Volume 82, Issue 3 p. 251-256
Free Access

Intraoperative surgical complication during cesarean section: an observational study of the incidence and risk factors

Thomas Bergholt

Thomas Bergholt

From the Department of Obstetrics and Gynecology in Gentofte,

Search for more papers by this author
Jens Karl Stenderup

Jens Karl Stenderup

Herlev and

Search for more papers by this author
Agnete Vedsted-Jakobsen

Agnete Vedsted-Jakobsen

Glostrup, University of Copenhagen, Denmark

Search for more papers by this author
Peter Helm

Peter Helm

From the Department of Obstetrics and Gynecology in Gentofte,

Search for more papers by this author
Carsten Lenstrup

Carsten Lenstrup

From the Department of Obstetrics and Gynecology in Gentofte,

Search for more papers by this author
First published: 11 April 2003
Citations: 77

Abstract

Background. The study was intended to estimate the incidence of intraoperative surgical complications with the impact of the educational level of the surgeon and a history of previous cesarean section on intraoperative complications at cesarean childbirth.

Methods. In the period between August 1st 1995 and July 30th 1996, 7782 women gave birth at the three Obstetric Departments in Copenhagen County, Denmark, of which 929 (11.9%) were delivered by cesarean section. These women served as the study population, and their medical records were reviewed and data obtained immediately after delivery.

Results. The overall intraoperative complication rate was 12.1%. The rate of complications in emergency cesarean sections was 14.5% compared with 6.8% in the elective group. The educational level of the surgeon and a history of previous cesarean section were not found to be significantly associated to intraoperative complications. Low station of the presenting part of the fetus, high fetal birth weight, fetal distress and dystocia as indications and increasing maternal age were significant risk factors of lacerations. Placenta previa and placental abruption as indications, increasing prepregnancy body mass index, as well as low and high birth weight were significant risk factors for intraoperative blood loss more than 1 l. Duration of regular painful contractions had a preventive effect.

Conclusion. Utero-cervical lacerations and blood loss of more than 1 l were the most frequent intraoperative complications in cesarean section in the present study. The educational level of the surgeon or history of a previous cesarean section were not significantly related to these complications.

Abbreviations:

  • BMI
  • body mass index (kg/m 2 )
  • OR
  • odds ratio
  • CI
  • confidence interval
  • SD
  • standard deviation.
  • Kerr introduced the retrovesical, transverse, lower segment uterine incision in 1926, primarily in an attempt to reduce the incidence of uterine rupture in subsequent pregnancies. He also pointed out other advantages, such as the protective effect against blood loss, wound infection and laceration of the bladder, compared with his experience with the lower vertical incision (1). Today, the lower transverse technique is the most commonly employed type of uterine incision in cesarean section. Even though abdominal delivery is today's most common gynecologic-obstetric operation, few publications on the incidence and risk factors of intraoperative surgical complications have been presented. In surgical intervention it is decisive to perform a continuous audit of morbidity and mortality to evaluate if clinical practice is acceptable and if changes should be implemented to increase patient safety. We decided to describe the incidence of morbidity related to cesarean section in Copenhagen County. As the three participating departments are tertiary teaching hospitals and cesarean section rates have increased throughout the last decades, we wanted to evaluate if cesarean morbidity was affected by the educational level of the surgeon, and if women with a previous uterine scar would have an increased risk of interoperative morbidity.

    Materials and methods

    The University Hospitals in Gentofte, Herlev and Glostrup are tertiary teaching hospitals with 24-h anesthetic, pediatric and obstetric staff, with a neonatal intensive care unit in Glostrup. They serve a population of approximately 600 000 citizens in Copenhagen County. In the period between August 1st 1995 and July 30th 1996, 7782 women gave birth at the three Obstetric Departments, of which 929 (11.9%) were delivered by cesarean section. Of these, 636 (68.5%) were emergency operations. The modified Phannenstiel abdominal and lower segmental transverse uterine incision was the standard operative procedure in the three departments during the study (2). The medical records and parthograms of the women were reviewed by the author. The following outcome variables were obtained: lacerations of the uterus and vagina defined as all sizes of unintended digital or surgical injury of the uterus and the vagina in relation to the uterotomy, lacerations of the bladder and bowels, estimated intraoperative blood loss and blood transfusion, intraoperative hysterectomy and preoperative uterine rupture. For use in the multiple logistic regression modelling process, the following potential explanatory variables were additionally obtained; maternal age, prepregnancy BMI, parity, previous cesarean section, induction of labor, full hours of regular uterine contractions and ruptured membranes, gestation age in completed weeks according to last menstrual period or early ultrasound assessment, cephalic or noncephalic presentation, numbers of fetuses, fetal station defined in relation to the ischial spines, cervical dilatation at the time of cesarean section, use of oxytocin for augmentation, fetal distress, dystocia, placental abruption, placenta previa as indication for cesarean section stated by the obstetrician in charge, if the cesarean section was planned less (emergency) or more (elective) than 8 h before the operation actually took place, type of anesthesia, uterine incision, fetal weight, fetal length and the educational level of the surgeon initiating the operation. Educational level was classified into three categories, pre-educational, educational or specialist, according to the Danish educational system in Obstetrics and Gynecology, and the surgeon was categorized according to his occupational status in the department at the time of the cesarean section.

    Two multivariate models were established to estimate the risk factors for laceration of the uterus, cervix, vagina, bladder, or bowels and intraoperative blood loss ≥€1000 ml using the logistic regression method described by Hosmer and Lemeshow (3). In both models, the above-mentioned explanatory variables were evaluated with univariate likelihood ratio tests. As low segmental transverse uterine incision was used in 926 cases, this variable was extracted from further analysis. One woman was excluded in model 1 and six women in model 2 because of missing data. Chi-square test was used for the univariate analysis in Table I. All multivariate statistical analyses were performed using PROC LOGISTIC in SAS© for Windows, version 6.12 (SAS Institute Inc., Cary, NC, USA). Estimates are presented as the odds ratio with a 95% confidence interval. If the interval does not include 1.0, the estimate of the variable is statistically significant (P-value < 0.05). The study was accepted by the local Ethics Committee for the Sciences in Copenhagen County.

    Table I. Incidence of intraoperative surgical complications
    Type of complication Total Emergency Elective Statistical significance
    Cervical laceration 3.6% (33) 4.6% (29) 1.4% (4) <0.05
    Corporal laceration 0.3% (3) 0.3% (2) 0.3% (1) NS
    Vaginal laceration 1.2% (11) 1.7% (11) 0.0% (0) <0.05
    Bladder laceration 0.5% (5) 0.8% (5) 0.0% (0) NS
    Bowel laceration 0.0% (0) 0.0% (0) 0.0% (0) NS
    Lacerations in total 5.2% (48) 6.8% (43) 1.7% (5) <0.05
    Blood transfusion 1.0% (9) 1.1% (7) 0.7% (2) NS
    Estimated blood loss ≥1000ml 9.2%(77) 9.0%(57) 6.8% (20) NS
    Uterine rupture 0.3% (3) 0.5% (3) 0.0% (0) NS
    Hysterectomy 0.2% (2) 0.2% (1) 0.3% (1) NS
    Total 12.1% (112) 14.5% (92) 6.8% (20) <0.001
    • Figures in parentheses denote the number of patients. Note that the number of complications exceeds the number of patients because patients can have more than one complication.

    Results

    The study population is described in Table II. As presented in Table I, a total of 143 complications were seen in 112 women, resulting in an overall intraoperative complication rate of 12.1%. The rate of complications in emergency cesarean sections was 14.5% compared with 6.8% in the elective group. Laceration of the uterus, cervix, vagina, bladder or bowel was seen in 52 cases in 48 women. Blood transfusions were given during the cesarean section in nine cases, and uterine rupture was seen in three cases, all with a history of previous cesarean section. Hysterectomy was performed intraoperatively because of uncontrollable bleeding, in two cases with placenta previa and in one mother with placental abruption. An estimated blood loss more than or equal to 1000 ml was seen in 77 cases. Of the operations 10.4% were performed by surgeons before specialist education, 46.8% by surgeons during specialist education and 42.8% by specialists. A history of one or more cesarean sections was seen in 25.7% of the women.

    Table II. Demographic and obstetric characteristics of the study population
    Variable Number
    Age (mean) 30.3years (SD: 4.9)
    Parity (mean) 0.7 (SD 0.9)
    Pre-pregnancy body mass index (mean) 23.3kg/height2 (SD: 4.3)
    Previous cesarean section 25.7% (237/929)
    Gestational age (mean) 38.7weeks (SD: 2.7)
    Birth weight (mean) 3.315 kilograms (SD: 0.799)

    As shown in Table III, the educational level of the surgeon performing the cesarean section in this study was statistically insignificant in relation to the risk of intraoperative laceration. Furthermore, a history of previous cesarean section demonstrates a positive but insignificant association with intraoperative laceration (OR = 2.1; 95% CI: 1.0–4.4). The strongest risk factors for laceration in this study were the station of the presenting part of the fetus at or below the ischial spines (OR = 5.9; 95% CI: 2.5–13.9), and birth weight of 4000 g or more (OR = 5.3; 95% CI: 1.8–15.8), both of which demonstrate statistical significance. A significant effect modification appeared between these two risk factors. Other significant risk factors were fetal distress (OR = 3.7; 95% CI: 1.7–8.0) and dystocia (OR = 2.3; 95% CI: 1.1–5.1) as indications. Moreover, maternal age was a significant risk factor (OR = 2.1; 95% CI: 1.1–4.1) for an increase of 10 years. Emergency cesarean section was a significant risk factor in the univariate analysis (OR = 4.2; 95% CI: 1.6–10.7), but not in the adjusted analysis (OR = 2.3; 95% CI: 0.8–6.7).

    Table III. Risk of intraoperative laceration of the cervix, vagina and bladder during cesarean section
    Variables n (%) Crude OR Adjusted OR Adjusted 95% CI
    Educational level of surgeon
    not under specialist education 97 (10.5) 1.0 1.0
    under specialist education 434 (46.8) 1.2 1.2 (0.4–3.8)
    specialist 397 (42.7) 0.9 1.2 (0.4–3.8)
    Previous cesarean section
    no 689 (74.2) 1.0 1.0
    yes 239(25.8) 1.3 2.1 (1.0–4.4)
    Emergency cesarean section
    no 293 (31.5) 1.0 1.0
    yes 635 (68.5) 4.2 2.3 (0.8–6.7)
    Maternal age(10years) 928 (100) 1.8 2.1 (1.1–4.1)
    Station of the presenting part
    above ischial spines 820 (88 4) 1.0 1.0
    ischial spines or below 108 (11.6) 4.3 5.9 (2.5–13.9)
    Fetal distress as indication
    no 687 (74.0) 1.0 1.0
    Yes 241 (26.0) 2.3 3.7 (1.7–8.0)
    Dystocia as indication
    no 623 (67.1) 1.0 1.0
    yes 305 (32.9) 2.6 2.3 (1.7–5.1)
    Birth weight (g)
    <€3000 249 (26.8) 1.0 1.0
    3000–3999 507 (54.7) 0.9 2.4 (0.9–6.1)
    ≥4000 172 (18.5) 2.1 5.3 (1.8–15.8)
    Interaction between birth weight ≥€4000g
    and station at ischial spines or below
    no 895 (96.4) 1.0 1.0
    yes 33 (3.6) 1.9 0.2 (0.1–0.8)
    • Adjusted estimates are mutually adjusted for all variables in the model.

    Table IV presents the estimations of the risk factors for intraoperative blood loss more than or equal to 1000 ml. The relation between the educational level of the surgeon as well as a history of previous cesarean section and intraoperative blood loss was statistically insignificant. The strongest risk factors in this model were placenta previa (OR = 9.2; 95% CI: 3.2–26.2) and placental abruption (OR = 7.3; 95% CI: 3.2–16.9) as indications, both being highly significant. In addition, increasing prepregnancy BMI demonstrated an increasing risk, with indices greater than 25 being significant (OR = 3.8; 95% CI: 1.5–10.0). Duration of regular painful contractions of more than 1 h demonstrated a significant protective association (duration from 1 to 8 h with OR = 0.2; 95% CI: 0.1–0.6, and duration of more than 8 h with OR = 0.3; 95% CI: 0.1–0.8). Birth weight less than 3000 g was a significant risk factor with respect to intraoperative blood loss (OR = 2.4; 95% CI: 1.2–4.6), while birth weight greater than 4000 g also demonstrated a significant positive association (OR = 4.1; 95% CI: 2.0–8.1).

    Table IV. Risk of intraoperative blood loss ≥€1000 ml during the cesarean section
    Variable n (%) Crude OR Adjusted OR Adjusted 95% CI
    Educational level of surgeon
    not under specialist education 96 (10.4) 1.0 1.0
    under specialist education 432 (46.8) 1.1 1.3 (0.5–3.3)
    specialist 395 (42.8) 1.1 1.3 (0.5–3.5)
    Previous cesarean section
    no 686 (74.3) 1.0 1.0
    yes 237 (25.7) 0.6 0.6 (0.3–1.2)
    Emergency cesarean section
    no 291 (31.5) 1.0 1.0
    yes 632 (68.5) 1.3 1.6 (0.7–3.4)
    Pre-pregnancy BMI
    <20 177 (19.2) 1.0 1.0
    20–25 518 (56.1) 2.5 2.3 (0.9–5.7)
    >25 228 (24.7) 3.9 3.8 (1.5–10.0)
    Duration of regular painful contractions (h)
    0 445 (48.2) 1.0 1.0
    1–8 232 (25.1) 0.5 0.2 (0.1–0.6)
    ≥8 246 (26.7) 1.1 0.3 (0.1–0.8)
    Placental abruption as indication
    no 879 (95.2) 1.0 1.0
    yes 44 (4.8) 7.1 7.3 (3.2–16.9)
    Placenta previa as indication
    no 904 (97.9) 1.0 1.0
    yes 19 (2.1) 7.0 9.2 (3.2–26.2)
    Birth weight (g)
    <€3000 247 (26.8) 1.5 2.4 (1.2–4.6)
    3000–3999 505 (54.7) 1.0 1.0
    ≥4000 171 (18.5) 2.7 4.1 (2.0–8.1)
    • Adjusted estimates are mutually adjusted for all variables in the model.

    Discussion

    The incidence of intraoperative laceration in the present study is 5.2%, nearly half the incidence found in the studies by Nielsen and Hökegård and Van Ham et al. after adjustment for fetal complications (4,5), whereas the incidence of blood loss ≥€1000 ml and peroperative transfusion was quite similar. Nielsen and Hökegård found that the risk of intraoperative laceration, blood transfusion and injury to the fetus was significantly higher in an emergency compared with an elective cesarean section (4), and that the skills of the surgeon and history of a previous cesarean section, among other factors, were significantly associated with intraoperative complications in emergency cesarean sections. Van Lam et al. reported in their study the same correlation between emergency cesarean section and the risk of intraoperative laceration and blood loss (5). Both publications utilized univariate statistical analysis, and the effects of the risk factors were not estimated. Furthermore, the possibility of confounding and interaction by other variables were not discussed. As most biological and medical relations are complex mechanisms involving several causal components, multiple regression analysis is a useful tool to evaluate and control variables under consideration for confounding and interaction.

    The most influential preoperative risk factors with respect to intraoperative laceration were high birth weight and low station of the presenting part of the fetus. Both these risk factors confirm that difficulty in extracting the head and the shoulders throughout the uterine incision could be one causal mechanism of intraoperative laceration. Landesman and Graber presented a modification of the cesarean section technique in order to facilitate delivery of the impacted head in the second stage of labor, thereby reducing the injury to the fetal head and the uterus. An assistant lifts the fetal head after gentle disengagement with a cupped hand through the vagina. When the presenting part of the fetus has been lifted to the level of the uterine incision, the head should be directed through the incision (6). Another possible way to extract the head gently through the uterine incision could be by the use of the ventouse or forceps in situations with low station or high fetal weight. As fetal distress as an indication is a risk factor in this model, the speed at which a cesarean section is performed could be another causal mechanism in intraoperative laceration. In a study by Andersen et al., the authors focused on the relation between skin incision to delivery time, and uterine incision to delivery time, in 204 patients undergoing cesarean delivery. Based on stepwise multiple linear regression analysis, they concluded that these time intervals did not influence the neonatal outcome, and suggested that the surgical technique should always be directed at a gentle and nontraumatic delivery (7). Nielsen and Hökegård and Van Ham et al. in their studies found that an emergency cesarean section was a risk factor compared with elective operations (4,5). These findings are likely to be caused by the same basic mechanism regarding the importance of speed. This significant correlation was also seen in univariate analysis in the present study, but after controlling for other influential variables in the multivariate analysis this positive association became insignificant. The presence of increasing maternal age as a significant risk factor could be caused by an increasing prevalence of nonsymptomatic cardio-vascular morbidity, making the myometrial tissue more susceptible to lacerations in relation to cesarean section. It accentuates the imperative necessity of always implementing age as a variable in biological and medical models, as age is often related to morbidity and mortality.

    In the logistic regression model concerning intraoperative blood loss, the most influential risk factors were placental abruption and placenta previa as indications. This finding should not be a surprise, as the indication for cesarean section in these cases is based on the risk of severe intrapartum bleeding. Other risk factors were increasing prepregnancy BMI and high birth weight. Again, a plausible common causal mechanism behind these risk factors could be the difficult extraction of the fetus. It also seems that uterine contractions have a protective effect on the intraoperative blood loss. This could be the result of an improved vasoconstriction in the uterine wall with regular contractions after placenta removal. In addition, the lower uterine segment is thinner in most cases of cesarean section in women with uterine contractions compared with women undergoing elective operations. This could also explain the positive effect of uterine contractions on blood loss. Enkin and Wilkinson concluded in their review that manual removal of the placenta compared with spontaneous expulsion resulted in a clinically important and statistically significant increase in maternal blood loss (8). Uterine exteriorization before closure of the uterotomy has also been demonstrated to reduce the blood loss significantly during cesarean section (9). The technique used to remove the placenta in this study was not described in the medical records, and consequently was not applied in the analysis. The importance of speed as a risk factor in this model did not demonstrate a significant impact on the intraoperative blood loss, as an emergency section and fetal distress as indications did not contribute significantly to the final model. In future, spontaneous expulsion and exterioization of the uterus combined with minimal traumatic surgical techniques like the Misgav Ladach method (10) could possibly reduce the blood loss associated with cesarean section. These less invasive operative procedures were introduced in the three departments following the study period. Evaluation of the impact of these operative alterations on the incidence of intraoperative complications during a cesarean section is under consideration.

    The possibility of confounding might occur in the educational level of the surgeon. In obstetric practice, if any complication is expected before surgery the operation is usually allocated to a senior obstetrician. Consequently, a higher incidence of complications should be expected in these women. To control for this potential confounding, several explanatory variables were introduced in the logistic regression modelling process, but it is difficult to assess the influence of possible subjective mechanisms that could affect the preoperative risk evaluation and the decision concerning who should perform the operation. Furthermore, if the operation was performed by a resident, the surgeon was usually supervised by a senior obstetrician. If complications occurred during surgery the operation could be taken over by the supervising obstetrician. To control for this differential misclassification of the surgeon, the educational level of the surgeon who initiated the operation was used in the models. The outcome variables of this study were written in the medical records by the surgeon. In all surgical interventions, the surgeon describes the indication, findings, procedures, possible complications that might have occurred, and the estimated blood loss during surgery at the end of the operation. Estimating the blood loss is especially difficult in the case of a cesarean section, as the blood could be assimilated with an unknown amount of amniotic fluid. However, being an established procedure following surgery, it seems unlikely that the classification of intraoperative complications should be dependent on any of the explanatory variables.

    In obstetrics, as in other areas of medicine, an indication for intervention should only be established after careful weighing of the benefits and risks of the treatment. When a cesarean section is under consideration for a woman before or in labor, potential complications for the mother must be part of the obstetrician's concern. Even though cesarean section is the most frequently used surgical intervention in women today, few publications have addressed this topic. Perinatal audit has been introduced and established in Denmark for many years (11,12). Systematic audit of maternal morbidity and mortality in relation to birth has not received much attention. The method of auditing in labor wards concerning maternal aspects of childbirth has been described (13,14). The present study demonstrates how the systematic collection of data in relation to abdominal delivery can be used for descriptive and analytic purposes. No relation between the educational level of the surgeon and the risk of intraoperative complication was found, and we recommend that the three departments should continue their educational practice without maternal risk.

    Papers have been published stressing the autonomy of the mother and her wish for the mode of delivery, advocating for the mother's right to give birth by cesaresan section without any obstetric or medical indications (15–17). Thus the mode of delivery should be made after carefully presenting balanced information about the risks and benefits in relation to vaginal and abdominal deliveries. If the intention about well-informed patients should be substantial, labor ward units must address more attention to record and present their data concerning maternal outcomes (18). In addition, this will give each labor unit the possibility of evaluating and, if necessary, improving obstetric management to increase maternal and fetal care. In our opinion, this will enable us to meet maternal and political wishes in the future.

    Address for correspondence:
    Thomas Bergholt
    Gentoftegade 41
    2tv 2820 Gentofte
    Denmark
    e-mail: [email protected]