Contents - Previous - Next

Discussion

In discussion of the relationships between size at birth and later growth the first question that arose was the precision with which length can be measured in neonates. It was agreed that the measurement is not easy; it needs two people and duplicate readings, and in Guatemala mothers refused to accept the procedure (Martorell). Routine measurements are usually not very reliable, even in Europe. On the other hand, many participants felt that if health professionals were adequately trained and used proper apparatus, length could be measured at birth with an acceptable degree of accuracy. Some studies have reported a coefficient of reliability of 0.99, which is as good as can be achieved in adults. In Sweden, where birth length is recorded regularly, an analysis of 5000 records showed that, in spite of the errors, length at birth was a slightly better predictor of shortness at 7 years than weight at birth. Other measures may be considered as proxies for length; for example, arm length measured with a tape correlated well with total length (Martorell). Foot length might also be useful. The new technique of knemometry was also mentioned, but it seems that even those who are using it every day are not happy about its reliability in very young children, certainly not for routine measurements (Lodeweyckx). Finally, it was suggested that the measured length may be affected by the state of hydration of the soft tissues. In Jamaica children with kwashiorkor lost 2-3 cm in length during the first week, as they lost fluid (Golden). The neonate has a higher water content than the infant, and this could affect the measurement.

Ultrasound studies, such as those described by Falkner, give information about the length of the long bones in utero. From a biological point of view, for assessing the activity or intensity of bone growth, i.e. the rate of cellular multiplication, it may be more useful to look at fractional rather than absolute gains. For example, in puberty the absolute rates of gain are quite large, but this gives a misleading impression of the intensity of bone growth, because the fractional or proportional rate of gain is much less than in the young infant. Looked at in this way, there is a deceleration of growth during the later stages of pregnancy, which continues into the smooth curve of deceleration that is seen after birth.

Falkner made the distinction between symmetrical or proportional SGA, in which weight, length and head circumference are all reduced, and asymmetrical or disproportionate SGA, in which head circumference, and to a lesser extent, length are spared. It would be an oversimplification to suppose that an insult early in pregnancy affects length growth and an insult late in pregnancy affects weight growth. What is important for subsequent performance, according to Falkner, is whether growth of the brain, and hence the head, is impaired. Placental dysfunction tends to occur later in pregnancy; head growth is spared and the result is an asymmetrical SGA infant. There is ample experimental evidence for the importance of the timing of the insult. For example, in Widdowson's classical experiments on rats, long-term underfeeding in pregnancy resulted in permanent stunting of the offspring, whereas with short-term restriction followed by refeeding they all caught up (Widdowson & McCance, 1960). (Note: in the inter-generational studies of Stewart et al. (1980) on rats, marginally undernourished for some 20 generations, the pups were always small, but of normal weight for length. Ed.) In other experimental studies, irradiation of the uterus early in pregnancy resulted in retardation of longitudinal growth without catch-up, whereas later irradiation did not have this effect (Golden). Follow-up of children from mothers in Hiroshima who had been exposed to radiation early in pregnancy showed that they were short and remained short.

Failure to distinguish between symmetrical and asymmetrical SGA may account for some of the contradictions that came up. For example, in one of the infant studies in Pakistan, only 3% were below -2 SD in length, whereas 30% had a low weight for length (Karlberg). On the other hand, in the CRSP studies in Mexico and Kenya, length was more reduced than weight (Allen).

Discussion followed on the determinants of size at birth. One contention was that low birth weight, which is so common in developing countries, should not be regarded as abnormal; the size of the baby is appropriate for the size of the mother. An inter-generational effect has been described. A baby born small becomes a small woman; if her baby is then small, it reduces the risk of complications during delivery. On the other side it was argued that Asians in the US have an unusually low proportion of SGA infants (Martorell). This would suggest that the size of the infant at birth reflects the nutritional state of the mother ¹ rather than her size per se. It would not be profitable in the present state of knowledge to speculate about particular nutrients that may be involved. Infections and toxins may also play a part.