Modeling maternal weight and height in studies of pregnancy outcome among Hispanic women

The objective of this study was to evaluate methods of using maternal weight and height in studies of pregnancy outcome for Hispanic women. Reference anthropometric data came from 1166 Mexican-American women in the Hispanic Health and Nutrition Examination Survey (HHANES). Prospective data on maternal anthropometry and infant birth weight came from 1362 Hispanic women in the Kaiser-Permanente Contraceptive Drug Study and 12,786 women in the Guatemalan Cooperative Perinatal Study. Five methods of standardizing weight for height were evaluated, including power-type indexes and weights relative to HHANES reference data. In linear- and logistic-regression analyses, these methods were practically interchangeable, with no evident advantage of Hispanic reference data. However, if weight was not height-standardized the effect of height was underestimated; if height was omitted and weight was not height-standardized the effects of weight were exaggerated. Therefore, analyses of pregnancy outcome should include both height and height-standardized weight.