Sunday, 24 February 2013

Keeping childbirth normal

The variation in cesarean section operation (C-section) rates during childbirth is well documented; between hospitals, states and countries. The rates also vary over time, but here the pattern is at least consistent: the rate is increasing. So what? Sure there are complications, but isn’t that worth it if we are saving lives? As usual, it turns out that we have been overestimating the benefits and underestimating the harms of C-sections, which may explain the overuse of this treatment.

(for those short on time, feel free to skip to The Bottom Line)

The rates
In 1985, the World Health Organisation said that countries with the best neo-natal mortality rates had C-section rates of about 10% and suggested that the C-section rate should not be more than 10-15%. That was challenged by the obstetricians and has since been withdrawn. The rate in developed countries is now around 30% (higher in the USA and Australia) and around 50% in China and Brazil. Rates have roughly doubled in the last 20 years. The US rate in 1970 was 5%.

Let’s first look at how the benefits are overestimated. The reasons given by obstetricians for the increasing C-section rates are as follows: (they are also discussed in this review but you need access)

Increasing obesity in mothers
Obesity is associated with increased complications during childbirth (link). C-sections are more likely to be done if the woman is obese (link). But C-sections in obese women are also associated with higher complication rates, such as infection (link). Obesity on its own is not an indication for a C-section.

Increasing maternal age
C-section rates increase with increasing maternal age (links here and here), but often because the woman chooses the procedure, rather than it being necessary.

Breech presentation
Again, vaginal breech births are associated with increased complications, but that doesn’t mean that the current situation of performing a C-section on all women with breech presentation is necessary. Firstly, the improved outcomes from breech presentation that have been recorded are not necessarily due to increased C-section rates. Secondly, C-section can often be avoided by manually repositioning the fetus before labour (ECV: external cephalic version). (overview here).

What if the fetus can't be repositioned? The Term Breech Trial, published in The Lancet in 2000 was a randomised trail comparing C-section to a trial of labour. It showed better outcomes for the baby with C-section. Although guidelines published since (good one here), based on further evidence, have concluded that a trial of labour is reasonable and safe, the use of C-section as default has meant that midwives and doctors have become deskilled lost confidence at handling vaginal breech deliveries.

Again, induction is associated with increased C-section rates (link) but, again, we need to ask the question: why are the induction rates so high? Induction may be overdone for many reasons, as the benefits only apply when the birth is well overdue (after 41 weeks) not at 38 and 39 weeks, and certainly not to make it convenient or in some way easier for the women or doctors. In fact ½ - 2/3 of inductions are done for non-recognised conditions. (Time article, Science Daily article)

Fetal heart rate monitoring
CTG (electronic fetal monitoring) is one of those things that sounds like a no brainer: why wouldn’t you monitor the baby’s heart during labour? Fortunately, we have the answer to that question: because it does not lead to an improvement in mortality for the foetus in uncomplicated cases, and merely increases the rate of medical intervention (such as C-sections and instrumented births). Yet another “reason” for the high C-section rates that doesn’t stack up (Cochrane review here).

Failure to progress
The failure of labour to adequately progress is not a very objective measure. Several studies have shown that the diagnosis of 'failure to progress' is increasing, and is responsible for some of the increase in C-section rates.

Previous C-section
This is the main reason for having a C-section. Talking about making business for yourself. If you do a C-section on first timers, you will be doing one on everyone. It is uncommon and challenging to have a vaginal birth after a C-section due to the fear of uterine rupture. This is partly driven by defensive medicine and partly left over from when C-sections were done with a vertical incision (increasing the likelihood of later rupture). Now, the risk of uterine rupture with a trial vaginal birth is 1 in 500, and of those, the risk of perinatal death is less than 1 in 1,000. That’s roughly a 1 in 500,000 risk of death with a trial of normal birth.

Despite recommendations from the NIH (here) and other government bodies suggesting a trial of (normal) vaginal birth, and good evidence that this is likely to be successful and safe, the rate of C-section following previous C-section is proving hard to budge.

Now we will turn our attention to the underestimated harms.

Immediate harms
The direct, early harms are well documented and are much less than they were 100 years ago (maternal mortality is now as low as 2 in 100,000). It’s not perfect, but it’s pretty safe. It is still an operation, so it is prone to well-known but rare complications and the occasional complication out of left field, like catching fire (link).

Also, as with any operation, complications occur such as excessive bleeding, anaesthetic complications, wound infection, endometritis, urinary tract infection, damage to surrounding structures (e.g. urinary incontinence), blood clots, hysterectomy and wound breakdown.

Later harms
This is what everybody misses. After mum and baby have gone home and no longer need to see the doctor, problems can and do occur.

Firstly, the scarring that results from the C-section can lead to serious complications in over 2% of later pregnancies (link) due to uterine rupture (the main reason for caesarean sections: a previous C-section), placenta previa (low lying placenta) and placenta accreta  (where it grows into the uterus and doesn’t peel away, causing bleeding and often requiring surgery). These scarring issues make each subsequent C-section surgery more difficult, prolonged and more prone to complications.

Secondly, many complications to the baby have been reported, including problems from abnormal gut flora, depressed immune function, obesity, respiratory compromise, poor maternal-infant bonding and breast feeding issues.

Thirdly, there are a lot of other complications such as persistent pain at the site of surgery, endometriosis, adenomyosis, bowel adhesions and obstructions, abnormal periods and numbness around the scar.

Fourthly, there is a cost involved. C-sections are more expensive than normal births. That money could be spent on, say, providing more midwives in order to prevent more C-sections.

What to do?
1. Firstly recognise the problem. There is significant practice variation. There are different rates of intervention tied to socioeconomic and insurance status (here, here and here). And as described above, C-section have more problems than we usually consider, and are not always necessary.

2. Don’t use an obstetrician for an uncomplicated birth. They are more likely to get you started on the cascade of interventions (monitoring, scanning, inducing etc.) that culminate in C-section. This study from Australia showed that the Caesarean section rate (and the epidural rate, and the episiotomy rate, and the use of neonatal intensive care) can be reduced in low risk women by using a midwife instead of an obstetrician.

3. Change the culture in the maternity unit by providing training on vaginal breech birthing techniques, provide increased midwife support (one to one), and provide multidisciplinary review of C-section cases to see what can be learned and if future cases can be avoided.

4. Change the culture of the community. Provide the information to the public about the relative risks of management options for pregnant women, and provide data on clinical outcomes for each maternity unit.

This Cochrane review of interventions to reduce the unnecessary C-section rate, such as getting a second opinion, providing women with more information, reviewing practice and trialling labour, shows that the C-section rate can be significantly reduced.

The bottom line
  • The caesarean section rate varies considerably between practices, socioeconomic populations and countries.
  • The indications for surgery are being over-applied, such that the caesarean section may not provide any significant benefit in many cases.
  • The rate can be reduced without harm, therefore minimising the (often underestimated) early and late complications associated with caesarean section.
In my opinion, this problem is representative of much of modern medicine. There is a distorted perception of risk that leads to a tendency to go for the knife when there is doubt. This is reinforced by the knowledge that you won't get sued for doing a C-section, but you might if you recommend a normal birth.

Further reading. 
This article provides a great overview, but without access you will have to pay for it. 
This earlier BMJ article on the medicalisation of childbirth is free.
This Atlantic Monthly article covers all aspects of technological interference in childbirth.
This Harvard Magazine review is comprehensive and well written, but lacks specific references. 
'Costing the cascade' provides information on how much these interventions are costing society. 
The NSW state government has also put out a good document titled Towards Normal Birth


  1. Great overview - thank you for doing this! I wrote about the difference between scientific and technological birth (including from a first person point of view) here:

    1. Thanks Alice,

      Your article is excellent. I had read it, but I have now added it to the blog post under "Further reading". You cover many aspects of technological interference, not just C-section, and explore the important issue of insurance company and society attitudes. I highly recommend it to my readers.

    2. Thanks! And I"m so happy to have discovered your blog.

  2. wow... great take down. Well done!

  3. I understood (2nd hand only, I'm sorry) that continuous fetal monitoring led to a decreased rate of neonatal seizures and ??something else that also was probably a surrogate marker for better neurological outcomes

    1. Thanks,
      Yes, according to the Cochrane review, monitoring halved the seizure rate, but I am not sure of the significance of that. It can't be a surrogate for brain injury because the cerebral palsy rate was higher in the monitored group: RR 1.74, 95% CI 0.97 to 3.11, n = 13,252.

      Mainly, it led to more instrumented births and C-sections. Intervention begets intervention.

    2. Right, thanks! Looks like I need to add that to my long list of papers to read

  4. Ah, caesareans. The comparison between countries in terms of caesarean rate and related mortality is pretty useless, as it assumes that except for the caesarean rate all other factors between countries were similar. Can you say with certainty that if you reduce the rate of caesars in Brasil from the current 50% to your "ideal" 10% related mortality will not increase?

    There are couple reviews showing that the morbidity related to elective caesarean is pretty much similar to that of vaginal birth. It is the emergency ones that are real evil, and they should be avoided. Ironically, one way to do it is to increase the number of elective caesareans.

    Another fallacy is to assume that a vaginal birth is always better than surgery. I covered one example of this in my article on my site. It is about the trial of forceps before caesarean: even though forceps inflict way more damage than a caesarean every midwife seem to be very happy that the baby came out "naturally".

    Longer term complication profile of an elective caesarean is pretty good. The risk of abnormal placentation increases with every consecutive pregnancy/caesarean. However, as most women these days have no more than two to three children, the rate is pretty manageable. The connection between caesarean and breastfeeding is observed in the emergency caesarean setting, but it is probably a secondary variable, as breastfeeding if affected by myriads of factors. As far as chronic pain is concerned, one study has demonstrated similar rates between vaginal and surgical birth (reference in the article on my site).

    Lastly, check your facts before making statements. The epidural accident you mentioned occurred in labour ward, and the woman was given the block for delivery, not caesarean. It was the substitution error, and this topic is covered in another article on my site. Even the figures of caesarean mortality cannot be taken at face value: was the caesarean the direct cause of death or was it associated with it? For example, if a woman starts bleeding and then has the caesarean, can you blame the surgery for the outcome?

    1. Thanks Eugene,

      I think all your points are good, but I have two comments. I don't have an 'ideal' cesarean rate, but in answer to your question, no, I can't be certain, but I am pretty certain that the cesarean rate in Brazil could be lowered without causing much harm, and probably reducing some harms.

      Regarding the epidural case I mentioned, you are right. I didn't check the facts because I was working there at time, but because it was the anaesthetic staff talking about it, and because I work in theatres, I always assumed it was in theatres. I have removed it from the post. Thanks for picking that up.

  5. My remark came out too harsh I think, I apologize for that. As far as Brazilian CS rate goes, if it works well in that place, why would you want to reduce the caesarean rate?

    1. Thanks Eugene, a very good question. Allow me to generalise, because I am asked that question (the gist of it, anyway) a lot, whether it be about arthroscopies, fracture fixation, spine fusion etc. The gist of it is: if it works, why do less? Without having a long discussion about what we mean by "works", I guess it comes down the the core of my message, that benefits are overestimated and harms are underestimated. If x% of C-sections are necessary and (x+y)% are being done, then y% are being done unnecessarily, costing money and exposing people to risk.

      It also comes down to practice variation. If the rate is 10% in one country and 50% in another then, barring any unusual differences in the population, someone is wrong (10% is too low or 50% is too high). The only other alternative is that they are both wrong.

  6. Thanks for your post on this important topic. I would like to challenge the assertion that older women are 'choosing' surgical birth. I suggest it is more appropriate to say that older woman have been socialised and programmed to think that surgical birth is safer. Younger women are also having their views on birth constructed negatively with predicable results. As to breastfeeding issues, it is very true that breastfeeding is affected by multiple variables, but you only have to work as a midwife in a postnatal ward to see the immense difficulties that women struggle with post caesarean in that crucial first few hours after birth to understand the trajectory of diminishing breastfeeding rates.

    1. Thanks Carolyn,

      I agree with your opinion on women 'choosing' surgical birth. It is similar to other surgical procedures and medical interventions: patients may be free to choose, but their choice is heavily influenced by what they are told by their doctor and the current opinion in society. These are often biased towards the intervention because, as usual, the benefits are overstated and the harms understated.

  7. This comment has been removed by a blog administrator.

    1. Advertising-only comments are removed.