Azara Blog: Caffeine is allegedly linked with miscarriages in early pregnancy

Blog home page | Blog archive

Google   Bookmark and Share

Date published: 2008/01/21

The BBC says:

Pregnant women should consider avoiding caffeine, say researchers who found even moderate consumption in early pregnancy raises the miscarriage risk.

Currently, the Food Standards Agency sets an upper limit during pregnancy of 300mg - or four cups of coffee a day.

But an American Journal of Obstetrics and Gynaecology study found more than 200mg of caffeine a day doubled the risk compared to abstainers.
For the latest research, Dr De-Kun Li and colleagues at the Kaiser Permanente Division of Research, studied 1,063 women who had become pregnant in the last month or two.

They asked the women to provide a detailed diary about their caffeine intake up to their 20th week of pregnancy.

When they compared this information with how many of the women had miscarried by 20 weeks gestation, 172 of the women in total, they found a link.

Compared with non-users, women who consumed up to 200mg of caffeine a day had an increased risk of miscarriage - 15% versus 12%.

For women who drank more than 200mg, the risk increased to 25%.

A classic confusion of correlation and causation, so the research, like most health-related research, really proves nothing. It's a very easy game to play. You come up with a list of things you don't like. You come of with a list of negative effects. You have a bunch of people fill in a questionnaire asking about both lists. You look for correlations. You are bound to find some. You publish your correlations. The BBC and other media pick up the story and forget that it is a correlation and imply it must be a causation. You get publicity and money to do more research. Is it at all surprising that nobody does health studies properly (with random samples) when you can be perpetually funded for doing such weak research? Of course here there could be a causation, as in theory there could be a causation whenever there is a correlation. It's just that the researchers have no evidence there is a causation, so the wink, wink, nudge, nudge implication that there is a causation is just wrong.

All material not included from other sources is copyright For further information or questions email: info [at] cambridge2000 [dot] com (replace "[at]" with "@" and "[dot]" with ".").