Objective
Blunt vs sharp expansion of the uterine incision at cesarean delivery has been investigated as a technique primarily to reduce intraoperative blood loss. The objective of this systematic review was to compare the effects of either intervention on maternal outcomes.
Study Design
A systematic review with metaanalyses that used the DerSimonian and Laird random effects model was performed. The Cochrane Central Register of Controlled Trials (The Cochrane Library 2012, Issue 4), MEDLINE (1948–Apr 2012), EMBASE (1947–Apr 2012), and the reference lists/citation history of articles were searched. Only randomized controlled trials were included.
Results
Four trials (1731 patients) were evaluated. Data from one recently completed trial (535 patients) were not yet available. Metaanalyses revealed a trend towards reduced maternal blood loss with blunt expansion of the uterine incision that was statistically significant when measured by surgeon’s estimation of volume lost, but not by comparison of pre- and postoperative hematocrit and hemoglobin levels or a requirement for blood transfusion. There was a trend towards fewer unintended extensions in the blunt group and no difference in the incidence of endometritis.
Conclusion
Blunt dissection of the uterine incision at cesarean delivery appears to be superior to sharp dissection in minimizing maternal blood loss. However, this conclusion could change when data from a new unpublished large trial are available.
The global rate of cesarean deliveries is approximately 15% ; the rate in some developed countries reaches as high as 1 in 3. It generally is accepted that a greater amount of blood loss is likely in an operative, compared with vaginal, delivery. As obstetric hemorrhage remains a leading cause of maternal morbidity and death, techniques such as manual placental extraction, in situ uterine repair in the place of exteriorization, and blunt traction in the cephalocaudad, rather than transverse, direction for uterine incision have been proposed to minimize intraoperative blood loss during cesarean delivery.
Blunt expansion of the uterine incision with fingers, rather than scissors, is one other such suggestion. Previously, proponents of either the blunt or sharp method would defer to training protocols, personal experience, or theoretic reasoning to explain their choice of technique. To date, a small number of studies specifically have examined the impact of the hysterotomy expansion technique on maternal blood loss during cesarean delivery. The main suggested advantage of the blunt approach includes decreased trauma to the vasculature with less bleeding and ooze from the dissected myometrial edge. Secondary potential benefits include less risk of injury to the neonate and cord and increased speed of delivery.
However, there are concerns about reduced control of length and direction of the uterine incision that potentially could cause damage of lateral uterine and parametrial blood vessels and increased risk of unintended extensions that could contribute further to hemorrhage. The effect of blunt division of the uterine wall on the incidence of endometritis postcesarean delivery is also of concern.
The aim of this review was to compare the impacts of sharp vs blunt hysterotomy on the primary outcome of maternal blood loss and the secondary outcomes of unintended extension, incidence of postoperative endometritis, injury to the neonate, postoperative pain, and operative time/time to delivery.
Materials and Methods
The Cochrane and Preferred Reporting Items for Systematic Review and Metaanalyses (PRISMA) guidelines were followed for the performance and reporting of this systematic review.
Types of studies and outcomes considered
All prospective randomized controlled trials (RCTs) that compared blunt vs sharp expansion of the initial uterine incision at transverse lower segment cesarean delivery were considered. Quasirandomized trials and studies that assessed vertical lower-segment or classic upper-segment uterine incisions were excluded. Ongoing or recently completed trials with no data yet available were noted for future analyses. Our primary outcome was maternal blood loss. Our secondary outcomes were incidence of extension, endometritis, neonatal morbidity, postoperative pain, and time to delivery.
Search strategy
A literature search of the Cochrane Central Register of Controlled Trials (The Cochrane Library 2012, Issue 4), Medline (via Ovid; 1948–April 2012), and Embase (via Ovid) (1947–April 2012) was performed in week 4, April 2012. The prospective search protocol for each database is given in Table 1 . No language restrictions were used. All titles were assessed; where the abstract suggested a potentially eligible study, the full text was retrieved. Scopus was used to cross-reference the references and citation history of full-text articles. A search for ongoing or recently completed trials was performed in week 4, April 2012, with the Australia/New Zealand, United Kingdom, and United States Clinical Trials registries ( www.anzctr.org.au , www.controlled-trials.com , www.clinicaltrials.gov , respectively). Studies were evaluated critically for design and risk of bias, according to criteria set out in the Cochrane handbook for systematic reviews of interventions.
| No. | Central | Medline | Embase |
|---|---|---|---|
| 1 | MeSH descriptor cesarean delivery explode all trees | exp cesarean section or caesarean section.mp or exp uterus | exp cesarean section or caesarean section.mp or exp uterus |
| 2 | blunt or finger* or digital or sharp or instrument* or dissection or expansion or extension | blunt.mp or finger$.mp or digital.mp or sharp.mp or instrument$.mp or dissection.mp or expansion.mp or extension.mp | blunt.mp or finger$.mp or digital.mp or sharp.mp or instrument$.mp or dissection.mp or expansion.mp or extension.mp |
| 3 | #1 and #2 | randomized controlled trial.pt or controlled clinical trial.pt or randomized.ab or placebo.ab or clinical trials as topic.sh or randomly.ab or trial.ti | randomized controlled trial or controlled clinical trial or randomized.ab or placebo.ab or exp clinical trial(topic) or randomly.ab or trial.ti |
| 4 | and/1-3 | and/1-3 | |
| 5 | exp animals/not humans.sh | exp animals/not humans.sh | |
| 6 | 4 not 5 | 4 not 5 |
Statistical analysis
Data were extracted onto a standardized collection form by 2 independently working authors (L.X., A.C.) and entered into RevMan (version 5.1, 2011; The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). Data were analyzed with the use of a random-effects metaanalysis (DerSimonian and Laird model) with risk ratio as the pooled estimate for dichotomous data and mean difference for continuous data. Mantel-Haenszel and inverse variance methods were used, respectively. Analysis was performed on an intention-to-treat basis. Substantial statistical heterogeneity was considered to be present when there was inconsistency between trials in the direction or magnitude of effects. This was assessed visually from the forest plots or when the I 2 statistic was >50%, respectively. Statistical significance was defined as a probability value < .05.
Results
Evidence base
The literature search returned 495 articles ( Figure 1 ). From these, 4 RCTs (1731 patients) that reported data appropriate for the clinical question ( Tables 2 ) were included in the study. One recently completed RCT had not yet reported the results, so it could not be included. Two studies were not RCTs.
| Study and location | Women randomly assigned, n | Mean age at delivery, y | Inclusion criteria | Exclusion criteria | Intervention |
|---|---|---|---|---|---|
| Rodriguez et al, Florida | 286 | Blunt, 25.8; sharp, 25.7 | All patients undergoing primary or repeat transverse lower segment cesarean delivery; indications included arrest, malpresentation, previous cesarean delivery, fetal distress | Patients unable to give informed consent because of emergency surgery or communication difficulties | Uterine incision extended by fingers or control extended with bandage scissors; operator: 2nd- to 4th-year resident under direct supervision |
| Magann et al, Mississippi | 975 | Blunt, 24.7; sharp, 24.4 | All patients undergoing primary or repeat transverse lower segment cesarean delivery | Women declining participation, emergency surgery, use of vertical lower segment or classical upper segment uterine incision | Uterine incision extended by fingers or control extended with scissors; 20 units oxytocin in 1000 mL Ringer’s lactate rapid infusion after placental delivery; operator: 2nd- to 4th-year residents with assistance of attending staff |
| Hidar et al, Sousse, Tunisia | 300 | Blunt, 31.52; sharp, 31.6 | All patients with a singleton pregnancy undergoing transverse lower segment cesarean delivery at gestation of >36 wk | Age <20 y; known coagulopathy; placenta previa; transverse presentations; preoperative bleeding | Uterine incision extended by fingers or control extended with scissors; manual placental delivery/in situ uterine repair; operator: 3rd- or 4th-year resident under senior supervision |
| Sekhavat et al, Yazd, Iran | 200 | Blunt, 24.3; sharp, 25.1 | Primiparous patients undergoing elective transverse lower segment cesarean delivery with fundal placenta | Severe medical and surgical disorders; blood disorder/anemia; known thromboembolic disorder; multiple gestation; fetal macrosomia; polyhydramnios; emergency surgery for placental abruption; placenta previa, and severe preeclampsia | Uterine incision extended by fingers or control extended with scissors; manual delivery of placenta, 10 units oxytocin in 500 mL normal saline solution >10 minutes; all patients underwent general anesthesia; operator: 2nd-year resident under supervision |
Study design
Hysterotomy in the blunt expansion groups was performed after an initial 1-2cm incision was made through the uterine wall and then extended by insertion of the surgeon’s index fingers laterally and cephalad. In the sharp expansion groups, the extension was achieved by cutting with bandage scissors laterally and cephalad.
There were some differences in study design between the trials. Of the studies that specified cesarean technique, 1 study used the Pfannenstiel incision and the other Joel-Cohen incision. Two studies described the implementation of manual delivery of the placenta. An oxytocin infusion after placental delivery was used in 2 studies (20 units in 1000 mL Ringer’s lactate; 10 units in 500 mL normal saline solution). Prophylactic antibiotics were given to all patients in 1 study and to patients in active labor after cord clamping in another.
Women who underwent lower-segment cesarean delivery were included in these trials; however, there were differing additional enrolment criteria. Three studies excluded women who underwent emergency surgery. One study enrolled exclusively nulliparous women because of the risk of uterine inertia with multiparity. General anesthesia was also used in all patients in this study. Two studies excluded patients with known coagulopathy and multiple gestations. From the 2 studies that provided data, the most common indications for cesarean delivery were arrest of labor, malpresentation, previous cesarean delivery, and fetal distress.
Critical appraisal of included RCTs
All 4 trials were judged to have low overall risk of bias ( Figure 2 ). Two trials had a high risk of detection bias in subjective outcome measures because of lack of blinding. One study had an unclear risk of selection bias because the authors did not specify the method of randomization; however, there were no statistically significant differences between the experimental and control groups in all recorded characteristics.
Primary outcome: impact on maternal blood loss
Mean drop in hematocrit level
Three studies evaluated maternal blood loss by reduction in hematocrit level before and after the operation.
Magann et al compared immediate preoperative hematocrit level with that 48 hours after the operation. A statistically significant reduction in hematocrit drop was found favoring the blunt group ( P = .003). Consistent with this, the incidence of women who experienced a >10% decrease in hematocrit level was also significantly smaller in the blunt, rather than the sharp, group ( P = .03; author calculated with χ 2 test).
Hidar et al compared immediate preoperative hematocrit level with that 24-48 hours after the operation and found no difference in the average reduction in hematocrit level between the 2 groups ( P = .58). The incidence of >10% reduction in hematocrit level was also comparable ( P > .05).
Sekhavat et al measured hematocrit level immediately before the operation and 24 hours after the operation. A decreased drop in pre- to post-operative hematocrit level in the blunt group that was detected when the data were compared with the sharp group was significant ( P < .05).
Combined results from the 3 trials revealed a trend that favored blunt hysterotomy for reduced drop in hematocrit level after the operation ( Figure 3 ) ; however, this did not reach statistical significance (mean difference [MD], −0.86%; 95% CI, −2.04 to 0.32; 3 trials; 1445 patients). There was substantial statistical heterogeneity across studies (I 2 = 91%).
Mean drop in hemoglobin
Three studies recorded change in pre- and postoperative hemoglobin levels.
Rodriguez et al found no difference in mean hemoglobin level decrease when they compared the hemoglobin level that was measured at admission and the 24 hours postoperative measurement in the 2 groups ( P = .08; author calculated with unpaired t test).
Hidar et al found the 2 groups comparable in mean hemoglobin level drop when they compared the values that were measured immediately before and 24-48 hours after surgery ( P = .51).
Sekhavat et al found a significantly smaller mean hemoglobin level difference when they compared the values measured before and 24 hours after surgery in the blunt group ( P < .05).
Pooled results showed a trend toward a reduced drop in hemoglobin level that favored the blunt dissection group ( Figure 4 ), although this was not statistically significant (MD, 7.41 g/L; 95% CI, −20.53 to 5.72; 3 trials; 786 patients). There was substantial heterogeneity across studies (I 2 = 98%).
