The potential effect of the UK 1990 height centile charts on community growth surveillance.

The UK 1990 height charts are derived from an up to date dataset and introduce a change in the centile lines, particularly the addition of the 0.4th centile. This study examined the likely impact of these changes. Height data from London school children (1990-1993) were examined using Tanner and Whitehouse (TW) and UK 1990 charts. Numbers of children with height below TW 3rd centile were compared with numbers below the UK 1990 3rd and 0.4th centiles. The TW charts identified only 1% of children below the TW 3rd centile, while the UK 1990 charts identified 3% below the 3rd and 0.4% below the 0.4th centiles. If the 3rd centile remains as the referral 'cut off' for short stature, the introduction of the UK 1990 charts would increase current workload two- to three-fold, while a change to the 0.4th centile would reduce it by 50%. A significant number of children with abnormalities may be excluded from further assessment as a result of this latter change. In this small scale community study it is not possible to assess the consequences of this change. The heights at diagnosis of children with growth hormone (GH) deficiency (peak GH < 20 mU/l during a standard provocation test) were therefore compared to the 0.4th centile (UK 1990 charts). Sixty eight children with heights < 2nd centile (UK 1990 charts) currently receiving GH replacement (17 female, 51 male, aged 9.7, SD 3.5, years) were assessed, and of these, 28 (41%) had heights at diagnosis between 0.4th and 2nd centile, with a mean height standard deviation score of -2.32 (SD 0.21). This suggests that if the 0.4th centile were to be used as the sole criterion for referral for slow growth, a significant proportion of children with abnormality would not be referred for further assessment. The UK 1990 2nd centile should replace the TW 3rd centile. Children below this should undergo an intermediary medical assessment to confirm height measurement, to exclude from referral children with mild familial short stature and to identify concerns regarding the child.