Do Epidurals Decrease Maternal Morbidity?

I recently saw an article that stated that epidural use could decrease severe maternal morbidity risk. They found that the use of an epidural in vaginal deliveries was associated with a decrease in the risk for severe maternal morbidity — and this was true regardless of the woman’s race. That seems really significant. What is your take on this study?

—Caren

This is a very nice study; you can see it in full here. The authors’ bottom line is that use of an epidural decreases severe maternal morbidity by 14%, mostly driven by a 9% decrease in the risk of postpartum hemorrhage. By way of background: postpartum hemorrhage is defined as a loss of more than 1,000 milliliters of blood, which can often be treated but is scary and dangerous. This complication occurs in about 2% of births; this paper therefore implies that an epidural would reduce the risk by about 0.2%, or 2 in 1,000. This effect is significant, but it is small in magnitude.

The paper is an observational study — they compare women who had epidurals with those who did not — and as you would expect, the samples are quite different. Epidurals are more common among women who give birth in teaching hospitals, for example, and those hospitals have better resources on other dimensions, too. These differences are a concern but what is effective about their analysis, though, is that they have a huge sample: seven years of all births in New York, for a sample size of almost 600,000. This allows them to better match the women with epidurals to those who do not have them (more on this strategy here).

When they do this matching, they are able to better compare across similar women, and they find this reduction in risk for those who use an epidural. The authors appropriately caveat that they cannot be sure it’s the epidural versus other unmeasured features of the two groups. And they cannot know why this is happening, although they suggest it may be that it’s faster to intervene on some risks if women already have an epidural (because some of these interventions hurt).

Is this a reason to have an epidural if you otherwise do not want one? Perhaps not, given the small effects. But I share the view of the authors that it is a reason that epidurals should be available to everyone if they do want them. And certainly this deserves more research.