatively lower in the second half of the 1st yr. Awareness of this growth difference should prevent overidentification of growth problems in breast-fed infants.

In an effort to set an internationally usable standard for optimal growth in young children, in 2006 the World Health Organization released growth charts based on the Multicenter Growth Reference Study (MGRS). Rather than describing the growth of typical children, the MGRS describes the growth of children who are predominantly breast-fed and raised under optimal conditions. Six study sites representing 5 continents were included: USA, Brazil, Norway, Ghana, Oman, and India. Use of the new charts in developing nations results in identification of many more children as malnourished and eligible for therapeutic feeding programs. Their use in the USA results in many fewer infants being identified as underweight (comparison of curves shown in Fig. 13-2). Charts are available online at www.who.int/childgrowth/standards/en/.

Figure 13-2 Comparison of the WHO and CDC length/height-for-age z score curves for boys.

(From de Onis M, Garza C, Onyango AW, et al: Comparison of the WHO child growth standards and the CDC 2000 growth charts, J Nutr 137[1]:144–148, 2007.)

For adolescents, caution must be used in applying cross-sectional charts. Growth during adolescence is linked temporally to the onset of puberty, which varies widely. By using cross-sectional data based on chronological age, the charts combine youth who are at different stages of maturation. Normal variations in the timing of the growth spurt can lead to misdiagnosis of growth abnormalities. The data for 12 yr old boys include both early-maturing boys who are at the peak of their growth spurts and late-maturing ones who are still growing at their prepubertal rate. The net result is to artificially level off the growth peak, making it appear that adolescents grow more gradually and for a longer period than they do. When additional precision is necessary, growth charts derived from longitudinal data, such as the height velocity charts of Tanner and colleagues, are recommended.

Specialized charts have been developed for U.S. children with various conditions, including very low birthweight and prematurity; Down, Turner, and Klinefelter syndromes; cerebral palsy; and achondroplasia. In addition, growth charts for children of distinct ethnic groups or nationalities may be found on the World Wide Web.

Body mass index for age complements the standard growth charts for children over 2 yr of age. BMI can be calculated as weight in kilograms/(height in meters)² or weight in pounds/(height in inches)² × 703, with fractions of pounds and inches expressed as decimals. Values may be plotted on standard BMI charts (see Fig. 13-1). These calculations can be easily performed electronically using a variety of desktop and hand-held devices. BMI percentile varies with age over childhood: a 6 yr old girl with a BMI of 21 is overweight, whereas a 16 yr old girl with the same BMI is just above the 50th percentile.

Height velocity charts, which evaluate the rate of growth per yr, are considered by many to give a more sensitive and specific indicator of abnormal growth. They are used primarily by pediatric endocrinologists.

Although many parents think it is important to see growth charts, a recent study shows they often misinterpret their meaning. Clinicians are cautioned to provide clear interpretation when using growth charts as visual aids.

Analysis of Growth Patterns

Growth is a process rather than a static quality. An infant at the 5th percentile of weight for age may be growing normally, may be failing to grow, or may be recovering from growth failure, depending on the trajectory of the growth curve. Infants may lose up to 10% of their birth weight in the 1st wk of life and regain it by the end of the 2nd wk. They will then gain steadily at a rate of 20-30 g/day for the first 3 mo. Table 13-1 gives typical growth and calorie requirements for children through age 6 yr. Formulas are available for the estimation of average height and weight and height for children of various ages, but given their complexity and the easy availability of growth charts, use of the latter is preferable.

Despite the facts that the National Center for Health Statistics (NCHS) charts represent cross-sectional rather than longitudinal data and that children tend to grow in spurts, most children tend to track along a percentile, referred to as “following the curve.” A normal exception commonly occurs between 6 and 18 mo of life. For full-term infants, size at birth reflects the influence of the uterine environment; however, size at 2 yr correlates with mean parental height, reflecting the influence of genes. Between 6 and 18 mo of age, infants may shift percentiles upward or downward toward their genetic potential. Thereafter, most children will track along a growth percentile, with variation within two large percentile bands (a small infant might track between the 5th and 25th percentiles, a large one between the 75th and 95th). This tracking often represents the mid-parental height and a corresponding weight, where mid-parental height is calculated in inches as follows:

It is important to correct for various factors in plotting and interpreting growth charts. For premature infants, overdiagnosis of growth failure can be avoided by using growth charts developed specifically for this population. A cruder method, subtracting the weeks of prematurity from the postnatal age when plotting growth parameters, does not capture the variability in growth velocity that very low birthweight (VLBW) infants demonstrate. While VLBW infants may continue to show catch-up growth through early school age, most achieve weight catch-up during the 2nd yr and height catch-up by 2.5 yr, barring medical complications (Chapter 91). For children with particularly tall or short parents, there is a risk of overdiagnosing growth disorders if parental height is not taken into account or, conversely, of underdiagnosing growth disorders if parental height is accepted uncritically as the explanation.

The analysis of growth patterns and the detection of aberrant growth patterns provide critical information for the detection of pathologic conditions. Calculation of daily and monthly growth, such as weight gain in g/day (see Table 13-1), allows more precise comparison of growth rate to the norm. Weight loss, or failure to gain normally, is often the first sign of pathology.

The diagnosis of failure to thrive (Chapter 38), usually a diagnosis of children under 3 yr of age, is considered if a child’s weight is below the 5th percentile, if it drops down more than 2 major percentile lines, or if weight for height is less than the 5th percentile. Weight for height below the 5th percentile remains the single best growth chart indicator of acute undernutrition. A BMI less than the 5th percentile also indicates that a child is underweight. Brief periods of weight loss or poor weight gain are usually rapidly corrected and do not permanently affect size. Children who have been chronically malnourished may be short (stunted) as well as thin, so that their weight-for-height curves may appear relatively normal. Chronic, severe undernutrition in infancy may depress head growth, which may be an ominous predictor of later cognitive disability. Low weight for age or height or weight loss may be referred to as wasting.

When growth parameters fall below the 5th percentile, values can be expressed as percentages of the median, or standard, value. A 12 mo old girl weighing 7.1 kg is at 75% of the median weight (9.5 kg) for her age.

Another way to evaluate weight is to determine the ideal body weight for height and compare the current weight to the ideal body weight for length or height. A 15 mo old boy who is 79 cm is at the 50th percentile. The ideal weight is 12 kg. If he weighs 8 kg (<5th percentile), he is 67% of ideal body weight, an indication of severe wasting. Table 13-2 provides interpretation of percent ideal body weight from obese to severe wasting.

Table 13-2 INTERPRETATION OF PERCENT OF IDEAL BODY WEIGHT

Extremes of height or weight can also be expressed in terms of the age for which they would represent the standard or median. For instance, a 30 mo old girl who is 79 cm (<5%) is at the 50th percentile for a 16 mo old. Thus the height age is 16 mo. Weight age can also be expressed this way.

Linear growth deficiency (stunting) is more likely to be due to congenital, constitutional, familial, or endocrine causes than to nutritional deficiency (Chapter 43). In endocrine disorders, length or height declines first or at the same time as weight; weight for height is normal or elevated. In nutritional insufficiency, weight declines before length, and weight for height is low (unless there has been chronic stunting). Figure 13-3 depicts typical growth curves for 4 classes of decreased linear growth. In congenital pathologic short stature, an infant is born small and growth gradually tapers off throughout infancy. Causes include chromosomal abnormalities (Turner syndrome, trisomy 21; Chapter 76), perinatal infection (TORCH), extreme prematurity, and teratogens (phenytoin, alcohol) (Chapter 90). In constitutional growth delay, weight and height decrease near the end of infancy, parallel the norm through middle childhood, and accelerate toward the end of adolescence. Adult size is normal. In familial short stature, both the infant and the parents are small; growth runs parallel to and just below the normal curves.

Figure 13-3 Height-for-age curves of the four general causes of proportional short stature: postnatal onset pathologic short stature, constitutional growth delay, familial short stature, and prenatal onset short stature.

(From Mahoney CP: Evaluating the child with short stature, Pediatr Clin North Am 34:825, 1987.)

Obesity affects large numbers of children. Growth charts can confirm an impression of obesity if the weight for height exceeds 120% of the standard (median) weight for height. According to the CDC, a BMI over the 95th percentile indicates obesity and a BMI between the 85th and 95th percentiles indicates overweight. Although widely accepted as the best clinical measure of under- and overweight, BMI may not provide an accurate index of adiposity, because it does not differentiate lean tissue and bone from fat. Measurement of the triceps, subscapular, and suprailiac skinfold thickness can be used to estimate adiposity; considerable experience is needed for accuracy. The American Academy of Pediatrics Nutrition Handbook, 6th edition, questions the use of fat folds to estimate total body fat, noting that the method has not been validated in young children and that basic assumptions of the method, that subcutaneous fat is a marker of total fat and that measured sites represent average skin fat thickness, are not true. Other methods of measuring fat, such as hydrodensitometry, bioelectrical impedance, and total body water measurement are used in research, but not in clinical evaluation.

Other Indices of Growth