Nonetheless, the incidence of preeclampsia (3.8%) in this study is similar to that reported in other, recent studies.33 Furthermore, elimination of women refusing to participate in our study, if they differed both by exposure and outcome from those enrolled, might have biased our results.