Charlton Rachel A, Cunnington Marianne C, de Vries Corinne S, Weil John G
Worldwide Epidemiology, GlaxoSmithKline, Harlow, Essex, UK.
Drug Saf. 2008;31(1):39-51. doi: 10.2165/00002018-200831010-00004.
Pregnancy registries are the most commonly used data resource for the post-marketing surveillance of drug teratogenicity. However, the limited sample size and potential selection bias in these registries has led us to investigate the potential of the UK General Practice Research Database (GPRD) as an alternative data source for monitoring drug safety during pregnancy. In addition, a literature review identified further observational data sources that monitor pregnancy outcomes for future evaluation. Initial feasibility studies focused on the ability of the GPRD to capture pregnancy outcomes for a range of drug class exposures, all of which are currently under investigation in pregnancy registries, during pregnancy. The comparator pregnancy registries were identified via a MEDLINE search, whilst eligible pregnancies, in which women received one or more prescriptions for the drug of interest during pregnancy, were identified in the GPRD using the mother-baby link. The number of pregnancy outcomes following exposure to medication for a range of conditions with varying prevalence, including depression, migraine, epilepsy, herpes simplex and HIV, captured by the two data sources were compared. For depression, a relatively prevalent condition, the GPRD recorded the same number of mean annual intrauterine exposures to fluoxetine as the pregnancy registry (118 exposures/year). Ascertainment of intrauterine exposure to drug treatments for less prevalent conditions was found to be higher for the pregnancy registries than the GPRD; for the older antiepileptic drugs (valproate and carbamazepine), the pregnancy registry recorded between four and five times as many mean annual exposures as the GPRD. Virtually no antiretroviral exposures (three) were identified during the time period of interest on the GPRD, compared with 3946 in the Antiretroviral Pregnancy Registry. Data from the GPRD meet established criteria for evaluating outcomes of pregnancy. For prevalent conditions, it has the potential to replace or work alongside standard pregnancy registries and the alternative data sources identified. Further studies are now needed to assess its ability to replicate known teratogenic associations.
妊娠登记是药物致畸性上市后监测最常用的数据来源。然而，这些登记中样本量有限和潜在的选择偏倚促使我们研究英国全科医疗研究数据库（GPRD）作为孕期药物安全性监测替代数据源的潜力。此外，文献综述确定了其他监测妊娠结局以供未来评估的观察性数据源。初步可行性研究聚焦于GPRD捕捉一系列药物类别暴露的妊娠结局的能力，目前所有这些暴露情况在妊娠登记中均正在研究。通过MEDLINE检索确定了对照妊娠登记，而使用母婴关联在GPRD中确定了符合条件的妊娠，即妇女在孕期接受了一种或多种所关注药物的处方。比较了两个数据源捕捉到的暴露于治疗一系列患病率各异病症（包括抑郁症、偏头痛、癫痫、单纯疱疹和艾滋病毒）药物后的妊娠结局数量。对于抑郁症这种相对常见的病症，GPRD记录的每年平均宫内暴露于氟西汀的数量与妊娠登记相同（每年118例暴露）。发现妊娠登记对于不太常见病症的宫内药物治疗暴露的确定率高于GPRD；对于较老的抗癫痫药物（丙戊酸盐和卡马西平），妊娠登记记录的每年平均暴露数量是GPRD的四至五倍。在感兴趣的时间段内，GPRD几乎未识别出抗逆转录病毒暴露（3例），而抗逆转录病毒妊娠登记中为3946例。GPRD的数据符合评估妊娠结局的既定标准。对于常见病症，它有潜力替代标准妊娠登记或与之协同工作，并替代已确定的其他数据源。现在需要进一步研究来评估其复制已知致畸关联的能力。
