Pills, pregnancy and sons' fertility

“Common painkillers could be linked to male reproductive disorder,” The Guardian reported today, saying that scientists have found a link between the use of mild painkillers in pregnancy and the risk of giving birth to a son with undescended testes. Many newspapers reported on the research behind this news, with The Independent suggesting that painkillers may even be linked ‘to global sperm decline’.

The study featured both an analysis of pregnant women’s medication use and animal research looking at the development of rats. During some of its analyses the study found the timing and duration of mild painkiller use during pregnancy was linked to the risk of undescended testes in male babies.

There are some shortcomings to the research, including small sample sizes and the number of analyses involved reducing the certainty of its finding. However, it does highlight an important avenue for more research.

It should be noted that the rates of undescended testes seen were still relatively low.

Current advice states that pregnant women should avoid ibuprofen and aspirin during pregnancy, although there is no evidence that occasional use of paracetamol is harmful. The results of this study are unlikely to change those recommendations, but women should seek advice from their GP or midwife before taken any medications during their pregnancy.

Where did the story come from?

The study was carried out by researchers from the University of Copenhagen and other academic and medical institutions in Denmark, Finland and France. The research was funded by the European Commission, the Villum Kann Rasmussen Foundation, the Novo Nordisk Foundation, Inserm and France’s Ministère de l’Enseignement Supérieur et de la Recherche. The study is soon to be published fully in the peer-reviewed medical journal, Human Reproduction.

The papers have generally covered this research well although they focus mainly on the top line summaries of the study and do not highlight the shortcomings of the research that should be kept in mind when interpreting its findings. The study also included a number of subgroup analyses, and different news sources have each focused on the varying findings from these subanalyses. For example, BBC News reported a ‘seven-fold’ increase in risk, and The Guardian a 16-fold risk, while others report more-than-doubled risk.

What kind of research was this?

The researchers were investigating the theory that there is a link between exposure to mild painkillers and reduced masculinisation, which originated from the results of animal studies in the 1980s. They say that there has been an increase in human male reproductive disorders in recent decades and that studies suggest lifestyle and environmental factors may play a key role.

There were two main phases to this study: a human study phase and a second phase of animal research. In the human study, the researchers undertook a cohort study that initially surveyed 2,297 Danish and Finnish pregnant women, asking about details of their pregnancy and assessing their pregnancy outcomes. This was carried out to establish whether there was any link between exposure to mild painkillers during pregnancy and undescended testes in their newborn boys, a condition known as ‘congenital cryptorchidism’.

In the animal experiments, the researchers assessed exposure to painkillers on the developing foetuses, measuring markers of their masculinisation and their levels of testosterone.

What did the research involve?

The human study enrolled 2,297 pregnant women from two hospitals, one in Denmark, one in Finland. The women were questioned using a written questionnaire or a telephone interview (some of the Danish women only) during their third trimester. After birth their babies were then examined for cryptorchidism. The researchers based their main analysis on the 491 women who were carrying boys and who had completed the telephone interview.

The questionnaire asked about their health and use of medicines during pregnancy. It asked if women had taken any medication during their current pregnancy and, if so, to specify what they took, for what reason, how much and during which week of the pregnancy.

Women completing the telephone interview were asked a more targeted question: “Have you taken any pain-relief during this pregnancy, e.g. normal painkillers or stronger brands?” If they answered yes, they were asked further questions about the type of product and when they took it. Some women completed both the telephone interview and the questionnaire.

More women reported using painkillers during the telephone interview so the researchers decided to perform an analysis of only the Danish women who were interviewed over the phone (491 women). As Finnish women were only assessed using the written questionnaire, all 1286 were analysed separately.

Several different analyses were performed, assessing both general use of painkillers and use of specific types of painkillers (paracetamol, ibuprofen, aspirin, and ‘simultaneous use of >1 compound’). The analyses also looked at the women’s use of painkillers across the full term of pregnancy and during just the first and second trimesters. The analyses were adjusted for diseases, use of other medications, carrying twins and gestational age of the child.

Further analyses were undertaken according to non-use of painkillers, use for one to two weeks and use for more than two weeks during first and second trimester.

The animal study involved feeding pregnant rats with different doses of paracetamol and aspirin and then investigating the effects of these substances on the masculinisation of developing foetuses. The masculinisation was measured in two ways, first using an anatomical feature commonly used for this purpose, called the ‘anogenital distance’. This is the distance between the anus and the base of the penis, a distance that is shorter when in-utero testosterone levels have been lower. The concentration of testosterone in extracted testes was also measured.

What were the basic results?

Analysis of the 491 Danish women found no significant link between overall use of mild analgesics during pregnancy overall and congenital cryptorchidism compared to reporting no use. There was also no link between use of painkillers in the first trimester and the condition, although use in the second trimester increased the likelihood by 2.3 times (95% CI 1.12 to 4.73). The risk of the disorder was increased in women reporting painkiller use for more than two weeks compared with no use.

A number of different subgroup analyses were performed.These found the abnormality was linked to simultaneous use of more than one compound over the whole pregnancy, use of aspirin in the first trimester, use of mild painkillers (any) in the second trimester, plus the use of aspirin, ibuprofen or more than one compound in the second trimester.

In the animal studies, exposure to painkillers reduced the anogenital distance more than in a control rat, suggesting reduced exposure to testosterone. Maternal body weight gain, litter size and number of live foetuses were not affected. Painkiller exposure also reduced the levels of testosterone in the males in the litters.

How did the researchers interpret the results?
The researchers say that collectively their results point to a scenario where the use of mild painkillers in pregnancy has a “possible effect on feotal development”. They say that more investigation is urgently needed and that they intend to follow up their participants as the boys are now entering puberty.

Conclusion

This cohort study with supplementary evidence from animal research provides some evidence that the use of painkillers could have an effect on rates of undescended testes in boys. It is preliminary but important evidence, although at the moment, it is unlikely to change the current recommendations to pregnant women. These are to: avoid medications in general and to use paracetamol rather than ibuprofen or aspirin if painkillers are needed. However, advice should always be sought from a GP or midwife before taking medication.

There are several shortcomings with this research that should be kept in mind when interpreting the findings. These have not been highlighted by the news articles:

• The different response rates about painkillers use suggest that the questionnaires used a less accurate method of asking about these drugs. This was so much so that the researchers limit their analyses of Danish women to just those who were interviewed over the phone and do not draw conclusions from their results from Finnish women, which showed no link between painkiller use and crytorchidism.
• The researchers perform several subgroup analyses on their data and do not adjust for multiple comparisons. This increases the likelihood that they have found false positive associations, i.e. they conclude there is a link between painkiller use and cryptorchidism when in fact there is not one.
• Of the 17 different subgroup analyses presented in their main results table, only six results showed a significant association. All of these had wide confidence intervals (meaning the result is not precise) and were based on small samples. Only 10 interviewed women reported taking more than one compound during their entire pregnancy, for example.
• Many of the subgroups have only small numbers of women in them. Overall, the analyses were based on quite a small group of boys because of the exclusion of women responding to the questionnaire and the fact that only 42 boys from 491 interviewed mothers had cryptorchidism. Some of the subgroups were analysing very small numbers of people. The analysis of small groups means the results are not robust, as demonstrated by the lack of precision in the results’ wide confidence intervals.
• As reported by The Guardian , the study found a 16-fold increased risk if women took more than one type of painkiller in their second trimester. However, once again the small numbers involved hamper the reliability of this finding as only seven women of the 491 analysed reported this type of painkiller use. (95% CI 3.29 to 78.6).
• Women responded to questions in their third trimester so they may not have accurately remembered all the medication they took throughout their pregnancy and when they used it. They would be more likely to remember medicine that they had recently taken.
• The researchers explain the lack of significance in the Finnish sample by saying that this aspect of their study may have been underpowered because fewer boys in Finland are born with this disorder.
• There was no adjustment for the health of the baby. Cryptorchidism can occur alongside other genetic abnormalities, and maternal alcohol consumption is also considered a risk factor.
• Even if the risk increases suggested by these results are accurate, the overall risk of cryptorchidism is quite low (about 8% of the population).

Overall, this study does not prove a link between use of painkillers in pregnancy and this reproductive disorder in boys but it does suggest an important avenue for more research. The evidence is certainly not strong enough yet to say that the global decline in sperm counts may be due to the use of mild painkillers. The limitations of this exploratory research, most notably the small numbers involved in some analyses, mean that mothers and pregnant women should not be worried by these results.

For women wondering what to take away from these results, the advice is that it is safest not to take these drugs during pregnancy. Specifically, the advice is to avoid ibuprofen and aspirin during pregnancy but there is no evidence that occasional use of paracetamol is harmful. The results of this study are unlikely to change those recommendations, but pregnant women should always seek advice from their GP or midwife before taken any medications, including painkillers.