Wednesday, November 9, 2011

Study Finds NSAIDs Linked to Miscarriage

NSAIDs In a newly reported study, using nonsteroidal anti-inflammatory drugs (NSAIDs) during pregnancy more than doubled the odds for a miscarriage. Risks were greatest with diclofenac, and aspirin was not examined; however, the results suggest that all NSAIDs should be used cautiously in women who are or who may become pregnant.

Writing in an early online edition of the Canadian Medical Association Journal (CMAJ), a research team in Quebec, Canada, report building a pregnancy registry by combing information on women from 3 large healthcare databases in the region [Nakhai-Pour, et al. 2011]. Using a retrospective case-control design, they then analyzed information for 4,705 cases (mean age, 29 years) of women from the registry who experienced spontaneously aborted pregnancies (miscarriage); this excluded women who had a planned abortion or preterm delivery after 20 weeks of gestation, or who were exposed to misoprostol, NSAID suppositories, or known teratogens. For each case, the researchers randomly selected 10 women from the database as controls; these women did not experience miscarriage but were otherwise matched by demographic characteristics and approximate dates of pregnancy.

Among cases, 7.5% had been prescribed non-aspirin NSAIDs prior to their miscarriages, compared with only 2.6% of control subjects. Overall, use of such NSAIDs increased the odds of miscarriage more than 3-fold; and, after adjusting the data to account for various possible confounders — eg, comorbidities, comedications, other factors — the odds ratio was 2.43 (95% Confidence Interval, 2.12-2.79). In general, women who had miscarriages were slightly older, had received social assistance, had more comorbidities in the year before pregnancy, and had used antidepressants, oral corticosteroids, or anti-infective agents to a significantly greater extent, but antiemetic agents to a lesser extent.

All types of non-aspirin NSAIDs significantly increased the odds for miscarriage; however, the highest risk occurred among women who used diclofenac (Adjusted Odds Ratio [OR]=3.09; 95% CI, 1.96-4.87). This was followed by naproxen (OR=2.64; 95% CI, 2.13-3.28), celecoxib (OR=2.21; 95% CI, 1.42-3.45), ibuprofen (OR=2.19; 95% CI, 1.61-2.96), and rofecoxib (OR=1.83; 95% CI, 1.24-2.70). Women had an increased risk for miscarriage with each agent alone or when used in combination with others (OR=2.64), but there was no dose-response effect observed.

The investigators conclude that exposure to any type or dosage of non-aspirin NSAIDs during early pregnancy may increase the rate of miscarriage, and such analgesics should be used cautiously in these women. In a related news item [here], one of the study authors, Anick Bérard, of the University of Montreal, stated, “If a patient has pain or headache, acetaminophen is a safer alternative during pregnancy. However, if NSAIDs are needed to treat chronic conditions, such as rheumatoid arthritis or lupus, women should carefully plan their pregnancy and discuss treatment options with their physicians.”

COMMENTARY: Unfortunately, as with most studies of NSAIDs, aspirin was not included in this research, so its potential risks are unknown. Furthermore, there are several other limitations to consider…

  • Reasons for NSAIDs being prescribed for the women were not reported, so it is possible that an underlying condition in some cases, rather than the analgesic itself, was a primary precipitating factor in miscarriage.

  • The authors concede that, while their study was large, it only covered about a third of pregnant women in Quebec during the data collection period; therefore, the outcomes might not be generalizable to the wider population.

  • While the database indicated filled prescriptions for NSAIDs, this did not necessarily reflect actual intake of the medications. The best the researchers could surmise was that the women took at least one dose; consequently, it is not known whether problems arise more with higher doses of NSAIDs and if occasional use would be acceptably safe.

  • As a matter of statistical probability, the researchers dutifully acknowledge that there is at least a 5% chance that any of their findings could be false due to random effects. However, the same could be said of any research in which P-values for determining statistical significance are set to ≤0.05 and 95% Confidence Intervals are calculated.

  • Another statistical point is that results are reported in this study as Odds Ratios, which can be somewhat misleading [as discussed in an UPDATE here]. For example, the relative risk of miscarriage for NSAIDs overall would be considerably less that the reported odds ratio of 2.43. Furthermore, data are not provided to calculate absolute risks associated with individual NSAIDs, which might be small, and this also precludes calculating the NNH (number-needed-to-harm) for each drug. So, the outcomes data are not as helpful for clinical application purposes as they could have been.

All of this raises an important question: What are the safest analgesic choices for women during pregnancy?

From this research study it appears that rofecoxib incurs the lowest risk of miscarriage, but this drug was withdrawn from the market due to cardiac safety concerns. Bérard recommends acetaminophen, and at least one review [here] has affirmed the relative safety for the fetus of properly dosed, single-ingredient (but not combination) acetaminophen products during pregnancy. However, a large epidemiological study in Denmark [here] concluded that acetaminophen use during pregnancy significantly increased maternal risk of preeclampsia, chronic hypertension, pulmonary embolism, and deep vein thrombosis. Specific doses were not mentioned.

In an earlier UPDATE [here], we described a large, population-based, case-control investigation in the United States that found an association between maternal use of opioid analgesics during early pregnancy and certain uncommon birth defects, including some types of congenital heart malformations that are important contributors to infant morbidity and mortality. However, the researchers also acknowledged that the absolute risks of these defects occurring for any individual woman and her baby are relatively quite small. In this regard, oxycodone and meperidine demonstrated the lowest risks.

As with all other medications, rarely, if ever, are pregnant women deliberately included in clinical trials of analgesics. So most of the current evidence is based on retrospective assessments of databases that often have limitations or deficiencies. It seems that more definitive research and guidelines are still needed regarding the safest analgesics and their dosages during pregnancy. Meanwhile, practitioners will need to weight relative risks-benefits on a case-by-case basis, as they probably have done all along.

REFERENCE: Nakhai-Pour HR, Broy P, Sheehy O, Bérard A. Use of nonaspirin nonsteroidal anti-inflammatory drugs during pregnancy and the risk of spontaneous abortion. CMAJ. 2011(Oct);183:1713-1720 [PDF here].