Changing practice in growth monitoring

No evidence exists that monitoring height velocity is useful

The French pay child benefits only on production of a school medical certificate confirming that height measurements are up to date. If the French, why not the British? Height is easily measured, and height velocity is claimed to be a sensitive index of many disorders, genetic, metabolic, and psychosocial. More than just a marker for health, growth could be a useful tool for picking up silent disease. All that we lack is the evidence.

The dilemma of pursuing a screening programme based on a scientifically plausible but unproved hypothesis has recently been raised.1 In particular, the use of growth data to screen for silent disease in children has been the subject of lengthy debate, but there have been few cohort studies on which to judge its effectiveness, and consensus guidelines for referral have yet to emerge. Much of the debate currently relies on evidence from the Wessex growth study, which since 1986 has monitored a cohort of very short children in the community.2 This study has been unable to show the ability of repeated height measurements to identify new disease, has found height velocity to be unreliable, and has concluded that height screening at school entry is the best means of identifying silent disease in school age children.

Stature alone can be a simple and useful index of disease. Further investigation of the 147 children from the Wessex growth study who had been passed as “short normal” at school entry identified eight cases of previously unidentified disease.3 In only four was it remediable, but in all it was informative. Predictably, the proportion of children with organic disease increased with the degree of short stature. Use of the new 0.4th instead of the third centile should reduce to a minimum the numbers referred, but this has to be weighed against missed diagnoses. Indeed, two of the eight cases of silent disease in the Wessex cohort lay above the current 0.4th centile cut off for referral at school entry.

In less extreme cases of short stature referrals are sometimes controlled by a “wait and see” policy, using height velocity, often over a very short period, as a secondary screening tool.4 Children apparently growing well can then be dismissed and the rest referred for specialist advice. Growth hormone therapy may even be started on the basis of the auxological data.5

Height measurement is inevitably imprecise.6 It is unlikely, a priori, that velocity—which is not measured but only derived from height measurement—could be more informative. The Wessex study has clearly shown that, while successive heights are highly correlated, successive 12 month velocities are not.7 Velocity thus fails both to reflect previous growth and to predict future growth. Its interpretation is further complicated by the variable onset of the pubertal growth spurt and the fact that “satisfactory” growth, at any age, is conditional on height and age.8 The evidence is that the height velocity chart, often promoted as a better means of evaluating growth than height chart alone, has no place in community paediatrics.

The insecurity felt by many in abandoning familiar measures is understandable. In addition to the short children, however, the Wessex study has followed a cohort of 140 average height controls over the past 12 years. Only three have acquired disease—two diabetes and one hypothyroidism. At no time did the growth of either falter, nor did any of the remainder of the cohort cross a height centile band, at least not before puberty. This should lay to rest any concern that important new disease will be missed after the age of 5 without further measurements. The most recent guidelines concede that velocity estimates are unreliable but nevertheless propose that the height charts of short children should be checked for movement across centile bands.9 The evidence, however, is that the sensitivity and specificity of these measures in identifying silent disease is inadequate.10

It remains unclear whether more frequent height checks from infancy, a greater awareness of signs and symptoms other than short stature, or a single measurement of height would best identify growth related disease. The evidence so far suggests that the single measurement at school entry is the most sensitive anthropometric marker for silent disease.3 At that age very short stature must result from sustained slow growth. Further proof of growth failure is unnecessary and any concern about disease missed is best addressed, not by awaiting additional measurements, but by improving clinical acumen. Short, but otherwise healthy, school entrants can also be reassured that they are no more likely to become ill than their taller peers.8 To be spared repeated height checks in school would come as a relief to many. The debate will undoubtedly continue unless resolved by a large scale prospective community study. Until then there is no evidence that growth monitoring, as opposed to height screening, is a cost effective use of scarce resources.