Contents - Previous - Next

Prenatal influences on postnatal growth: Overview and pointers for needed research

1. Introduction and background
2. Fetal growth
3. Intrauterine growth retardation
4. Small-for-gestational-age infants
5. Genetic and environmental factors
6. Reference values for fetal growth

F. Falkner 1, W. Holzgreve 2 and R.H. Schloo 2

1 Maternal and Child Health Program, and Department of Pediatrics, University of California, Berkeley and San Francisco, CA, USA;

2 Department of Obstetrics and Gynecology, University of Münster, Germany

Correspondence to: F. Falkner.

1. Introduction and background

In many discussions of linear growth retardation there is a need to address arguably the most important phase of human growth - growth from conception to term. Only comparatively recently have studies focused on fetal growth and outcome after gleaning important data from intrauterine study - difficult though such studies are.

Questions that come to mind include: Can children be programmed in utero to be linearly growth retarded after birth? What are the relationships between intrauterine growth retardation (IUGR) and postnatal growth? Depending on its growth in utero, the newborn baby can be categorized as a healthy full-term (FT) infant within the normal birth weight range; a macrosomic infant above the normal birth weight range; or an infant of low birth weight (ILB). Regarding the latter, WHO estimates are of great interest: they show that approximately two-thirds of all infants of low birth weight born in the developed world are true pre-term (PT) infants and one-third are small-for-gestational-age (SGA). This relationship is reversed in the developing world, where about 75% of ILB are SGA. The much higher proportion of SGA infants in the Third World seems to be due primarily to malnutrition and infection and should therefore be preventable.

2. Fetal growth

During the embryonic period the rate of weight gain of the fetus is not very high. Its subsequent increase and decrease are mainly a reflection of cell multiplication. A period of linear growth peak velocity, as in adolescence, can be assumed, but the gestational age at which it occurs is still somewhat uncertain. Thence, linear growth deceleration occurs and continues in the postnatal period. Weight gain of the fetus follows the same general pattern as growth in length, except that peak weight velocity is thought to occur later.

There is a discontinuity between fetal and infant growth curves. Adaptation appears to occur. There is considerable evidence that, beginning at 34-36 weeks, fetal growth slows down, owing to space constraints within the uterus. Twins slow down demonstrably earlier when their combined weight is approaching that of a 36-week singleton fetus.

A hypothetical curve of weight gain velocities connecting a peak weight velocity point at, say, 32 weeks, with a weight gain velocity point 8 weeks post partum, clearly takes no account of the slowing in growth just described. Infants usually lose some weight immediately after birth which they then rapidly regain. On average, catch-up growth is greatest in those infants most delayed in utero. As a result, there is a significant negative correlation between birth weight and weight gain in the early postnatal months. The same phenomenon occurs for length; the smaller the infants, the more, on average, they grow in the postnatal period.

It might be expected that any maternal factor regulating fetal growth would act protectively to prevent serious depletion of maternal nutrient and mineral supplies by slowing down fetal growth. But near normal fetal growth even in severe maternal malnutrition would indicate that there are placental and fetal growth-controlling factors that operate normally under these circumstances. How is it, then, that especially the nutrition component of poor and adverse environmental conditions is associated with lowered human birth weight? This is largely due to factors causing a marked reduction in growth rate after 36 weeks, for the mean birth weight of infants born at 36 weeks in many parts of the world under different circumstances is quite similar.

A mother not achieving her own growth potential, due to malnutrition or other adverse environmental factors, is likely to have smaller fetuses and newly born infants than had she grown up in good circumstances. Clearly, two generations or more may be needed to reverse the effects of a poor environment on size at birth. Guatemalan mothers of short stature had babies of smaller size than did mothers of medium stature. A food supplement given to both samples of mothers during pregnancy increased the birth size of the small mothers' infants more than the size of infants of larger mothers. It did not eliminate the difference completely, but nearly so (Lechtig & Klein, 1980).

3. Intrauterine growth retardation

It is not intended to discuss causation of IUGR, but only to consider its possible relationships with postnatal growth.

Utero-placental and umbilical blood flow can be important factors, as can the transfer of glucose through the placenta or the production of fetal insulin. The roles of fetal pituitary growth hormone and fetal thyroid hormone need continuing study; anencephalic and athyroid fetuses, for example, seem not to exhibit fetal growth retardation (Vorherr, 1982). Placental lactogen and somatomedin and a somatostatin-like substance of the placenta and fetus may well influence fetal growth, and animal experiments certainly suggest the assistance of placental-fetal growth-promoting and - controlling factors. Their identification in the human would contribute much to our knowledge of fetal growth, health and adaptation.

Regarding the placenta, Lechtig et al. (1975; 1977) indicated that nutritional supplementation of the mother during pregnancy is associated with improved placental weight and higher levels of alkaline ribonuclease activity. Leaf et al. (1992), in a very recent study of essential fatty acids at birth, showed that placental function is important in the transfer of some fatty acids from mother to fetus, and that these fatty acid levels were correlated with fetal growth and maturation in the premature infants they studied. Clearly, the placenta has very complex metabolic and endocrine functions. IUGR associated with placental dysfunction tends to occur later in pregnancy and head growth appears spared, whereas early onset IUGR tends not to spare head growth (see later discussion on SGA infants).

It is well established that malnutrition and infection can cause IUGR. During the past decade it has been documented that prenatal intrauterine infection increases fetal IgM. Mekki et al. (1988) and Lechtig et al. (1974) sampled cord blood and showed that IgM levels were higher in newly born infants from very poor environmental backgrounds.

There is certainly a pressing need for detailed studies on the role of infection in IUGR, particularly in tropical and underdeveloped countries. Pardi et al. (1993) showed that techniques for sampling cord blood in utero (cordocentesis) offer the opportunity to assess a fetus's metabolic environment before parturition. In fetuses with IUGR, the detection of, for example, hypoxia, acidemia, low amino-acidemia, and endocrine abnormalities, creates a greater opportunity to study many aspects of not only IUGR but also fetal growth. Kempley, Gamso & Nicolaides (1993), using Doppler ultrasound, measured left renal artery blood flow in the first postnatal week of very low birth weight SGA infants. Compared with weight-and gestation-matched controls, SGA infants had significantly lower blood flow velocity. Thus, abnormalities of blood flow velocity appear to persist after delivery in those SGA infants.

Doyle et al. (1990) found positive correlations between maternal nutrient intake, assessed during one week towards the end of the first trimester of pregnancy on the one hand, and weight, length and head circumference of the newborns on the other, especially in infants with birth weights below 2500 g. Vitamin/mineral supplementation of mothers during the last two trimesters of pregnancy, however, had no significant effects on birth dimensions.

4. Small-for-gestational-age infants

Prenatal intervention that would prevent low birth weight infants who are SGA would be a major contribution towards perinatal health improvement. We need, then, to search not only for causes of SGA (of which there are clearly many), but also for methods for predicting their linear growth. On the one hand, there are SGA infants growing very poorly after birth; on the other there are SGA infants who grow very well and achieve their genetic potential. It is commonly, and quite wrongly, taught that all SGA infants with growth retardation tend to catch-up in early months, though often not completely. Some do and some do not; clearly SGA infants are by no means an aetiologically homogeneous group.

Gruenewald (1963) 30 years ago, and Warshaw (1992) recently, together with Miller (1992), point out that it is inappropriate to diagnose SGA on birth weight alone. They describe what they call symmetrical and asymmetrical intrauterine growth retardation (perhaps the terms proportional/disproportional are more appropriate). In the symmetric type, body weight, body length and head circumference are all proportionately small at birth. In these SGA infants postnatal growth was poor/sluggish.

Asymmetric SGA infants have low weight and length, but normal head circumference. Most of these SGA infants showed postnatal catch-up growth and grew well. A small proportion exhibited some catch-up but, at least by two years, had not yet reached the 5th centile.

An important factor in all SGA infants is whether the soft tissue mass is reduced. Measurement of the above dimensions in all SGA infants at birth is clearly important as is, of course, the calculation of body mass index.

5. Genetic and environmental factors

The Louisville Longitudinal Study of Twins (Falkner & Matheny, 1993) provides evidence of a strong positive correlation between placental weight and birth weight. In monozygous twins, who are phenotypically similar, differences in birth weight and in postnatal growth may be determined by placental weight. It was found that the concentrations of several biochemical and nutritional substances were the same in the placental segments belonging to each twin, but because of the difference in placental weights the absolute amounts were different. Could it be that there is a critical placental mass, below which postnatal growth deficit occurs?

Fig. 1. Six monthly increments in length and height for Twins A & B.

A pair of male monozygous twins from the Louisville Study illustrates many features relevant to this discussion. The smaller twin's (Twin A) birth weight was 52% of that of his larger brother (Twin B). Twin A's placental segment represented only 46% in mass of Twin B's. Most importantly, with both twins born at 39 weeks' gestation, Twin A was a SGA infant.

Fig. 1 shows that Twin A exhibited marked catch-up growth in length during the first nine months to one year in an attempt (hypothetical) to compensate his length deficit of 7.0 cm at birth. Thereafter, both twins grew at approximately the same velocity. However, Twin A never exhibited sufficient catch-up for a long enough period to achieve his co-twin's stature. Both twins started their adolescent growth spurts at the same age (16 years). The within-pair difference was then 8.0 kg in weight and 5.4 cm in stature.

Because head circumference is a good indicator of brain mass (Brandt, 1986; Dobbing & Sands, 1978), it needs to be taken into account in any discussion of IUGR. Placental factors should also always be considered.

6. Reference values for fetal growth

Having noted the absence of any longitudinal growth data on the human fetus, we decided to construct curves of fetal growth. Usher (Usher & McLean, 1969) and Lubchenko (Lubchenko et al., 1970) curves are useful clinically, but as Naeye wrote in 1978: "At present there seems to be no completely valid method available to determine normal fetal growth using measurements from neonates because there is no assurance that prematurely born neonates are normally grown". We have added, after reference to the 'prematurely born' the category of small-for-gestational-age infants. Naeye went on to say: "Ultrasound and other modalities may eventually solve this problem by providing accurate, sequential in utero measurements of normal fetuses".

Regarding fetal length and body weight, was there for example a period of peak length velocity and peak weight velocity comparable to the adolescent velocity peaks of height and weight? Tanner (1950) had suggested that fetal peak growth velocity in length occurs around 20 weeks' gestation, and in weight a good deal later - around 30 weeks' gestation (Fig. 2).

We wanted to verify this and calculated, first, a fetal growth velocity curve for weight from Lubchenko's cross-sectional distance curves (Lubchenko et al., 1970). (We recognised that such calculations disobey good biometric principles, but we were searching for possible indications only.)

Fig. 3 shows indeed that there seems to be a peak weight velocity at 33 weeks. But using the same biometric procedure of calculating velocity from O'Brian & Queenan's (1981) cross-sectional data for fetal femur length, measured ultrasonically, suggests a decelerating velocity curve following a possible peak already at 16 weeks of gestation. They showed too, in a small longitudinal sample, a gradually decelerating curve from 16 to 25 weeks. Deter et al. (1987), in a similar measurement of one fetus, found a peak velocity at 16-19 weeks. Munsick's (1984) cross-sectional leg length data also suggested an acceleration of growth from 9 weeks' gestation to a peak at 15-16 weeks, then a deceleration. After calculating crown-rump velocity from pooled cross-sectional data, a peak velocity at 15-16 weeks appeared (Iffy & Deter, 1975). We started a longitudinal fetal growth study 1 measuring femur length ultrasonically. The fetal femur can be measured from 15 weeks' gestation to term, and was chosen because from this early gestational age to term it provides a stable measure of linear growth. Measures of postnatal growth outcome were also planned, together with the recording of possible growth factor influences both pre- and postnatally.

1 The project start happened to coincide with the conclusions and recommendations of a U.S.A. National Institutes of Health Consensus Conference on Ultrasonography. One of the resulting recommendations (without summarising the conferees' reasons) was that ultrasonography should not be performed on pregnant women without "a clear clinical reason". We felt, immediately that it would be inappropriate to start our project in California (U.S.A.), as originally planned.

Fig. 2. Crown-heel length velocity, cm per 4 weeks of gestational age. From Tanner (1950).

Fig. 3. Fetal weight velocity, g per 2 weeks of gestation. From Lubchenko et al. (1970) cross-sectional data. Numbers shown at each gestational age.

All pregnant women attending antenatal clinics in Germany had, and are still offered, routine ultrasonography at each visit. The Department of Obstetrics and Gynaecology, University of Münster, undertook the present study as a cooperative project with the Departments of Paediatrics, MCH, and Obstetrics and Gynaecology, University of California. The study is incomplete, but certain results and conclusions are presented here, not only as a contribution, but hopefully to show the need for further research and to provide the stimulation for such.

In addition to providing reference values for fetal linear growth, emphasis will be placed on the SGA infant, since intrauterine growth retardation (of whatever cause) may or may not affect linear growth. To date, 257 fetuses and children have been studied longitudinally from the 11th week of gestation to two years after birth. A total of 2293 separate sets of intrauterine measurements have been made on these subjects. The sample of mothers from Münster and its suburbs was reasonably homogeneous in both middle socioeconomic background and racial origin (German-Caucasian stock). Originally we were not all that surprised to note a low number of SGA infants in the study, but lately there has been an increasing proportion of SGA in immigrant mothers, not all of whom have the usual prenatal clinic visits.

6.1. Methodology

The following ultrasonographic measures were made on a fetus at each antenatal visit:

- Femur length (this measure was chosen as one representing linear growth that could be measured reasonably accurately from 15 weeks' gestation to term)

- Biparietal diameter

- Abdomino-thoracic diameter

Head circumference and body weight were calculated by formulae.

6.2. Results and discussion

Fig. 4 shows the mean reference values for femur length fitted from individual measures. Fig. 5 shows the same with seven centile curves. There were three individual ultrasonography measurers, and Fig. 6 shows the velocity curve in cm per week as a mean from all three measurers. The curve was almost identical to that of one measurer after performing 54.01% of all measurements.

Fig. 4. Fitted distance curve of femur length in cm for weeks of gestational age.

Fig. 5. Distance curves of femur length in cm for weeks of gestational age with seven centile lines.

Fig. 6. Velocity curves of femur length in cm per week of gestational age.

Fig. 7 shows fetal femur length, very close to birth, plotted against neonatal length, with a correlation coefficient of 0.813.

Fig. 7. Femur length in cm, close to birth, plotted against newborn length.

In 61 fetuses, who had femur length measurements made within one week prior to birth, we found that multiplying the femur length by 7.0538 would give a very good estimate of total body length, with a standard deviation of 0.0493. Thus, as a new predictive formula, we suggest that multiplying femur length by seven during the second half of gestation will give a good indication and prediction of total fetal and neonatal length.

Further, Falkner & Roche (1987) measured femur length on radiographs taken within two days of birth on neonates entering the Fels Longitudinal Study. Such films were available of 238 full-term infants, together with their total length measurement obtained by the study team on visiting the hospital, also within two days of birth. The correlation coefficients between femur length and total length were highly significant, both at birth and two months later. Since postnatal linear growth outcome is of importance, fetal growth reference values will be enhanced by continuing the linear growth curves into postnatal life.

6.3. Postnatal growth

This was assessed by paediatricians carrying out standard follow-up examinations at well-baby clinics.

Fig. 8 shows prenatal femur length data from 15 weeks to term and the body length of these fetuses as infants from birth to two years. Both phases show a steady deceleration. The postnatal curves of length (and weight and head circumference) are similar to those of other reference values for this age group in the Münster area.

Fig. 8. Distance curves of fetal femur lengths in cm for gestational weeks, and the body lengths of the same babies in cm for postnatal days.

6.4. Fetal body weight

We derived predictive curves for fetal weight from femur length very close to birth, neonatal biparietal diameter, length and weight, and we constructed velocity curves for fetal weight based on Hadlock's (Hadlock et al., 1985) formula for the estimation of fetal weight from fetal femur length, head and body measurements. Fig. 9 shows the mean curve with a fetal peak weight velocity around 30 weeks' gestation.

6.5. Small-for-gestational-age infants

There were 33 SGA infants, designated as such by being below the 10th centile for birth weight. This corresponds to an expected approximate 10% of the total sample. There were 21 fetuses whose femur length curves remained consistently more than 2 SD below the mean until birth. Interestingly, only 10 of these 21 ended up as SGA infants.

Eventually, we should like to be able to diagnose prospectively an SGA infant by inspection of the femur length curve. At this stage of the study, with its corresponding sample numbers, 10 of 21 newborns could be predicted to be SGA by femur length measures, but 11 could not.

Fig. 9. Velocity curve of fetal weight in g per gestational weeks.

Fig. 10. Distance curves of biparietal diameter in cm for gestational age in weeks, with the 50th centile and 2 standard deviations.

Searching for a single measure that would be a good prognosticator for SGA infants, it appears at this stage that the abdominal-thoracic diameter has the best predictive power. It is interesting to point out that this prenatal measure is very largely influenced by liver size. This raises the question: Is it possible that fetal liver size/function is related to eventual size at birth?

If one adds biparietal diameter to abdominal-thoracic diameter and femur length in a predictive equation for SGA infants, the prognostic value improves considerably. Fig. 10 shows the evolution of biparietal diameter, Fig. 11 of head circumference during gestation.

Here we need to recall earlier discussions on intrauterine growth retardation and the difference in growth outcome for those fetuses found to be proportionately small in utero and those found to be disproportionately small (disproportionately small because of having 'normal' size fetal heads). We hope eventually to be able to identify proportionately and disproportionately small fetuses by distance and velocity measures of femur length and biparietal diameter-head circumference. We can present already acceptable reference values for some measures of fetal growth, and these will continue to improve. The key, however, is to relate individual children's growth pre- and postnatally. At this stage, sample sizes are still too small to allow us to relate postnatal longitudinal growth to differences in the specific somatic effect of intrauterine growth retardation, but our investigation continues.

Fig. 11. Distance curves of head circumference in cm for gestational age in weeks, with the 50th centile and 2 standard deviations.

Acknowledgements - We have drawn considerably on the inaugural dissertation for the degree of Doctor of Medicine of Rudiger Hans-Joachim Schloo: "Neue Aspekte zur pränatalen Messung der Femurlänge und ihre Beziehung zum postnatalen Wachstum des Kindes". Westfälische Wilhelms-Universität Münster, 1992. Referent: Univ. Prof. Dr. Med. W. Holzgreve.


Brandt I (1986): Growth dynamics of low birth weight infants with emphasis on the perinatal period. In Human growth, 2nd edn, vol. 1, eds F Falkner & JM Tanner, pp. 415-486. New York: Plenum Press.

Deter RL, Rossavik IK, Hill RM, Cortissaz C & Hadlock FP (1987): Longitudinal studies of femur growth in normal fetuses. J. Clin. Ultrasound 15, 299-305.

Dobbing J & Sands J (1978): Head circumference, biparietal diameter, and brain growth in fetal and post-natal life. Early Hum. Dev. 2, 81.

Doyle W, Crawford MA, Wynn AHA & Wynn SW (1990): The association between maternal diet and birth dimensions. J. Nutrit. Med. 1, 9-17.

Falkner F & Roche AF (1987): Relationship of femoral length to recumbent length and stature in fetal, neonatal and early childhood growth. Hum. Biol. 59, 769-773.

Falkner F & Matheny A (in press): The long-term development of twins: Anthropometry and cognition. In Multiple pregnancy: Epidemiology, gestation and perinatal outcome, eds L Keith, E Papiernik & B Luke. New York: Little Brown.

Gruenewald P (1963): Chronic distress and placental insufficiency. Biol. Neonate 5, 215-265.

Hadlock FP, Harrist RB, Sharman RS, Deter RL & Park SK (1985): Estimation of fetal weight with the use of head, body and femur measurements - a prospective study. Am. J. Obstet. Gynecol. 151, 333-337

Iffy R & Deter RL (1975): Ultrasound crown-rump measures. Pediatrics 56,173-179.

Kempley ST, Gamso HR & Nicolaides KH (1993). Renal artery blood flow velocity in very low birth weight infants with intrauterine growth retardation. Arch. Dis. Child 68, 588-590.

Leaf AL, Leighfield MJ, Costeloe KL & Crawford MA (1992): Early Hum. Dev. 30, 183-191.

Lechtig A & Klein RE (1980): Maternal food supplementation and infant health: Results of a study in rural areas of Guatemala. In Maternal nutrition during pregnancy and lactation, eds H Aebi & R Whitehead, pp. 285-313. Bern: Hans Huber.

Lechtig A, Mata LJ, Habicht JP, Urrutia JJ, Klein RE, Guzman G, Caceres A & Alford C (1974): Levels of IgM in cord blood of Latin American newborns of low socioeconomic status. Ecol. Food Nutr. 3, 171-178.

Lechtig A, Yarbrough C, Delgado H, Martorell R, Klein RE & Behar M (1975): Effect of moderate maternal malnutrition on the placenta. Am. J. Obstet. Gynecol. 123, 191-201.

Lechtig A, Rosso P, Delgado H, Bassi J, Martorell R, Yarbrough C, Winick M & Klein RE (1977): Effect of moderate maternal malnutrition on the levels of alkaline, ribonuclease activity in the human placenta. Ecol. Food Nutr. 6, 83-90.

Lubchenko LO, Hansman C, Dressler M & Boyd E (1970): Intrauterine growth as estimated from liveborn birth-weight data at 24 to 42 weeks of gestation. Pediatrics 45, 937-944.

el-Mekki A, Deverajan LV, Soufi S, Stannegard O & Al-Nakig W (1988): Specific and non-specific serological markers in the screening for CMV infection. Epidemiol. Infect. 101, 495-501.

Miller HC (1992): Intra-uterine growth retardation (1992): Past, present and future. Growth. Genetics and Hormones 8, 2, 5-6.

Munsick AA (1984): Human fetal extremity lengths in the interval from 9 to 21 menstrual weeks of pregnancy. J. Obstet. Gynecol. 149, 883-887.

O'Brien GD & Queenan JT (1981) Growth of the ultrasound fetal femur length during normal pregnancy. Am. J. Obstet. Gynecol. 141, 833.

Pardi G et al. (1993): Diagnostic value of blood sampling in fetuses with growth retardation. N. Engl. J. Med. 328, 692-696.

Tanner JM (1950): Foetus into man. Cambridge, MS: Harvard University Press.

Usher R & McLean F (1969): Intra uterine growth of liveborn Caucasian infants at sea level. Standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J. Pediatr 74, 910-910.

Vorherr H (1982): Factors influencing fetal growth. Am. J. Obstet. Gynecol. 142, 577-588.

Warshaw JB (1992): Intra-uterine growth restriction revisited. Growth. Genetics and Hormones 8, 2, 5-8.

Contents - Previous - Next