39,44 This study has some limitations. Because of low exposure numbers, subanalyses for the individual triptans apart from sumatriptan were not feasible. The study was restricted to major congenital malformations detected at delivery or the following days in hospital because of the fact that ascertainment of minor malformations often occurs only after discharge from hospital. However, it is reassuring that several validation Galunisertib in vitro studies have shown a high accuracy of the registration of major malformations.45,46 The study was based upon self-reported

migraine pharmacotherapy with a possible underreporting of drug use. However, there is no reason to believe this reporting was differential among the women as the data were collected prospectively before the pregnancy outcome was known. The second questionnaire only covered triptan use up to gestational PLX-4720 mw week 30, and this may have led to the loss of data on triptan

therapy beyond this point in time. Migraine diagnosis has not been validated, and the categorization of the 3 study groups depended on the accuracy of the women’s reporting. This could have led to an underestimation of the strengths of the associations found during the logistic regression analyses. Finally, only 42% of the invited mothers agreed to participate in this study during the period between 1999 and 2006. However, as the differences MCE between the participants in the Norwegian Mother and Child Cohort Study and the general population of pregnant women are only minor,47,48 the estimates of the various associations are most likely to be valid also in the general population of pregnant women. The study also has several strengths. The study population consisted of more than 65,000 pregnant women and their infants, representing the largest study population on triptan use during pregnancy. The vast spectrum

of health-related and sociodemographic data derived from both the Medical Birth Registry of Norway and the Norwegian Mother and Child Cohort Study enabled analyses to be performed on the associations between triptan use prior to and during pregnancy and various pregnancy outcomes while controlling for several important potential confounders. Because of the prospective nature of data collection, the risk of recall bias was avoided. The inclusion of the second control group, consisting of women who only used triptan therapy prior to pregnancy, allowed for the controlling for any possible disease effects. No previous studies have included such a control group. Finally, despite the relatively low frequency of the use of triptans other than sumatriptan, this study also presents the possible effect of these drugs on pregnancy outcome – information that has so far been quite limited or even nonexistent.