J. Perinat. Med. 2018; aop Niels Rochow*, Malak AlSamnan, Hon Yiu So, Dirk Olbertz, Anna Pelc, Jan Däbritz, Roland Hentschel, Ursula Wittwer-Backofen and Manfred Voigt Maternal body height is a stronger predictor of birth weight than ethnicity: analysis of birth weight percentile charts https://doi.org/10.1515/jpm-2017-0349 Middle East and North Africa, and Asia excluding Middle Received November 9, 2017. Accepted April 27, 2018. East). Percentile charts were calculated for each maternal Abstract height group. Results: The average BW and maternal height differ sig- Background: Anthropometric parameters such as birth nificantly between ethnic groups. On current percentile weight (BW) and adult body height vary between ethnic charts, newborns of taller mothers (≥176 cm) have a low groups. Ethnic-specific percentile charts are currently rate of SGA and a high rate of LGA, whereas newborns of being used for the assessment of newborns. However, shorter mothers (≤157 cm) have a high rate of SGA and a due to globalization and interethnic families, it is unclear low rate of LGA. When the BW data are stratified based on which charts should be used. A correlation between a the maternal height, mothers of similar height from dif- mother’s height and her child’s BW (1 cm accounts for a ferent ethnic groups show similar average BWs, SGA and 17 g increase in BW) has been observed. The study aims LGA rates. to test differences in small for gestational age (SGA) and Conclusion: Maternal body height has a greater influence large for gestational age (LGA) rates, employing BW per- on BW than maternal ethnicity. The use of BW percentile centile charts based on maternal height between ethnic charts for maternal height should be considered. groups. Keywords: Birth weight percentiles; growth restriction; Methods: This retrospective study of 2.3 million mother/ large for gestational age; small for gestational age. newborn pairs analyzed BW, gestational age, sex, mater- nal height and ethnicity from the German perinatal survey (1995–2000). These data were stratified for mater- nal height (≤157, 158–163, 164–169, 170–175, ≥176 cm) and Introduction region of origin (Germany, Central and Northern Europe, North America, Mediterranean region, Eastern Europe, The assessment of intrauterine growth is a daily routine of perinatal care to identify fetuses and newborns that are intrauterine growth restricted, small for gestational *Corresponding author: Niels Rochow, MD, Department of Pediatrics, age (SGA) or large for gestational age (LGA) [1]. The SGA McMaster University, 1280 Main Street West, Room 4F, Hamilton, and LGA are often defined as birth weights (BWs) below ON L8S 4K1, Canada, E-mail: [email protected] th th Malak AlSamnan, Hon Yiu So and Anna Pelc: Department of the 10 and above the 90 percentile, respectively, for Pediatrics, McMaster University, Hamilton, ON, Canada sex and gestational age. Proper identification of SGA or Dirk Olbertz: Neonatology, Clinic Südstadt, Rostock, Germany LGA is imperative, as these fetuses and newborns exhibit Jan Däbritz: Department of Pediatrics, University Medicine Rostock, higher rates of mortality and morbidity. Consequently, Rostock, Germany; and Centre for Immunobiology, Blizard Institute, diagnostics, treatment and admission are often sug- Barts Cancer Institute the Barts and The London School of Medicine gested [2–5]. and Dentistry, Queen Mary University of London, London, UK Roland Hentschel: Department of General Pediatrics, Division of To assess intrauterine growth, BW charts for sex and Neonatology/Intensive Care Medicine, Medical Center – Faculty of gestational age were developed from perinatal surveys Medicine, University of Freiburg, Freiburg, Germany based on millions of datasets. It has been shown that BW Ursula Wittwer-Backofen: Centre for Medicine and Society, cutoffs for SGA and LGA differ between populations and University of Freiburg, Freiburg, Germany ethnic groups. The general perception is that ethnic-spe- Manfred Voigt: Department of General Pediatrics, Division of Neonatology/Intensive Care Medicine, Medical Center – Faculty of cific characteristics, such as differences in body height, Medicine, University of Freiburg, Freiburg, Germany; and Centre for are related to differences in BWs. This led to the adoption Medicine and Society, University of Freiburg, Freiburg, Germany and use of ethnic-specific growth charts [6–8]. Brought to you by | Washington University in St. Louis Authenticated Download Date | 6/6/18 2:15 PM 2 Rochow et al., Birth weight percentiles based on maternal height However, due to globalization and migration, there Northern Europe (excluding Germany) and North America, 4.6% from is an increasing number of interethnic families. Based Mediterranean countries, 3.6% from Eastern Europe, 6.9% from Mid- dle East and North Africa, 1.1% from Asia (excluding Middle East) and on this observation, it could be hypothesized that ethnic- 1.5% from other regions. Data from all neonates were included in our specific cutoffs for SGA and LGA may not apply. analysis irrespective of the presence or absence of neonatal disorders Findings in the literature support the hypothesis that and irrespective of the length of gestation. The gestational age was maternal body height is a predictor of BW. For instance, estimated according to Naegele’s rule and was corrected by sono- in animal experiments, it has been demonstrated that graphic assessment in weeks 8–14 when there was a discrepancy of a fetus was smaller at birth than another fetus with the ≥14 days. BW percentiles were calculated for gestational age and sex. The empirical percentile values were determined using a weighted same genetics if the former was gestated in the uterus of means method, as expressed by the following formula: a smaller breed mare [9]. In humans, it has been shown that BW is related to the maternal height [10–15]. The rela- pp kn≤⋅ <+kd1, ,=⋅nk−= (xgpk+⋅dgkk+1 − g ) tion between maternal height and BW has been attributed 100100 to the fact that mothers of shorter stature have a smaller where n is the number of births for a specific gestational day, uterine environment, which ultimately limits the space g1 ≤ … ≤ gn are the sorted BWs for that gestational day, p is the percen- th available for intrauterine growth [11]. This concept is tile, xp is the weight representing the p percentile and k is the floor of further supported by the observation that the BWs of half- p n⋅ . For k < 1, xp was set to the smallest weight, g1 [18]. siblings are more strongly correlated when the common 100 Cutoffs for SGA, AGA (appropriate for gestational age) and parent is the mother [16]. Furthermore, in cases of ovum LGA were calculated for the complete dataset and selected cohorts donation, it has been shown that the maternal anthro- for five maternal height groups ≤( 157 cm, 158–163 cm, 164–169 cm, pometry was related to the BW. The BWs of the infants 170–175 cm and ≥176 cm). Neonates with BW less than the 10th per- were closely associated with the body anthropometry of centile were classified as SGA, while those with BW above the 90th the recipient mother but unrelated to the BW of the ovum percentile were classified as LGA. All newborns between the 10th and donor’s own child [17]. 90th BW percentiles were considered AGA. For this analysis, we also Recently, we found a linear relationship between classified the data based on the mother’s region of origin: Germany, Central and Northern Europe, North America, Mediterranean region, the maternal height and the infant’s BW over a wide Eastern Europe, Middle East and North Africa and Asia excluding the range. For each 1 cm increase in the maternal height, Middle East. the average BW increased by 17 g [15]. We have also Multivariate linear regression models were carried out to study shown that cutoffs for LGA, as well as SGA, differ by up the effects on BW using the “lm” function of R software package. to 600 g [12]. This was demonstrated by comparing term Model 1 tests the effect of either the region of origin or maternal height on BW. A variance analysis for the full and reduced models newborns of short and light-weight mothers (≤161 cm was performed. The F statistic was used to indicate the effect of either and ≤57 kg) with those of tall and heavy-weight mothers the region of origin or maternal height on BW. (≥172 cm and ≥106 kg) [12]. Equation model 1a: As evidence shows that there is a correlation between maternal height and BW, this study aims to test whether BW ~region of origin + maternal height group (full model) SGA and LGA rates differ between ethnic groups when BW ~region of origin (reduced model) reference values for BW percentiles based on maternal height groups are used. It is hypothesized that BW is more Equation model 1b: strongly affected by sex, gestational age and maternal BW ~region of origin + maternal height group (full model) height than by the newborn’s ethnicity. BW ~maternal height group (reduced model) Model 2 analyzes the effects of the region of origin, maternal Methods height, gestational age and sex on BW. Boys born of German origin and a maternal height of 164–169 cm were the comparator. The data for this observational cohort study were taken from the Equation model 2: routine data collection of the German perinatal survey performed between 1995 and 2000. Between 1995 and 1997, all German federal BeWs~region of origin ++maternal height group gestational age + x states except Baden-Württemberg contributed to our dataset, while between 1998 and 2000, only Bavaria, Brandenburg, Hamburg, Meck- The analysis was assisted by IBM SPSS Statistics for Windows, lenburg-Western Pomerania, Lower Saxony, Saxony, Saxony-Anhalt Version 21.0. Armonk, NY, USA, Microsoft Excel®, Microsoft Corp., and Thuringia contributed data.