Uterine exteriorisation at caesarean section: influence on maternal morbidity


Correspondence: Dr E. C. O. Edi-Osagie, IVF Unit, Department of Reproductive Medicine, St Mary's Hospital, Hathersage Road, Whitworth Park, Manchester M13 0JH, UK.


Objective To compare the influence on caesarean section morbidity of uterine exteriorisation or in

Design Randomised controlled trial.

Setting Princess Anne Maternity Unit of the Royal Bolton Hospital, UK.

Population One hundred and ninety-four women undergoing delivery by caesarean section.

Methods Two intra-operative readings of arterial pulse rate, mean arterial blood pressure, and arterial haemoglobin oxygen saturation were obtained. Pre-operative and day-3 haemoglobin concentrations were determined. Intra- and post-operative complications, puerperal pain scores, and febrile and infectious morbidity were assessed. A postal questionnaire was used to assess morbidity six weeks after delivery.

Main outcome measures Intra-operative changes in pulse rate, mean arterial blood pressure and oxygen saturation; peri-operative changes in haemoglobin concentration; incidence of intraoperative vomiting, pain, intra- and post-operative complications, and febrile and infectious morbidity; immediate and late puerperal pain scores; satisfaction with the operation.

Results No clinically significant differences between uterine exteriorisation and in situ repair were found in pulse rate, mean arterial pressure, oxygen saturation and haemoglobin changes. Likewise, the incidence of vomiting and pain was similar. Vomiting occurred in 10% of all the women, and 57% of all pain complaints occurred at the initial skin incision. There was a trend towards higher immediate and late pain scores in the exteriorisation group, reaching statistical significance on day 3. Overall, pain scores averaged 6/10 on day 1 despite patient-controlled analgesia, and three-quarters of all women reported persisting pain on day 42. Intra- and postoperative complications, febrile and infectious morbidity, and duration of hospital stay were similar in both groups.

Conclusions We have demonstrated that uterine exteriorisation and in situ repair have similar effects on peri-operative caesarean section morbidity. Intra-operative pain reflected adequacy of anaesthesia, while vomiting reflected adequacy of pre-operative preparation of patients. Exteriorising the uterus at caesarean section is a valid option.


The technique of uterine exteriorisation at caesarean section was described by Sanger in 1882, and modified by Leopold two years later1. In this century Portes and Phaneuf advocated it2. Regional anaesthesia was introduced into obstetric practice in this country in the early 1960s and popularised by Pitkin3. Its popularity engendered unpopularity for the technique of uterine exteriorisation, because of hypothesised dangers of the technique4. These include intra-operative haemodynamic instability, vomiting, pain and greater puerperal morbidity. The two studies that addressed these issues directly did not substantiate these fear5,6. However, their conclusions were criticised by a subsequent systematic review which concluded they were of poor methodology and power, respectively8.

Rare but often life-threatening complications of caesarean section have been reported, with some authors suggesting a link with uterine exteriorisations8–12. The studies that have investigated caesarean section morbidity, report an incidence varying widely, from as low as 4.2% to > 90%13,14 The most common morbidities reported are pyrexia (3%-58%), urinary tract infection (1.2%–19%), endometritis (4%-17.8%), wound infection (0.9%-7%), blood transfusion (1.9%-6.3%), and chest infection (< 4%)5,13–17. Only one study compared puerperal morbidity following uterine exteriorisation or in situ repair, reporting a reduction in febrile days following exteriorisation5. Its findings were criticised by a subsequent systematic review because of major flaws in its methodology7. While obstetricians remain convinced of the surgical merit of the technique of uterine exteriorisation and continue to employ it, its safety remains a disputed mater. The aim of the present study was to compare the influence on caesarean section morbidity of uterine exteriorisation or in situ repair.


This was a randomised controlled trial for which ethical approval was granted by the research ethics committee of the Royal Bolton Hospital. The setting was the Princess Anne Maternity Unit of the hospital (a district general hospital with 4000 deliveries and 600 caesarean sections a year). One hundred and ninety-four women who needed caesarean section for delivery (22% of caesarean sections) were recruited between January 1996 and June 1997. Women with placenta praevia, placental abruption and chorioamnionitis were excluded because of the potential effects of these conditions on outcome measures. Women were actively recruited from the antenatal clinics, maternity wards and delivery suite, and signed informed consent was obtained from all subjects. Randomisation was by a table of random numbers and the randomised allocations kept secure in sealed envelopes. One hundred women were randomised to uterine exteriorisation and 94 to in situ repair. The study was not blinded for obvious reasons, and all operations were performed by four registrar equivalents who were comfortable with both methods of uterine repair.

All the women had the following procedures performed:

  • 1Two estimations of haemoglobin concentration (Hb) pre-operatively and on day 3.
  • 2Peri-operative bladder catheterisation.
  • 3Lower uterine segment caesarean section through a Pfannenstiel incision.
  • 4Repair of the uterine incision in two layers with through a pertonisation f the lower uterine segment with the uterus either exteriorised or in situ.
  • 5Intra-operative antibiotics (usually a combination of cefuroxime 1.5 g and metronidazole 500 mg) after clamping of the umbilical cord.
  • 6Two readings of pulse rate, mean arterial blood pressure and arterial haemoglobin oxygen saturation (measured by pulse oximetry). These were obtained at the time of the initial skin incision, and during repair of the first layer of the uterine incision. Both readings were recorded by the anaesthetist or anaesthetic technician when prompted by the surgeon.
  • 7Pain relief by patient-controlled analgesia, utilising morphine for 24–36 hours, supplemented by rectal diclofenac or codeine-paracetamol combination as required.
  • 8Four-hourly temperature assessment.
  • 9Daily pain score assessments for five days on a Likert scale (of 1–10, increasing with severity of pain).
  • 10High vaginal swab, mid-stream urine specimen and wound swab on the third post-operative day.
  • 11A postal questionnaire at six weeks to assess lochial loss, persisting pain, urinary, wound and pelvic infection, and satisfaction with the operation.

The outcome measures studied were intra-operative changes in pulse rate, mean arterial blood pressure and arterial haemoglobin oxygen saturation; incidence and severity of intra-operative vomiting and pain; peri-operative changes in haemoglobin concentration; intra- and post-operative complications; immediate and late puerperal pain scores; febrile and infectious morbidity; and, satisfaction with the operation.

Statistical methods

The power of the study was based on a projected sample size of 200 women, with 100 randomly allocated to each study group. This was sufficient to detect, at the conventional 5% statistical level, a mean difference of one unit in puerperal pain score (SD 2.1) with 89% power. These numbers were also sufficient to detect a mean difference of one day in length of hospital stay (SD 2) with 94% power, and an odds ratio of 3 for post-operative morbidity rates with 86% power. Continuous variables following normal distributions were analysed by two-tailed Student t test (paired and unpaired for intra-group and inter-group comparisons, respectively), and those not following normal distributions by the Mann-Whitney U test. Proportions were compared using χ2 tests, or Fisher's exact test if numbers were small; 95% confidence intervals were produced for differences, and the Pearson coefficient was used to determine correlation (normally distributed variables). Analysis was by intention to treat.


Table 1 shows the characteristics of subjects in both groups. The groups were comparable with regard to maternal age, parity and gestation at delivery. Regional anaesthesia was used in 54 women (54%) who were exteriorised and 63 (67%) who underwent in situ repair. These formed the denominators for assessing intra-operative vomiting and pain. The postal questionnaires sent out at six weeks were returned by 70 women (70%) who had exteriorisation and by 69 (73%) who had in situ repair; these formed the denominators for assessing late puerperal pain, infectious morbidity and satisfaction with the operation. The mean duration of surgery was 29.9 minutes (95% CI 28.3, 31.5) in the exteriorisation group and 31.4 minutes (CI 29.3,33.4) in the in situ repair group (P= 0.259).

Table 1.  Characteristics of the women in the exteriorisation (EXT) and in situ repair (ISR) groups. Values are given as n (%) or mean [SD] {range}.
CharacteristicsEXT (n = 100)ISR (n = 94)
Maternal age (years)28.01[5.71]{16–44}29.02[6.29]{16–41}*
Parity1.06 [1.4] {0–8}0.94 [1.0] {0–4}
Gestation at delivery (weeks)38.45[1.56]{34–41}38.02[1.98]{29–42}
Anaesthesia for delivery
  General46 (46)31 (33)
  Regional54 (54)63 (67)
Questionnaires returned70 (70)69 (73)

Changes in pulse rate were very variable in both groups, with wide fluctuations evidenced by wide confidence intervals (Table 2). Initial pulse rates were comparable, and while the intra-operative drops were statistically insignificant in the exteriorisation group and significant in the in situ repair group, neither were of clinical significance. Inter-group comparison of the pulse differences (-3 versus −8) did not reach statistical significance (P = 0.151). Bradycardia (pulse < 60 bpm), occurred in one woman in each group: exteriorisation: 56 bpm rising to 58 bpm; in situ repair: 109 bpm falling to 40 bpm.

Table 2.  Intra-operative haemodynamic parameters. Values are given as mean [95% CI]. MAP = mean arterial pressure.
ParametersEXT (n = 100)ISR (n = 94)P
  1. *Unpaired Student's t test.

  2. **Paired Student's t test.

  3. † Mann-Whitney U test.

  Initial99.4 [95.6, 103.1]98.7 [94.4, 103.1]0.826*
  Uterine repair96.1 [92.6, 99.6]90.4 [87.4, 93.4]0.015*
  Difference−3.26 [−6.84, 0.32]−8.30 [−12.66, −3.94]0.151†
  P0.074**< 0.001**
  Initial96.0[92.6, 99.4]92.2[88.7,95.5]0.112*
  Uterine repair86.8 [83.6, 90.0]86.3 [83.0, 89.6]0.830*
  Difference−9.18 [112.40, −5.95]−5.91 [−9.60, −2.23]0.221*
Oxygen saturation
  Intial97.4[97.1, 97.6]97.4[97.1,97.6]0.952*
  Uterine repair97.2[96.9,97.4]97.1[96.,97.3]0.709*

The intra-operative drops in mean arterial blood pressure were similar in both groups, with wide fluctuations evidenced by wide confidence intervals (Table 2). Inter-group comparison of differences in mean arterial blood pressure did not reach statistical significance (P = 0.221). The incidence of hypotension (mean arterial blood pressure < 70 mmHg) was similar, occurring in 15 women (15%) in the exteriorisation group and 14 (1 5%) in the in situ repair group. The lowest exteriorisation value was 52.6 mmHg (two women showed an intra-operative increase and 13 a drop), while the lowest in situ repair value was 51.3 mmHg (five women showed an increase and nine a drop). All hypotensive women were treated with rapid fluid infusions and 3–10 mg boluses of ephedrine, with prompt restoration of blood pressure.

The initial oxygen saturation values and corresponding intra-operative drops were similar in both groups. Inter-group comparison of oxygen saturation differences did not reach statistical significance (P= 0.195). For the purpose of this study, an oxygen saturation value < 96% was considered low. Fourteen women (14%) in the exteriorisation group and six (6%) in the in situ group had a low oxygen saturation value. Seven women in the exteriorisation group showed an intra-operative increase (average initial oxygen saturation 94.4%, rising to 96.1%), and seven a decrease (average initial oxygen saturation 96.7%, falling to 94.7%), the lowest value being 94%. Four women in the in situ repair group showed an increase (average initial oxygen saturation 94%, rising to 95.75%), and two a decrease (average initial oxygen saturation 96.5%, falling to 94%), the lowest value being 91%. No adverse clinical consequences of the low oxygen saturation were observed and all subjects responded to oxygen therapy.

Pre-operative and day 3 haemoglobin concentrations were similar in both groups. The women in the exteriorisation group showed a mean peri-operative haemoglobin drop of 1.01 g/dL (95% CI 0.64, 1.38), with mean pre-operative and day 3 levels of 11.34 g/dL (95% CI 11.11, 11.53) and 10.33 g/dL (95% CI 10.05, 10.62), respectively (P < 0.001). The corresponding peri-operative drop among those with in situ repair was 0.66 g/dL (95% CI 0.39,0.93), with mean pre-operative and day 3 levels of 11.09 g/dL (95% CI 10.90, 11.29) and 10.43 g/dL (95% CI 10.20, 10.66), respectively (P < 0.001). Comparison of the mean peri-operative differences between the groups did not reach statistical significance (P= 0.162).

There were no significant differences in the incidence of nausea (4/54 and 2/63, respectively; P = 0.413), retching (3/54 and 2/63, respectively; P = 0.661) and vomiting (7/54 and 5/63, respectively; P = 0.543) between the groups (Fisher's exact test). In women in the exteriorisation group vomiting occurred in 6/41 elective deliveries (15%) compared with 1/13 emergency deliveries (8%). The corresponding figures for those in the in situ group were 5/54 (9%) and 0/9 (0%), respectively. One woman in each group experienced vomiting when the uterus was handled for repair; all other cases occurred before the uterine incision. There were no significant differences in the incidence of intra-operative pain (5/54 and 2/63, respectively; P= 0.246, Fisher's exact test). Pain occurred at the initial skin incision in 2/54 of those exteriorised and 2/63 of those undergoing in situ repair. Pain was felt by women in the exteriorisation group at incision of the rectus sheath (1/54) and and at uterine exteriorisation (2/54); of 63 women undergoing in situ repair, none of them reported pain during either of these procedures. Thus, the occurrence of pain coincided with incision of the skin or rectus sheath in 517 women (71%) and with uterine exteriorisation in 2/7 (29%).

Other intra-operative complications were similar in the two groups (P= 0.115, Fisher's exact test). Haemorrhage occurred in 2/54 women in the exteriorisation group and 4/63 women having in situ repair, while uterine angle tear occurred in one woman in each group. There was failure of procedure in one woman in the exteriorisation group (1/54) (exteriorisation not achieved) and in four (4/63) women in the in situ repair group (exteriorisation necessary to control bleeding). These five subjects were analysed with the groups to which they were originally randomised.

Immediate post-operative pain scores decreased steadily from days 1 to 5 (Table 3). The difference in scores between the groups was significantly higher in the exteriorisation group on day 3 (P= 0.046). Seventy-six percent and 77% of women in the exteriorisation and in situ repair groups, respectively, reported persisting pain six weeks after surgery, with the difference in mean scores outside statistical significance (P = 0.059). In 10/70 exteriorised women (14%) and 4/69 (6%) of those who had in situ repair, the score at six weeks was ≥ 5. The Pearson coefficients of correlation (r) between pain scores on days 1 and 5 were 0.396 for women in the exteriorisation group (P = 0.003) and 0.241 for those in the in situ repair group (P= 0.098). On days 1 and 42 r = 0.178 for women in the exteriorisation group (P = 0.193) and 0.191 for those who had in situ repair (P= 0.193). Between days 5 and 42, r = 0.378 (P = 0.040) and 0.172 (P = 0.243) for exteriorised women and those with in situ repair, respectively.

Table 3.  Mean puerperal pain scores. Values are given as mean [95% CI].
DayEXT (n= 100)ISR (n = 94)P*
  1. *Mann-Whitney U tests

  2. **n=70 (questionnaires returned).

  3. n=69 (questionnaires returned).


For the purpose of this study febrile morbidity was defined as a temperature of 38°C occurring after the first 24 h of surgery, and persisting for at least 24 hours (Table 4). The groups did not show any significant difference in febrile morbidity (P = 1). Urinary tract infection was the most common infectious morbidity, occurring more commonly in the immediate than late puerperium in both groups. Its incidence in the immediate or late puerperium was not significantly different between the groups (Tables 4 and 5). Likewise, the incidence of positive bacteriology (wound and high vaginal swabs), and wound and pelvic infections was similar in both groups (Tables 4 and 5). Wound infections in the late puerperium were reported by 18/70 (26%) of those having exteriorisation and 16/69 (23%) women with in situ repair (P = 0.844), compared with only 12 (12%) and 13 (14%) respectively who had positive wound swabs in the immediate puerperium (P = 0.831). Most women who reported puerperal wound infections required antibiotic treatment [8/18 (44%) and 9/16 (56%), respectively]. The incidence of other complications was similar in both groups (P= 1), as was the length of hospital stay (P = 0.187).

Table 4.  Immediate puerperal morbidity (in hospital). Values are given as mean [95% CI] and n (%). UTI = urinary tract infection; HVS = high vaginal swab; DVT = deep vein thrombosis.
ParametersEXT (n= 100)ISR (n= 94)P
  1. *Mann-Whitney U test.

  2. † Fisher's exact test.

Hospital stay (days)5.8[5.5, 6.2]5.6[5.3, 5.9]0.187*
Fever (T > 38°C > 24 h)4(4)3(3)1 †
Positive HVS16(16)15(16)1 †
UTI15 (15)12 (13)0.683 †
Positive wound swab12 (12)13 (14)0.831 †
Other complications
Blood transfusions3 (3)2 (2) 
Inflammation0 (0)1(1) 
TOTAL6 (6)5 (5)1 †
Table 5.  Late puerperal morbidity (in community). Values are given as n (%). UTI = urinary tract infection.
ParametersEXT (n= 70)*ISR (n= 69)*P
  1. *Numbers of returned questionnaires

  2. **χ2 test.

  3. † Fisher's exact test.

Lochia duration (days)
  0–206(9)13 (19) 
  21–4250(71)42 (61) 
  NS0 (0)2 (3) 
  TOTAL70(100)69 (100)0.186**
Lochia pattern
  Symptomatic13(19)12(17)1 †
  Iron therapy41(59)35(51)0.396 †
UTI5 (7)6 (9)0.764 †
Wound infection18 (26)16 (23)0.844 †
Pelvic infectionl(1)5 (7)0.116 †

There was no significant difference between the groups in the duration (P = 0.186) or amount (P= 0.746) of puerperal lochial loss (Table 5). The duration of loss was three to six weeks in approximately half the number of women in both groups, and approximately 80% of all the women had stopped losing lochia by six weeks. Just under 90% reported the amount of loss as varying from spotting to menstrual type. Both groups were again similar in the incidence of reported symptoms of anaemia (P = I), and iron therapy (P= 0.396) six weeks into the puerperium.

Six weeks after surgery, 58/70 (82%) women who had exteriorisation and 57/69 (83%) of those having had in situ repair assessed their abdominal scar quality as excellent or good (P= 1), and satisfaction by the two groups of women with the caesarean section experience was reported by 60/70 (86%) and 64/69 (93%), respectively (P = 0.274); mean satisfaction scores (out of 10) were 8.0 (95% CI 7.5, 8.5) and 8.5 (95% CI 7.9, 9.0), respectively (P = 0.089).


Hershey and Quilligan5 reported similar blood loss, duration of surgery, hospital stay, and rates of puerperal febrile and infectious morbidity in their groups of women who underwent either uterine exteriorisation or in situ repair. They also reported a higher rate of vomiting among women who had exteriorisation, and higher mean haematocrit drops in those who had in situ repair. Magann et al.6 did not find any differences in blood loss or haematocrit drops between their two groups. The conclusions of these studies were criticised by a subsequent systematic review7 which found they were of poor methodology and power, respectively. Obstetricians generally believe exteriorisation of the uterus facilitates caesarean section in various ways, one of which is by improving access to the lower uterine segment, leading to a shorter duration of surgery. This study did not substantiate this as it revealed no statistically (or clinically) significant difference in operating time between the groups.

The vagus nerve supplies cholinergic parasympathetic nerve fibres to the heart, stimulation of which causes a decrease in heart rate18. At rest, there is a good deal of tonic vagal discharge to the heart (vagal tone), opposing the moderate amount of tonic discharge in the cardiac sympathetic nerves. It has been suggested that uterine exteriorisation stimulates a profound vagal discharge, with consequent bradycardia which may even be life threatening8. This has not been substantiated by our study. We did not demonstrate any significant difference in pulse rate between the groups (P= 0.151), and the intra-operative drops in both groups were of no clinical significance. The wide pulse fluctuations observed in both groups (evidenced by wide confidence intervals) followed no clear pattern (Table 2). It appears fluctuations in pulse rate are a common feature of caesarean sections irrespective of the uterine repair technique, and hardly pose a danger to patients. Perhaps the haemodynamic alterations that occur at delivery contribute to this benign phenomenon. The low incidence of bradycardia (1%) was reassuring, as was the response to intravenous atropine.

Exteriorising the uterus is also thought to cause a decrease in mean arterial blood pressure, although the exact mechanisms are less certain. Vagal stimulation causes a decrease in heart rate, which the body ordinarily compensates for by increasing peripheral resistance. However, vagal stimulation could also decrease peripheral resistance by causing vasodilatation of the splanchnic, pulmonary, renal and uterine blood vessels18. All these effects are compounded by the functional sympathetic inhibition that accompanies regional anaesthesia. There is no documented evidence that exteriorising the uterus does indeed stimulate a vagal response. This study found no difference in mean arterial blood pressure changes between women with exteriorisation and those with in situ repair (P= 0.221). Both groups, however, showed large intra-operative drops in mean arterial blood pressure which followed no clear pattern, with large variations evidenced by wide confidence intervals (Table 2). The haemodynamic changes that occur at parturition should affect mean arterial blood pressure, and include sudden uterine decompression with consequent auto-transfusion of blood from the uteroplacental bed into the systemic circulation, as well as sudden loss of It is reasonable to conclude that a decrease in mean arterial blood pressure is a feature of caesarean sections (especially under regional anaesthsia), and is not influenced by the uterine repair technique. Hypotension, when it occurs (15% in this study), responds favourably to treatment.

There was no significant difference in arterial groups (P = 0.195). The intra-operative drops in each group were small (0.22% and 0.25%, respectively) and of no clinical consequence. Although low oxygen saturation occurred in 14% and 6.4%, respectively, the lowest pre-operative and intra-operative values occurred in women who had in situ repair (91% and 93% respectively, different women). None of these women had any untoward effects of low oxygen saturation, and all responded to oxygen therapy. Vartikar et al.8 reported Doppler changes consistent with venous air embolism in 65% of 78 caesarean sections, 47% showing a decrease in oxygen saturation averaging 5.2%, thus suggesting a link between venous air embolism and a fall in oxygen saturation. These Doppler changes were most common at the time of the hysterotomy repair (n= 47) compared with uterine exteriorisation (n= 16), and the hysterotomy incision (n = 9). There have been other reports of sometimes life threatening venous air embolism complicating caesarean section, but no report has suggested a link between venous air embolism and uterine exteriorisation9,11. Of interest is the common occurrence of placenta praevia or abruption with the condition. A review of this condition by Nelson11 reported 45 fatalities in labour and delivery: 7 during caesarean section, 12 in the first stage, 12 in the second and 14 in the third. Placenta praevia occurred in 24% of the cases reviewed. Venous air embolism is a rare but potential complication of any stage of pregnancy, labour and delivery, antepartum haemorrhage being the major predisposing factor. There is no evidence to suggest that the uterine repair technique has any influence on its occurrence, but caution in the presence of antepartum haemorrhage is advised.

Wilcox et al.21 reported the difficulties of estimating blood loss at caesarean section, and concluded it was impossible to perform this accurately. To avoid those pitfalls, we studied changes in haemoglobin concentration between pre-operative and day 3 estimations. There was no peri-operative difference in haemoglobin change between the groups (P= 0.162). The magnitude of the perioperative drops in each group (1.01 g/dL in the exteriorisation group, and 0.67 g/dL in the in situ repair group), is interesting because many obstetricians believe less blood is lost when the uterus is exteriorised for repair. This belief may owe its origin to the erroneous picture of a relatively bloodless field during exteriorisation, since the blood flows away from the operating field onto drapes, unlike the pooling of blood and obscuring of the operating field that occurs with in situ repair. Further blood loss might occur on returning the uterus to the abdomen from an exteriorised position, possibly because of dislodged haemostatic plugs, or re-opening of previously occluded venous sinuses, these small bleeds adding up to a larger total loss.

There was no significant difference in the incidence of nausea (P = 0.413), retching (P = 0.661) and vomiting (P = 0.543) between the groups. The predisposition to nausea and vomiting in pregnancy, coupled with increased intra-gastric pressure, significantly increase the risks of vomiting at caesarean section. Interestingly, vomiting occurred immediately after regional anaesthesia (before the onset of surgery) in a third of all women who vomited. Hypotension was thought to play a role in this, as the restoration of mean arterial blood pressure to preanaesthetic levels coincided with resolution of nausea, retching and/or vomiting. It is disturbing that 10% of our study population vomited during surgery, 90% of this occurring during elective procedures, in women supposedly with an empty stomach. We discovered an inconsistent antiemetic practice duringthe course of this study, the consequence of which was that all emergency cases received a parenteralanti-emetic (usually metoclopramide), whereas elective cases did not. We now strongly recommend a universal anti-emetic policy for caesarean sections, in addition to the usual safeguards of avoiding hypotension, administering antacids, and maintaining adequate oxygenation. Vomiting is a potential complication of caesarean section, especially when performed under regional anaesthesia, irrespective of the uterine repair technique.

Pain during caesarean section has been attributed to various factors, including uterine exteriorisation. Vartikar et al.9 reported chest pain in 26% of their 78 caesarean sections, but failed to indicate at what stage of the operation this occurred. This study did not demonstrate any significant difference between the groups in the incidence of intra-operative pain (P= 0.246). Interestingly, 57% of all pain complaints occurred at the initial skin incision, 14% at incision of the rectus sheath, and 29% on uterine exteriorisation. These data suggest the occurrence of pain depended more on the adequacy of anaesthesia utilised for surgery, than the individual techniques of the operation. Russell2 has suggested that achieving spinal anaesthesia, up to the level of at least T5, will reduce the likelihood of intra-operative pain. It is doubtful whether this is universally practised, and individual anaesthetists may vary the level of block to reduce the risk of hypotension. The pattern of pain continued into the late puerperium (six weeks), when similar percentages (76% vs 77%) in both groups reported persistent pain, with scores that were not significantly different (P= 0.059). The poor correlation between pain scores on days 1, 5 and 42 was surprising and suggests different factors contribute differing proportions to pain at various stages of the puerperium. The peritoneal stretch associated with uterine exteriorisation may have contributed to higher pain scores in the immediate puerperium.

It was remarkable that pain scores averaged 6/10 on day 1 in both groups (often as high as 10/10), despite the unit's policy of multi-modal post-operative analgesia (morphine by patient-controlled analgesia and diclofenac). More remarkable was the finding that three-quarters of women had persistent pain six weeks after surgery, with scores of ≥ 5 or more in 14% and 6%, respectively. Unfortunately, it was not possible to determine the origins of this pain (abdominal scar or pelvis) because of the source of the data (postal questionnaire). It has not been reported previously that pain is a significant complication six weeks after caesarean section. This clear evidence of inadequacy of current post-caesarean analgesia policy calls for its comprehensive review, with due consideration to the prolonged nature of the pain. The challenge is to design more effective analgesia policies, which at the same time cater for the needs of the new mother in bonding with and safely breastfeeding her baby.

There were 56 incidents of puerperal morbidity in 43 women (43%) who had uterine exteriorisation, and 59 incidents in 39 women (41%) who had in situ repair. The uterine repair method did not influence the rates of febrile or infectious morbidity in this study. Febrile morbidity rates (4%vs 3%, P= 1) were impressive compared with previous reports where similar criteria were used14. This probably reflects the benefits of giving prophylactic intra-operative antibiotics and utilising regional anaesthesia which encourages early mobilisation, and reduces postoperative metabolic stress reaction, as well as bowel and lung complication23. Hillan14 reported puerperal pyrexia in 58% of 619 caesarean sections, using the same criteria. Only 26.7% of her patients received antibiotics, with a mere 0.3% receiving it intra-operatively. Unfortunately, no mention was made of theanaesthetic methods utilised in that study.

Expectedly, the rates of urinary tract infection in the community were less than hospital rates, similar in both groups and comparable to previous reports14–17. The rates of wound infections in the community were surprisingly high (26% and 23%, respectively), compared with hospital rates (0% and 1.1%, respectively), and rates of positive wound swabs (12% and 14%, respectively). Although the data were obtained by postal questionnaire with its recognised shortcomings, the numbers of these women requiring antibiotic treatment (approximately half) adds merit to the rates. This high rate of later wound infection could reflect wound care in the community. Perhaps, recent trends towards earlier discharge from hospital played a role, as this might have occurred before wound healing was satisfactorily achieved. The rates of wound infection in hospital compare favourably with previous studies, but no published data pertaining to community wound infection rates after caesarean section are available for comparison. Perhaps patient education, by altering habits, may help to lower this high rate of community wound infection, as may treating all positive wound swabs in hospital even in the absence of overt infection. In contrast, a high percentage of women in both groups (82%vs 83%, P= 1) reported an excellent or good abdominal scar quality at six weeks.

There were no significant differences between the groups, in the incidence of other puerperal complications. The rates of positive high vaginal swab (16% in both groups) was not reflected in rates of pelvic infection in the community (1.4% vs 7.3%, respectively). This may be due to the bactericidal properties of uterine granulation tissue during regeneration and the cleansing role of lochial24. Or perhaps, they are totally unrelated phenomena. The rates of blood transfusion and deep vein thrombosis were comparable with previous reports14,17. The duration and amount of lochia were comparable to previous reports. Expectedly, there was a high level of satisfaction with the operation (86%vs 93%, P= 0.274).


We found no evidence that exteriorisation of the uterus at caesarean section has a more harmful influence on maternal morbidity than in situ repair. We recommend that the choice of uterine repair technique should depend on the surgeon's experience and preference. Caution is, however, advised in certain situations, including presence of large uterine fibroids and major degree placenta praevia, as well as in women with pain at the onset of surgery. Providing adequate anaesthesia for women undergoing caesarean section is of paramount importance, and the need for suitably qualified anaesthetists cannot be over-emphasised. A universal anti-emetic policy is advised to reduce the risk of vomiting, with its attendant danger of aspiration. Most of the women in this study were satisfied with their experience of caesarean section. Overall, morbidity rates were low and, other than pain, were uninfluenced by the uterine repair technique. Caesarean section appears to be associated with a significant amount of post-operative pain, which persists well into the puerperium. Our evidence indicates that current post-operative analgesia is inadequate. There is an urgent need for its review, with special consideration to the protracted nature of this pain. With adequate pre-operative preparation, effective anaesthesia, and observing the cautions above, the practice of uterine exteriorisation at caesarean section should remain a valid option.


The authors would like to thank all the women who participated in this study and the Royal Bolton Hospital which provided material support; the assistance of the obstetricians and midwives who contributed to this work; Dr Farrand for useful advice in the planning stages; the pathology department staff who provided pathology services; the anaesthetic staff for their assistance, and especially the technicians who were instrumental in obtaining accurate records of intra-operative haemodynamics; the medical records and audit departments; and Dr B. Faragher for providing statistical advice.