Pulmonary embolism is a potentially preventable perioperative complication. In a recent cohort study of surgical, urologic, and gynecologic oncology patients, the number one cause of postoperative death was fatal pulmonary embolism.1 Although there is literature demonstrating that prophylaxis does decrease the incidence of venous thromboembolism (pulmonary embolism and/or deep venous thrombosis) in gynecologic cancer patients,2–5 the incidence of pulmonary embolism with single-agent prophylaxis remains between 1.2% and 4%,2,6–9 and as high as 6.8% in patients with ovarian cancer.9 There is no consensus among gynecologic oncologists as to the optimal regimen for pulmonary embolism prevention. In a 2006 survey of the Society of Gynecologic Oncologists membership, 22–41% of respondents reported using sequential compression devices alone, whereas at least 42% employed dual prophylaxis with an anticoagulant and sequential compression devices.10
Dual-agent prophylaxis has been shown to be superior to single-agent prophylaxis in neurosurgery,11 orthopedic surgery,12 and general surgery.13 Extended prophylaxis with enoxaparin has been shown to decrease venous thromboembolism in surgical oncology patients (of which a small percentage were gynecologic) when compared with no prophylaxis after hospital discharge.14 In one study of cancer patients, up to 40% of clots were diagnosed more than 21 days after surgery.1 In 2004, the American Association of Chest Physicians (ACCP) published guidelines for venous thromboembolism prophylaxis. One of the regimens supported by the ACCP is combined prophylaxis in all gynecologic cancer patients, and prolonged prophylaxis for 2–4 weeks after hospital discharge in patients with malignancy who are over the age of 60 years or have a history of prior venous thromboembolism.4
In 2005, the surgeons in the Division of Gynecologic Oncology met and discussed a plan to improve surgical care by implementing a dual prophylaxis regimen for all patients undergoing major surgery based on one of the regimens proposed in the published ACCP guidelines.4 The primary objective of this quality improvement project was to examine the feasibility of implementing a perioperative venous thromboembolism prevention protocol of dual prophylaxis with sequential compression devices plus heparin or low molecular weight heparin in all patients, with 2 weeks of extended prophylaxis after hospital discharge in patients meeting high-risk criteria. The secondary objective was to evaluate the effect of the protocol on the incidence of venous thromboembolism, bleeding complications, and heparin-induced thrombocytopenia in patients undergoing major surgery for presumed gynecologic cancer.
MATERIALS AND METHODS
On January 1, 2006, we initiated an evidenced-based protocol mandating all patients admitted to the gynecologic cancer service for major surgery at the University of Wisconsin Hospital have dual venous thromboembolism prophylaxis with sequential compression devices and an anticoagulant (Fig. 1).4 Previously, all patients were given sequential compression devices starting before the induction of anesthesia and continuing until discharge from the hospital. Low molecular weight heparin or heparin was added at the discretion of the faculty surgeon. The preprinted gynecologic oncology order sheets were changed to reflect the new venous thromboembolism prevention protocol. One to 2 hours before induction of anesthesia, all patients were given 5,000 units of heparin subcutaneously; in patients receiving an epidural, heparin was administered after epidural placement. Sequential compression devices were used as sole prophylaxis in only those patients determined to have a contraindication to anticoagulation: active bleeding, uncorrected coagulopathy, patients on thrombolytic medications, ocular or neurologic surgery within 10 days, thrombocytopenia, or recent hemorrhagic stroke. Patients who had a history of heparin-induced thrombocytopenia received fondaparinux, a substitution that was based on the expert opinion of a hematologist colleague.
Postoperatively, all patients continued sequential compression devices until discharge. Patients who had benign intraoperative findings received heparin 5,000 units every 12 hours, starting 12 hours after the preoperative dose and continuing at least until hospital discharge. Patients with malignancy received heparin 5,000 units every 8 hours, starting 8 hours after the preoperative dose and continuing at least until hospital discharge. The initiation of postoperative heparin or low molecular weight heparin could be delayed at the surgeon's discretion in patients who were at extremely high risk for postoperative bleeding. Serial hematocrits were drawn in those patients overnight, and if stable, anticoagulant prophylaxis was started the morning of postoperative day one. Platelet counts and hematocrit were drawn at regular intervals in all patients on anticoagulation throughout hospitalization. The preprinted order set allowed the pharmacist to substitute low molecular weight heparin (dalteparin 5,000 units subcutaneously or enoxaparin 40 mg subcutaneously daily) for heparin after making the team aware. To make the substitution, patients needed to have not had an epidural for the last 12 hours, and in patients who were going home on anticoagulation, insurance coverage was evaluated.
Patients who met high-risk criteria went home on extended duration prophylaxis with three times daily heparin or daily low molecular weight heparin for 2 weeks after discharge. Women with malignancy who were either over the age of 60 years or had a history of prior venous thromboembolism were considered high risk based on the ACCP guidelines.4 Patients were taught to self-administer injections by nurses on the floor. Weekly hematocrit and platelet counts were obtained during the 2 weeks after discharge.
Patients were seen and examined daily in the hospital. No other screening for venous thromboembolism was performed. Patients with possible signs or symptoms of deep vein thrombosis (DVT), including lower extremity edema, tenderness, or palpable masses, were evaluated with lower extremity Doppler imaging. Patients with otherwise unexplained tachypnea, dyspnea, tachycardia, or oxygen desaturation were evaluated for pulmonary embolism using chest computed topography. In patients who could not have contrast, ventilation-perfusion scan was performed.
This study was evaluated by the University of Wisconsin Institution Review Board and considered exempt. Cases were reviewed to compare the incidence of venous thromboembolism in patients admitted to the gynecologic cancer service for major surgery in 2005 compared with 2006 (the protocol was implemented January 1, 2006). Patients who were actively being treated for a pulmonary embolism or DVT, had a history of heparin-induced thrombocytopenia, and/or were already on anticoagulation for another purpose (such as valvular disease or an inferior vena cava filter) at the time of admission were excluded from this analysis. The protocol did not apply to patients who were on other services but were operated on by a gynecologic oncologist or those who had minor procedures, including port placement; therefore, they were also excluded from this analysis. If patients were admitted more than once during the study period, only data from the first admission and the ensuing 6-week follow-up period were included.
Records of all patients who fit inclusion criteria in 2005 and 2006 were reviewed for baseline demographics and perioperative details. Any DVT or pulmonary embolism diagnosed within 6 weeks of surgery was identified through review of the electronic medical record. The primary outcome was compliance with the protocol, as measured by 90% of patients receiving an anticoagulant in the hospital and 90% of those who should have gone home on an anticoagulant having one prescribed in the year 2006. Secondary endpoints included venous thromboembolism, pulmonary embolism, and DVT incidence, as well as the rate of bleeding complications and heparin-induced thrombocytopenia.
Continuous outcomes were compared between groups using the Wilcoxon two-sample rank test, and categorical outcomes were compared using the χ2 for association. Incidence of venous thromboembolism, pulmonary embolism, and DVT was compared between groups using logistic regression, both in univariable models and in multivariable models, including potential baseline confounders as covariates. The multivariable logistic regressions given in Table 3 were checked for validity by performing exact versions of them.15 The two sets of results were nearly identical. A multivariable analysis for DVT was not performed due to the small numbers of events. The result of a two-tailed test at the 5% level was regarded as significant.
In 2005 and 2006, a total of 644 patients were admitted to the gynecologic oncology service for major surgery. Thirty-nine patients were excluded from this analysis: two had an active thrombus, 35 were on anticoagulation for another purpose, and two had a prior diagnosis of heparin-induced thrombocytopenia. This left a total of 605 evaluable patients, 294 in 2005 and 311 in 2006. Patients were followed until 6 weeks postoperatively or the closest visit to that interval. The median follow-up in 2005 was 32 (interquartile range [IQR] 22–45) days and in 2006 was 30 days (IQR 21–39) (P=.04).
Patient demographics were similar between the two years (Table 1). The number of laparoscopic surgeries increased from 2005 (7.4%) to 2006 (14.8%, P=.01), whereas the median operative time decreased from 3.5 hours (IQR 2.8–4.3) in 2005 to 2.9 hours (IQR 2.2–3.9) in 2006 (P<.001) (Table 2). The median hospitalization time also decreased, from 4 days (IQR 3–6) in 2005 to 3 days (IQR 2–4) in 2006 (P<.001) (Table 2).
Adherence to the protocol was determined by the percentage of patients who received an anticoagulant in the hospital and the percentage of women who went home on an anticoagulant of those who met criteria for prolonged prophylaxis. The target for both was 90%. Of the 311 patients who met inclusion criteria in 2006, 305 (98.1%) received an anticoagulant while in the hospital. The initiation of postoperative heparin or low molecular weight heparin could be delayed until the first postoperative day at the surgeon's discretion in patients who were at extremely high risk of postoperative bleeding. This occurred relatively infrequently—in only 22 of 305 patients who received an anticoagulant (7.2%) was the first postoperative dose delayed. In 2005, 153 (52%) of patients received heparin or low molecular weight heparin at some point during their hospitalization, which is significantly less than in 2006 (P<.001). The schedule and dosing of anticoagulant prophylaxis when given in 2005 was not consistent. Of the 153 patients who received an anticoagulant in 2005, 59 (38.6%) had the postoperative dose started the day after surgery or later, as compared with 7.2% in 2006 (P<.001). Most of the patients who got prophylaxis had it continued until the time of discharge only.
Of the 90 women in 2006 meeting criteria to go home on prolonged prophylaxis, 91.1% were prescribed an anticoagulant at the time of discharge. Ten patients were sent home on anticoagulation who did not meet criteria. They were all patients with malignancy aged younger than 61 years, including four women who were aged 60 years and may represent a misinterpretation of the age threshold on the preprinted orders by the ordering physician. None of the patients had bleeding complications.
Dual prophylaxis with three-times-daily heparin (or daily low molecular weight heparin) followed by prolonged prophylaxis in patients who met high-risk criteria resulted in a significant decrease in the incidence of venous thromboembolism from 6.5% (19 of 294) in 2005 to 1.9% (6 of 311) in 2006 on univariable analysis (odds ratio 0.29, 95% confidence interval 0.11–0.72). In 2005 there were 14 pulmonary embolism and seven DVT, and in 2006 there were four pulmonary embolism and two DVT. We performed a multivariable analysis based on the differences between the two groups at baseline: operative time, days of hospitalization, and laparoscopic surgery. The decrease in venous thromboembolism remained significant, with the incidence in 2006 being one third of that in 2005 (odds ratio 0.33, 95% confidence interval 0.12–0.88) (Table 3).
The median estimated blood loss in 2005 was 250 mL, and was 200 mL in 2006 (P=.22). Forty-one patients (13.9%) received blood transfusions in 2006, and 48 (15.4%) received blood transfusions in 2005 (P=.69). The median number of units transfused was two in both years (P=.25). Two patients (0.7%) were reoperated on for bleeding in 2005, and three (1%) were reoperated on in 2006 (P=.95). Of the 305 women who received an anticoagulant in the hospital, 11 (3.6%) had the anticoagulant stopped before discharge from the hospital due to bleeding concerns. There was no difference in the number of patients who received ketorolac (Toradol, Roche Laboratories, Nutley, NJ) in 2005 and 2006 (48% compared with 43.7%, P=.34). No patient in either group was diagnosed with heparin-induced thrombocytopenia during the follow-up period.
A quality improvement protocol of dual prophylaxis, with extended duration prophylaxis in high-risk patients, was successfully implemented and was associated with a significant decrease in venous thromboembolism in patients undergoing surgery for gynecologic cancer. The 70% reduction in the incidence of venous thromboembolism is statistically and clinically significant. This translates into a number needed to treat of 22, and given the fact that no increase in bleeding complications were seen in 2006, we feel that the favorable risk/benefit ratio supports use of this protocol.
Venous thromboembolism has been identified as the number one potentially preventable complication by the Agency for Healthcare Research and Quality (AHRQ).16 It is also one of four areas of focus for the Surgical Care Improvement Project,17 which is run by Medicare and has the goal of decreasing operative complications by 25% by the year 2010. This is reflective of the national movement toward standardization of venous thromboembolism prevention regimens to promote quality of care and minimize iatrogenic error.
Our intention was to create an evidence-based protocol that decreased venous thromboembolism, encouraged compliance, and could be transferable to other surgical services. We based our protocol on a literature review and the ACCP recommendations, which treat heparin and low molecular weight heparin as interchangeable.4 We expected that once-daily dosing would increase compliance, which was especially relevant in patients receiving prolonged prophylaxis.
The success of the implementation of this protocol can be attributed at least in part to its multidisciplinary nature. Attending physicians, residents, nurses, and pharmacists worked in concert to ensure compliance and monitor safety. A preprinted order set, developed collaboratively and approved by a multidisciplinary committee, simplified the process. The University of Wisconsin Hospital is transitioning to electronic physician order entry, and there may be a role for automatic prompts to alert physicians to patients who meet criteria for prolonged prophylaxis. Protocols, specifically, preprinted order sets, have been shown to decrease medication errors18,19 and have been espoused by regulatory agencies, including the AHRQ20 and the Institute of Medicine.
The change in venous thromboembolism seen between 2005 and 2006 could be multifactorial. The rate of pulmonary embolism seen in our institution in 2005 (4.8%) was at the higher end of the range that might be expected in patients receiving single-agent prophylaxis based on the literature.2,7,9,21 This may be because compliance with sequential compression devices is imperfect, making this method less effective than it is when employed in a prospective trial, where compliance is monitored closely. The criteria one uses to prompt evaluation with computed topography or ventilation/perfusion scans22 can vary between different clinicians. Our practices did not change between 2005 and 2006 in this regard, so we do not expect that the drop in pulmonary embolism is related to a different threshold for evaluation, although it may explain the discrepancy in pulmonary embolism incidence seen in our study as compared with other authors.
We cannot say conclusively what decreased the rate of venous thromboembolism: implementing a widespread protocol, the use of preoperative heparin, the use of dual prophylaxis, the use of prolonged prophylaxis in high-risk patients, or a combination of these factors. Dual prophylaxis has been shown to be beneficial when compared with single-agent prophylaxis in other surgical specialties, including neurosurgery,11 orthopedic surgery,12 and general surgery.13 Prolonged prophylaxis has been shown to be of benefit in a large prospective trial of surgical oncology patients who got enoxaparin daily for 4 weeks after discharge,14 although it has not been studied specifically in gynecologic cancer patients. Another explanation for the significant decrease in venous thromboembolism seen in 2006 relative to 2005 could be the dosing schedule used in 2006. Heparin three times daily has been shown to decrease venous thromboembolism in gynecologic cancer patients,3 whereas twice-daily dosing is ineffective.6 Both schedules were employed in 2005 when heparin was administered. Also, most studies of anticoagulation in gynecologic cancer patients include a preoperative dose, which we incorporated into the protocol.4,23 This was rarely being done in 2005.
Although the use of anticoagulants was significantly higher in 2006 than 2005, 52% of patients got some form of anticoagulant in 2005. The fact that venous thromboembolism decreased in 2006 despite what was likely the use of anticoagulant with sequential compression devices in higher-risk patients in 2005, seems to support the superiority of a universal thromboprophylaxis regimen over stratification based on surgeon preference.
A strength of this study is its applicability to the debate over venous thromboembolism prophylaxis, because it is a real-world look at the efficacy of dual prophylaxis and prolonged prophylaxis in high-risk populations. The University of Wisconsin is reflective of a traditional university hospital setting, which makes it likely that the protocol could be transferred to another academic institution. The one exception to this generalization is that pharmacists at the University of Wisconsin are located on each patient floor, and every patient has both an admission and discharge interview with a pharmacist to review medications. The main weakness of this study is its retrospective design. There may have been other differences between the two years that we did not identify or measure that could account for the reduction in venous thromboembolism. We used logistic regression to control for known confounders, but others may exist. Also, while over 97% of patients had a postoperative visit at UW and the median follow up time was 32 days in 2006 and 30 days in 2005, it is possible that an event could have been missed if it occurred at a local facility after the postoperative visit.
We have demonstrated a significant decrease in the incidence of venous thromboembolism in patients undergoing major gynecologic cancer surgery associated with a universal protocol of dual prophylaxis and prolonged prophylaxis in high-risk patients. Given the low incidence of associated complications and the potential dangers associated with venous thromboembolism, we advocate that consideration be given to implementing this standard prophylaxis protocol.
1. Agnelli G, Bolis G, Capussotti L, Scarpa RM, Tonelli F, Bonizzoni E, et al. A clinical outcome-based prospective study on venous thromboembolism after cancer surgery: the @RISTOS project. Ann Surg 2006;243:89–95.
2. Clarke-Pearson DL, Synan IS, Hinshaw WM, Coleman RE, Creasman WT. Prevention of postoperative venous thromboembolism by external pneumatic calf compression in patients with gynecologic malignancy. Obstet Gynecol 1984;63:92–8.
3. Clark-Pearson DL, DeLong E, Synan IS, Soper JT, Creasman WT, Coleman RE. A controlled trial of two low-dose heparin regimens for the prevention of postoperative deep vein thrombosis. Obstet Gynecol 1990;75:684–9.
4. Geerts WH, Pineo GF, Heit JA, Bergqvist D, Lassen MR, Colwell CW, et al. Prevention of venous thromboembolism: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004;126 suppl:338S–400S.
5. Einstein MH, Hartenbach EM. Venous thromboembolism prophylaxis: is two better than one? Gynecol Oncol 2007;106:437–8.
6. Clarke-Pearson DL, Coleman RE, Synan IS, Hinshaw W, Creasman WT. Venous thromboembolism prophylaxis in gynecologic oncology: a prospective, controlled trial of low-dose heparin. Am J Obstet Gynecol 1983;145:606–13.
7. Clarke-Pearson DL, Dodge RK, Synan I, McClelland RC, Maxwell GL. Venous thromboembolism prophylaxis: patients at high risk to fail intermittent pneumatic compression. Obstet Gynecol 2003;101:157–63.
8. Clarke-Pearson DL, Synan IS, Creasman WT. Anticoagulation therapy for venous thromboembolism in patients with gynecologic malignancy. Am J Obstet Gynecol 1983;147:369–75.
9. Martino MA, Borges E, Williamson E, Siegfried S, Cantor AB, Lancaster J, et al. Pulmonary embolism after major abdominal surgery in gynecologic oncology. Obstet Gynecol 2006;107:666–71.
10. Martino MA, Williamson E, Rajaram L, Lancaster JM, Hoffman MS, Maxwell GL, et al. Defining practice patterns in gynecologic oncology to prevent pulmonary embolism and deep venous thrombosis. Gynecol Oncol 2007;106:439–45.
11. Agnelli G, Piovella F, Buoncristiani P, Severi P, Pini M, D'Angelo A, et al. Enoxaparin plus compression stockings compared with compression stockings alone in the prevention of venous thromboembolism after elective neurosurgery. N Engl J Med 1998;339:80–5.
12. Eisele R, Kinzl L, Koelsch T. Rapid-inflation intermittent pneumatic compression for prevention of deep venous thrombosis. J Bone Joint Surg Am 2007;89:1050–6.
13. Wille-Jorgensen P, Rasmussen MS, Andersen BR, Borly L. Heparins and mechanical methods for thromboprophylaxis in colorectal surgery. The Cochrane Database of Systematic Reviews 2003,Issue 4. Art. No.: CD001217. DOI: 10.1002/14651858.CD001217.
14. Bergqvist D, Agnelli G, Cohen AT, Eldor A, Nilsson PE, Le Moigne-Amrani A, et al. Duration of prophylaxis against venous thromboembolism with enoxaparin after surgery for cancer. N Engl J Med 2002;346:975–80.
15. Hirji KF, Mehta CR, Patel NR. Computing distributions for exact logistic regression. J Am Stat Assoc 1987;82:1110–7.
16. Shojania KG, Duncan BW, McDonald KM, Wachter RM, Markowitz AJ. Making health care safer: a critical analysis of patient safety practices. Evid Rep Technol Assess (Summ) 2001;(43):i-x, 1–668.
18. Donihi AC, DiNardo MM, DeVita MA, Korytkowski MT. Use of a standardized protocol to decrease medication errors and adverse events related to sliding scale insulin. Qual Saf Health Care 2006;15:89–91.
19. Kozer E, Scolnik D, MacPherson A, Rauchwerger D, Koren G. Using a preprinted order sheet to reduce prescription errors in a pediatric emergency department: a randomized, controlled trial. Pediatrics 2005;116:1299–302.
20. United States Department of Health and Human Services. Agency for Healthcare Research and Quality. AHRQ Patient Safety Network. Kozer E, Scolnik D, MacPherson A, Rauchwerger DG. Using a preprinted order sheet to reduce prescription errors in a pediatric emergency department: a randomized, controlled trial. Pediatrics 2005;116:1299–302. Available at: http://psnet.ahrq.gov/resource.aspx?resourceID=3105
. Retrieved January 11, 2008.
21. Clarke-Pearson DL, Synan IS, Dodge R, Soper JT, Berchuck A, Coleman RE. A randomized trial of low-dose heparin and intermittent pneumatic calf compression for the prevention of deep venous thrombosis after gynecologic oncology surgery. Am J Obstet Gynecol 1993;168:1146–53.
22. DeBernardo RL Jr, Perkins RB, Littell RD, Krasner CN, Duska LR. Low-molecular-weight heparin (dalteparin) in women with gynecologic malignancy. Obstet Gynecol 2005;105:1006–11.
23. Einstein MH, Pritts EA, Hartenbach EM. Venous thromboembolism prevention in gynecologic cancer surgery: a systematic review. Gynecol Oncol 2007;105:813–9.
Figure. No caption available.