Rantakallo P, Hartikainen-Sorri A L
Am J Epidemiol. 1981 May;113(5):590-5. doi: 10.1093/oxfordjournals.aje.a113136.
To test the hypothesis that the lower birth weight of infants whose mothers smoke during pregnancy is mediated by lesser weight gain during pregnancy, the authors examined data from an original series consisting of 12,068 births in Northern Finland. Weight gain data were collected for every 10th case chosen by systematic sampling in such a way that every group of abnormal births was treated separately to assure correct representation in the sample. The weight gain of the smokers was 200 g less than that of the controls matched with them for age, parity, place or residence and marital status, but the difference in birth weight and placental weight accounted for about 85% of the total sample in order to quantify the separate and joint contributions to birth weight of smoking, maternal weight gain, pre-pregnant weight, height and duration of gestation. The effect of smoking alone was highly statistically significant in all models. Its effect was independent of weight gain, but when all the variables were introduced simultaneously, the effect was reduced to half. Another set of regression analyses were performed with weight gain as the dependent variable and the others as explanatory variables. The effect of smoking on weight gain was not statistically significant. The result was not affected by including parity in the explanatory variables.
为了验证孕期母亲吸烟的婴儿出生体重较低是由孕期体重增加较少所介导的这一假设，作者研究了来自芬兰北部12068例分娩的原始系列数据。通过系统抽样，每10例选取1例收集体重增加数据，对每一组异常分娩进行单独处理，以确保样本中有正确的代表性。吸烟者的体重增加比与其年龄、产次、居住地和婚姻状况相匹配的对照组少200克，但为了量化吸烟、母亲体重增加、孕前体重、身高和孕期持续时间对出生体重的单独和共同影响，出生体重和胎盘重量的差异约占总样本的85%。在所有模型中，单独吸烟的影响在统计学上具有高度显著性。其影响独立于体重增加，但当同时引入所有变量时，影响降低了一半。以体重增加作为因变量，其他变量作为解释变量进行了另一组回归分析。吸烟对体重增加的影响在统计学上不显著。在解释变量中纳入产次并不影响结果。
