The importance of critical evaluation of the literature regarding safety of antidepressant use in pregnancy

The important question about antidepressant use in pregnancy is reviewed by Byatt and colleagues in this issue of Acta Psychiatrica Scandinavica 1. The authors main conclusions following a review of the literature focusing on controversial results, including 44 studies with (n = 18) reporting on major malformations, (n = 18) on neonatal behavioral symptoms, and (n = 8) on persistent pulmonary hypertension of the newborn (PPHN), are that while some studies suggest an increased risk of specific major malformations, the findings are inconsistent, although the overall risks appear to be small. Neonatal behavioral symptoms can occur in up to 30% of neonates exposed to antidepressants, and in some studies, PPHN has been weakly associated with in utero antidepressant exposure. They also recommend taking into account untreated maternal illness, as this may also exert its own adverse effects on the infant. In addition, they offer some valuable advice in interpreting results of studies, so as to assist the clinician in the treatment of a pregnant and depressed woman. A substantial number of women will become pregnant while taking an antidepressant and may require pharmacological treatment throughout their pregnancy. As almost half of all pregnancies are unintended, this represents a large number of pregnant women taking an antidepressant. Prior to 2005, research using observational designs conducted on the use of SSRIs (Serotonin Reuptake Inhibitors) in pregnancy reported no association between SSRI use and congenital malformations 2, which was based on a meta-analysis of the current data at that time. They appeared to be relatively safe to take during pregnancy and there was no evidence of concern in both the scientific literature and the media. However, this all changed in late 2005, when GlaxoSmithKline (GSK) published on their website, pregnancy outcomes of 815 infants exposed to paroxetine during pregnancy and reported a 2% incidence of cardiovascular malformations (where 1% is expected in the general population). Subsequently, based on this report and data from two unpublished abstracts presented at scientific meetings, The Food and Drug Administration (FDA) issued a warning about the possible adverse effects associated with paroxetine in the first trimester of pregnancy. Subsequently, as Byatt et.al 1 reported, many studies and reviews have been published with ‘conflicting results’ reporting on thousands of pregnancy outcomes. A recent (August 2nd 2012) GOOGLE search using the keywords ‘antidepressants, pregnancy, birth defects’ revealed 1 550 000 results, many describing how ‘dangerous/harmful’ antidepressants are to take in pregnancy and warning women ‘not to take them if they are pregnant.’ In addition, there are many websites encouraging women who had a baby who had PPHN or a birth defect and took an antidepressant during pregnancy, to seek the advice of a lawyer to sue the drug company. Even ‘experts’ disagree regarding the teratogenicity of antidepressants, most notably paroxetine as illustrated by 2 commentaries in a prominent teratology journal, which were published along with a meta-analysis and presented opposing opinions. Scialli concluded that ‘the scientific evidence does not support the conclusion that paroxetine causes cardiovascular defects,’ while Bérard stated that ‘evidence-based literature shows consistent epidemiologic evidence that paroxetine use during pregnancy increases the risk of cardiac malformations in newborns.’3 From these statements, one is prompted to question how it is that two experts in the same field have offered such opposing conclusions based on their evaluation of the same data. Understandably, this mixed information has caused anxiety for pregnant women, who may require pharmacological treatment for the depression and for their healthcare providers from whom they seek advice. Due to the ethical restrictions of randomized controlled trials (RCTs) in pregnant women, studying the safety of drugs in pregnancy is a complex process. Consequently, observational study designs (i.e., case reports, case series, cohort studies, case–control and nested case–control studies and administrative database studies) are currently used, which have many limitations. Recent years have seen an increase in the number of computerized databases, which were not designed for scientific studies, especially prescription data bases, where it is only known if the woman redeemed the prescription, not if she actually took the drug. Nevertheless, researchers have used these databases to conduct complex analyses of data, resulting in a substantial increase in such studies. Together with other observational studies using different methodologies, seemingly conflicting results have been published in the peer-reviewed literature and subsequently in the media regarding the safety of antidepressant use in pregnancy. This brings up a very important question of how are these results disseminated to the scientific community and subsequently to their patients. Most clinicians have a very rudimentary knowledge of epidemiology and statistics, so do not have the skills to carefully examine the often complex methodology of these studies, so rely on the conclusions of the authors, which do not always match their results, most often just from reading the abstract in the journal. When individual studies regarding the safety of antidepressants in pregnancy are published, much is made of very small increased odds ratios (OR), usually <2 (which most epidemiologists consider relatively unimportant) and are explained in a way that it appears much more significant than it really is. Small but statistically significant risks are the key at the population level, but most often are not clinically important, which is the information that a clinician requires to inform the patient of their individual risk. In addition, studies that did not find an increased risk are frequently ignored by both the scientific literature and in the media. It is rare to see in the media that a new study reporting on the safety/risk of an antidepressant did not find an increased risk for birth defects or other adverse outcomes. Consequently, studies with marginally significant results can have a far reaching impact on events such as precedent setting lawsuits. In the case of GlaxoSmithKline, the company was ordered to pay $1.5 million to a couple whose baby was born with a heart defect following exposure to Paxil® in pregnancy. By July 2010, the company reportedly settled about 800 Paxil® birth defects lawsuits for approximately $1 billion 4. In addition, widely disseminated frightening headlines in the media can also cause women to stop taking a needed antidepressant during pregnancy, sometimes with serious consequences, such as contemplating suicide 5. In summary, this is observational research, and consequently, there are some deficiencies in study design and analysis among all of the studies. However, this does not mean that the information provided from the results of these studies is not valuable, as long as the methodology and analysis are critically evaluated. It is unlikely that in the near future, pregnant women will be included in randomized controlled trials, so this reinforces the need to improve the rigor of the available study methods. However, the bottom line is that current differing research designs regarding the safety of antidepressants in pregnancy did not produce conflicting results per se. Apparent small differences appear to have been likely because of the way selected results were disseminated. Finally and of great importance, improved knowledge transfer and translation will ensure that pregnant women and their healthcare providers receive the most accurate evidence-based information, for decision-making regarding the use of antidepressants during pregnancy. Adrienne Einarson for The Motherisk Program received an unrestricted educational grant to study the safety of Cymbalta®(duloxetine) in pregnancy from Eli Lilly Inc. Canada.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.