Commentary on Raitasalo et al. (2015): the great potential in nation-wide registers to study prescription drug use and abuse.

Abstract

In their paper, Raitasalo et al. 1 identify mothers who are at greater risk for benzodiazepine use; namely, those with substance use problems and psychiatric problems. A second important finding is that the use of benzodiazepines was lowest during pregnancy for all groups, followed by a sharp increase during the years after the children were born. While, from our perspective, the study is important in several ways, this commentary can address only a few of these. One point is that information about the use of addictive prescription drugs among women of childbearing age is scarce; there is a lack of information on both the effect of benzodiazepine use during pregnancy and the effect that maternal use may have on the children. Observational studies using data from prescription registries are therefore an appropriate approach to map the prevalence of benzodiazepine use, as well as the use of other prescription drugs such as z-hypnotics and opioids. By studying any benzodiazepine use, Raitasalo et al.’s study captures the full spectrum of use. Some studies have shown that the misuse of benzodiazepines is associated with several negative risk factors and behaviours 2, 3. While less is known about the consequences of the therapeutic use of benzodiazepines, a recent study refuted any strong association between prenatal therapeutic use of benzodiazepines and lower language competence in offspring at age 3 years 4. This study adds the important knowledge that many female benzodiazepine users of childbearing age belong to groups with other risk factors, such as substance abuse, psychiatric disease or poverty. Thus, Raitasalo et al. show that benzodiazepine use is another factor that might be important in identifying ‘at-risk’ families. In order to identify families that should be offered additional follow-up as a preventive measure, it is therefore important to also include questions about addictive prescription drug use. When clinicians receive a positive response to questions on the use of such prescription drugs there is reason to ask follow-up questions, as they would for a positive response to questions on drinking or depression. Benzodiazepine users are often co-medicated with other prescription drugs, such as other hypnotics and opioids 4-8. As co-occurring use is common, this might suggest that mothers who use hypnotics or opioids should receive the same attention from clinicians trying to identify at-risk families as do mothers using benzodiazepines. Further studies on the effects of various patterns of benzodiazepine use in different life periods are warranted. Raitasalo et al. point out that previous research is inconsistent with regard to whether or not in-utero exposure to benzodiazepines increases the risk of negative birth outcomes 1, 9-11. Clearly, this should be given further attention. While randomized controlled trials are the gold standard with which to study outcomes from prescription drug use, pregnant women are often excluded from clinical trials. The large, prospective observational data available in nation-wide registers may therefore be the most effective way to investigate possible teratogenic effects from prescription drugs 12. The sharp increase in benzodiazepine use during the years after childbirth is cause for concern, as parental prescription drug use is associated with prescription drug use in their children as adults 13. For instance, repeated use of both benzodiazepines and opioids among mothers is associated with the increased risk that their children also become repeated users of benzodiazepines and opioids, respectively 14. The study by Raitasalo et al. 1 underscores some of the opportunities offered by nation-wide registers. One advantage is the inclusion of whole populations; a second advantage is that registry data allow for large samples, well suited to study rare outcomes: for instance, short- and long-term outcomes in children exposed to benzodiazepines in utero. A third advantage is that when personal identification numbers are assigned to all residents, such as in Nordic countries, this enables the matching of person-level data from several registers. This allows for controlling for important confounding variables such as education, employment and marital status, children, mental and physical health, and allows for data matching for parents and children. Another advantage is that as the registers are established and in use and data are available, they represent a relatively cost- and time-effective way of conducting large-scale prospective studies. Registry studies such as the one by Raitasalo and colleagues are important, because they provide both a complete description of drug use that might be problematic and needing further research attention, and as a background from which to enable clinicians to identify families at risk of different factors and behaviours. None.