An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
. 2001 Jun 2;322(7298):1365.
Risk of miscarriage in pregnant users of NSAIDs
More information is needed to be able to interpret study's results
Editor—Nielsen et al found that the use of non-steroidal anti-inflammatory drugs in early pregnancy was associated with an increased risk of miscarriage but not with congenital abnormality, low birth weight, or preterm delivery.1 Unfortunately, the study has several deficiencies that make interpretation of the results difficult.
Firstly, the study was based on the Danish birth registry and the prescription database; the clinical information available was limited, and a random sample confirmed the prescription in only 71% of the pregnancies. Although the authors state that the data had high validity, Kristensen et al reported that the birth registry under-reported the incidence of preterm delivery by over 50%.2
Secondly, in assessing the risk of miscarriage the authors stratified the study group according to the time between the prescription of non-steroidal anti-inflammatory drugs and the date of discharge from hospital after miscarriage. How the control group was stratified is not clear. Without this information, any conclusion about the association between exposure to non-steroidal anti-inflammatory drugs and miscarriage cannot be substantiated.
Thirdly, to establish the association between exposure to non-steroidal anti-inflammatory drugs in the first trimester and the risk of miscarriage the possible confounding factors must be controlled for. Although the prescriptions were “mostly for benign conditions of the muscles and skeleton,” the exact frequency and indications for prescription were not clearly stated. A patient with systemic lupus erythematosus is more likely to take up a prescription for these drugs, and carries a higher risk of miscarriage because of the underlying disease. The possibility of confounding by indications for treatment with non-steroidal anti-inflammatory drugs cannot be excluded in this study.
Lastly, the study looked only at the prescription of non-steroidal anti-inflammatory drugs, not at the actual intake of the drugs. No information is given on compliance. Women tend to avoid any drug treatment during pregnancy, so the actual consumption was probably low. The conclusion that use of non-steroidal anti-inflammatory drugs during pregnancy is associated with an increased risk of miscarriage requires further confirmation.
References
1.Nielsen GL, Sørensen HT, Larsen H, Pedersen L. Risk of adverse birth outcome and miscarriage in pregnant users of non-steroidal anti-inflammatory drugs: population based observational study and case-control study. BMJ. 2001;322:266–270. doi: 10.1136/bmj.322.7281.266. . (3 February.) [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—Use of non-steroidal anti-inflammatory drugs is associated with an increased risk of miscarriage, according to Nielsen et al.1-1 Their result may be biased because induced abortions were not considered in establishing the study base. It is based on registration of 63 drug prescriptions (1.5%) in 4268 miscarriages being compared with 318 first trimester prescriptions (1.1%) in 29 750 live births. The appropriate comparison should rather be all (or a sample of) pregnancies at risk of miscarriage, including abortions.1-2
Miscarriages may occur in women who had intended to have an abortion. In Denmark in the 1990s this might have been a fifth of the miscarriages, or even more if miscarriage and the intention to have an abortion were associated. Some of the women who had a miscarriage could have been subject to a prescribing practice not influenced by the pregnancy, because they did not intend to continue with the pregnancy.
If a quarter of women with miscarriages intended to have an abortion, of whom 3% had a prescription for a non-steroidal anti-inflammatory drug, this bias could explain the apparent association. This systematic difference between the group who had a miscarriage and the group who had live births could also explain why prescriptions tended to decrease more steeply in the livebirth group than the miscarriage group over the first weeks of pregnancy, as indicated by data provided in table 3 of the paper.
References
1-1.Nielsen GL, Sørensen HT, Larsen H, Pedersen L. Risk of adverse birth outcome and miscarriage in pregnant users of non-steroidal anti-inflammatory drugs: population based observational study and case-control study. BMJ. 2001;322:266–270. doi: 10.1136/bmj.322.7281.266. . (3 February.) [DOI] [PMC free article] [PubMed] [Google Scholar]
1-2.Olsen J. Calculating risk ratios for spontaneous abortions: the problem of induced abortions. Int J Epidemiol. 1984;13:347–350. doi: 10.1093/ije/13.3.347. [DOI] [PubMed] [Google Scholar]
Editor—As we pointed out in our paper, the analyses were based solely on linking of registry data; we did not take into account the indication for which the drugs were used or compliance. The information about exposure and outcome was collected independently; misclassification due to low compliance and use of over the counter preparations was therefore probably similar in the cases and controls and would tend to underestimate our observed risk estimates.
In the discussion of the miscarriages we emphasised that we could not control for confounding by indication. The odds ratio of 6.99, however, is higher than is usually ascribed solely to confounding. The register unfortunately does not include gestational age in cases of miscarriage. As most miscarriages occur in the first trimester, the best available control group comprised women in the first trimester who went on to have live births.
A written confirmation of drug prescription in about 70% of the patients is within the limits most often seen when dealing with pharmacoepidemiology.2-1
We would emphasise our reassuring observation of lack of an association between the taking up of prescriptions of non-steroidal anti-inflammatory drugs and adverse pregnancy outcome. We agree with Chan and Yuen that the association between non-steroidal anti-inflammatory drugs and miscarriage must be confirmed in other studies before definitive conclusions can be drawn.
We deliberately restricted our analysis to pregnancies with spontaneous miscarriage to avoid bias from including women who had chosen to terminate their pregnancies by an induced abortion. Women who choose to terminate their pregnancies may differ from those intending to carry them through in several ways—for example, smoking and alcohol habits. In some cases exposure to non-steroidal anti-inflammatory drugs may have been an additional factor in deciding to have an induced abortion. As the waiting time for getting a legally induced abortion is short, we believe that the number of spontaneous miscarriages while waiting for an abortion is low. We think it unlikely that a change in prescribing habits in this short period will substantially change our risk estimates.
References
2-1.West SL, Strom BL. Validity of pharmacoepidemiology drug and diagnosis data. In: Strom BL, editor. Pharmacoepidemiology. Chichester: John Wiley; 1995. [Google Scholar]
