- Email: [email protected]
Intra-operative haemorrhage by blunt versus sharp expansion of the uterine incision at caesarean delivery: a randomised clinical trial
BJOG: an International Journal of Obstetrics and Gynaecology April 2002, Vol. 109, pp. 448 –452
Intra-operative haemorrhage by blunt versus sharp expansion of the uterine incision at caesarean delivery: a randomised clinical trial Everett F. Maganna,*, Suneet P. Chauhanb, Laura Bufkina, Karen Fielda, William E. Robertsa, James N. Martin Jra Objective To determine whether the method used to expand the uterine incision for caesarean delivery affects the incidence of intra-operative haemorrhage. Design A prospective randomised study of women undergoing a low segment transverse caesarean delivery. Participants were assigned to have their uterine incision either sharply or bluntly expanded. Participants Between June 1998 and June 2000, 470 women drew assignments to the sharp expansion group and 475 to the blunt group. Results The maternal demographics of age, race, nulliparity, and body mass index as well as pre-operative haematocrit were similar between groups. Compared with the blunt group, the estimated blood loss (886 versus 843mL, P ¼ 0.001), change in the mean haematocrit (6.1% versus 5.5%, P ¼ 0.003), incidence of postpartum haemorrhage (13% versus 9%; relative risk ¼ 1.23, 95% CI 1.03, 1.46) and need for a transfusion (2% versus 0.4%; relative risk ¼ 1.65, 95% CI 1.25, 2.21) were significantly greater in the sharp group. Conclusion In caesarean delivery, sharply expanding the uterine incision significantly increases intra-operative blood loss and the need for subsequent transfusion. INTRODUCTION The most recent figures of maternal mortality in the United States are 7.5 maternal deaths per 100,000 live births1. Gherman et al.2 found that obstetric haemorrhage, as well as pre-eclampsia and thromboembolic disease, accounts for the majority of these deaths. Obstetric haemorrhage can occur in association with operative abdominal delivery3. In a recent investigation at our institution, we discovered that 8% of our caesarean deliveries are accompanied by blood loss in excess of 1500mL and/or a drop in the haematocrit of 10%4. Several techniques have been proposed to reduce blood loss at operative abdominal delivery, including placental management (spontaneous compared with manual extraction) and in situ uterine incision repair compared with exteriorisation5 – 7. Another technique is uterine expansion (blunt versus sharp) of the uterine incision in low segment transverse caesarean deliveries8. Rodriguez et al.8 reported no difference between the methods of uterine expansion with
a
Department of Obstetrics and Gynaecology, University of Mississippi Medical Centre, Jackson, USA b The Spartanburg Regional Healthcare System, Spartanburg, Mississippi, USA * Correspondence: Professor E. F. Magann, Department of Obstetrics and Gynaecology, University of Mississippi Medical Center, 2500 North State Street, Jackson, Mississippi 39216-4505, USA. D RCOG 2002 BJOG: an International Journal of Obstetrics and Gynaecology PII: S 1 4 7 0 - 0 3 2 8 ( 0 2 ) 0 1 2 9 6 - X
respect to post-operative decrease in maternal haemoglobin: blunt (1.8g/dL (0.1)) compared with sharp (2.2g/dL (0.2)). Although in Rodriguez’s study the blood loss was greater in the sharp group (147 patients) compared with the blunt group (139 patients), the difference was not significant. In light of the relatively small size of the investigation, our group wondered if a larger study group might exhibit the same findings. The purpose of this investigation was to determine which method of expanding the uterine incision at the time of operative delivery is associated with an increased risk of postpartum haemorrhage (decrease in haematocrit of 10% or more) or need for a packed red blood cell transfusion.
METHODS All patients admitted to the University of Mississippi Medical Centre that were destined to deliver abdominally by primary or repeat low segment transverse uterine incision were invited to participate in this investigation. Exclusion criteria included women who declined to participate, women undergoing emergency surgery with insufficient time to properly counsel the patient, and women in whom either a low segment vertical uterine or a classical upper segment were utilised. The Institutional Review Board of the University of Mississippi Medical Centre, Jackson, Mississippi, approved this investigation. Following study enrolment and completion of the informed consent, the women were randomised to either www.bjog-elsevier.com
BLUNT VS SHARP UTERINE EXPANSION
449
Fig. 1. Profile of the randomised clinical trial.
Group 1 (sharp expansion of the uterine incision) or Group 2 (blunt expansion of the uterine incision). Assignment to one of the two treatment groups was generated using a table of random numbers. The assigned treatment was written on a card and sealed in secure opaque envelopes numbered in sequence. All operative procedures were performed similarly before the uterine incision9. A transverse uterine incision in the lower uterine segment of approximately 2cm in length was made with the scalpel and the uterine incision was then expanded by the designated method. Sharp expansion of the primary incision was developed by cutting laterally and cephalad using bandage scissors. Blunt expansion of the primary incision was derived by placing the index fingers of the operating surgeon into the incision and pulling the fingers apart laterally and cephalad. After delivery of the placenta, 20 units of oxytocin were placed in 1000mL of lactated ringers solution and rapidly infused into the patient to assist in the adequate contraction of the postpartum uterus. Additional use of oxytocin, prostaglandins, or intravenous fluids to assist to assist in the management of intraoperative haemorrhage was recorded in all patients. The remainder of the operation was performed similarly for all patients9. All surgical procedures were performed in the same institution by second, third or fourth year obstetric and gynaecologic residents with the assistance and/or the direct D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
supervision of the attending staff. Blood loss estimation was undertaken using several methodologies. First, it was estimated by the operating surgeon and the attending staff by evaluating the amount of blood collected in the suction apparatus, the plastic steridrapes, and the lap pads and sponges. Secondly, the blood loss was assessed by comparing the immediate pre-operative haematocrit with a second haematocrit obtained 48 hours after the operative procedure. The total number of patients from each group with a fall in the maternal haematocrit of 10% was calculated. In addition, the number of patients from each group who were transfused was determined. Uniform criteria for all transfused patients included an haematocrit of <24% and haemodynamic instability. An extension of the primary uterine incision was defined as any defect which was observed outside the limits of the original incision. Extension of the primary incision into the broad ligament(s) that involved the uterine vessels were recorded. Cervical extensions of the primary uterine incision were also recorded. A sample size and power calculation was undertaken, and it was determined that 474 women in each group was sufficient power (power of 0.80, a of 0.05, and b ¼ 0.20) to demonstrate a difference of 5% between groups. These calculations are based on a previous study in which the risk of obstetric haemorrhage at caesarean delivery was
450
E.F. MAGANN ET AL.
approximately 10%4. Assuming that the decision to make the uterine incision as transverse, vertical or classical might be mandated before surgery and that some patients would refuse study participation, a target study population of 500 women were recruited for each group. The data were analysed for statistical significance by using the Student t test, the x2 test, and Fisher’s exact test. P < 0.05 was considered significant. Relative risk (RR) and 95% confidence interval (CI) were calculated. All analyses were on an intent-to-treat basis and included all randomised patients.
Table 1. Maternal demographics. MgSO4 ¼ magnesium sulfate; HCT ¼ hematocrit; BMI ¼ body mass index; CS ¼ caesarean section. Sharp (n ¼ 470)
Blunt (n ¼ 475)
Mean [SD] age (years)
24.4 [6.2]
24.7 [6.3]
Race White African-American Others
81 347 42
84 351 40
Nulliparous
153
157
Mean [SD] BMI
34.2 [8.7]
33.7 [8.5]
Prior CS delivery
263
278
Anesthesia General Regional Pre-eclampsia on MgSO4
51 419 78
89 386 82
Cervical dilation at CS <4 cm 4 – 9 cm 10 cm
316 128 26
311 142 22
Mean [SD] pre-op HCT
34.3 [3.9]
34.1[3.8]
RESULTS During the 24-month study interval there were 6583 deliveries at the University of Mississippi Medical Centre of which 1514 were abdominal deliveries. Forty-three women of the 1000 women eligible for the study were not registered because of the decision of the operating surgeon to perform a low segment vertical or classical caesarean section. This decision was based on gestational age, fetal position or known uterine pathology. This left a total of 957 women eligible to participate in this study. Twelve women declined participation in the study, leaving a total of 945 women who drew assignment to this prospective randomised trial. A total of 470 women were enrolled in the sharp expansion group and 475 in the blunt expansion group (Fig. 1). As shown in Table 1, demographics including maternal age ( P ¼ 0.46), race ( P ¼ 0.951), nulliparity (RR 0.98; 95% CI 0.86, 1.13), body mass index ( P ¼ 0.372) chorioamnionitis at caesarean delivery (RR 0.94; 95% CI 0.75, 1.17) and birthweight ( P ¼ 0.421) were similar between the two groups. Antenatal and intrapartum events that have been shown to influence blood loss in obstetric patients10 – 12 and in our specific obstetric patient population3 were assessed (Table 1). The number of women in the sharp group with chronic hypertension and superimposed pre-eclampsia or women with pre-eclampsia treated with magnesium sulphate was not significantly different from the number of such women in the blunt group (RR 0.97, 95% CI, 0.82, 1.16). Likewise the number of women who had had a previous caesarean delivery was also similar between the groups (RR 0.87; 95% CI 0.67, 1.13). The mean cervical dilation at the time of caesarean delivery was similar between the sharp and blunt groups (2.8cm [3.2] vs 3.2cm [3.1]) ( P ¼ 0.051). The number of participants who were dilated less than 4cm, 4cm to 9cm and 10cm at the time of caesarean delivery was similar between groups ( P ¼ 0.585). As illustrated in Table 1, more women in the sharp group received conduction anaesthesia compared with the blunt group (RR 1.42, 95% CI 1.13 – 1.79). Chorioamnionitis was present in a similar number of patients at the time of caesarean delivery whether in the sharp or blunt groups (RR 0.94, 95% CI 0.75, 1.17). Birthweight and was also similar between groups (Table 1).
Chorioamnionitis Mean [SD] birthweight (g)
48 2937 [829]
54 2980 [814]
The mean pre-operative haematocrit was similar between the groups 34.3 [3.9] in the sharp compared with 34.1 [3.8] in the blunt ( P ¼ 0.42). However, the estimated blood loss at operative delivery was greater in the sharp group compared with blunt group ( P ¼ 0.001). The use of additional oxytocin, haemabate and intravenous fluids beyond the 1000mL of lactated ringers solution with 20 units of oxytocin was also similar between the sharp and blunt groups. Consistent with this, the change in the haematocrit from the pre-operative to post-operative assessment was greater in the sharp group compared with the blunt group ( P ¼ 0.003). A significantly greater number of women in the sharp group compared with the blunt group had >10% decrease in the pre-operative to the post-operative haematocrit (RR 1.23; 95% CI 1.03 –1.46). Many more women in the sharp group were transfused post-operatively than were women in the blunt group (RR 1.65; 95% CI 1.25 –2.21). A greater number of women in the sharp group had a cervical laceration than did the women in the blunt group but the difference was not significant (RR 1.32; 95% CI 0.9– 1.79). The total number of extensions was also greater in the sharp group compared with the blunt group (RR 1.66; 95% CI 1.47 –1.86). The length of the extensions were also assessed. A greater number of women in the sharp group had 1cm and 2cm extensions of their primary incisions, 3cm extensions (RR 0.48; 95% CI 0.34– 0.69) but not D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
BLUNT VS SHARP UTERINE EXPANSION
451
Table 2. Outcomes.
Mean [SD] estimated blood loss* Oxytocin þ >1 liter of fluid Haemabate Mean [SD] HCT change** >10% decrease in HCT Transfusion Uterine scar extension 1 – 2 cm 3 cm Broad ligament Cervical lacerations Postpartum endometritis
Sharp (n ¼ 470)
Blunt (n ¼ 475)
886 [197] 35 22 6.1 [3.2] 62 9
843 [164] 31 19 5.5 [3] 42 2
57 69 16 15 66
20 24 7 8 51
RR (95% CI)
1.07 (0.84, 1.35) 1.08 (0.80, 1.45) 1.23 (1.03, 1.46) 1.65 (1.25, 2.21) 0.49 0.48 0.59 1.32 1.16
(0.33, (0.34, (0.32, (0.90, (0.97,
0.73) 0.69) 1.12) 1.79) 1.38)
* P ¼ 0.001. ** P ¼ 0.003.
extensions into the broad ligament[s] compared with blunt expansion with 1cm and 2cm extensions (RR 0.49; 95% CI 0.33, 0.73), 3cm extensions (RR.48; 95% CI 0.34, 0.69) and extensions into the broad ligament (RR.59; 95% CI 0.32, 1.12) (Table 2).
DISCUSSION The purpose of this investigation was to determine whether a specific surgical technique, blunt compared with sharp expansion of the transverse lower uterine incision, was related to an increased intraoperative blood loss at caesarean delivery. In order to draw valid conclusions between the surgical techniques used and relate these to blood loss the investigator must ensure that other reasons for obstetric haemorrhage are similar between the studied patient groups. A number of specific risk factors for excessive bleeding at the time of operative abdominal delivery has been reported, including general anaesthesia, emergency delivery compared with elective caesarean section, infection such as chorioamnionitis, labour disorders (i.e. protracted active phase of labour) and delivery technique (manual versus spontaneous) used to remove the placenta10 – 12. In addition, there are other recognised risk factors that increase the risk of obstetric haemorrhage in women delivering at the University of Mississippi Medical Centre, including pre-eclampsia, disorders of active labour, being of Native-American (Choctaw) ethnicity, previous postpartum haemorrhage and obesity (maternal weight >300 pounds)3. In this study population, all factors were equally distributed between the groups with the single exception of an increased number of women in the blunt group receiving general anaesthesia. Blood loss with general anaesthesia is reported to be increased compared with conduction anaesthesia10. Thus, at the onset one would anticipate that the blood loss would be greater in the blunt versus the sharp group, but this is not what we observed. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
Blood loss at the time of operative delivery was assessed using a variety of procedures. The first method was to measure the blood loss at the time of operative delivery. We have shown previously that as long as the calculations are done meticulously that they are reasonably accurate and reflect actual blood loss13. Using this methodology, blood loss was significantly greater in the sharp group compared with the blunt group. This calculation of the blood loss was done by assessing the blood in the suction apparatus, in the plastic pouches of the drapes and the saturation of the lap pads and sponges by the operating surgeon and circulating nurse. The authors acknowledge that since the operating surgeon is obviously aware of the technique being used that this could influence his final estimation of the blood loss but this would not influence the change in the haematocrit or the post-operative need for a blood transfusion. A second methodology used to estimate blood loss was determination of the change in haematocrit with that obtained 48 hours after the operative procedure. The factors that have been determined to influence intraoperative blood loss were similar between the two groups. Likewise, the use of additional oxytoxic agents and intravenous fluids used to treat intraoperative uterine atony was also not different between groups despite the similarity in risk factors. The change in haematocrit was significantly greater in the sharp group. The American College of Obstetricians and Gynecologists has defined postpartum haemorrhage as a change of 10% in the haematocrit between admission and the postpartum period or need for a red cell transfusion3. The number of women with a negative change in the haematocrit >10% between the pre-operative haematocrit and that obtained 48 hours after surgery was also greater than in the sharp compared with the blunt group. Furthermore, the need for transfusion of packed red blood cells because of an haematocrit which was <24% and haemodynamic instability was assessed. This parameter was assessed since blood loss differences between
452
E.F. MAGANN ET AL.
groups might be considered to be more statistically significant but not clinically significant. However, the administration of packed red blood cells is a significant medical event and it clearly singles out those women who have had significant bleeding and have increased their morbidity because of that bleeding. The two commonly used techniques to expand the uterine incision are the blunt technique using the operator’s fingers and the sharp technique using scissors to expand the uterine incision. The expected advantages of the blunt technique are speed and decreased bleeding from the myometrial edges14. The disadvantage of the blunt technique is the lack of control which may cause damage laterally to major vessels15. The reported advantage of the sharp technique is the precise expansion of the uterine incision thereby protecting the uterine artery and parametrial veins16. Both techniques are used at the University of Mississippi Medical Centre but the blunt technique is the primary method of expanding the uterine incision. The only randomised controlled trial found by the authors in the English language literature which has assessed expanding uterine incisions is the study by Rodriguez et al.8 In that investigation, although the decrease in haemoglobin was greater in the sharp group, it was not statistically significant. Would the trend in the haemoglobin have been significant if they had sampled a larger patient population? Unfortunately, in their study no information is presented about the use of blood transfusions. Why is blood loss increased when sharp extension of the uterine incision is used? The most logical reason is that there is increased bleeding from the incised incision itself. This increased bleeding could result from muscle ooze, traumatised vasculature or secondary to a greater forward extension of the distal incision. In our study we observed that the number of 1cm – 2cm and 3cm extensions of the primary incision were greater in the sharp compared with the blunt group. Although not statistically significant the number of extensions into the broad ligament(s) including the uterine vessels and cervical extensions was also greater in the sharp compared with the blunt expansion group. We postulate that the increased bleeding is due to the greater number of extensions in the sharp group. The blunt expansion of the uterine incision appears to offer some degree of protection to the uterine vessels. The overall reduction in the number of extensions appears to significantly impact the amount of blood lost at abdominal delivery. The results of this investigation suggest that sharply expanding a low segment transverse uterine incision at the time of operative delivery significantly increases the risk of intraoperative postpartum haemorrhage. The number of patients who will require a transfusion because of an
haematocrit <24 with haemodynamic instability also is increased.
Acknowledgements This project was supported in part by the Vicksburg Hospital Medical Foundation, Vicksburg, Mississippi.
References 1. Maternal mortality: United States 1982 – 1996. MMWR 1998;47: 705 – 707. 2. Gherman RB, Macri CJ, Gerstenfeld T, Goodwin TM. Maternal mortality at LAC/ USC Women’s Hospital: a 17-year review. Obstet Gynecol 2000;95:S44 – S45. 3. American College of Obstetricians and Gynecologists. Postpartum hemorrhage. Technical Bulletin No. 243. Washington, DC: ACOG, 1998. 4. Naef III RW, Chauhan SP, Chevalier SP, Roberts WE, Meydrech EF, Morrison JC. Prediction of hemorrhage at caesarean delivery. Obstet Gynecol 1994;83:923 – 926. 5. McCurdy Jr CM, Magann EF, McCurdy CJ, Saltzman AK. The effect of placental management at caesarean delivery on operative blood loss. Am J Obstet Gynecol 1992;167:1363 – 1367. 6. Magann EF, Dodson MK, Allbert JR, McCurdy Jr CM, Martin RW, Morrison JC. Blood loss at time of caesarean section by method of placental removal and exteriorization versus in situ repair of the uterine incision. Surg Gynecol Obstet 1993;177:389 – 392. 7. Wilkinson C, Enkin MW. Manual removal of placenta at caesarean section. Cochrane Database Systematic Review 2, CD000130, 2000. 8. Rodriguez AI, Porter KB, O’Brien WF. Blunt versus sharp expansion of the uterine incision in low-segment transverse caesarean section. Am J Obstet Gynecol 1994;171:1022 – 1025. 9. Depp R. Cesarean delivery. In: Gabbe SG, Niebyl JR, Simpson JL, editors. Obstetrics: Normal and Problem Pregnancies. New York: Churchill Livingston, 1996:576 – 582. 10. Combs CA, Murphy EL, Laros RK. Factors associated with hemorrhage in caesarean deliveries. Obstet Gynecol 1991;77:77 – 82. 11. Gilstrap III LC, Hauth JC, Hankins GD, Patterson AR. Effect of type of anesthesia on blood loss at caesarean section. Obstet Gynecol 1987;69:328 – 332. 12. Kamani AA, McMorland GH, Wadsworth LD. Utilization of red blood cell transfusion in an obstetric setting. Am J Obstet Gynecol 1988;159:1177 – 1181. 13. Dodson MK, Magann EF, Chauhan SP, Harris RL, Martin Jr JN, Morrison JC. Accuracy of blood loss estimation and measurement at caesarean birth. J Matern Fetal Med 1994;3:171 – 174. 14. Clark SL. Cesarean delivery. In: Hankins GDV, Clark SL, Cunningham FG, Gilstrap LC, editors. Operative Obstetrics. Norwalk, Connecticut: Appleton and Lange, 1995:301 – 322. 15. Jananoul R. Incisions of the pregnant uterus and delivery of low birth weight infants. Am J Obstet Gynecol 1985;152:971 – 974. 16. Burchell RC. Cesarean section. In: Iffy L, Charles D, editors. Operative Perinatology. New York: McMillian Publishing, 1984:706 – 730. Accepted 22 January 2002
D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
Intra-operative haemorrhage by blunt versus sharp expansion of the uterine incision at caesarean delivery: a randomised clinical trial Everett F. Maganna,*, Suneet P. Chauhanb, Laura Bufkina, Karen Fielda, William E. Robertsa, James N. Martin Jra Objective To determine whether the method used to expand the uterine incision for caesarean delivery affects the incidence of intra-operative haemorrhage. Design A prospective randomised study of women undergoing a low segment transverse caesarean delivery. Participants were assigned to have their uterine incision either sharply or bluntly expanded. Participants Between June 1998 and June 2000, 470 women drew assignments to the sharp expansion group and 475 to the blunt group. Results The maternal demographics of age, race, nulliparity, and body mass index as well as pre-operative haematocrit were similar between groups. Compared with the blunt group, the estimated blood loss (886 versus 843mL, P ¼ 0.001), change in the mean haematocrit (6.1% versus 5.5%, P ¼ 0.003), incidence of postpartum haemorrhage (13% versus 9%; relative risk ¼ 1.23, 95% CI 1.03, 1.46) and need for a transfusion (2% versus 0.4%; relative risk ¼ 1.65, 95% CI 1.25, 2.21) were significantly greater in the sharp group. Conclusion In caesarean delivery, sharply expanding the uterine incision significantly increases intra-operative blood loss and the need for subsequent transfusion. INTRODUCTION The most recent figures of maternal mortality in the United States are 7.5 maternal deaths per 100,000 live births1. Gherman et al.2 found that obstetric haemorrhage, as well as pre-eclampsia and thromboembolic disease, accounts for the majority of these deaths. Obstetric haemorrhage can occur in association with operative abdominal delivery3. In a recent investigation at our institution, we discovered that 8% of our caesarean deliveries are accompanied by blood loss in excess of 1500mL and/or a drop in the haematocrit of 10%4. Several techniques have been proposed to reduce blood loss at operative abdominal delivery, including placental management (spontaneous compared with manual extraction) and in situ uterine incision repair compared with exteriorisation5 – 7. Another technique is uterine expansion (blunt versus sharp) of the uterine incision in low segment transverse caesarean deliveries8. Rodriguez et al.8 reported no difference between the methods of uterine expansion with
a
Department of Obstetrics and Gynaecology, University of Mississippi Medical Centre, Jackson, USA b The Spartanburg Regional Healthcare System, Spartanburg, Mississippi, USA * Correspondence: Professor E. F. Magann, Department of Obstetrics and Gynaecology, University of Mississippi Medical Center, 2500 North State Street, Jackson, Mississippi 39216-4505, USA. D RCOG 2002 BJOG: an International Journal of Obstetrics and Gynaecology PII: S 1 4 7 0 - 0 3 2 8 ( 0 2 ) 0 1 2 9 6 - X
respect to post-operative decrease in maternal haemoglobin: blunt (1.8g/dL (0.1)) compared with sharp (2.2g/dL (0.2)). Although in Rodriguez’s study the blood loss was greater in the sharp group (147 patients) compared with the blunt group (139 patients), the difference was not significant. In light of the relatively small size of the investigation, our group wondered if a larger study group might exhibit the same findings. The purpose of this investigation was to determine which method of expanding the uterine incision at the time of operative delivery is associated with an increased risk of postpartum haemorrhage (decrease in haematocrit of 10% or more) or need for a packed red blood cell transfusion.
METHODS All patients admitted to the University of Mississippi Medical Centre that were destined to deliver abdominally by primary or repeat low segment transverse uterine incision were invited to participate in this investigation. Exclusion criteria included women who declined to participate, women undergoing emergency surgery with insufficient time to properly counsel the patient, and women in whom either a low segment vertical uterine or a classical upper segment were utilised. The Institutional Review Board of the University of Mississippi Medical Centre, Jackson, Mississippi, approved this investigation. Following study enrolment and completion of the informed consent, the women were randomised to either www.bjog-elsevier.com
BLUNT VS SHARP UTERINE EXPANSION
449
Fig. 1. Profile of the randomised clinical trial.
Group 1 (sharp expansion of the uterine incision) or Group 2 (blunt expansion of the uterine incision). Assignment to one of the two treatment groups was generated using a table of random numbers. The assigned treatment was written on a card and sealed in secure opaque envelopes numbered in sequence. All operative procedures were performed similarly before the uterine incision9. A transverse uterine incision in the lower uterine segment of approximately 2cm in length was made with the scalpel and the uterine incision was then expanded by the designated method. Sharp expansion of the primary incision was developed by cutting laterally and cephalad using bandage scissors. Blunt expansion of the primary incision was derived by placing the index fingers of the operating surgeon into the incision and pulling the fingers apart laterally and cephalad. After delivery of the placenta, 20 units of oxytocin were placed in 1000mL of lactated ringers solution and rapidly infused into the patient to assist in the adequate contraction of the postpartum uterus. Additional use of oxytocin, prostaglandins, or intravenous fluids to assist to assist in the management of intraoperative haemorrhage was recorded in all patients. The remainder of the operation was performed similarly for all patients9. All surgical procedures were performed in the same institution by second, third or fourth year obstetric and gynaecologic residents with the assistance and/or the direct D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
supervision of the attending staff. Blood loss estimation was undertaken using several methodologies. First, it was estimated by the operating surgeon and the attending staff by evaluating the amount of blood collected in the suction apparatus, the plastic steridrapes, and the lap pads and sponges. Secondly, the blood loss was assessed by comparing the immediate pre-operative haematocrit with a second haematocrit obtained 48 hours after the operative procedure. The total number of patients from each group with a fall in the maternal haematocrit of 10% was calculated. In addition, the number of patients from each group who were transfused was determined. Uniform criteria for all transfused patients included an haematocrit of <24% and haemodynamic instability. An extension of the primary uterine incision was defined as any defect which was observed outside the limits of the original incision. Extension of the primary incision into the broad ligament(s) that involved the uterine vessels were recorded. Cervical extensions of the primary uterine incision were also recorded. A sample size and power calculation was undertaken, and it was determined that 474 women in each group was sufficient power (power of 0.80, a of 0.05, and b ¼ 0.20) to demonstrate a difference of 5% between groups. These calculations are based on a previous study in which the risk of obstetric haemorrhage at caesarean delivery was
450
E.F. MAGANN ET AL.
approximately 10%4. Assuming that the decision to make the uterine incision as transverse, vertical or classical might be mandated before surgery and that some patients would refuse study participation, a target study population of 500 women were recruited for each group. The data were analysed for statistical significance by using the Student t test, the x2 test, and Fisher’s exact test. P < 0.05 was considered significant. Relative risk (RR) and 95% confidence interval (CI) were calculated. All analyses were on an intent-to-treat basis and included all randomised patients.
Table 1. Maternal demographics. MgSO4 ¼ magnesium sulfate; HCT ¼ hematocrit; BMI ¼ body mass index; CS ¼ caesarean section. Sharp (n ¼ 470)
Blunt (n ¼ 475)
Mean [SD] age (years)
24.4 [6.2]
24.7 [6.3]
Race White African-American Others
81 347 42
84 351 40
Nulliparous
153
157
Mean [SD] BMI
34.2 [8.7]
33.7 [8.5]
Prior CS delivery
263
278
Anesthesia General Regional Pre-eclampsia on MgSO4
51 419 78
89 386 82
Cervical dilation at CS <4 cm 4 – 9 cm 10 cm
316 128 26
311 142 22
Mean [SD] pre-op HCT
34.3 [3.9]
34.1[3.8]
RESULTS During the 24-month study interval there were 6583 deliveries at the University of Mississippi Medical Centre of which 1514 were abdominal deliveries. Forty-three women of the 1000 women eligible for the study were not registered because of the decision of the operating surgeon to perform a low segment vertical or classical caesarean section. This decision was based on gestational age, fetal position or known uterine pathology. This left a total of 957 women eligible to participate in this study. Twelve women declined participation in the study, leaving a total of 945 women who drew assignment to this prospective randomised trial. A total of 470 women were enrolled in the sharp expansion group and 475 in the blunt expansion group (Fig. 1). As shown in Table 1, demographics including maternal age ( P ¼ 0.46), race ( P ¼ 0.951), nulliparity (RR 0.98; 95% CI 0.86, 1.13), body mass index ( P ¼ 0.372) chorioamnionitis at caesarean delivery (RR 0.94; 95% CI 0.75, 1.17) and birthweight ( P ¼ 0.421) were similar between the two groups. Antenatal and intrapartum events that have been shown to influence blood loss in obstetric patients10 – 12 and in our specific obstetric patient population3 were assessed (Table 1). The number of women in the sharp group with chronic hypertension and superimposed pre-eclampsia or women with pre-eclampsia treated with magnesium sulphate was not significantly different from the number of such women in the blunt group (RR 0.97, 95% CI, 0.82, 1.16). Likewise the number of women who had had a previous caesarean delivery was also similar between the groups (RR 0.87; 95% CI 0.67, 1.13). The mean cervical dilation at the time of caesarean delivery was similar between the sharp and blunt groups (2.8cm [3.2] vs 3.2cm [3.1]) ( P ¼ 0.051). The number of participants who were dilated less than 4cm, 4cm to 9cm and 10cm at the time of caesarean delivery was similar between groups ( P ¼ 0.585). As illustrated in Table 1, more women in the sharp group received conduction anaesthesia compared with the blunt group (RR 1.42, 95% CI 1.13 – 1.79). Chorioamnionitis was present in a similar number of patients at the time of caesarean delivery whether in the sharp or blunt groups (RR 0.94, 95% CI 0.75, 1.17). Birthweight and was also similar between groups (Table 1).
Chorioamnionitis Mean [SD] birthweight (g)
48 2937 [829]
54 2980 [814]
The mean pre-operative haematocrit was similar between the groups 34.3 [3.9] in the sharp compared with 34.1 [3.8] in the blunt ( P ¼ 0.42). However, the estimated blood loss at operative delivery was greater in the sharp group compared with blunt group ( P ¼ 0.001). The use of additional oxytocin, haemabate and intravenous fluids beyond the 1000mL of lactated ringers solution with 20 units of oxytocin was also similar between the sharp and blunt groups. Consistent with this, the change in the haematocrit from the pre-operative to post-operative assessment was greater in the sharp group compared with the blunt group ( P ¼ 0.003). A significantly greater number of women in the sharp group compared with the blunt group had >10% decrease in the pre-operative to the post-operative haematocrit (RR 1.23; 95% CI 1.03 –1.46). Many more women in the sharp group were transfused post-operatively than were women in the blunt group (RR 1.65; 95% CI 1.25 –2.21). A greater number of women in the sharp group had a cervical laceration than did the women in the blunt group but the difference was not significant (RR 1.32; 95% CI 0.9– 1.79). The total number of extensions was also greater in the sharp group compared with the blunt group (RR 1.66; 95% CI 1.47 –1.86). The length of the extensions were also assessed. A greater number of women in the sharp group had 1cm and 2cm extensions of their primary incisions, 3cm extensions (RR 0.48; 95% CI 0.34– 0.69) but not D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
BLUNT VS SHARP UTERINE EXPANSION
451
Table 2. Outcomes.
Mean [SD] estimated blood loss* Oxytocin þ >1 liter of fluid Haemabate Mean [SD] HCT change** >10% decrease in HCT Transfusion Uterine scar extension 1 – 2 cm 3 cm Broad ligament Cervical lacerations Postpartum endometritis
Sharp (n ¼ 470)
Blunt (n ¼ 475)
886 [197] 35 22 6.1 [3.2] 62 9
843 [164] 31 19 5.5 [3] 42 2
57 69 16 15 66
20 24 7 8 51
RR (95% CI)
1.07 (0.84, 1.35) 1.08 (0.80, 1.45) 1.23 (1.03, 1.46) 1.65 (1.25, 2.21) 0.49 0.48 0.59 1.32 1.16
(0.33, (0.34, (0.32, (0.90, (0.97,
0.73) 0.69) 1.12) 1.79) 1.38)
* P ¼ 0.001. ** P ¼ 0.003.
extensions into the broad ligament[s] compared with blunt expansion with 1cm and 2cm extensions (RR 0.49; 95% CI 0.33, 0.73), 3cm extensions (RR.48; 95% CI 0.34, 0.69) and extensions into the broad ligament (RR.59; 95% CI 0.32, 1.12) (Table 2).
DISCUSSION The purpose of this investigation was to determine whether a specific surgical technique, blunt compared with sharp expansion of the transverse lower uterine incision, was related to an increased intraoperative blood loss at caesarean delivery. In order to draw valid conclusions between the surgical techniques used and relate these to blood loss the investigator must ensure that other reasons for obstetric haemorrhage are similar between the studied patient groups. A number of specific risk factors for excessive bleeding at the time of operative abdominal delivery has been reported, including general anaesthesia, emergency delivery compared with elective caesarean section, infection such as chorioamnionitis, labour disorders (i.e. protracted active phase of labour) and delivery technique (manual versus spontaneous) used to remove the placenta10 – 12. In addition, there are other recognised risk factors that increase the risk of obstetric haemorrhage in women delivering at the University of Mississippi Medical Centre, including pre-eclampsia, disorders of active labour, being of Native-American (Choctaw) ethnicity, previous postpartum haemorrhage and obesity (maternal weight >300 pounds)3. In this study population, all factors were equally distributed between the groups with the single exception of an increased number of women in the blunt group receiving general anaesthesia. Blood loss with general anaesthesia is reported to be increased compared with conduction anaesthesia10. Thus, at the onset one would anticipate that the blood loss would be greater in the blunt versus the sharp group, but this is not what we observed. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452
Blood loss at the time of operative delivery was assessed using a variety of procedures. The first method was to measure the blood loss at the time of operative delivery. We have shown previously that as long as the calculations are done meticulously that they are reasonably accurate and reflect actual blood loss13. Using this methodology, blood loss was significantly greater in the sharp group compared with the blunt group. This calculation of the blood loss was done by assessing the blood in the suction apparatus, in the plastic pouches of the drapes and the saturation of the lap pads and sponges by the operating surgeon and circulating nurse. The authors acknowledge that since the operating surgeon is obviously aware of the technique being used that this could influence his final estimation of the blood loss but this would not influence the change in the haematocrit or the post-operative need for a blood transfusion. A second methodology used to estimate blood loss was determination of the change in haematocrit with that obtained 48 hours after the operative procedure. The factors that have been determined to influence intraoperative blood loss were similar between the two groups. Likewise, the use of additional oxytoxic agents and intravenous fluids used to treat intraoperative uterine atony was also not different between groups despite the similarity in risk factors. The change in haematocrit was significantly greater in the sharp group. The American College of Obstetricians and Gynecologists has defined postpartum haemorrhage as a change of 10% in the haematocrit between admission and the postpartum period or need for a red cell transfusion3. The number of women with a negative change in the haematocrit >10% between the pre-operative haematocrit and that obtained 48 hours after surgery was also greater than in the sharp compared with the blunt group. Furthermore, the need for transfusion of packed red blood cells because of an haematocrit which was <24% and haemodynamic instability was assessed. This parameter was assessed since blood loss differences between
452
E.F. MAGANN ET AL.
groups might be considered to be more statistically significant but not clinically significant. However, the administration of packed red blood cells is a significant medical event and it clearly singles out those women who have had significant bleeding and have increased their morbidity because of that bleeding. The two commonly used techniques to expand the uterine incision are the blunt technique using the operator’s fingers and the sharp technique using scissors to expand the uterine incision. The expected advantages of the blunt technique are speed and decreased bleeding from the myometrial edges14. The disadvantage of the blunt technique is the lack of control which may cause damage laterally to major vessels15. The reported advantage of the sharp technique is the precise expansion of the uterine incision thereby protecting the uterine artery and parametrial veins16. Both techniques are used at the University of Mississippi Medical Centre but the blunt technique is the primary method of expanding the uterine incision. The only randomised controlled trial found by the authors in the English language literature which has assessed expanding uterine incisions is the study by Rodriguez et al.8 In that investigation, although the decrease in haemoglobin was greater in the sharp group, it was not statistically significant. Would the trend in the haemoglobin have been significant if they had sampled a larger patient population? Unfortunately, in their study no information is presented about the use of blood transfusions. Why is blood loss increased when sharp extension of the uterine incision is used? The most logical reason is that there is increased bleeding from the incised incision itself. This increased bleeding could result from muscle ooze, traumatised vasculature or secondary to a greater forward extension of the distal incision. In our study we observed that the number of 1cm – 2cm and 3cm extensions of the primary incision were greater in the sharp compared with the blunt group. Although not statistically significant the number of extensions into the broad ligament(s) including the uterine vessels and cervical extensions was also greater in the sharp compared with the blunt expansion group. We postulate that the increased bleeding is due to the greater number of extensions in the sharp group. The blunt expansion of the uterine incision appears to offer some degree of protection to the uterine vessels. The overall reduction in the number of extensions appears to significantly impact the amount of blood lost at abdominal delivery. The results of this investigation suggest that sharply expanding a low segment transverse uterine incision at the time of operative delivery significantly increases the risk of intraoperative postpartum haemorrhage. The number of patients who will require a transfusion because of an
haematocrit <24 with haemodynamic instability also is increased.
Acknowledgements This project was supported in part by the Vicksburg Hospital Medical Foundation, Vicksburg, Mississippi.
References 1. Maternal mortality: United States 1982 – 1996. MMWR 1998;47: 705 – 707. 2. Gherman RB, Macri CJ, Gerstenfeld T, Goodwin TM. Maternal mortality at LAC/ USC Women’s Hospital: a 17-year review. Obstet Gynecol 2000;95:S44 – S45. 3. American College of Obstetricians and Gynecologists. Postpartum hemorrhage. Technical Bulletin No. 243. Washington, DC: ACOG, 1998. 4. Naef III RW, Chauhan SP, Chevalier SP, Roberts WE, Meydrech EF, Morrison JC. Prediction of hemorrhage at caesarean delivery. Obstet Gynecol 1994;83:923 – 926. 5. McCurdy Jr CM, Magann EF, McCurdy CJ, Saltzman AK. The effect of placental management at caesarean delivery on operative blood loss. Am J Obstet Gynecol 1992;167:1363 – 1367. 6. Magann EF, Dodson MK, Allbert JR, McCurdy Jr CM, Martin RW, Morrison JC. Blood loss at time of caesarean section by method of placental removal and exteriorization versus in situ repair of the uterine incision. Surg Gynecol Obstet 1993;177:389 – 392. 7. Wilkinson C, Enkin MW. Manual removal of placenta at caesarean section. Cochrane Database Systematic Review 2, CD000130, 2000. 8. Rodriguez AI, Porter KB, O’Brien WF. Blunt versus sharp expansion of the uterine incision in low-segment transverse caesarean section. Am J Obstet Gynecol 1994;171:1022 – 1025. 9. Depp R. Cesarean delivery. In: Gabbe SG, Niebyl JR, Simpson JL, editors. Obstetrics: Normal and Problem Pregnancies. New York: Churchill Livingston, 1996:576 – 582. 10. Combs CA, Murphy EL, Laros RK. Factors associated with hemorrhage in caesarean deliveries. Obstet Gynecol 1991;77:77 – 82. 11. Gilstrap III LC, Hauth JC, Hankins GD, Patterson AR. Effect of type of anesthesia on blood loss at caesarean section. Obstet Gynecol 1987;69:328 – 332. 12. Kamani AA, McMorland GH, Wadsworth LD. Utilization of red blood cell transfusion in an obstetric setting. Am J Obstet Gynecol 1988;159:1177 – 1181. 13. Dodson MK, Magann EF, Chauhan SP, Harris RL, Martin Jr JN, Morrison JC. Accuracy of blood loss estimation and measurement at caesarean birth. J Matern Fetal Med 1994;3:171 – 174. 14. Clark SL. Cesarean delivery. In: Hankins GDV, Clark SL, Cunningham FG, Gilstrap LC, editors. Operative Obstetrics. Norwalk, Connecticut: Appleton and Lange, 1995:301 – 322. 15. Jananoul R. Incisions of the pregnant uterus and delivery of low birth weight infants. Am J Obstet Gynecol 1985;152:971 – 974. 16. Burchell RC. Cesarean section. In: Iffy L, Charles D, editors. Operative Perinatology. New York: McMillian Publishing, 1984:706 – 730. Accepted 22 January 2002
D RCOG 2002 Br J Obstet Gynaecol 109, pp. 448 – 452