Is obstetric and neonatal outcome worse in fetuses who fail to reach their own growth potential?

Sir, Danielian et al. (1992) draw attention to the important distinction between smallness-for-gestational age (SGA) and growth retardation but their argument is blunted by inaccurate methodology for calculating birthweight centiles. They use Altman & Coles’ (1980) nomograms without paying attention to maternal size, even though these nomograms, as well as the original Aberdeen data which these were based on (Thomson et al. 1968) were published together with adjustment tables for maternal height and weight for correct calculation of term birthweight centiles. The lack of regard for these tables when assessing previous birthweights may be understandable in a busy booking clinic; their disregard in a research project comparing centiles with outcome is not. A further, well documented variable affecting birthweight is race or ethnic group, and adjustment for this factor before using the nomograms has also been advocated (Grundy et al. 1978). In the absence of pathology such as smoking or maternal malnutrition (which should be excluded by history, examination and ponderal index if suspected), such physiological variation is important to recognise for the accurate assessment of birthweight. We have shown that disregard for maternal height, booking weight and ethnic group makes the assessment of the birthweight centile inaccurate, and overor under-diagnoses SGA in at least a quarter of babies (Gardosi et al. 1992). This has recently been confirmed in a large audit of deliveries in this region (Wilcox et al. in press). Third trimester ultrasound-estimated fetal weight centiles, as used in the study by Danielian et al. (1992), would be subject to constitutional variation due to maternal characteristics. Therefore, projecting these for comparison with unadjusted term birthweight centiles may be misleading and detract from the main issue, ie the importance of growth over absolute size. Another problem with this study is its small sample size, resulting for example in only 10 SGA babies available for analysis. The authors fail to explain why only 197 ‘unselected’ cases were included from the 8 month trial period of this retrospective study, this being perhaps a tenth of the actual deliveries normally expected during an interval in their unit. The question of growth v. perinatal outcome is an important subject and its study requires thorough investigation. J. Gardosi Senior Lecturer/Honorary Consultant Department of Obstetrics & Gynaecology Queen’s Medical Centre Nottingham NG7 2UH