Use of Proton-Pump Inhibitors in Early Pregnancy and the Risk of Birth Defects
Björn Pasternak, M.D., Ph.D., and Anders Hviid, Dr.Med.Sci.
N Engl J Med 2010; 363:2114-2123November 25, 2010
DISCUSSION
This nationwide cohort study showed that there were no significant associations between the use of PPIs during the first trimester and the risk of major birth defects. In subgroup analyses, estimates of risk were similar across the individual PPIs, with the exception of rabeprazole, for which there were limited data. In secondary analyses, we found no significant associations between the use of PPIs during the first trimester and subgroups of major birth defects according to organ system.
Our initial analysis defined exposure to PPIs as the filling of a prescription at some time between 4 weeks before conception and the end of the first trimester, since we wanted to include data from women who had started taking PPIs before conception and continued taking them during early pregnancy. When exposure within 4 weeks before conception was included in the main analysis, there was a significant association between exposure to PPIs and major birth defects. When separate analyses were performed on data from women who filled prescriptions within 4 weeks before conception and from those who filled prescriptions in the first trimester, only women who received PPIs within 4 weeks before conception were at increased risk for having infants with birth defects. Furthermore, when the analyses accounted for the daily doses of PPIs in prescriptions filled before conception, only infants whose mothers took PPIs before the estimated date of conception but did not receive enough doses to have a theoretical chance of continued exposure beyond conception were at increased risk for birth defects. Analyses of data from alternative exposure-time windows supported the main results. Although some misclassification of the timing of exposure is possible17 given the uncertainty regarding the dates of conception, the results of analyses focusing on exposures during weeks 3 to 8 after presumed conception were consistent with the main results involving exposure during the 12 weeks after presumed conception.
The observed association between exposure to PPIs within the month before conception and major birth defects may represent unmeasured confounding or may have been due to chance. The plasma half-lives of PPIs range between 1 and 2 hours; therefore, it is unlikely that there was a carryover effect from exposure before conception into early pregnancy.
Our results with respect to the effect of exposure to PPIs during the first trimester are consistent with findings from previous research involving various types of cohorts, including registry-based historical cohorts, prospective cohorts of mothers seeking advice from teratology services, and cohorts identified from pregnancy registries.7-9,11,23,24 A meta-analysis of these studies, which was based on 1530 women exposed to PPIs in early pregnancy, showed that there was no significant increase in the risk of birth defects associated with exposure to PPIs (odds ratio, 1.12; 95% CI, 0.86 to 1.45).6 Our study, which had a substantially larger cohort than the samples in all the previous reports collectively, confirms the findings from previous reports regarding omeprazole and extends those findings to other PPIs.
Our study covered a period of 13 years and involved a nationwide population and independent ascertainment of exposure and outcome. Several previous studies7,8,23 did not follow children beyond the immediate neonatal period, and defects diagnosed later would not have been included in the analyses, whereas our study had a 1-year follow-up period for birth defects. Although we adjusted for several potential confounders, it is possible that there were confounding factors that we did not identify. Our primary concern would be factors that could have masked a risk of birth defects associated with the use of PPIs. Given the size of the cohort, such factors would have had to be common or would have had to be strongly associated with both the use of PPIs and a reduced risk of defects. We consider it unlikely that there were unmeasured confounding factors that met these criteria.
We used a registry-based case-finding strategy for the identification of birth defects, and the outcome may have been subject to minor misclassification.18 Although any misclassification was probably random, it could bias the results toward no effect. We used filled prescriptions as proxies for exposure to PPIs. If women did not take the dispensed PPIs, however, the results would be biased toward no effect, and teratogenic effects, if present, could be obscured. Omeprazole and lansoprazole became available as over-the-counter drugs during the last years of the study. Misclassification of exposure on the basis of the use of over-the-counter preparations by women who were classified as not having been exposed to PPIs could similarly bias the results toward no effect. However, the analysis that was restricted to the period when PPIs were available by prescription only provides further reassurance that misclassification of exposure was not an important source of bias.
The main outcome measure — all major birth defects combined — may have some shortcomings. Teratogens typically cause specific defects or groups of defects and do not necessarily increase the rate of birth defects overall.25 Nevertheless, composite outcomes are often used in studies of birth defects26,27 and offer opportunities to identify previously unknown or unsuspected associations. A limitation of cohort studies of births defects is that specific defects are uncommon, and therefore the power to detect associations with individual defects is limited.25 Our analyses of subgroups of birth defects should therefore be interpreted with caution. Large case–control studies28,29 may provide opportunities to investigate specific defects with sufficient power.
Most studies of the safety of PPIs during pregnancy have involved women exposed to omeprazole. The results of our study add to the data supporting the safety of this drug with regard to birth defects. Moreover, research that showed no increase in the risk of spontaneous abortions or preterm births associated with the use of PPIs was also based largely on exposure to omeprazole.6 Further study is needed to address the safety of PPIs with regard to perinatal outcomes, as well as the outcomes when PPIs are taken during lactation, and to address specific birth defects and potential long-term risks associated with individual PPIs.30
In conclusion, in this nationwide cohort study, we found no significant association between the use of PPIs during the first trimester of pregnancy and the risk of major birth defects. These results provide reassurance that PPIs, and omeprazole in particular, can be used relatively safely during the first trimester.
No comments:
Post a Comment