Venous thromboembolism is an uncommon but potentially fatal complication of surgery.1,2 Assessing the risk of venous thromboembolism and providing appropriate prophylaxis are critical for surgical planning and affect everyday practice for the gynecologic surgeon.
General surgery literature has shown mode of surgery affects venous thromboembolism risk. Risk of venous thromboembolism for laparoscopy is decreased compared with open surgery.3,4 This is attributed to enhanced postoperative recovery in minimally invasive surgery, because patients ambulate earlier, are discharged sooner, and resume baseline activities faster. A small number of studies that assess the risk of venous thromboembolism among patients undergoing hysterectomy also observed a lower venous thromboembolism incidence among laparoscopic hysterectomies compared with open hysterectomies.5–11
Current venous thromboembolism risk assessment tools such as the Caprini Risk Assessment Model do not account for differences between minimally invasive surgery and laparotomy. Similarly, professional societies such as the American College of Obstetricians and Gynecologists12 or the American College of Chest Physicians13 have not yet incorporated mode of surgery into clinical practice guidelines for assessing venous thromboembolism risk or prescribing thromboprophylaxis.
The importance of identifying patients at risk for venous thromboembolism preoperatively cannot be understated because appropriate selection of thromboprophylaxis is based on each individual patient's risk profile. The aim of our study is to determine the 3-month postoperative incidence of venous thromboembolism among patients undergoing gynecologic surgery. Additionally, we seek to stratify the incidence of venous thromboembolism by mode of surgery and compare the risk of venous thromboembolism in laparotomy with minimally invasive approaches (laparoscopy and vaginal surgery).
MATERIALS AND METHODS
We conducted a retrospective cohort study of all gynecologic surgeries at Beth Israel Deaconess Medical Center and the University of Massachusetts Memorial Health Care. This study was approved by the institutional review boards at both institutions. International Classification of Diseases, 9th Revision (ICD-9) procedure codes assigned at discharge were used to identify all surgeries between May 2006 and June 2015 at Beth Israel Deaconess Medical Center and October 2007 through December 2014 at the University of Massachusetts Memorial Health Care (Appendix 1, available online at http://links.lww.com/AOG/B153). The ICD-9 procedure codes were used rather than Current Procedural Terminology (CPT) procedure codes because ICD-9 codes exist for both diagnoses and procedures, are more comprehensive than CPT codes, and avoid bundling of procedures often observed in CPT coding. Study periods were chosen as a result of data and resource availability.
At Beth Israel Deaconess Medical Center, ICD-9 procedure codes are assigned by “encounter.” For most patients, an encounter is the same as an admission; however, if the patient is readmitted to the hospital within 72 hours of discharge for related reasons, the second admission is combined with the original encounter. To maintain consistency between the two study sites, a similar encounter was created at of the University of Massachusetts Memorial Health Care.
Our study used the electronic medical record system maintained by the individual hospitals otherwise used for billing and hospital records. Trained hospitals’ medical coders are responsible for entering data into the system based on notes and dictations by clinicians caring for the patients. Throughout the course of the study, distributions of individual variables were checked before use to ensure face validity. For any questionable data points, coauthors reviewed individual medical records for clarification (H.-C.H., A.L., T.A.M.S.).
Mode of surgery was assigned based on the available ICD-9 procedure codes. If multiple ICD-9 procedure codes existed for each surgery or there were multiple surgeries during an encounter, the highest mode of incision was assigned using the following hierarchy: abdominal, laparoscopic, vaginal surgery, and other. The laparoscopic group encompasses conventional laparoscopy, robotic-assisted laparoscopy, and combined laparoscopic-assisted vaginal surgery. The vaginal group was the most heterogeneous ranging from minor cases such as hysteroscopy and dilation and curettage to major cases such as vaginal hysterectomy. Surgeries designated as “other” included cases such as intraperitoneal port removal, InterStim (Medtronic) placement, and perirectal abscess surgery.
A subset of surgeries, including myomectomies, could not be assigned a mode of surgery based on ICD-9 procedure codes alone because the ICD-9 codes were nonspecific. To ensure accurate assignments for myomectomies, the free-text descriptions of all procedures in the electronic surgical database were reviewed and used to assign mode of myomectomy. A manual chart review was performed for any cases requiring further clarification. At the University of Massachusetts Memorial Health Care, in addition to reviewing free-text descriptions, CPT codes (58140, 58146, 58145, 58545, and 58546) were used to assign the mode of surgery for myomectomies. This strategy was not used at Beth Israel Deaconess Medical Center given institutional differences in the use of CPT codes.
Postoperative venous thromboembolisms were defined as deep vein thromboses, pulmonary embolisms, or both that occurred within 90 days of the index surgery. Only clinically recognized cases that were treated with anticoagulation, vena cava filter, or both were included. Deep vein thromboses and pulmonary embolisms were identified using ICD-9 diagnosis codes (ICD-9 453.4–453.9, 415.1s). Deep vein thrombosis cases were limited to those presenting in the lower extremities. Thromboses in other locations such as ovarian vein thromboses or upper extremity venous thromboembolisms were excluded. In addition, chronic venous thromboembolisms (ICD-9 code 453.5) and preexisting venous thromboembolisms, that is, venous thromboembolisms present before the surgery date, were excluded. If a patient had more than one surgical encounter within a 90-day period of time and a venous thromboembolism occurred, it was attributed to highest mode of incision during those 90 days. All identified venous thromboembolisms were reviewed in the medical record by one of the coauthors (A.L.) for confirmation.
Patient demographics and clinical and surgical characteristics were retrieved with the electronic medical record including age, type of admission (inpatient, observation, outpatient), American Society of Anesthesiologists score, emergent surgery, use of thromboprophylaxis (mechanical, pharmacologic, dual prophylaxis), duration of surgery (incision to close, in minutes), length of hospital stay (in days), cancer diagnosis, and type of surgery. At Beth Israel Deaconess Medical Center, ICD-9 codes were used to identify diagnoses of gynecologic malignancies during the encounter. At the University of Massachusetts Memorial Health Care, the institutional tumor registry was accessed to identify these patients. All data were combined and subsequently deidentified.
The incidence of venous thromboembolism by mode of surgery was compared among all gynecologic surgeries. Continuous variables are presented as mean and SD or median and interquartile range and were compared using t tests or Wilcoxon rank-sum tests, as appropriate. Categorical variables were compared using the χ2 or Fisher exact test, as appropriate. Risk ratios and 95% CIs were calculated using log-binomial regression and generalized estimating equations with an independent correlation matrix to account for multiple encounters for the same patient. Risk ratios were adjusted for age at surgery, race and ethnicity, presence of cancer, presence of prophylaxis (mechanical or pharmacologic), American Society of Anesthesiologists score 3 or greater, surgical time, and admission status. Each covariate was adjusted for individually.
In addition, the incidence of venous thromboembolism by mode of surgery was analyzed within hysterectomies and myomectomies separately, because they are the two most common gynecologic surgeries performed by multiple modes of surgery.
Demographic and surgical variables of the study cohort are presented in Table 1. There were a total of 43,751 gynecologic surgical encounters (24,903 at Beth Israel Deaconess Medical Center and 18,848 at University of Massachusetts Memorial Health Care) among 37,485 individual patients during the study period. The majority underwent vaginal surgery (59.4%) such as dilation and curettage followed by laparoscopy (24.2%) and laparotomy (16.5%) (Table 2).
Among the 43,751 gynecologic surgical encounters, 8,273 were for hysterectomy with the following breakdown by mode of surgery: 56.8% laparotomy, 33.9% laparoscopy, and 9.3% vaginal approach (Table 2). The mode of hysterectomy shifted over the course of the study with a notable rise in laparoscopy and decline in laparotomy (Fig. 1A). Comparing the first year of the study with its final year, there was an increased proportion of laparoscopic hysterectomies (11.6% [n=31/268] to 55.0% [n=181/329], P<.001) and a decreased proportion of hysterectomies performed by laparotomy (74.3% [n=199/268] to 41.3% [n=136/329], P<.001) and vaginal approach (14.2% [n=38/268] to 3.7% [n=12/329], P<.001). Each institution contributed a comparable number of hysterectomies to the study cohort; (4,198 [50.7%] at Beth Israel Deaconess Medical Center vs 4,089 [49.4%] at the University of Massachusetts Memorial Health Care); however, institutional variations were noted for mode of hysterectomy (Appendix 2, available online at http://links.lww.com/AOG/B153).
There were 2,851 surgical encounters for myomectomy with the following breakdown by mode of surgery: 22.6% laparotomy, 15.2% laparoscopy, and 62.2% vaginal or hysteroscopic approach (Table 2). Over the course of the study period, laparoscopic myomectomy became more common, from 5.0% of all myomectomies in 2006 to 20.9% of all myomectomies in 2015. However, in 2014, there was a decrease in laparoscopic myomectomy and increase in abdominal myomectomy (Fig. 1B). One possible explanation for this phenomenon is that this is the same year that the U.S. Food and Drug Administration made their safety communication regarding electric power morcellation, because other groups have also noted an increase in abdominal myomectomies in 2014 in response to the U.S. Food and Drug Administration safety communication.14,15 Overall, the number of myomectomies from each institution was similar (1,432 [50.2%] at Beth Israel Deaconess Medical Center vs 1,419 [49.8%] at the University of Massachusetts Memorial Health Care); however, the breakdown by mode of surgery varied by institution (Appendix 2, http://links.lww.com/AOG/B153).
A total of 96 patients were diagnosed with postoperative venous thromboembolism among all gynecologic surgeries, resulting in a 0.2% overall venous thromboembolism incidence (Table 2). The venous thromboembolism incidence was 1.1% among patients with cancer and 0.2% in patients who underwent gynecologic surgery for benign indications (Tables 3 and 4). The greatest incidence of venous thromboembolism was among patients with cancer undergoing laparotomy (2.2%).
Among all patients undergoing gynecologic surgery, those with venous thromboembolism were more likely to have undergone inpatient surgery rather than outpatient surgery, with admissions longer than 24 hours in 85.4% of patients with venous thromboembolism compared with 30.4% of patients without venous thromboembolism (P<.001). Additionally, patients with venous thromboembolism were older (average age 55.7 vs 44.2 years, P<.001), more likely to have cancer (30.2% vs 5.8%, P<.001), and American Society of Anesthesiologists score 3 or greater (40.6% vs 9.3%, P<.001). The racial makeup of the venous thromboembolism and nonvenous thromboembolism groups was different (P=.01) with 19.8% of the venous thromboembolism group identifying as black compared with 10.3% of patients without venous thromboembolism. More patients in the venous thromboembolism group had a longer total surgical time compared with patients without venous thromboembolism (P<.001), with patients with venous thromboembolism having fewer surgeries shorter than than 30 minutes (8.3% vs 45.5%) and more surgeries longer than 3 hours (34.4% vs 7.7%) (Table 5).
The majority of patients with venous thromboembolism received some form of thromboprophylaxis before surgery. Among the venous thromboembolism group, 90.6% received mechanical thromboprophylaxis, pharmacologic thromboprophylaxis, or both compared with the nonvenous thromboembolism group in which only 55.0% received some form of thromboprophylaxis (P<.001). Of all gynecologic patients in our cohort, 3,112 (7.1%), received pharmacologic prophylaxis with or without mechanical prophylaxis. Among the patients with venous thromboembolism, 77 (80.2%) received pharmacologic prophylaxis with or without mechanical prophylaxis compared with 7.0% of patients who did not have a venous thromboembolism (Table 5). The high rate of prophylaxis among patients who developed postoperative venous thromboembolism may reflect surgeons' abilities to preoperatively identify patients at increased risk, guiding appropriate selection of thromboprophylaxis.
Patients who underwent minimally invasive gynecologic surgery (laparoscopic or vaginal) were less likely to develop venous thromboembolism than patients who had a laparotomy. Compared with laparotomy, laparoscopic surgery was associated with a 78% lower risk of venous thromboembolism (risk ratio [RR] of 0.22, 95% CI 0.13–0.37) and vaginal surgery was associated with a 93% lower risk of venous thromboembolism (RR 0.07, 95% CI 0.04–0.12). This association persisted even after adjusting the data for known venous thromboembolism risk factors such as age, race, cancer, use of pharmacologic thromboprophylaxis, admission status, surgical time, and American Society of Anesthesiologists score (Table 6).
Among hysterectomies, there were 55 cases of venous thromboembolism with a 0.7% venous thromboembolism incidence in this cohort (Table 2). The incidence of venous thromboembolism differed by mode of surgery (P<.001) and was highest among patients undergoing hysterectomy who underwent laparotomy (1.0%) followed by laparoscopic hysterectomy (0.3%) and vaginal hysterectomy (0.1%).
None of the patients who underwent outpatient hysterectomy (n=427) developed postoperative venous thromboembolism. All patients with venous thromboembolism after hysterectomy had an inpatient admission with greater than 24 hours’ stay. The venous thromboembolism incidence was 1.8% among patients undergoing hysterectomy with cancer and 0.4% among patients who underwent hysterectomy for benign indications. Patients who developed venous thromboembolism compared with no venous thromboembolism after hysterectomy were more likely to be older (average age 60.3 years vs 53.1 years, P<.001) and have cancer (47.3% vs 17.5%, P<.001) or concomitant medical comorbidity (American Society of Anesthesiologists score 3 or greater 36.4% vs 16.1%, P<.001). Surgical time was not different between patients undergoing hysterectomy with or without venous thromboembolism (P=.65) (Table 5).
Laparoscopic hysterectomy was associated with 71% lower risk of venous thromboembolism (RR 0.29, 95% CI 0.14–0.62) and vaginal hysterectomy was associated with 87% lower risk of venous thromboembolism (RR 0.13, 95% CI 0.02–0.96) compared with patients who underwent abdominal hysterectomy. The association of minimally invasive hysterectomy (laparoscopic and vaginal) with decreased risk of venous thromboembolism persisted when the data were adjusted for surgical time and age (Table 6). Furthermore, when adjusting for other covariates such as race, cancer, thromboprophylaxis, and American Society of Anesthesiologists score 3 or greater, laparoscopic hysterectomy compared with laparotomy was associated with a lower risk of venous thromboembolism (race RR 0.32, 95% CI 0.15–0.67; cancer RR 0.33, 95% CI 0.16–0.70; thromboprophylaxis RR 0.23, 95% CI 0.11–0.49; American Society of Anesthesiologists score 3 or greater RR 0.29, 95% CI 0.13–0.60). Although vaginal hysterectomy was also associated with lower risk of venous thromboembolism compared with laparotomy when adjusting for these covariates, the findings were not significant for vaginal hysterectomy, because there was only one patient in our cohort who had venous thromboembolism after vaginal hysterectomy (Table 6).
There were five cases of venous thromboembolism among all patients who underwent myomectomy for a 0.2% venous thromboembolism incidence in the myomectomy cohort (Table 2). None of the venous thromboembolism cases were among patients with a cancer diagnosis or those who had undergone a vaginal or hysteroscopic myomectomy. Given the small numbers, there was insufficient data to perform any other analyses within the myomectomy group.
We found the overall incidence of venous thromboembolism among patients undergoing gynecologic surgery is low and varies by mode of surgery. The effect of mode of surgery persisted when data were corrected for other known venous thromboembolism risk factors such as age, race, cancer, thromboprophylaxis, American Society of Anesthesiologists score, and length of surgery.
The American College of Obstetricians and Gynecologists’ Practice Bulletin on prevention of venous thromboembolism stratifies risk of venous thromboembolism by length of surgery, but not by mode of surgery.12 The American College of Chest Physicians recognized the effect of mode of incision with the 2008 update to their Clinical Practice Guidelines; however, this approach was revised in their 2012 update in which mode of surgery was no longer included and instead the choice of prophylaxis was determined by estimating preoperative venous thromboembolism risk stratification based on Caprini score.13 Moreover, the Caprini score has not been specifically validated in a gynecologic population.13
Currently, most clinicians use the Caprini score and guidelines set forth by the American College of Chest Physicians and the American College of Obstetricians and Gynecologists to assess a patient's venous thromboembolism risk profile and select thromboprophylaxis. These guidelines stratify venous thromboembolism risk based on patient characteristics, but do not differentiate by mode of surgery.12,13,16 Thus, many surgeons apply the same parameters for selection of venous thromboembolism prophylaxis for laparoscopic or vaginal surgery as they do for laparotomy, although patients undergoing minimally invasive surgery ambulate earlier and resume baseline activities sooner, which presumably diminishes one's risk of postoperative venous thromboembolism. However, any benefit from a minimally invasive approach could be offset by increased length of surgery during laparoscopy or robotic surgery.5,17,18 Moreover, some known risk factors for venous thromboembolism such as obesity or cancer are also associated with an increased likelihood of laparotomy over minimally invasive surgery.18 The true effect of each facet of laparoscopy in gynecologic surgery on venous thromboembolism risk remains unknown.
In our study, nearly all (90.6%) patients who developed postoperative venous thromboembolism had received some form of thromboprophylaxis (mechanical, pharmacologic, or both) with 80.2% (n=77) having received pharmacologic thromboprophylaxis. Our findings are consistent with the findings of other groups,7 suggesting that gynecologic surgeons may be accurately predicting patients at risk for venous thromboembolism. However, 7% (n=3,035) of patients without venous thromboembolism also received pharmacologic thromboprophylaxis; low-risk patients undergoing low-risk surgery may be unnecessarily receiving a potentially high-morbidity medication. Use of pharmacologic thromboprophylaxis is associated with increased intraoperative blood loss, postoperative hemorrhage or hematoma, transfusion, and reoperation.7,10,19 To limit overuse of this high-morbidity medication, there is a need for more tailored recommendations for venous thromboembolism prevention with room for improvement in the currently available guidelines.20
Strengths of this study include the large number of surgical encounters (N=43,751) and long 9-year study period. This study includes venous thromboembolism incidence and risk factors for not only hysterectomy, but also myomectomy and gynecologic surgery in general, for both benign and malignant indications. Prior studies have used databases to capture large case volumes because of the low incidence of venous thromboembolism. Our study assesses venous thromboembolism incidence using local data from two large tertiary care centers with the ability to directly perform chart review. As a result, we were able to assess the accuracy of the collected data and extend the follow-up period to 3 months, whereas prior studies using national databases could assess venous thromboembolism incidence for only 1 or 2 months postoperatively given the limitations and set parameters of the databases.5–7,9 Our data are generalizable; changes in surgical approach over the study period were consistent with national trends.21
Limitations of this study include its retrospective design and reliance on electronic medical record systems with inherent limitations of electronic data mining. An example of this is our inability to capture body mass index. Furthermore, we are unable to capture patients who went to other hospitals for their postoperative venous thromboembolism diagnosis and treatment. However, our extended 3-month postoperative period does increase the opportunity to detect patients with venous thromboembolism complications diagnosed at outside hospitals because this allows more time for patients with venous thromboembolism to schedule a visit to see their health care provider. Other limitations include the small number of myomectomies. Although we assess the incidence of venous thromboembolism in patients undergoing myomectomy, we remain underpowered to draw further conclusions about risk factors given the rarity of this surgical complication despite the large multicenter study design. Additionally, despite our large sample, because of the rarity of venous thromboembolism, we were unable to construct a statistical model to simultaneously adjust for multiple factors.
In conclusion, this study adds to data demonstrating that venous thromboembolism is rare in gynecologic surgery, particularly when a patient undergoes a minimally invasive procedure for benign indications. Because current practice guidelines do not account for mode of surgery, we find them to be insufficient for the modern gynecologic surgeon to counsel patients on their individual venous thromboembolism risk or to make ideal decisions regarding selection of thromboprophylaxis. We call for further research on the role of gynecologic surgery in risk of venous thromboembolism with the goal of fine-tuning venous thromboembolism prediction models and tools for surgeons to reflect emerging data showing that a minimally invasive approach is associated with a decreased risk of venous thromboembolism.
1. Geerts WH, Bergqvist D, Pineo GF, Heit JA, Samama CM, Lassen MR, et al. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133(suppl):381–453S.
2. Clarke-Pearson DL, DeLong ER, Synan IS, Coleman RE, Creasman WT. Variables associated with postoperative deep venous thrombosis: a prospective study of 411 gynecology patients and creation of a prognostic model. Obstet Gynecol 1987;69:146–50.
3. Nguyen NT, Hinojosa MW, Fayad C, Varela E, Konyalian V, Stamos MJ, et al. Laparoscopic surgery is associated with a lower incidence of venous thromboembolism compared with open surgery. Ann Surg 2007;246:1021–7.
4. Shapiro R, Vogel JD, Kiran RP. Risk of postoperative venous thromboembolism after laparoscopic and open colorectal surgery: an additional benefit of the minimally invasive approach? Dis Colon Rectum 2011;54:1496–502.
5. Barber EL, Neubauer NL, Gossett DR. Risk of venous thromboembolism in abdominal versus minimally invasive hysterectomy for benign conditions. Am J Obstet Gynecol 2015;212:609.e1–7.
6. Barber EL, Gehrig PA, Clarke-Pearson DL. Venous thromboembolism in minimally invasive compared with open hysterectomy for endometrial cancer. Obstet Gynecol 2016;128:121–6.
7. Swenson CW, Berger MB, Kamdar NS, Campbell DA Jr, Morgan DM. Risk factors for venous thromboembolism after hysterectomy. Obstet Gynecol 2015;125:1139–44.
8. Kahr HS, Thorlacius-Ussing O, Christiansen OB, Skals RK, Torp-Pedersen C, Knudsen A. Venous thromboembolic complications to hysterectomy for benign disease: a nationwide cohort study. J Minim Invasive Gynecol 2018;25:715–23.e2.
9. Ritch JM, Kim JH, Lewin SN, Burke WM, Sun X, Herzog TJ, et al. Venous thromboembolism and use of prophylaxis among women undergoing laparoscopic hysterectomy. Obstet Gynecol 2011;117:1367–74.
10. Brummer TH, Heikkinen A, Jalkanen J, Fraser J, Mäkinen J, Tomás E, et al. Pharmaceutical thrombosis prophylaxis, bleeding complications and thromboembolism in a national cohort of hysterectomy for benign disease. Hum Reprod 2012;27:1628–36.
11. Wallace SK, Fazzari MJ, Chen H, Cliby WA, Chalas E. Outcomes and postoperative complications after hysterectomies performed for benign compared with malignant indications. Obstet Gynecol 2016;128:467–75.
12. Prevention of deep vein thrombosis and pulmonary embolism. ACOG Practice Bulletin No. 84. American College of Obstetricians and Gynecologists. Obstet Gynecol 2007;110:429–40.
13. Gould MK, Garcia DA, Wren SM, Karanicolas PJ, Arcelus JI, Heit JA, et al. Prevention of VTE in nonorthopedic surgical patients: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141(suppl):e227–77S.
14. Stentz NC, Cooney LG, Sammel M, Shah DK. Changes in myomectomy practice after the U.S. Food and Drug Administration safety communication on power morcellation. Obstet Gynecol 2017;129:1007–13.
15. Lum DA, Sokol ER, Berek JS, Schulkin J, Chen L, McElwain CA, et al. Impact of the 2014 Food and Drug Administration warnings against power morcellation. J Minim Invasive Gynecol 2016;23:548–56.
16. Caprini JA, Arcelus JI, Hasty JH, Tamhane AC, Fabrega F. Clinical assessment of venous thromboembolic risk in surgical patients. Semin Thromb Hemost 1991;17(suppl 3):304–12.
17. Nezhat C, Lavie O, Hsu S, Watson J, Barnett O, Lemyre M. Robotic-assisted laparoscopic myomectomy compared with standard laparoscopic myomectomy—a retrospective matched control study. Fertil Steril 2009;91:556–9.
18. Shah DK, Vitonis AF, Missmer SA. Association of body mass index and morbidity after abdominal, vaginal, and laparoscopic hysterectomy. Obstet Gynecol 2015;125:589–98.
19. Dahl OE, Quinlan DJ, Bergqvist D, Eikelboom JW. A critical appraisal of bleeding events reported in venous thromboembolism prevention trials of patients undergoing hip and knee arthroplasty. J Thromb Haemost 2010;8:1966–75.
20. Sandadi S, Lee S, Walter A, Gardner GJ, Abu-Rustum NR, Sonoda Y, et al. Incidence of venous thromboembolism after minimally invasive surgery in patients with newly diagnosed endometrial cancer. Obstet Gynecol 2012;120:1077–83.
21. Morgan DM, Kamdar NS, Swenson CW, Kobernik EK, Sammarco AG, Nallamothu B. Nationwide trends in the utilization of and payments for hysterectomy in the United States among commercially insured women. Am J Obstet Gynecol 2018;218:425.e1–18.
Supplemental Digital Content
© 2018 by the American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.