Like all invasive medical procedures, they have risks and benefits.
I'm profoundly grateful for the existence of this type of pain relief because it can change the experience of labour from one which is utterly traumatic to one which is positive and calm, and would like all women who want one to have access to it on request.
HOWEVER
I myself decided to try and avoid having another one after having one (which worked very well) with my first. My main reason is a respect for the normal functioning of the hormonal physiology of birth and the immediate postnatal period. I felt very different after my epidural birth in comparison to my non-epidural births, despite my first and third labour being similar in length (ie, loooong!). I definitely felt 'flatter' after my epidural birth. I also didn't want to have to labour in the sort of setting necessary to access epidural analgesia. In other words, it's not just epidural that was the problem for me - it's whole care package that comes with it.
"Epidural do NOT increase the risk of intervention. It is now thought that epidural are correlated with intervention - women who ask for epidurals are more likely having challenging births and need pain relief because their births are challenging, not need intervention because of the epidural."
There is some truth in this. But unpicking the situation is difficult if not impossible, as in all studies looking at the impact of epidural analgesia on labour there is huge cross-over between the epidural and non-epidural arms of trials. This makes a bit of a nonsense of the findings IMO. Plus about 95% of the studies only compare mothers who have had opioids (who are often supine for much of their labour and may be more likely to have an abnormal fetal trace, as are the unborn babies of women who have had an epidural) with mothers who have opted for an epidural. They're not comparing women having an unmedicated, active birth with women who have an epidural, who mostly have a supine labour.
Interestingly, the Birthplace Study 2011 flagged up that healthy, low risk first time mothers who opt to labour in environments where epidurals are readily available have about double the rates of emergency c-section and instrumental deliveries when compared to the outcomes for similar low risk women who have chosen a setting for birth where access to an epidural requires transfer by road. The figures for mums having a second or subsequent baby are much starker - giving birth in a setting where epidurals are more readily accessible appears to increase the c-section rate by a factor of x 5 compared to similar healthy multiparous women giving birth in settings where access to epidurals require transfer.
It does leave me asking a question about how much higher rates of epidural use contribute towards much higher rates of c/s and assisted delivery in obstetric settings, and what other factors could contribute towards such a dramatic difference in outcomes.