Single Data Point Experiments: Birth and Male Methotrexate Users

There’s a line about astronomers using a single data point with huge error bars to fit lines to that you hear in many physics talks and lectures. This is usually to cover up the fact that the data being discussed has only a few points with even larger errors bars, but it does pose the question about how you interpret real world limited data sets. In the physical sciences it’s usually straightforward to run the experiment again (unless you are looking at a supernova), but in the life sciences gaining data can be far more problematic.

An example of this difficultly has been evident in my home life recently, with the birth of my daughter last week. I have fairly mild (earlish onset) rheumatoid arthritis, for which I have taken various different medications over the years. The most current one is called methotrexate and doubles up as a chemotherapy drug if you happen to have cancer. There are lots of potential side effects, as with most drugs, but with methotrexate they are a little more probable than with others. You have to have continuing blood tests to make sure your liver function is OK, for example. One significant warning is that you should not get pregnant while taking it, and that this warning applies to men as well as women. And then we found out we were going to have a baby…

Initially there is the concern that something very bad is likely to happen to your baby. The warnings are clear, DO NOT get pregnant while taking this drug. The possible side effects for the baby are also clear. Methotrexate can cause fetal abnormalities, especially head and limb deformities. The reasons seem to be that folic acid is an antagonist for folic acid, and a reduction in folic acid during pregnancy and prior to pregnancy leads to such deformities. The other major problem is that methotrexate would appear to lead to spontaneous abortions in some cases, and it can be used as part of a planned termination, for example.

At times like these, you turn to the internet, and you find that the answers are not as clear cut as the labels on the bottle make out. Because of the different uses of methotrexate there are different dosage regimes, and for arthritis (and it should be noted that it is not medically clear why methotrexate works for arthritis, which I always find odd) you are typically on a much lower dose compared to someone with cancer. In addition, while there seems to be a body of data dealing with females taking methotrexate while in pregnancy, there are hardly any trials or significant data on what happens if the male is taking it when conception happens. For women, I think the data points to a significant risk of birth deformities should you get that far, for men it is hard to say. Certainly for low doses there seems to be no evidence (either way) that there is any adverse effect. The problem is that there is really no way to carry out a trial, other than by monitoring those births that take place. And as far as I am aware there is no-one doing this. So I am very much a single data point.

The consultant that we saw when we discussed these issues didn’t seem to have much of an clear idea bout the risks. She looked up some papers, as I did  I, and showed them to us. She suggested that if the pregnancy got past the third month all such be well, as any chromosomal damage would likely be significant and would cause auto-termination. She had another patient who had done the same thing and all had been well. Her thinking seemed sound, and was more or less the same conclusion my wife and I came to having read the literature that we could access. But could the evidence be stronger? My thinking is that this data is valuable, and if every consultant has,say, one or two such cases a year, they could pool that data and produce a clear analysis. What good would that do? Well, it would bring clarity to the risks and outcome eventualities. What if no child born to an adult male taking low dose methotrexate in the UK ever had deformities (or at least no more than is statistically normal)? We could then alter the drug safety advice and, I suspect, make a a good number of pregnancies a lot more stress free, which has to be good for mother, baby and father alike.

Having written this, I suppose I should try and do something about collating this data myself. So maybe watch this space for a male methotrexate fathers data collection website. More generally, it would be really great if the NHS could be more joined up about this, and where unusual medical issues arise they should be collated. Open data is what is needed, and then people to mine it.

Our baby is three days old, is doing well and showing no ill-effects of her stress filled start to life.