Portal de Periódicos da CAPES

Do Customized Birth Weight Charts Add Anything but Complexity to the Assessment of Fetal Growth?

2014; Elsevier BV; Volume: 36; Issue: 2 Linguagem: Inglês

10.1016/s1701-2163(15)30652-6

2665-9867

Jennifer A. Hutcheon,

Gestational Diabetes Research and Management

In this issue, Melamed and colleagues1.Melamed N. Ray J.G. Shah P.S. Berger H. Kingdom J.C. Should we use customized fetal growth percentiles in urban Canada?.J Obstet Gynaecol Can. 2014; 36: 164-170Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar draw attention to an important clinical and population health issue: the need to distinguish constitutionally small infants from infants who are small because their growth has been pathologically restricted in utero. Although it has long been recognized that small size alone (small for gestational age, SGA, often defined as a birth weight < 10th percentile) may not be a sufficient marker of abnormal fetal growth,2.Altman D.G. Hytten F.E. Intrauterine growth retardation: let's be clear about it.Br J Obstet Gynaecol. 1989; 96: 1127-1132Crossref PubMed Scopus (105) Google Scholar the challenge has been to come up with a better alternative. In recent years, the idea of customized birth weight charts has gained popularity. Customized birth weight charts are birth weight-for-gestational-age charts that have been adjusted to account for factors such as maternal height, parity, and ethnicity.3.Gardosi J. Chang A. Kalyan B. Sahota D. Symonds E.M. Customised antenatal growth charts.Lancet. 1992; 339: 283-287Abstract PubMed Scopus (633) Google Scholar It is hoped that using a regression model to predict each infant's optimal weight, then comparing an infant's actual weight to its predicted weight, will better distinguish infants who are constitutionally small from infants with intrauterine growth restriction (IUGR). The rationale for customization has great intuitive appeal, and as a result, a number of authors, including Melamed and colleagues, have argued for its adoption into clinical practice.1.Melamed N. Ray J.G. Shah P.S. Berger H. Kingdom J.C. Should we use customized fetal growth percentiles in urban Canada?.J Obstet Gynaecol Can. 2014; 36: 164-170Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar, 4.Resnik R. One size does not fit all.Am J Obstet Gynecol. 2007; 197: 221-222Abstract Full Text Full Text PDF PubMed Scopus (35) Google Scholar, 5.Royal College of Obstetricians and Gynaecologists Green-top Guidelines N0.31: the investigation and management of the small for gestational age fetus. Royal College of Obstetricians and Gynaecologists, London2002Google Scholar However, new screening tools should not be adopted into clinical practice without evidence that they are effective in predicting the health outcome of interest. When assessed using standard criteria for evaluating diagnostic and screening tests (such as likelihood ratios or sensitivity and specificity), customized birth weight charts have very poor predictive value in identifying infants with IUGR. As a result, adopting customized birth weight charts into clinical care would essentially be replacing an existing tool that has a poor predictive ability with one that is more complex, but still has poor predictive ability. Rather than focusing our energy and resources on attempting to refine and implement an ineffective tool, we should assess birth weight using a simple, universal standard that reflects the growth of healthy infants (the approach proposed by the World Health Organization [WHO] and used by pediatricians in Canada and internationally to assess growth). This would allow us to focus our energy instead on developing new approaches for diagnosing IUGR that will stand up to current standards for screening and diagnostic tools. Early studies evaluating the predictive ability of customized birth weight charts appeared to produce promising results. These large, population-based studies compared the risks of adverse perinatal outcomes among infants classified by customized birth weight charts and by conventional population birth weight charts.6.Clausson B. Gardosi J. Francis A. Cnattingius S. Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.BJOG. 2001; 108: 830-834Crossref PubMed Google Scholar, 7.Ego A. Subtil D. Grange G. Thiebaugeorges O. Senat M.V. Vayssiere C. et al.Customized versus population-based birth weight standards for identifying growth restricted infants: a French multicenter study.Am J Obstet Gynecol. 2006; 194: 1042-1049Abstract Full Text Full Text PDF PubMed Scopus (167) Google Scholar, 8.McCowan L.M. Harding J.E. Stewart A.W. Customized birthweight centiles predict SGA pregnancies with perinatal morbidity.BJOG. 2005; 112: 1026-1033Crossref PubMed Scopus (187) Google Scholar They found that the relative risks of perinatal adverse outcomes among SGA infants (vs. non-SGA infants) were higher among infants classified with a customized chart than those classified with a conventional population birth weight chart, suggesting that the customized charts were better able to distinguish between high- and low-risk infants. Unfortunately, these earlier studies suffered from a key methodological limitation: it is well recognized that birth weight charts are biased at preterm ages because the weights of preterm newborns do not reflect healthy growth.9.Ding G. Tian Y. Zhang Y. Pang Y. Zhang J. Application of a global reference for fetal-weight and birthweight percentiles in predicting infant mortality.BJOG. 2013; (Jul 17, [Epub ahead of print.])https://doi.org/10.1111/1471-0528.12381Crossref PubMed Scopus (33) Google Scholar, 10.Hutcheon J.A. Platt R.W. The missing data problem in birth weight percentiles and thresholds for "small-for-gestational-age.".Am J Epidemiol. 2008; 167: 786-792Crossref PubMed Scopus (85) Google Scholar Customized charts, however, do not suffer from this bias because their reference values at preterm ages are derived from an ultrasound estimated fetal weight (EFW)-based chart, which reflects the weights of healthy, ongoing pregnancies at preterm gestational ages.3.Gardosi J. Chang A. Kalyan B. Sahota D. Symonds E.M. Customised antenatal growth charts.Lancet. 1992; 339: 283-287Abstract PubMed Scopus (633) Google Scholar As a result, studies comparing population birth weight charts with customized charts were not just evaluating the process of customizing birth weight charts for maternal characteristics but were also comparing the use of EFW charts with birth weight charts in prediction of adverse outcome. This made it impossible to determine if the apparent benefits of customization were due to the adjustment for maternal characteristics or the use of EFW-based charts at preterm ages. More recent studies have resolved the design flaw of these earlier studies by comparing the customized chart with a non-customized, but EFW-based chart when including preterm ages, or by restricting comparisons to term births.11.Carberry A.E. Raynes-Greenow C.H. Turner R.M. Jeffery H.E. Customized versus population-based birth weight charts for the detection of neonatal growth and perinatal morbidity in a cross-sectional study of term neonates.Am J Epidemiol. 2013; 178: 1301-1308Crossref PubMed Scopus (29) Google Scholar, 12.Larkin J.C. Hill L.M. Speer P.D. Simhan H.N. Risk of morbid perinatal outcomes in small-for-gestational-age pregnancies: customized compared with conventional standards of fetal growth.Obstet Gynecol. 2012; 119: 21-27Crossref PubMed Scopus (35) Google Scholar, 13.Mikolajczyk R.T. Zhang J. Betran A.P. Souza J.P. Mori R. Gülmezoglu A.M. et al.A global reference for fetal-weight and birthweight percentiles.Lancet. 2011; 377: 1855-1861Abstract Full Text Full Text PDF PubMed Scopus (333) Google Scholar, 14.Sjaarda L.A. Albert P.S. Mumford S.L. Hinkle S.N. Mendola P. Laughon S.K. Customized large-for-gestational-age birthweight at term and the association with adverse perinatal outcomes.Am J Obstet Gynecol. 2013; (Sep 10, S0002-9378(13)00947-2, [Epub ahead of print.]): iihttps://doi.org/10.1016/j.ajog.2013.09.006Abstract Full Text Full Text PDF Scopus (27) Google Scholar, 15.Zhang J. Mikolajczyk R. Grewal J. Neta G. Klebanoff M. Prenatal application of the individualized fetal growth reference.Am J Epidemiol. 2011; 173: 539-543Crossref PubMed Scopus (49) Google Scholar, 16.Hutcheon J.A. Zhang X. Cnattingius S. Kramer M.S. Platt RW Customised birthweight percentiles: does adjusting for maternal characteristics matter?.BJOG. 2008; 115: 1397-1404Crossref PubMed Scopus (100) Google Scholar These studies have failed to show convincing benefits from adjusting for maternal characteristics. For example, in a re-analysis of Swedish Medical Birth Registry data, classifying infants using a customized chart produced a relative risk of stillbirth among SGA infants that was significantly higher than that obtained from classifying infants using a conventional birth weight chart (relative risks of 6.1; 95%CI 5.6 to 6.7 vs. 3.8; 95% CI 3.4 to 4.1), but no different than the relative risk obtained when infants were classified with an EFW-based reference (RR=6.2; 95% CI 5.7 to 6.7).16.Hutcheon J.A. Zhang X. Cnattingius S. Kramer M.S. Platt RW Customised birthweight percentiles: does adjusting for maternal characteristics matter?.BJOG. 2008; 115: 1397-1404Crossref PubMed Scopus (100) Google Scholar In other words, simply using an EFW-based chart provided the same improved prediction as using a customized chart, and additionally adjusting for maternal characteristics did little to further improve prediction. Using estimated fetal weights obtained during a routine 30- to 33-week scan from 6787 women participating in the RADIUS trial, Zhang and colleagues also found that the customized chart was no better at predicting perinatal mortality and morbidity than an EFW-based chart (likelihood ratios of 3.01; 95% CI 1.97 to 4.61 vs. 3.08; 95% CI 1.99 to 4.76, respectively).15.Zhang J. Mikolajczyk R. Grewal J. Neta G. Klebanoff M. Prenatal application of the individualized fetal growth reference.Am J Epidemiol. 2011; 173: 539-543Crossref PubMed Scopus (49) Google Scholar In data from a WHO study of 237 025 births from 24 countries worldwide, prediction of perinatal morbidity and mortality was only minimally improved by adjusting an EFW-based chart for country of origin (area under the curve [AUC] of 0.679 without adjustment vs. 0.699 with adjustment), and further adjustment for maternal characteristics provided no incremental value at all (AUC of 0.698).13.Mikolajczyk R.T. Zhang J. Betran A.P. Souza J.P. Mori R. Gülmezoglu A.M. et al.A global reference for fetal-weight and birthweight percentiles.Lancet. 2011; 377: 1855-1861Abstract Full Text Full Text PDF PubMed Scopus (333) Google Scholar Similar conclusions have been reached in studies that were restricted to term births or that compared infants of similar gestational age.11.Carberry A.E. Raynes-Greenow C.H. Turner R.M. Jeffery H.E. Customized versus population-based birth weight charts for the detection of neonatal growth and perinatal morbidity in a cross-sectional study of term neonates.Am J Epidemiol. 2013; 178: 1301-1308Crossref PubMed Scopus (29) Google Scholar, 12.Larkin J.C. Hill L.M. Speer P.D. Simhan H.N. Risk of morbid perinatal outcomes in small-for-gestational-age pregnancies: customized compared with conventional standards of fetal growth.Obstet Gynecol. 2012; 119: 21-27Crossref PubMed Scopus (35) Google Scholar, 14.Sjaarda L.A. Albert P.S. Mumford S.L. Hinkle S.N. Mendola P. Laughon S.K. Customized large-for-gestational-age birthweight at term and the association with adverse perinatal outcomes.Am J Obstet Gynecol. 2013; (Sep 10, S0002-9378(13)00947-2, [Epub ahead of print.]): iihttps://doi.org/10.1016/j.ajog.2013.09.006Abstract Full Text Full Text PDF Scopus (27) Google Scholar Of particular interest, Carberry and colleagues evaluated the body composition of 581 term newborns classified as SGA or LGA (large for gestational age) by customized and population birth weight charts, with the hypotheses that infants classified as SGA by the customized chart would be more likely to have low body fat (reflecting true IUGR) and infants classified as LGA by the customized chart would be more likely to have a high percent body fat (reflecting fetal overgrowth).11.Carberry A.E. Raynes-Greenow C.H. Turner R.M. Jeffery H.E. Customized versus population-based birth weight charts for the detection of neonatal growth and perinatal morbidity in a cross-sectional study of term neonates.Am J Epidemiol. 2013; 178: 1301-1308Crossref PubMed Scopus (29) Google Scholar Neither of these hypotheses were supported by their findings: infants classified as SGA by the customized standard were no more likely to have low body fat as those classified as SGA by the population chart (AUCs of 0.80; 95% CI 0.75 to 0.85 vs. 0.83; 95% CI 0.79 to 0.88, respectively), while infants classified as LGA by the customized standard were no more likely to have a high percentage of body fat than those classified as LGA by the population standard. Taken as a whole, the literature on customization suggests that while using an ultrasound EFW-based chart significantly improves prediction of adverse perinatal outcomes (particularly at preterm ages), there is little, if any, incremental gain from further adjusting for maternal characteristics.17.Zhang J. Sun K. Invited commentary: the incremental value of customization in defining abnormal fetal growth status.Am J Epidemiol. 2013; 178: 1309-1312Crossref PubMed Scopus (12) Google Scholar Further, the predictive value of customized charts as assessed using standard measures for evaluating diagnostic tests is very poor: typically, likelihood ratios are < 5, AUCs are < 0.7, and sensitivity < 50%.12.Larkin J.C. Hill L.M. Speer P.D. Simhan H.N. Risk of morbid perinatal outcomes in small-for-gestational-age pregnancies: customized compared with conventional standards of fetal growth.Obstet Gynecol. 2012; 119: 21-27Crossref PubMed Scopus (35) Google Scholar, 15.Zhang J. Mikolajczyk R. Grewal J. Neta G. Klebanoff M. Prenatal application of the individualized fetal growth reference.Am J Epidemiol. 2011; 173: 539-543Crossref PubMed Scopus (49) Google Scholar, 16.Hutcheon J.A. Zhang X. Cnattingius S. Kramer M.S. Platt RW Customised birthweight percentiles: does adjusting for maternal characteristics matter?.BJOG. 2008; 115: 1397-1404Crossref PubMed Scopus (100) Google Scholar, 17.Zhang J. Sun K. Invited commentary: the incremental value of customization in defining abnormal fetal growth status.Am J Epidemiol. 2013; 178: 1309-1312Crossref PubMed Scopus (12) Google Scholar Irrespective of whether customized charts are better than conventional charts, they are still not very good. Our efforts and resources would likely be better spent furthering research on more promising approaches such as combining weight measurements with assessments of placental health18.Proctor L.K. Toal M. Keating S. Chitayat D. Okun N. Windrim R.C. et al.Placental size and the prediction of severe early-onset intrauterine growth restriction in women with low pregnancy-associated plasma protein-A.Ultrasound Obstet Gynecol. 2009; 34: 274-282Crossref PubMed Scopus (118) Google Scholar or biomarkers.19.Benton S.J. Hu Y. Xie F. Kupfer K. Lee S.W. Magee L.A. et al.Can placental growth factor in maternal circulation identify fetuses with placental intrauterine growth restriction?.Am J Obstet Gynecol. 2012; 206 (e1-17): 163Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar An alternative to the customized percentiles derived through Gardosi and colleagues' regression-based approach3.Gardosi J. Chang A. Kalyan B. Sahota D. Symonds E.M. Customised antenatal growth charts.Lancet. 1992; 339: 283-287Abstract PubMed Scopus (633) Google Scholar is to use group-specific charts. Canadian birth weight charts are currently sex-specific20.Kramer M.S. Platt R.W. Wen S.W. Joseph K.S. Allen A. Abrahamowicz M. et al.A new and improved population-based Canadian reference for birth weight for gestational age.Pediatrics. 2001; 108: E35Crossref PubMed Scopus (1226) Google Scholar; Melamed et al. are proposing that charts be further stratified by ethnicity. The decision to adjust for ethnic-specific differences in birth weight is complex and requires reflection on the reasons why we construct group-specific charts. Customized or group-specific charts are based on the assumption that the differences in birth weight between groups are due to physiological rather than pathological influences on birth weight. For example, although infants of smokers are smaller on average than infants of non-smokers, they are also at increased risk of adverse outcomes. As a result, we would not produce separate charts for smokers because this would essentially be normalizing the pathologically small size of their infants (i.e., "your baby is small, but don't worry, it isn't smaller than the babies of other smokers"). While it is clear in the case of smoking that we should not create group-specific charts, the case of ethnicity is very challenging because it is unclear to what extent ethnicity reflects true differences in genetic growth potential and to what extent it captures environmental exposures or cultural practices within a group. For example, factors associated with suboptimal fetal growth such as lower socioeconomic status, smoking, or pregnancy complications such as preeclampsia may be more common in certain ethnic groups. Rates of SGA would be higher in such groups, but this would be appropriate, because the infants are "small and at increased risk," not "small but healthy." As a result, simply observing differences in birth weight between ethnic groups is not a sufficient reason to produce ethnic-specific charts and could actually do harm by inadvertently normalizing the weight of infants who are small, but at increased risk because of suboptimal intrauterine growth. In pediatrics, it is widely accepted that differences in growth between ethnic groups largely reflect environmental influences rather than differences in genetic growth potential.21.deOnis M. Garza C. Onyango A.W. Martorell R. WHO Child Growth Standards.Acta Paediatr. 2006; 450: 1-101Google Scholar, 22.Graitcer P.L. Gentry E.M. Measuring children: one reference for all.Lancet. 1981; 2: 297-299Abstract PubMed Scopus (128) Google Scholar Among well-nourished children, only 3% to 6% of variation in growth is explained by differences in ethnicity, while factors such as poverty and environment play a much larger role.23.Habicht J.P. Martorell R. Yarbrough C. Malina R.M. Klein R.E. Height and weight standards for preschool children. How relevant are ethnic differences in growth potential?.Lancet. 1974; 1: 611-614Abstract PubMed Scopus (392) Google Scholar For this reason, pediatric growth in Canada and elsewhere is assessed using WHO Growth Charts, which were derived from healthy children of widely different ethnic backgrounds and countries including the United States, Norway, India, Ghana, and Oman.21.deOnis M. Garza C. Onyango A.W. Martorell R. WHO Child Growth Standards.Acta Paediatr. 2006; 450: 1-101Google Scholar, 24.Marchand V. Nutrition and Gastroenterology Committee of the Canadian Paediatric Society Promoting optimal monitoring of child growth in Canada: using the new World Health Organization growth charts.Paediatr Child Health. 2010; 15: 77-79Crossref PubMed Google Scholar Work is currently underway to develop new standards that will extend the WHO Growth Charts to include fetal, neonatal, and preterm postnatal growth.25.Villar J. Altman D.G. Purwar M. Noble J.A. Knight H.E. Ruyan P. et al.The objectives, design and implementation of the INTERGROWTH-21st Project.BJOG. 2013; 120 (S2): 9-26Crossref PubMed Scopus (191) Google Scholar Using a similar methodology as the pediatric charts, the INTERGROWTH study, led by a team from Oxford University and funded by the Bill & Melinda Gates Foundation, is creating new fetal growth standards derived from healthy pregnancies representing diverse ethnicities and countries.25.Villar J. Altman D.G. Purwar M. Noble J.A. Knight H.E. Ruyan P. et al.The objectives, design and implementation of the INTERGROWTH-21st Project.BJOG. 2013; 120 (S2): 9-26Crossref PubMed Scopus (191) Google Scholar If we are to move away from using a national birth weight chart in Canada, we should be moving towards use of a simple, universal growth chart based a multi-ethnic population of healthy pregnancies, as proposed by the WHO and adopted by the pediatric community in Canada, rather than in the opposite direction of creating separate charts for population sub-groups for whom it is not fully understood if differences in weight are physiological or pathological.