Alcohol consumption before and during pregnancy comparing concurrent and retrospective reports

Background: Reported alcohol consumption during pregnancy has proven to be higher when reported retrospectively during or after the pregnancy, compared with concurrent reports. In a longitudinal, population-based questionnaire study, we investigated whether these findings could be replicated. Methods: A longitudinal, population-based questionnaire study. Among a representative half of pregnant women in Oslo, 92% accepted to join the study. Non-Norwegian speaking and/or immigrants from nonwestern countries were not invited. Questionnaires were answered at 17 (T1) and 30 weeks of pregnancy (T2) and 6 months after term (T3). The response rate was at T1 93% (N = 1,749), at T2 82%, and at T3 94% [representing 86% (T1), 70% (T2), and 64% of those invited to join the study]. Results: Significantly more alcohol consumption after pregnancy recognition was reported retrospectively at both T2 and T3 [T2 0.15 and T3 0.18 standard units per week (SU/wk)] than concurrently at T1 or T2 (T1 0.10 and T2 0.14 SU/wk). When comparing the 2 retrospective reports at T2 and T3, there was a significant increase over time. Predictive factors for higher retrospective reports were higher age, more SU per week before pregnancy, reported alcohol use after pregnancy recognition (at T1), and depression after the pregnancy. A predictive factor for lower retrospective reporting was anxiety during, but not after, the pregnancy. Contrary to previous suggestions, anxiety concerning abnormality in the 6-month-old child was not a predictive factor. Conclusion: Concurrently reported alcohol consumption during pregnancy is probably under-reported. Thus, concurrent reports appear to underestimate fetal exposure.