Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Ordeals for the fetal origins hypothesis
- D I W Phillips, C Osmond (30 April 1999)
-
Ordeal for fetal origins hypothesis?
- Pat Doyle, Dave Leon, Noreen Maconochie, Susan Moreton, Bianca de Stavola (14 May 1999)
-
Twins are specifically excluded from the fetal origins hypothesis
- D I W Phillips (26 May 1999)
-
The Path Analysis Is Inaproperly Used
- Clayman ZK Zhang (19 April 2002)
|
|
|||
|
D I W Phillips MRC Unit, University of Southampton, C Osmond
Send response to journal:
|
Sir - Williams and Poulton (April 3rd, p 897) show that their 22 adolescent twins have lower blood pressure than singletons. They interpret their data as being contrary to the fetal origins hypothesis, because they presume that twins, being small at birth, would tend to have higher rather than lower blood pressure in later life. However, as twins have different patterns of fetal growth from singletons they were specifically excluded from the fetal origins hypothesis.1 There are several reasons why the low birthweight of twins may not have the same significance as intrauterine growth retardation in singleton births.2 Ultrasound evidence suggests that twins downregulate their growth rate early in gestation, possibly during the first trimester.3 Studies in fetal lambs suggest that early downregulation of fetal growth protects against growth retardation induced by undernutrition in later gestation.4 Finally, the metabolic and endocrine changes associated with growth retardation in singletons, including hypoinsulinaemia, are not observed in twins.5 DIW Phillips C Osmond Medical Research Council (University of Southampton), Southampton General Hospital, Tremona Road, Southampton SO16 6YD. 1. Barker DJP. Fetal origins of coronary heart disease. Br Med J 1995;311:171-174. 2. Phillips DIW. Mortality among twins. BMJ 1995;310:1330-1331. 3. Leveno KJ, Santos-Ramos R, Duenhoelter JH, Reisch JS, Whalley PJ. Sonar cephalometry in twins: A table of biparietal diameters for normal twin fetuses and a comparison with singletons. Am J Obstet Gynecol 1979;135:727-730. 4. Harding J, Liu L, Evans P, Oliver M, Gluckman PD. Intrauterine feeding of the growth retarded fetus: can we help? Early Human Development 1992;29:193-197. 5. Van Assche FA, Aerts L, Holemans K. Low birthweight and ischaemic heart disease. Lancet 1994;343:731-732. |
|||
|
|
|||
|
Pat Doyle, Senior lecturer in epidemiology London School of Hygiene and Tropical Medicine, Dave Leon, Noreen Maconochie, Susan Moreton, Bianca de Stavola
Send response to journal:
|
Dear Sirs, Williams and Poulter (1) and an accompanying editorial by Susser and Levin (2) challenged the fetal origins hypothesis with some vigour. The paper and editorial place particular emphasis on the finding that twins had lower blood pressure than singletons. It was argued that this result provided crucial evidence against the foetal origins hypothesis, which according to them would predict the opposite effect given that twins compared to singletons have lower growth rates in utero and are lighter at birth. While we applaud the intent to subject the hypothesis to rigorous tests, the line of argument taken by the paper and editorial are unconvincing. Firstly, the conclusion rests on 22 twins only. Moreover, these may not be representative of all twins in the population. Indeed, examination of the birthweight distribution of these twins indicates that there may be missing individuals at very low birthweight (<1500gm). Secondly, the data as a whole (singletons and twins combined) do show a negative association between birthweight and blood pressure at age 9, and thus support the fetal origins hypothesis. This is given far less prominence than the twin finding. Thirdly, and most importantly, the notion that the growth impairment suffered by twins is similar to other forms of fetal growth impairment is questionable. Postnatal catch-up in twins is substantial, with the height and weight deficit of twins compared to singletons being almost entirely eliminated by 8 years of age (3). In contrast, size at birth in singletons (and in twins) is predictive of postnatal adult height and weight. Furthermore, divergence in fetal growth rates between twins and singletons occurs very early in gestation (4,5). Reduced size at birth in the population as a whole is importantly influenced by growth faltering in late gestation as rapidly increasing nutritional demands of the fetus exceed supply. If it is this later type of growth impairment that is associated with programming of blood pressure and cardiovascular disease, then the early down regulation of growth rate in twins may not be associated with increases in blood pressure, as this minimises the risk that their nutritional demands in late gestation will exceed supply. Given the weight of evidence in support of a negative association between birthweight and blood pressure (6), twins may be interesting exceptions to the rule. As such, rather than twins constituting an ordeal, studying why they are different to singletons may help reveal the mechanisms underlying fetal programming of later blood pressure. 1. Williams S , Poulton R. Twins and maternal smoking: ordeals for the fetal origins hypothesis? a cohort study. BMJ 1999;318: 897-900 2. Susser M, Levin B. BMJ 1999; 318: 885-886 3. Wilson, R.S. Twin growth : initial deficit, recovery, and trends in concordance from birth to nine years. Ann.Hum.Biol. 6:205-220, 1979. 4. Wilson, R. S.. Twins : Measures of birth size at different gestational ages. Ann.Hum.Biol. 1:57-64, 1974. 5.Taylor, G.M. , Owen,P. and Mires, G.J.. Foetal growth velocities in twin pregnancies. Twin.Res. 1 (1):9-14, 1998. 6. Law, C. M. and. Shiell, A. W. Is blood pressure inversely related to birth weight ? The strength of evidence from a systematic review of the literature. J.Hypertens. 14:935-941, 1996. |
|||
|
|
|||
|
D I W Phillips
Send response to journal:
|
Sir - Williams and Poulton (April 3rd, p 897) show that their 22 adolescent twins have lower blood pressure than singletons. They interpret their data as being contrary to the fetal origins hypothesis, because they presume that twins, being small at birth, would tend to have higher rather than lower blood pressure in later life. However, as twins have different patterns of fetal growth from singletons they were specifically excluded from the fetal origins hypothesis.1 There are several reasons why the low birthweight of twins may not have the same significance as intrauterine growth retardation in singleton births.2 Ultrasound evidence suggests that twins downregulate their growth rate early in gestation, possibly during the first trimester.3 Studies in fetal lambs suggest that early downregulation of fetal growth protects against growth retardation induced by undernutrition in later gestation.4 Finally, the metabolic and endocrine changes associated with growth retardation in singletons, including hypoinsulinaemia, are not observed in twins.5 DIW Phillips C Osmond Medical Research Council (University of Southampton), Southampton General Hospital, Tremona Road, Southampton SO16 6YD, 1 Barker DJP. Fetal origins of coronary heart disease. Br Med J 1995;311:171-174. 2 Phillips DIW. Mortality among twins. BMJ 1995;310:1330-1331. 3 Leveno KJ, Santos-Ramos R, Duenhoelter JH, Reisch JS, Whalley PJ. Sonar cephalometry in twins: A table of biparietal diameters for normal twin fetuses and a comparison with singletons. Am J Obstet Gynecol 1979;135:727-730. 4 Harding J, Liu L, Evans P, Oliver M, Gluckman PD. Intrauterine feeding of the growth retarded fetus: can we help? Early Human Development 1992;29:193-197. 5 Van Assche FA, Aerts L, Holemans K. Low birthweight and ischaemic heart disease. Lancet 1994;343:731-732. |
|||
|
|
|||
|
Clayman ZK Zhang, Resear staff MRC Environmental Epidemiology Unit,Southampton,SO16 6YD
Send response to journal:
|
Sir- Dr Williams has inaproperly used Path Analysis method in their research report(BMJ 1999;318:897).Their at leat two points need more concern:(1)Is it possible to introduce binary variables into a path diagram? path analysis is a generalized linear regression analysis,in which more than two bonary variable may cause the inflation of variance.In Dr Williams' report,their are at two binary variables,including Twin and Male sex.This may introduce more problem in the analysis.(2)Is so many path way neccessary? The main aim of path analysis is to find the real path way by the statistical power,and in this way to test a theoretical hypothesis.If the path coefficient(we can just see them as correlation parameters between two variables)is not significant,this path should be erased from the possible diagram.Usually the path do not stand as the parameter less than 0.10(than mean the correlation can only explain about one percent of the variance).From Fig2 in the paper we easily found more path coefficients less than 0.10.I can't imagine the roles played by these path in defending the hypothesis of this report. Clayman ZK Zhang MRC Epidemiology Unit Southampton General Hospital Southampton,SO16 6YD |
|||