From the Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, Canada.
Department of Medicine, Solna, Clinical Epidemiology Unit, Karolinska Institutet, Stockholm, Sweden.
Epidemiology. 2019 Mar;30(2):234-242. doi: 10.1097/EDE.0000000000000959.
Observational cohort studies have consistently shown that maternal weight gain in pregnancy is positively associated with fetal size, but it is unknown whether the association is causal. This study investigated the effect of pregnancy weight gain on fetal growth using a sibling comparison design to control for unmeasured confounding by genetic and shared environmental factors.
Our study population included 44,457 infants (21,680 women) with electronic medical records in the Stockholm-Gotland Obstetrical Database, 2008-2014. We standardized pregnancy weight gain into gestational age-specific z-scores. Fetal size was classified as birthweight (gram), and as small- and large-for-gestational-age birth (birthweight <10th or >90th percentiles, respectively). Our sibling comparison analyses used multivariable linear fixed effects models for birthweight and hybrid logistic fixed effects models for small- and large-for-gestational-age birth (SGA and LGA). We repeated analyses using conventional (unmatched) regression models.
Sibling comparison analyses showed a clinically meaningful association between weight gain and fetal size (e.g., adjusted difference of +89 g birthweight [95% CI = 82, 95 g]; adjusted risk ratios [aRR] for SGA of 0.80 [95% CI = 0.75, 0.86] per 1 z-score increase in weight gain for a woman of body mass index [BMI] = 25). These findings were consistent across the range of BMI. Estimates were only modestly attenuated compared with conventional approach (+97 g [95% CI = 92, 102 g], aRR for SGA of 0.70 [95% CI = 0.67, 0.73] per 1 z-score increase in weight gain).
The positive association between pregnancy weight gain and fetal size we found using a sibling comparison design suggests that this relation has minimal confounding by familial factors that remain constant between pregnancies.
观察性队列研究一致表明，孕期体重增加与胎儿大小呈正相关，但尚不清楚这种关联是否具有因果关系。本研究采用同胞比较设计，通过控制遗传和共享环境因素引起的未测量混杂，调查孕期体重增加对胎儿生长的影响。
我们的研究人群包括 2008 年至 2014 年斯德哥尔摩-哥德堡产科数据库中 44457 名有电子病历的婴儿（21680 名妇女）。我们将孕期体重增加标准化为特定胎龄的 z 分数。胎儿大小分为出生体重（克），以及小于胎龄儿（SGA）和大于胎龄儿（LGA）。我们的同胞比较分析使用多变量线性固定效应模型进行出生体重分析，混合逻辑固定效应模型进行 SGA 和 LGA 分析。我们使用传统（未匹配）回归模型重复了分析。
同胞比较分析显示，体重增加与胎儿大小之间存在临床意义上的关联（例如，体重增加 1 z 分数的调整差异为+89 g [95%置信区间（CI）= 82，95 g]；对于 BMI 为 25 的女性，体重增加每增加 1 z 分数，SGA 的调整风险比[aRR]为 0.80 [95% CI = 0.75，0.86]）。这些发现与 BMI 范围一致。与传统方法相比，估计值仅略有减弱（+97 g [95% CI = 92，102 g]，体重增加每增加 1 z 分数，SGA 的 aRR 为 0.70 [95% CI = 0.67，0.73]）。
我们使用同胞比较设计发现，孕期体重增加与胎儿大小之间存在正相关关系，这表明这种关系受孕期之间保持不变的家族因素的混杂影响最小。
