In a recent interesting article in the Journal, Sween et al reported a significant relationship between early pregnancy body fat mass and the risk of preeclampsia, particularly among obese individuals. We believe the results presented will be considered important among the readers of the Journal. Although a continuing argument exists regarding the validity of bioelectric impedance analysis (BIA) technique in estimation of body compositions, we needed to acknowledge some of our comments regarding methodologic issues.

The basic theory of BIA relies on the assessment of changes in alternating potential that are applied to the body by the device. This alternating current tends to pass predominantly through the conductive water-containing tissues. Therefore, total body water has a key role in determination of body composition. To minimize the influence of daytime total water change in the body, a “pretest” preparation is mandatory. This includes performance of the BIA at the same time of the day at all times, with no food or beverage intake for 4 hours before testing and with the bladder emptied immediately before the measurement. This issue might be more critical with clinicians who wish to use BIA test in their pregnant subjects; however, in research projects, greater care is required for such a pretest protocol that seems to be missing in the article by Sween et al.

Moreover, total body change during pregnancy is a potential confounder for BIA testing. Sween et al noted that they eliminated this effect by using pregnancy-specific BIA equations that were validated in this specific condition by Van Loan et al and Lukaski et al. However, the study by Van Loan et al was designed to generate equations by analyzing earlier longitudinal study results with well-nourished Dutch pregnant women. It should be emphasized that the use of general prediction equations across different (ethnic) population groups without previous testing of their validity should be avoided. Therefore, it is obviously hard to generate equations from the study of Van Loan et al for the pregnant women in the United States. Besides, that study by Van Loan et al was designed to generate equations, not to validate these equations. Also, Lukaski et al did not assess validity equations generated by Van Loan et al, but they studied only the accuracy of BIA to determine the total body water not the body composition. We hope these comments might improve the understanding of the results that were presented in this study.