Background
Hysterectomy is one of the most frequently performed major gynecological surgical procedures. Even when the indication for the procedure is benign, relatively high complication rates have been reported. Perioperative bleeding seems to represent the most common cause of complications and in 2004, 8% of all women in Denmark undergoing benign hysterectomy experienced a bleeding complication. Tranexamic acid is an antifibrinolytic agent that has shown to effectively reduce bleeding complications within other surgical and medical areas. However, knowledge about the drug’s effect in relation to benign hysterectomy is still missing.
Objective
To investigate the antihemorrhagic effect of prophylactic tranexamic acid in elective benign hysterectomy.
Study Design
A double-blinded randomized placebo-controlled trial was conducted at 4 gynecological departments in Denmark from April 2013 to October 2014. A total of 332 women undergoing benign abdominal, laparoscopic, or vaginal hysterectomy were included in the trial, randomized to either 1 g of intravenous tranexamic acid or placebo at start of surgery. Chi-square test and Student t test statistical analyses were applied.
Results
The primary outcome of intraoperative total blood loss was reduced in the group treated with tranexamic acid compared to the placebo group when estimated both subjectively by the surgeon and objectively by weight (98.4 mL vs 134.8 mL, P = .006 and 100.0 mL vs 166.0 mL, P = .004). The incidence of blood loss ≥500 mL was also significantly reduced (6 vs 21, P = .003), as well as the use of open-label tranexamic acid (7 vs 18, P = .024). Furthermore, the risk of reoperations owing to postoperative hemorrhage was significantly reduced in the tranexamic acid group compared to the placebo group (2 vs 9, P = .034). This corresponds to an absolute risk reduction of 4.2% and number needed to treat of 24. No incidence of thromboembolic events or death was observed in any of the groups.
Conclusion
The results support the hypothesis that prophylactic treatment with tranexamic acid reduces the overall total blood loss, the incidence of substantial blood loss, and the need for reoperations owing to postoperative hemorrhage in relation to benign hysterectomy. No incidences of serious adverse events occurred. Thus, tranexamic acid should be considered as a prophylactic treatment prior to elective benign hysterectomy.
Hysterectomy is one of the most frequently performed major gynecological surgical procedures. In the United States the rate of hysterectomy was 5.1 per 1000 women in 2004, but there is large variation in rates between countries. In Denmark, the annual rate of hysterectomy on benign indications has been slightly decreasing throughout the last decade, from 1.9 per 1000 women in 2002 to 1.5 in 2008. Uterine fibroids and menorrhagia account for the majority of these operations. The Danish Hysterectomy and Hysteroscopy Database (DHHD) is a national register that collects data on all women undergoing elective benign hysterectomy. Data from this register has revealed a relatively high complication rate of 16–18% from 1998 to 2006, of which perioperative bleeding complications represent the most common cause. From 2004 to 2006, 6–8% of all women undergoing benign hysterectomy in Denmark experienced a bleeding complication with both socioeconomic and individual implications.
Tranexamic acid (TA) is an antifibrinolytic agent approved for treatment of various types of hemorrhage. It inhibits fibrin degradation, thereby promoting the blood’s ability to form stable blood clots. In several countries, the drug is used as prophylactic treatment prior to major surgery. In a Cochrane review addressing TA’s efficacy in all types of surgery, a significant reduction of bleeding was found corresponding to a mean of 414 mL. Similar results have been found within traumatology.
The antihemorrhagic effect of TA regarding gynecological surgery has been investigated in only a limited number of clinical trials. In contrast, a substantial number of studies have been conducted on the effect of TA as treatment for women with menorrhagia. Early studies have found menorrhagia associated with elevated levels of fibrinolytic activity. This may contribute to the explanation of the relatively high rates of bleeding complications related to hysterectomy, as menorrhagia is one of the most common indications for the procedure. Consequently, women undergoing hysterectomy might benefit from prophylactic TA during surgery. On this basis the aim of this study was to investigate the antihemorrhagic effect of prophylactic TA in elective benign hysterectomy.
Materials and Methods
A prospective randomized double-blinded placebo-controlled trial was conducted at 4 hospitals in Denmark. Participants were enrolled between April 1, 2013, and October 30, 2014 at the Gynecological Departments of North Zealand Hospital, Rigshospitalet, Odense University Hospital, and Aarhus University Hospital. A total of 539 women were consecutively screened for eligibility ( Figure 1 ): 131 were excluded owing to exclusion criteria, 62 declined to participate, 1 had the operation cancelled by the physician and, 3 women were incorrectly not randomized. Subsequently, 342 women were successfully randomized, out of which 5 withdrew their consent to participate, 4 had the operation canceled, and 1 had the operation postponed, resulting in 332 participants being included in the trial. Distribution of the participants included was 153 at North Zealand Hospital, 78 at Rigshospitalet, 81 at Odense University Hospital, and 20 at Aarhus University Hospital. The protocol for the trial was registered at www.clinicaltrials.gov , id: NCT01940419 . The design, conduct, and report of the study are in accordance with the CONSORT 2010 statement.
All women scheduled for benign hysterectomy were potentially eligible to participate in the study. A physician or a project nurse in the outpatient clinic was responsible for the enrollment. Exclusion criteria were thrombophilia, previous or active thromboembolic disease, family history of thromboembolism, allergy to TA, renal impairment, ongoing hematuria, subarachnoid hemorrhage, malignant disease, or antithrombotic treatment. Women were also excluded if they were unable to fully understand project information. In case of cancellation of the hysterectomy after randomization, the women were excluded from further participation, since none of the predefined outcomes could be obtained.
Computer randomization was performed by a central, independent unit through the web-based clinical trial support system Trial Partner (Public Health and Quality Improvement, Central Denmark Region, Aarhus N, Denmark). The randomization was carried out in blocks of alternating sizes of 2, 4, and 6, and stratified by method of surgery (abdominal, vaginal, laparoscopic/robotic surgery). The patient, the physician, and the person collecting data were all blinded during the trial. Blinding was maintained throughout registration and analysis of all data.
Just before the operation was initiated, 1 g TA (10 mL of 0.1 g TA/mL) was administered intravenously (IV). A corresponding volume of sterile saline was used as placebo (10 mL). Both fluids were transparent and odorless. An independent project nurse prepared the study medicine in a room separated from the operating theatre after randomization. Subsequently the project nurse handed over the allocated study medicine to the anesthesiologist in the operating theatre, after which she took no further part in the study.
The primary outcomes were defined as intraoperative estimated blood loss by the surgeon (EBLs) and by weight (EBLw) in mL. EBLw was calculated by subtraction of the weight of saline irrigation, suction drain, and all used towels/pads after the operation from the preoperative weight; 1 g of fluid was converted to 1 mL of blood. EBLs was a subjective estimate of intraoperative blood loss made by the primary surgeon. The secondary outcomes included intraoperative blood loss ≥1000 mL or ≥500 mL, use of intraoperative open-label IV TA, use of intraoperative topical hemostatics, blood transfusion during primary hospitalization, duration of surgery, change in hemoglobin level from preoperatively to 4–12 hours postoperatively, overall need for analgesic >48 hours postoperatively, complicated condition of pain during primary hospitalization (including all diagnoses of abdominal pain), postoperative bleeding complications within 30 days postoperatively (including all diagnoses of hematoma and hemorrhagia), reoperations owing to bleeding complications within 30 days postoperatively, length of hospital stay (days), readmission owing to bleeding complications within 30 days of the operation (including all diagnoses of hematoma/hemorrhagia and abdominal pain), thromboembolic events, and mortality within 30 days postoperatively.
The surgeon estimated data regarding the primary outcomes immediately after termination of the procedure. Values were noted on a prespecified trial form. Additional data regarding the primary hospitalization, including adverse events of thromboembolism, were collected by the local investigators. They collected data from the local electronic patient records, and information was registered in an online trial database. Information concerning baseline characteristics and the 30-days-postoperative period were obtained from DHHD and the Danish National Patient Register and merged with the trial database. Data regarding education level were retrieved by an individual questionnaire supplied to the participants prior to the operation.
Retrospective data from DHHD estimated the mean intraoperative blood loss in women undergoing benign hysterectomy in Denmark in 2010 to be 155 mL. Based on the medical literature the anticipated effect of TA on intraoperative blood loss reduction was set to 25%. With a type 1 error of 0.05 and a type 2 error of 0.20, the power analysis showed that a total of 314 women had to be included in the trial. Initially we conducted descriptive univariate analyses of relevant patient characteristics to ensure effective randomization. Subsequently univariate statistical analyses concerning the primary and secondary outcomes were performed. The Student t test was applied on continuous variables and chi-square test on noncontinuous variables. When 50% of the counts for noncontinuous data were expected less than 5, the Fisher exact test was applied. Results of relative risk (RR) with 95% confidence intervals (CI) were presented for noncontinuous outcomes. If continuous variables were not normally distributed they were log-transformed before the t test was applied and results presented as geometric means, medians, and interquartile ranges.
It was decided that women with negative values on the EBLw were censored in the primary analyses. The correlation between the EBLs and EBLw was examined. A chi-square test exploring statistical difference in the number of censored women in the intervention and placebo groups was carried out. Furthermore, a t test regarding EBLs in the women with negative EBLw values was performed. Subgroup analyses regarding the primary outcomes were conducted on the 3 different methods of hysterectomy: vaginal hysterectomy, abdominal hysterectomy (AH), and laparoscopic/robotic assisted laparoscopic hysterectomy (LH/RH); Subgroup analyses were also performed on the group of women having menorrhagia/metrorrhagia as indication vs all other indications. The statistical analyses were conducted using SAS version 9.4 software (SAS Institute, Cary, NC).
The study was approved by the National Ethical Committee and the Danish Health and Medicine Authority: journal number 2011-425 and EudraCT-number 2012-005407-40. The Danish Data Protection Agency issued authorization for the study: journal number 2007-58-0015. The study was carried out according to the Declaration of Helsinki and Good Clinical Practice. Three independent Good Clinical Practice units were responsible for monitoring and assurance of quality during the trial. All women signed informed consent before inclusion in the study. The study was not supported by any pharmaceutical company.
Results
A total of 332 participants were included. All hysterectomies were performed under general anesthesia and all participants received preoperative antibiotic and anticoagulation treatment 4–12 hours postoperatively, according to national guidelines. All participating hospitals had residents under surgical training. The vast majority of the hysterectomies were performed by a senior gynecologist assisted by a resident under training, but no data concerning the surgeon’s experience were collected. Table 1 summarizes the participant characteristics. No statistically significant differences were observed between the TA and the placebo group.
| Characteristics | TA group n = 165 | Placebo group n = 167 | P |
|---|---|---|---|
| Age (years), mean ± SD (range) | 47.9 ± 8.9 (31–77) | 49.1 ± 9.9 (29–85) | .247 |
| BMI, median (IQR) | 25.0 (22.0–28.7) | 25.9 (22.4–28.5) | .824 |
| Smoker, n (%) | 36 (21.9%) | 31 (18.9%) | .494 |
| Alcohol ≥7 units/week, n (%) | 19 (11.6%) | 23 (14.0%) | .509 |
| Education level a ≤3, n (%) | 67 (47.6%) | 63 (42.0%) | .344 |
| ASA class b >2, n (%) | 1 (0.6%) | 0 (0.0%) | .492 |
| Comorbidity, n (%) | 31 (18.8%) | 36 (21.6%) | .530 |
| Preoperative NSAID, n (%) | 5 (3.0%) | 7 (4.2%) | .556 |
| Indication menorrhagia/metrorrhagia, n (%) | 73 (44.2%) | 81 (48.5%) | .436 |
| Uterine weight (g), median (IQR) | 242 (123–417) | 228 (104–430) | .593 |
| Planned AH, n (%) | 24 (14.5%) | 25 (15.0%) | .913 |
| Planned VH, n (%) | 26 (15.8%) | 28 (16.8%) | .803 |
| Planned LH/RH, n (%) | 115 (69.7%) | 114 (68.3%) | .778 |
| Preoperative hgb (g/dL), mean ± SD (range) | 8.3 ± 0.9 (5.2–10.5) | 8.2 ± 0.7 (4.9–10) | .444 |
a Education levels: 1–3 = ground school or short higher education; 4–5: long or medium-long higher education
b Anesthesia Physical Classification System classes: ASA1 = healthy patient; ASA2 = mild systemic disease; ASA3 = severe systemic disease; ASA4 = severe systemic life-threatening disease.
Both EBLs and EBLw were significantly reduced in the TA group compared to the placebo group (EBLs: 98.4 vs 134.8 mL, P = .006 and EBLw: 100 vs 166 mL, P = .004) ( Table 2 ). This corresponded to an absolute reduction of 27% concerning EBLs and 40% concerning EBLw. As the mean total blood loss was expected to differ between the 3 types of hysterectomy, a subgroup analysis was conducted. The results showed a significant reduction in the AH group and the LH/RH group. In the subgroup of women with menorrhagia/metrorrhagia as indication for hysterectomy, the mean total blood loss was correspondingly significantly reduced in the TA group compared with the placebo group. Data on EBLs contained two women with missing values. Forty-eight women had missing information on EBLw, 25 in the TA group and 23 in the placebo group. A total of 102 participants were calculated to have negative EBLw. A chi-square test showed no statistical difference between the number of censored participants in the 2 groups (54 vs 48 censored records in the TA and placebo group, respectively, P = .358). The mean EBLs of the corresponding censored EBLw values showed no difference between the TA and placebo group (53 vs 56 mL, P = .781). Figure 2 shows the correlation between the log-transformed data of EBLs and EBLw with exclusion of the negative EBLw values. The correlation coefficient was r = 0.78, P < .001.
| Blood loss estimated by the surgeon | |||
|---|---|---|---|
| Outcome | TA group n = 164 | Placebo group n = 166 | P |
| Total blood loss (mL) n = 330 | 98.4/100.0 (50–200) n = 164 | 134.8/150.0 (50–300) n = 166 | .006 |
| Total blood loss (mL), VH n = 54 | 74.9/75.0 (41–180) n = 26 | 81.3/95.0 (47–145) n = 28 | .757 |
| Total blood loss (mL), AH n = 49 | 204.2/216.5 (111–330) n = 24 | 331.1/350.0 (213–500) n = 25 | .031 |
| Total blood loss (mL), LH/RH n = 227 | 89.7/100.0 (50–200) n = 114 | 125.3/150.0 (50–300) n = 113 | .014 |
| Total blood loss (mL), indication menorrhagia/metrorrhagia, n = 152 | 94.9/100.0 (50–200) n = 72 | 150.1/155.0 (75–300) n = 80 | .005 |
| Blood loss estimated by weight | |||
|---|---|---|---|
| Outcome | TA group n = 86 | Placebo group n = 96 | P |
| Total blood loss (mL) n = 182 | 100.0/95.0 (52–251) n = 86 | 166.0/178.5 (74.5–397) n = 96 | .004 |
| Total blood loss (mL), VH n = 45 | 79.9/80.0 (44–180) n = 21 | 70.2/81.0 (47–135) n = 24 | .673 |
| Total blood loss (mL), AH n = 45 | 189.1/202.0 (100–333) n = 20 | 366.5/382.0 (213–617) n = 25 | .008 |
| Total blood loss (mL), LH/RH n = 92 | 83.7/82.0 (45–193) n = 45 | 168.9/178.0 (60–403) n = 47 | .006 |
| Total blood loss (mL), indication menorrhagia/metrorrhagia, n = 72 | 88.3/92.5 (45–225) n = 30 | 174.7/165.5 (68–479) n = 42 | .012 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
