In June 2007, NSW Health organised a statewide forum, Caesarean Section – Future Direction. The purpose of the multidisciplinary forum was to examine the rise in caesarean births within the public health system and to determine the future policy for caesarean birth in NSW. I, amongst others, was asked to speak at the forum on the implications of an increasing caesarean section rate for maternal and perinatal health. This is a shortened version of that paper.
We have a dilemma when we consider the evidence available regarding caesarean section. Here we have a group of people closely gathered around a pothole created by a leaking pipe, considering wisely the causes and solutions – believing they are in fact examining the Grand Canyon. In the background, several scattered, probably unrelated individuals are gazing at a scene that is so immense they cannot hope to really comprehend the magnitude before them. Water erosion in fact, caused both cavities, but the pothole is evidence of what happens on a small scale while the Grand Canyon is evidence of what happens on a grand scale, over millions of years. While their causes are somewhat similar, one could not say the two look remotely alike.
It is time we as health professionals that we took our gaze from the pothole to the Grand Canyon. When we look at maternal and perinatal health associated with caesarean section, we often measure outcomes associated with small numbers over a short period of time and this is potentially a different picture to when we look at large numbers over many years, especially when dealing with rare and cumulative outcomes. This is one of the limitations we find with randomised control trials.
There are 178 Cochrane Systematic Reviews (CSR) that mention caesarean section, so things are looking up at first glance. However, on closer examination we find the following: a CSR on term breech (available), small babies (numbers too small), planned caesarean section for women with a twin pregnancy (protocol only), caesarean section for non-medical reasons at term (no trials available), planned elective versus planned vaginal birth for women with a previous caesarean birth (no trials available) and non-clinical interventions for reducing unnecessary caesarean section (protocol only). So we have in fact a lot of instruction manuals on the ‘how to’, but little on the ‘when to’ and how ‘often to.’
The Term Breech Trial hit our maternity units in 2000 with instant effect (Hannah et al. 2000). In fact the results were not even published before many units changed their policy. In all honesty the change to practices predated the trial and herein lay part of the problem. A CSR including three trials (n=2396) – of which over 2000 women came from the Term Breech Trial – showed that in the caesarean section arm there was a decrease in perinatal or neonatal death and/or neonatal morbidity but no difference in death or neurodevelopment delay at two years. Mothers had increased short-term maternal morbidity (of nearly 30%). There was some decrease in incontinence at three months but not at two years and there was increased constipation in caesarean section group at 2 years (Hannah et al. 2004).
So what did this tell us? It told us that for the 4% of term breech singleton babies born, that vaginal birth was less safe for them than caesarean. It also told us it wasn’t better for women or babies long-term but I don’t think we actually heard this? Very little publicity of the long-term follow-up occurred. The effect the Term Breech Trial had in particular on the rise in caesarean section rates overall, I believe was significant. I can’t prove a causal relationship but we cannot explain this nearly 20% increase between 1999 and 2001 by an increase in caesarean sections for breech – as that was already at a low of 0.9% in 1999 and in 2005 sits at 0.4% (see figure 2). Did we extrapolate results inappropriately from a trial examining a complicated vaginal birth to all vaginal births? Is it perhaps at this point in clinical care that we stopped looking for every reason not to do a caesarean section and began to look for every reason to do a caesarean section?
So what now? If the randomized trials can’t give us answers, and perhaps never really will be able to on rare and long term outcomes, where do we turn? Have we ignored other very valid measures and are we careering towards disaster? Do we, should we, can we indeed turn back?
With the pothole revealing little about the impact a rising caesarean section rate would have on maternal and perinatal health let us look at the Grand Canyon. At first we are overwhelmed by its vastness. Where to look, how to see it all and what is really important? In the last two years the Grand Canyon has been coming into focus. I want to begin by looking at mortality and then moving onto morbidity.
Let us begin by looking at maternal mortality rates associated with elective caesarean sections in comparison to vaginal birth, as this is where the debate truly lies. The maternal mortality rate with elective caesarean section has been reported to be anywhere from two to eight times higher than vaginal birth (Lilford et al. 1990; Schuitemaker et al. 1997; Hall et al.; 1994; Hall et al. 1999). If we include emergency caesarean sections, rates much higher have been quoted. The problem with all these very large population based studies is the reason for the elective caesarean might add to the morbidity being observed.
A study published in the Lancet last year examined caesarean delivery rates and pregnancy outcomes for the 2005 World Health Organisation (WHO) global survey on maternal and perinatal health in Latin America, a country with a soaring caesarean section rate, particularly in the private sector (Villar et al. 2006). This was a prospective survey, comprising 24 geographic regions in eight countries in Latin America. Individual data was obtained on all women admitted for delivery over three months to 120 institutions randomly selected from 410. There were 97 095 births with a median caesarean rate of 33% (51% in private hospitals). The rate of caesarean delivery was positively associated with postpartum antibiotic treatment, severe maternal morbidity and mortality, increase in fetal mortality rates (including elective caesareans), increased numbers of babies admitted to intensive care for seven days or longer (adjusted for preterm) and rising rates of preterm delivery. Interestingly an increase risk of preterm delivery and neonatal mortality started when caesarean section rates were between 10-20%, approximating the WHO recommended rates. The authors concluded that high rates of caesarean delivery do not necessarily indicate better perinatal care and can be associated with harm (Villar et al. 2006).
Some have argued, that South America cannot be called a developed country, hence these outcomes.
A population based, case controlled study for a five-year period in France (1996-2000) examined postpartum maternal mortality and caesarean delivery (Deneux-Tharaux et al. 2006). In the study all maternal deaths due to non-pregnancy outcomes, multiple births, conditions existing prior to delivery and women hospitalised during pregnancy were excluded (269 deaths reduced to 65). This was then adjusted for potential confounders. They found the risk of maternal death following caesarean was 3.6 times higher than for vaginal birth. Complications of anaesthesia, puerperal infection and venous thrombosis were all significantly higher.
Again, the problem of truly having a low risk planned caesarean group can be argued. A study by Liu et al (2007) for the Maternal Health Study Group of the Canadian Perinatal Surveillance system examined maternal mortality and severe morbidity with low-risk planned caesarean section vs planned vaginal birth. They took all planned caesareans for breech presentation for a 14-year period (1991-2005) (n=46 766) in Canada and compared them to planned vaginal birth (n=292 420). Severe maternal morbidity was 27.3 per 1000 women in the caesarean section group and 9.0 per 1000 women in the vaginal birth group. The planned caesarean section group had a greater incidence of postpartum cardiac arrest (OR 5.1, 95% CI 4.1-6.3), wound haematoma (OR 5.1, 95% CI 4.6-5.5), hysterectomy (OR 3.2, 95% CI 2.2-4.8), major puerperal infection (OR 3.0, 95% CI 2.7-3.4), anaesthetic complications (2.3, 95% CI 2.0-2.6), venous thrombosis (OR 2.2, 95% CI 1.5-3.2), haemorrhage requiring hysterectomy (OR 2.1, 95% CI 1.2-3.8) and longer hospital stay (mean difference 1.47 days). Only haemorrhage requiring a blood transfusion was less in the caesarean section group (OR 0.4, 95% CI 0.2-0.8). The researchers found there was one excess postpartum readmission for every 75 caesareans.
But surely a maternal death is just the tip of the iceberg. If we aren’t reacting long before this then we can’t hope to reduce maternal mortality rates.
Waterson et al (2001) examined 25 maternity units over one year (1997-1998) in the UK (n=48, 865) and found 588 cases of severe obstetric morbidity and five maternal deaths. They then looked at severe maternal morbidity and its relationship to mortality. Severe morbidity was estimated to be 1.2% of deliveries and was significantly increased with caesarean delivery (for example, sepsis was 11 times greater with caesarean) and they estimated a severe morbidity/mortality ratio of 118:1. In other words it would take 118 episodes of severe maternal morbidity to lead to one maternal death.
Yang et al. (2007) published a paper exploring the association of caesarean delivery for first birth with placenta praevia and placental abruption in the second pregnancy. This was a retrospective cohort study in the USA from 1995 to 2000 (n= 5 146 742). They tried to quantify the risk in second pregnancies after a caesarean delivery in the first pregnancy. They found placenta praevia was 4.4 per 1000 (caesarean) and 2.7 per 1000 (vaginal birth), which was a 47% increase. Placental abruption was 6.8 per 1000 (caesarean) and 4.8 per 1000 (vaginal birth), which was a 40% increase. Both these complications bring with them serious implications for maternal and perinatal morbidity and mortality
A study by Silver et al (2006) examined maternal morbidity associated with multiple repeat caesarean deliveries. It was a prospective observational cohort of 30,132 women who had a caesarean without labour in 19 centres over 4 years. There was a significant rise in placenta accreta with subsequent caesareans.
Urinary Incontinence is often cited as a reason why women choose caesarean and in the National Sentinel Caesarean Section Audit (NSCSA) (2001) from the UK showed that 68% of consultant obstetricians believed caesarean section reduced chances of urinary incontinence. The dilemma is that a lot of pelvic floor weakening is due to pregnancy. Hereditary factors play a large role in urinary incontinence. Hereditary factors (collagen disorders) and pregnancy (antenatal stress incontinence) rather than parturition tend to reveal the defect. Caesarean has short-term protection if there is no history (20%), and only for the first birth, but less long-term protection. Menopause adds yet another dimension.
In the NSCSA, 78% of obstetricians believed faecal incontinence were decreased by having a caesarean. The problem is that 3-5% of women have faecal incontinence after elective caesarean (Lal et al., 2003). Long-term follow up does not show protective effect from caesarean section (MacArthur et al. 2005) and there is a similar incidence of fecal incontinence in elderly men as women (Read et al. 1995).
Preserving a woman’s sex life through caesarean section, referred to as ‘the honeymoon fresh vagina’ is a concerning belief. How much does a woman’s identity and self-esteem reside in her vagina? What is this saying about our understanding of relationships? The NSCSA (2001) showed 50% of consultant obstetricians agreed with the statement “elective caesarean will least affect the mother’s future sexual function.” Yet potential impact of birth on future sexual function did not impact strongly on women’s birth preferences (NSCSA, 2001). Obstetricians seem to worry more about women’s sex lives than they do! Barret et al. (2005) conducted a cross sectional study of 796 primiparous women by postal survey at 6 months. They found the protective effect of caesarean section on sexual function was limited to the early postnatal period (0-3 months) and at 6 months there was no difference.
Tokophobia (profound fear of labour and birth) is said to be another reason women choose caesarean section. Firstly, I wonder who’s Tokophobia we are talking about – women’s or clinicians? Fear of birth is increasing. Green et al. (2003) showed this in the UK. Reasons for fear of vaginal birth may be due to serious issues, such as sexual abuse and posttraumatic stress disorder. But can refusal of caesarean section lead to psychiatric disorders as has been suggested? With counseling as many as 56% of women with tokophobia opt for a vaginal birth and most are pleased with the experience (Sjogren et al. 1997). There is evidence that request for caesarean section due to fear of vaginal birth is accompanied by extensive psychological problems. Around 90% of women have anxiety and depression, 43% eating disorders, 63% are or have experienced abuse. Counseling increases vaginal birth and satisfaction (Nerum 2006).
Other maternal health implications:
What are the implications of an increasing caesarean section rate on perinatal mortality and morbidity?
It has been argued that we could potentially eliminate one intrauterine death between 38 weeks and birth per 600 pregnancies, one term intrapartum stillbirth per 5000 births, one case of hypoxic ischaemic encephalopathy per 1750 births, one case of cerebral palsy every 3000-5000 caesareans we perform, as well as reducing birth injuries, such as fractures and nerve injuries.
The problems with these statements are that a policy of caesarean section at 32 weeks would save more lives, but iatrogenic morbidity and mortality would outweigh this benefit. Perinatal deaths occur after elective caesarean (0.5%-1.6%). In 10 developed countries including the United States and Australia, despite a 5-fold increase in caesarean birth over recent decades, the incidence of cerebral palsy has remained steady, at about 1 in 500 births (MacLennan et al. 2005). MacLennan (2006) concluded, “As yet there is no obstetric policy that has been shown to reduce the rate of cerebral palsy.” Foley et al. (2005) found in Ireland that over a 12 year period, despite a greater than 2-fold rise in the caesarean section rate (from 6.9% to 15.1%), the overall term seizure rate and term peripartum death rate were not significantly altered. Gestational age is not accurate – up to 9% of babies born by elective caesarean are thought to be less than 38 weeks. As many as 1:18 babies (5.5%) born by elective caesarean vs 1:63 babies (1.6%) born vaginally cannot support themselves in room air (Morrison et al, 1995). This results in increased admission to neonatal intensive care units, resulting in maternal separation, anxiety, poor feeding, jaundice, cannulation and cross infection. Scalpel lacerations occur in 1-8% of caesareans. Long-term effects on health, such as neurodevelopment, psychological wellbeing and atopic disease unknown but new concerning evidence is emerging.
MacDorman et al. (2006) published a large study looking at the infant and neonatal mortality for primary caesarean and vaginal births, in women with “no indicated risk” in the United States. They used National linked data for 1998-2001. This included 5,762037 live births and 11,8977 (98%) infant deaths. The neonatal mortality rate was 2.9 times higher amongst infants delivered by caesarean section (1.77 per 1000 live births compared to those delivered vaginally 0.62).
As we bring the Grand Canyon into focus let us go up the valley and around the corner. What about the next pregnancy? What impact does a caesarean section have on the next baby, a thing we rarely think about at the time?
We know from recent research, and several major studies, that caesarean are associated with an increased risk of disorders of placentation in subsequent pregnancies. The large Scottish study by Smith et al. (2003) examined 120 633 singleton, second births between 1980-1998. The absolute risk of unexplained stillbirth at or after 39 weeks (1.1 per 1000/ 0.5 per 1000) was double the risk of stillbirth or neonatal death from intrapartum uterine rupture. Thus the very issue that appears most concerning to clinicians, fetal death from intrapartum uterine rupture, is in fact less of a concern than an unexplained stillbirth. We can’t do much to eliminate this because the risk is apparent from 34 weeks onwards.
During the NSCSA (2001) women said they would prefer more information about the risks (48%) and benefits (43%) of caesarean section. Nearly all women wanted a birth that was the safest option for the baby and least stressful for the baby. Women’s own safety and a quick recovery from the birth that would not impede breastfeeding were also strong preferences. When we take our gaze from the pothole to the Grand Canyon we cannot say we would be giving women what they want if the caesarean section rate continued to escalate. Intervening in childbirth is like throwing a pebble into a pond. The ripples keep on going and you don’t know where they will end up. But you can bet that on some distant shore there will be an effect. It is only relatively recently that we have been looking beyond the throw of the stone to the distant shore. What we see is very, very scary.
Figure 2. Caesarean section in NSW over ten years (Term Breech Trial 2000)