Higher order repeat caesarean sections: how safe are five or more?
Dr M. Rashid, Department of Obstetrics and Gynaecology, Queen's Hospital, Burton-upon-Trent, Staffordshire DE13 0RB, UK.
Objective To determine the maternal morbidity and mortality associated with multiple repeat caesarean sections.
Design Retrospective study.
Setting Security Forces Hospital serving Ministry of Interior and Security Forces personnel in Riyadh, Kingdom of Saudi Arabia.
Population Three hundred and eight case records undergoing between fifth and ninth caesarean section (mean 7) were studied and compared with a control group of 306 patients undergoing third or fourth caesarean section during the period January 1994–December 2002.
Main outcome measure Operative and post-operative complications and difficulties.
Results Five or more caesarean sections were associated with a longer operating time as well as an increased rate of severe adhesions. Blood transfusion rate was similar in the two groups but a drop of pre-operative to post-operative haemoglobin was significantly higher in the study group compared with the controls. There was no significant difference in the Apgar score of the baby, neonatal admission rate, incidence of caesarean hysterectomy, uterine scar rupture, placenta praevia, placenta accreta, bladder injury, incidence of postpartum pyrexia, wound infection and urinary tract infection between the two groups. There was no maternal death in the study group but one mother died in the control group.
Conclusion The higher order (5–9) repeat caesarean sections carry no specific additional risk for the mother or the baby when compared with the lower order (3 or 4) repeat caesarean sections.
A major obstetric hazard of repeat caesarean section is the increased risk of uterine scar rupture during pregnancy resulting in high fetal and maternal morbidity–mortality.1,2 In view of these risk factors, it is a common practice in the developed world to offer sterilisation to women after the third caesarean section. Women are given the chance of a fourth caesarean section in exceptional circumstances only. However, in countries where large families are encouraged by social and cultural factors, any attempt to limit the caesarean section to two or three is likely to be rejected. Saudi Arabia is such an example where patients are not deterred by the risk of multiple caesarean sections.
Although caesarean section is now safer than it has ever been in terms of sophistication in anaesthesia and surgery, little is known about the risks involved in multiple caesarean sections, especially when the number exceeds four. The studies describing patients undergoing five or more multiple caesarean sections have been too small for meaningful analysis.3–5 Individual case histories of women having caesarean sections (up to 13) have also been reported.6 The aim of the present study, the first of its kind on a large cohort, was to determine the outcome and associated risks for the mother and the fetus during the fifth or subsequent caesarean section.
A total of 318 pregnant women who had undergone four or more previous caesarean sections at the Department of Obstetrics and Gynecology, Security Forces Hospital, Riyadh, Saudi Arabia during January 1994 to December 2002 (inclusive) were included in the study. Of these, 10 patients were excluded from the analysis because of incomplete hospital records. The study group thus comprised 308 patients of which 1 had eight previous caesarean sections, 2 had seven previous caesarean sections, 17 had six previous, 79 had five previous and 219 patients had four previous caesarean sections.
A control group was formed by selecting the next woman delivered by caesarean section, with previous two or three elective caesarean sections, after each of the 318 women in the study group. Twelve women were excluded from the analysis because of incomplete hospital records. The control group thus comprised 306 patients of which 93 had three previous caesarean sections and 213 had two previous caesarean sections.
The case records of the study and control groups were analysed for the following parameters:
- (a)Demographic and clinical features, including age, parity, height and weight of the patient, the length of gestation, the mode of operation whether elective or not, type of anaesthesia used and whether tubal ligation was performed at the time of caesarean section or not.
- (b)Perinatal features including birthweight, number below 10th birthweight centile, Apgar score at 5 minutes, preterm birth below 36 weeks of gestation, incidence of multiple pregnancies in the two groups and number of admissions into neonatal intensive care unit (NICU).
- (c)Operative and post-operative course, including duration of operation, estimated blood loss during surgery, the severity of adhesions, incidence of placental abnormalities, rupture of scar, incidence of caesarean hysterectomy, bladder and bowel injury, blood transfusion, admission to surgical intensive care unit (SICU), fall in haemoglobin, incidence of wound infection, urinary tract infection, chest infection, pyrexia and length of hospital stay. The severity of pelvic adhesions was subjectively graded by the operating surgeon according to the American Fertility Society classification of adnexal adhesions.7 Filmy/avascular adhesions involving between 1% and 25% of the total area are classed as mild, dense/vascular involving 26–50% of pelvic area are moderate and cohesive adhesions involving greater than 50% of the area are severe.
The departmental policy is to perform elective caesarean section on patients who have had two or more previous caesarean deliveries between 37 and 38 weeks of gestation, unless there are other indications for early delivery. The operation is performed by a senior obstetrician or a well-trained assistant doctor. In general, Pfannenstiel incision is used to open the abdominal cavity and a transverse incision is made in the lower segment of the uterus. Longitudinal abdominal incision is given if the patient has already had a longitudinal incision (to avoid multiple scars) or if there is a history of severe pelvic adhesions obliterating the lower uterine segment and high attachment of bladder to the abdominal wall. In our unit, the visceral and parietal peritoneum are not generally closed after caesarean section. To reduce infectious morbidity following surgery, cephalosporin or augmentin as a single dose prophylactic antibiotic is given following cord clamping after delivery of the baby. Patients with a high risk of thromboembolism are given subcutaneous heparin in addition to the mechanical measures, and those with a moderate risk are treated either with subcutaneous heparin or managed with mechanical methods. χ2 test was used to analyse discrete variable and Mann–Whitney U test for analysing continuous variables.
The demographic and clinical features of the two groups are presented in Table 1.
Table 1. Demographic and clinical features in the study and control groups. Values are presented as mean [SD] or n (%) unless otherwise indicated.
|Maternal age (years)||35 ||33 |
|Height (cm)||152 ||152 |
|Weight (kg)||78 ||78 |
|Gestational age (weeks)||36.6 [1.4]||37 [1.7]|
|Mode of caesarean|
|Elective||281 (91)||257 (84)|
|In labour/SROM||22 (7)||45 (15)|
|Pregnancy complications||4 (1)||4 (1)|
|General||289 (94)||233 (76)|
|Regional (spinal/epidural)||19 (6)||73 (24)|
|Tubal ligation||88 (29)||25 (8)|
|Midline abdominal incision||264 (86)||54 (21)|
|Midline uterine incision||2||0|
|Apgar score <7 at 5 minutes||11 (4)||10 (3)|
|Preterm||26 (8)||33 (11)|
|NICU||60 (20)||56 (18)|
|Deaths||2 (1)||3 (1)|
Twenty-six patients in the study group had caesarean section before the scheduled time, 20 due to the onset of labour (within 2 hours of the start of labour), 2 due to placental abruption, 2 due to severe pre-eclampsia and 2 due to prelabour rupture of membranes. In the control group, 49 patients had emergency caesarean section, 26 of which were before the scheduled time, 20 due to onset of labour, 2 due to prelabour rupture of membranes, 2 due to severe antepartum haemorrhage and 2 due to non-reassuring cardiotocogram. Twenty-three patients in the control group came in labour after missing their appointment for elective caesarean section.
Intra-operative and post-operative course of the patients in the two groups is presented in Tables 2 and 3. There was no maternal death in the study group while one maternal death occurred in the control group. The patient was a 35 year old woman with two previous caesarean sections and placenta praevia undergoing an elective caesarean section at 37 weeks. She suffered massive bleeding from placenta accreta after the third stage and underwent total hysterectomy with internal iliac artery ligation.
Table 2. Operative course of patients in the two groups. Values are presented as mean [SD] or n (%) unless otherwise indicated.
|Duration of operation (minutes)||59 ||45 ||<0.001|
|Adhesion (severe)||167 (54)||47 (15)||<0.001|
|Praevia||11 (4)||9 (3)||0.62|
|Accreta||2 (1)||3 (1)||0.67|
|Caesarean hysterectomy||7 (2)||3 (1)||0.17|
|Bladder injury||4 (1)||2 (1)||0.45|
|Bowel injury||1 (0.3)||0||0.31|
|Blood loss during surgery (>1000 mL)||24 (8)||18 (6)||0.16|
|Rupture of scar at operation|
|Complete||4 (1)||2 (1)||0.45|
|Incomplete||7 (2)||5 (2)||0.53|
Table 3. Post-operative course of patients in the two groups. Values are presented as mean [SD] or n (%) unless otherwise indicated.
|Haemoglobin decrease (g/dL)||1.13 [0.5]||1.05 [0.4]||0.04|
|Blood transfusion||44 (14)||30 (10)||0.30|
|Wound infection||0||1 (0.3)||0.31|
|UTI||5 (2)||6 (2)||0.71|
|Chest infection||2 (1)||2 (1)||0.94|
|Temperature >37.5°C (unknown cause lasting > 48 hours)||14 (5)||13 (4)||0.86|
|SICU admission||9 (3)||15 (5)||0.28|
|Hospital stay (days)||5 ||3.5 [1.5]||<0.001|
|Maternal mortality||0||1 (0.3)||0.31|
There were 167 patients in the study group who had dense adhesions noted during the operation compared with 97 patients in the control group. Two hundred and sixty-four women (86%) required longitudinal incision to open the abdominal cavity and two cases required classical caesarean section in the study group due to severe adhesions. Also as a result of severe adhesions, the duration of the operation was longer in the study group compared with the control and hence greater blood loss and drop in post-operative haemoglobin in the study group. In the study group, uterine dehiscence was seen in seven patients (three were in labour), while four patients had complete uterine rupture (three during labour). There were five patients in the control group with uterine scar dehiscence (three were in labour) and two had complete uterine rupture (one during labour). All were clinically apparent apart from one incomplete rupture. All uterine rupture cases were successfully repaired apart from two complete ruptures in the study group and one incomplete rupture in the control group that required hysterectomy. There was no significant difference in the rest of the parameters between the two groups.
Seven patients in the study group required hysterectomy (four total and three subtotal hysterectomies) compared with three patients in the control group (two total and one subtotal hysterectomy). In the study group, the reason for hysterectomy included placenta praevia (two of which also had placenta accreta) and two had uterine rupture. All three cases of caesarean hysterectomy in the control group had placenta praevia (two of which also had placenta accreta, one with incomplete uterine rupture).
Intensive post-operative care was needed in nine patients in the study group and five patients in the control group. Cases requiring SICU admission in the study group included one patient who had severe pre-eclampsia, six cases of caesarean hysterectomies and two patients who had repair of ruptured uterus. In the control group, five patients needed SICU admission of which three had caesarean hysterectomy, one had repair of ruptured uterus and one had severe postpartum haemorrhage. Fourteen patients in the study group were advised to have classical caesarean section and bilateral tubal ligation in the next pregnancy due to dense adhesions between abdominal wall and lower uterine segment. Four patients were advised earlier operation at 35 weeks because of uterine dehiscence found at 37 weeks.
Multiple repeat caesarean sections are routinely performed in many parts of Saudi Arabia due to cultural opposition to limiting the number of caesarean section as this limits the size of the family. In our department, the average rate of caesarean section during the study period was 10% and previous caesarean section contributed to 30% of the total caesarean deliveries. Five or more caesarean sections made up 6% of the total caesarean section rate. The maximum number of caesarean section in this study was nine.
Rupture of gravid uterus either during pregnancy or labour is the most significant and catastrophic risk for both mother and baby. The presentation may vary from asymptomatic scar dehiscence to overt uterine rupture. Our results showed similar uterine rupture and dehiscence rate in the study group compared with the control group. The rate of 2–4% is higher than that reported after one previous caesarean section (<1%).8 The incidence of complete and incomplete scar dehiscence was also similar in the two groups. All cases of ruptured uterus were managed successfully and there were no serious consequences for the mother although one early neonatal death occurred in the control group with complete uterine rupture. Polyhydramnios and multiple gestation pose additional risks for scar dehiscence due to overdistension of the uterus. Fortunately, there was only one case of polyhydramnios (associated with twin pregnancy) in the study group that went into premature labour at 32 weeks.
Our results showed the incidence of placenta praevia and abnormal placenta (placenta accreta) to be similar in the study and control groups. An increased incidence of placenta praevia and accreta with multiple caesarean section has been reported, with rates of 10% for placenta praevia and 69% for placenta accreta.9
There was no significant difference in the incidence of caesarean hysterectomy between the two groups in our study. Uncontrollable haemorrhage due to placenta praevia and placenta accreta was the most common indication for hysterectomy, followed by rupture of lower uterine segment scar. Although emergency hysterectomy, a tragic complication for a woman of reproductive age, has become a relatively rare procedure in the developed world, it is not uncommon among the high risk repeat caesarean section patients in Saudi Arabia. Uncontrollable haemorrhage from the placental site and uterine rupture are the two most commonly reported indications for caesarean hysterectomy, with a post-operative morbidity of 35–60%.10,11 The single case of maternal death in our study following total hysterectomy was not directly related to the hysterectomy procedure, but a result of the development of coagulopathy. In the rest of the hysterectomy cases no major morbidity was noted. Patients who underwent subtotal hysterectomy were a result of the difficulty in dissecting the bladder from the cervix due to increased vascularity and adhesions from previous surgery. There was no difference in the SICU admission rate in the two groups.
It is noteworthy that the sole maternal death in the control group in our study was a result of massive haemorrhage from placenta praevia. Thromboembolism and anaesthesia-related accidents are the most common causes of death directly related to caesarean section procedure. However, there were no cases of thromboembolism in our study. Currently, there is no national database for maternal mortality rates in Saudi Arabia but a single report based on statistics acquired during 1989–1992 quotes a rate of 18/100,00012 with haemorrhage from placenta praevia as the dominant cause. Caesarean section reportedly contributed to 37% of direct maternal deaths, with more deaths due to emergency caesarean section (82%) compared with elective caesarean section (18%). Our results show a significantly lower emergency caesarean section rate in the study group compared with the control group. This is surprising considering the risks of uterine rupture and placental complications associated with initiation of labour in a uterus scarred by multiple repeat caesareans. However, with a combination of effective counselling of the women regarding risks and complications of multiple repeat caesarean sections, close antenatal follow up and early scheduling for caesarean section in selective cases, a lower rate of emergency caesarean section in the study group was indeed achieved. This general awareness of the risks associated with multiple repeat caesarean sections also led to a greater number of patients opting for tubal ligation in the study group compared with the control group despite the fact that sterilisation is generally rejected by Saudi women.
Significantly more severe adhesions were noted during operative delivery in the study group (54%) compared with the control group (15%). This higher incidence of severe adhesion in the study group was not unexpected as dense adhesion would tend to result from repeated surgery on the abdominal wall with increased chance of post-operative infection. It is probable that each additional caesarean section is at least as morbid as the first one.
From the operative point of view, a high number of previous caesarean sections is associated with technical difficulty during dissection of the abdominal wall and separation of bladder from the lower segment. This increased the average surgery time in the study group to 59 minutes compared with 45 minutes in the control group. However, this difficulty was of minor clinical importance as Apgar score of the newborn and the neonatal intensive care unit admission rate was not different in the two groups and primary healing of the abdominal wound was also good. High attachment of the bladder on abdominal wall is frequent in these patients and should be taken into consideration when opening the peritoneum. Alternatively, a midline incision should be considered in cases of severe adhesions, as illustrated by our results where a significant proportion of the patients in the study group (86%) had a midline incision to open the abdominal cavity compared with the control group (21%).
The incidence of bowel injury in our study was low; with only one case of bowel injury in the study group which happened while opening the abdominal cavity by midline incision as small bowels were adherent to the parietal peritoneum. Others have also reported a relatively low incidence of bowel injury during caesarean section as utero-intestinal adhesions are less common.6
The minor morbidity related to caesarean section such as UTI, wound infection, post-operative pyrexia and chest infection was comparable in the two groups. The incidence of blood transfusion rate was also similar in the two groups. However, the fall in the haemoglobin level from pre-operative to post-operative level was greater in the study group compared with the control group. This is consistent with a longer operating time observed in the study group compared with the control group, resulting in greater blood loss.
The success and the risk of trial of labour in women with several previous caesarean sections require further evaluation. At present, most obstetricians do not attempt trial of labour after three or more previous caesarean sections especially if the woman did not have previous vaginal delivery. Our study findings of impaired lower segment healing would discourage attempting vaginal delivery in such patients even though other indications for caesarean section may be absent.
The results from this large-scale study demonstrate that, in general, the maternal and fetal outcome associated with higher order multiple repeat caesarean section is comparable to the outcome with lower order repeat caesarean section. This is achievable through a combination of optimal prenatal care, effective patient counselling on the risks of multiple caesarean sections, adequate pre-operative preparation, maintenance of meticulous surgical techniques and careful post-operative follow up. Nevertheless, the general risks associated with operative delivery and frequently repeated pregnancies remain real and patients must be made aware of these.