Article Text
PDF
Abstract
Objective To determine our institutional rate of venous thromboembolism (VTE) following minimally invasive surgery for endometrial cancer and to perform a cost-effectiveness analysis of extended prophylactic anticoagulation after minimally invasive staging surgery for endometrial cancer.
Methods All patients with newly diagnosed endometrial cancer who underwent minimally invasive staging surgery from January 1, 2017 to December 31, 2020 were identified retrospectively, and clinicopathologic and outcome data were obtained through chart review. Event probabilities and utility decrements were obtained through published clinical data and literature review. A decision model was created to compare 28 days of no post-operative pharmacologic prophylaxis, prophylactic enoxaparin, and prophylactic apixaban. Outcomes included no complications, deep vein thrombosis (DVT), pulmonary embolism, clinically relevant non-major bleeding, and major bleeding. We assumed a willingness-to-pay threshold of $100 000 per quality-adjusted life year (QALY) gained.
Results Three of 844 patients (0.36%) had a VTE following minimally invasive staging surgery for endometrial cancer. In this model, no pharmacologic prophylaxis was less costly and more effective than prophylactic apixaban and prophylactic enoxaparin over all parameters examined. When all patients were assigned prophylaxis, prophylactic apixaban was both less costly and more effective than prophylactic enoxaparin. If the risk of DVT was ≥4.8%, prophylactic apixaban was favored over no pharmacologic prophylaxis. On Monte Carlo probabilistic sensitivity analysis for the base case scenario, no pharmacologic prophylaxis was favored in 41.1% of iterations at a willingness-to-pay threshold of $100 000 per QALY.
Conclusions In this cost-effectiveness model, no extended pharmacologic anticoagulation was superior to extended prophylactic enoxaparin and apixaban in clinically early-stage endometrial cancer patients undergoing minimally invasive surgery. This model supports use of prophylactic apixaban for 7 days post-operatively in select patients when the risk of DVT is 4.8% or higher.
- Venous thromboembolism
- anti-coagulation
- minimally invasive surgery
- endometrial cancer
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Statistics from Altmetric.com
What is already known on this topic
The rate of venous thromboembolism (VTE) after minimally invasive surgery for early-stage endometrial cancer is low. While there has been a cost-effectiveness analysis of prophylactic apixaban versus prophylactic enoxaparin for prevention of VTE in gynecologic oncology patients, there has not been a cost-effectiveness study evaluating the recommendation for extended prophylactic anticoagulation after minimally invasive staging surgery for clinically early-stage endometrial cancer.
What this study adds
This cost-effectiveness model found that no extended pharmacologic post-operative anticoagulation is cost saving following minimally invasive surgery for clinically early-stage endometrial cancer. In select patients at higher risk of post-operative VTE, prophylactic apixaban is preferred over prophylactic enoxaparin.
How this study might affect research, practice or policy
It is not cost-effective for clinically early-stage endometrial cancer patients undergoing minimally invasive surgery to have post-operative anticoagulation unless the patient is at elevated risk of VTE. When the deep vein thrombosis risk is 4.8% or higher, 7 days of prophylactic apixaban is cost-effective compared with no pharmacologic anticoagulation.
Introduction
Endometrial cancer is the most common gynecologic malignancy in the United States and the incidence is rising, likely related to an increase in obesity.1 Minimally invasive surgery is the standard of care for most cases of clinically early-stage endometrial cancers. Minimally invasive surgery has numerous benefits over open surgery, including faster recovery, decreased narcotic use, and fewer post-operative complications.2 3
Malignancy and surgery are known risk factors for venous thromboembolism (VTE), and post-operative VTE is one of the most common preventable causes of in-hospital deaths.4 5 The American College of Obstetricians and Gynecologists and the American College of Chest Physicians recommend that patients undergoing surgery for endometrial cancer should receive extended thromboprophylaxis.6 7 Despite these recommendations, a 2013 survey of Society of Gynecologic Oncology members found that the majority of gynecologic oncologists do not use extended anticoagulation in their post-operative minimally invasive endometrial cancer patients.8 This is supported by available clinical data, which has found that the rates of VTE after minimally invasive surgery for endometrial cancer are less than 1%.9–14 Moreover, post-operative pharmacologic anticoagulation is associated with bleeding complications, wound hematomas, and hematuria.15
Given that VTE risk is affected by a variety of individual patient factors, Barber et al developed a prediction model to identify those at highest risk of VTE and to guide selective use of post-operative prophylaxis.9 Points were assigned based on body mass index (BMI)>40 kg/m2 (2 points), operative time>180 min (2 points), and age>60 years (4 points). Those patients who scored greater than 2 points had a 7.8 times risk of VTE compared with those with a score of 2 or less. This algorithm advocates for extended prophylactic anticoagulation only in high-risk patients with scores greater than 2, but it has not been universally implemented. Thus, there may be an opportunity for no extended prophylaxis in a subset of patients undergoing minimally invasive surgery for endometrial cancer.
Historically, low molecular weight heparin has been the preferred form of post-operative anticoagulation, but recently direct oral anticoagulants have been gaining traction in post-operative gynecologic oncology patients.16 17 Several randomized controlled trials have proven the safety and efficacy of direct oral anticoagulants, as well as the high patient satisfaction rates.16 18 19 While there has been a cost-effectiveness analysis of apixaban versus enoxaparin for prevention of VTE in gynecologic oncology patients, there has not been a cost-effectiveness study specifically evaluating the recommendation for extended prophylactic anticoagulation after minimally invasive staging surgery for endometrial cancer.17
Given the lack of standardized guidelines in this patient population, as well as the need to balance the cost and risks of thromboprophylaxis with the low incidence rate of VTE, a cost-effectiveness study may assist in the development of a unified post-operative recommendation. We sought to evaluate our institutional rates of VTE after minimally invasive staging surgery for endometrial cancer and to evaluate the cost-effectiveness of extended prophylactic anticoagulation post-operatively based on published literature.
Methods
Institutional data review
Institutional review board approval was obtained from the University of Pittsburgh. All patients with newly diagnosed endometrial cancer who underwent minimally invasive staging surgery from January 1, 2017 to December 31, 2020 at a single academic institution were retrospectively identified. Clinicopathologic and outcome data were obtained through chart review, and data were compared using descriptive statistics.
Decision model
Since the cost-effectiveness model only used published de-identified data, institutional review board approval was not required. The Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement was used for this study. A decision model was created to compare three strategies in patients with endometrial cancer undergoing minimally invasive surgery: no extended pharmacologic prophylaxis, prophylactic enoxaparin, and prophylactic apixaban (Figure 1). The model assumes the use of sequential compression devices and one dose of pre-operative heparin per the enhanced recovery after surgery protocol.20 The decision for prophylactic apixaban usage was based on assumed equivalency in the prophylactic setting, recognizing limited data.16 For the model, we defined our patient population as having no history of VTE, no known thrombophilia, and no long-term anticoagulation usage.12
Cost-effectiveness model. DVT, deep vein thrombosis; MIS, minimally invasive surgery; pApix, prophylactic apixaban; PE, pulmonary embolism; pEnox, prophylactic enoxaparin.
A United States healthcare payor perspective was used, and outcomes were reported using cost-effectiveness ratios. A societal perspective was considered but did not change the original model given that the cost of prophylaxis was incorporated into the model and assuming no additional loss of productivity during the immediate post-operative period. TreeAge Pro 2022 software was used for model construction and analysis (TreeAge Software, Williamstown, MA, USA).
Event probabilities
Event probabilities were derived from the literature and included no complications, deep vein thrombosis (DVT), pulmonary embolism, clinically relevant non-major bleeding, and major bleeding (Table 1). We used an incidence of 0.25% for DVT and 0.54% for pulmonary embolism based on the average of two major studies.10 12 Notably, our own institutional incidence rate of VTE (of which all were pulmonary embolism) was 0.36%, which fell within the range used for the sensitivity analysis of this variable. While a higher rate of pulmonary embolism than DVT may seem counterintuitive, we hypothesize that a CT scan of the chest is more commonly performed than venous duplex scans. The risks of clinically relevant non-major bleeding and major bleeding were derived from a randomized controlled trial specifically conducted in gynecologic oncology patients.16 We assumed perfect adherence to post-operative anticoagulation.
Model probabilities, costs, and utility values
Costs
Medication costs were abstracted from the Federal Supply Schedule (https://www.va.gov/opal/nac/fss/pharmPrices.asp).21 Costs associated with complications of DVT, pulmonary embolism, clinically relevant non-major bleeding, and major bleeding were obtained from previously published studies and Medicare reimbursement data and were adjusted for medical inflation to 2019 USD. We opted to use the same cost estimates as were recently published by Ryan et al in the Journal of Clinical Oncology for a cost-effectiveness analysis of anticoagulation in ovarian cancer.22 Multiple additional sources were used to verify costs, including Centers for Medicare and Medicaid services 2022 Physician Fee Schedule (https://www.cms.gov/Medicare/Medicare), literature review, and expert opinion (Table 1). Total costs are per-patient, per-event including inpatient hospitalization, outpatient follow-up, and pharmaceutical costs. The upper and lower bounds for treatment costs were derived from the literature when feasible, otherwise they were varied by 25% in each direction.
Outcomes
Cost-effectiveness was reported in cost-effectiveness ratios that were calculated as cost per quality-adjusted life year (QALY) gained. Outcomes related to QALYs included no complications, DVT, pulmonary embolism, clinically relevant non-major bleeding treated outpatient, and major bleeding treated inpatient. The cost-effectiveness threshold in this study was $100 000 per QALY, which is a commonly cited United States benchmark; however, we recognize that varying thresholds may apply in the international setting and by limitations of practical healthcare spending.23 In the tree, QALY value was associated with each of the terminal nodes. QALYs were calculated as functions of the probability of the different paths outlined in the tree (Figure 1).
Quality of life
The values for utilities, or preference-derived quality of life-related values, were assigned between 0 (death) to 1 (perfect health). Utility values were derived from a literature review with expert input as needed (Table 1). While some cost-effectiveness analyses do not apply a utility value of 1 to cancer patients, we opted to start our patients at a utility of 1 as these patients tend to be clinically well.24 Therefore, we assumed a starting utility of 1 for no pharmacologic prophylaxis, 0.95 for prophylactic apixaban, and 0.85 for prophylactic enoxaparin.25 All patients then incurred a utility decrement of 0.19 to account for being in a post-operative state.26 Patients subsequently experienced a further utility decrement if they developed a complication (Table 1).
Clinical assumptions
Model assumptions included inpatient management for all major bleeding events as well as inpatient management for 25% of DVTs and 75% of pulmonary embolisms, while all cases of clinically relevant non-major bleeding were managed outpatient. We used the assumption that prophylactic apixaban and prophylactic enoxaparin were equally effective in preventing post-operative VTE in post-operative gynecologic oncology patients.16 The relative reduction in risk of VTE with the use of prophylactic apixaban and prophylactic enoxaparin in this patient population is not well defined, therefore a relative risk reduction of 50% was chosen for the base case analysis. A range of 0–100% risk reduction was used in the sensitivity analysis to account for the possibilities of no risk reduction, complete risk reduction, or a value in-between. We opted to exclude the rare and longer-term consequences of VTE of post-thrombotic syndrome and chronic thromboembolic pulmonary embolism as there were no data on the frequency of these events in the gynecologic oncology literature. This decision was consistent with recent cost-effectiveness analyses evaluating anticoagulation for gynecologic oncology patients and in randomized trials.27 ,22 28
Statistical analysis
To account for uncertainty and heterogeneity in clinical parameters and assumptions, we performed one-way sensitivity analyses and a probabilistic sensitivity analysis. In the one-way analyses, each parameter was varied to assess individual impact of variation on the model. Costs were varied by 25% in either direction due to the wide variation in healthcare costs. A probabilistic sensitivity analysis was conducted using a Monte Carlo simulation (5000 iterations) with random sampling of parameter distributions. The percentage of trials in which a strategy was considered cost-effective was reported over a series of willingness-to-pay thresholds. Beta distributions were used for clinical probabilities and quality of life utilities, and gamma distributions were used for costs.
Results
Institutional data
A total of 844 patients who underwent minimally invasive staging surgery for clinically early endometrial cancer were identified (Table 2). Per the standard enhanced recovery after surgery recommendations, all patients received unfractionated heparin pre-operatively and had routine use of sequential compression devices throughout surgery. There are no standardized guidelines for post-operative anticoagulation at our institution. The majority of patients received extended chemical post-operative VTE prophylaxis (n=615, 72.9%), whereas 16.2% (n=137) received no post-operative chemical prophylaxis (Online supplemental file 1). Three of 844 (0.36%) patients experienced a post-operative VTE. Two patients had a post-operative VTE on anticoagulation (0.33%) and one patient had a VTE without post-operative anticoagulation (0.73%). There was no difference in VTE rate between those who received and did not receive extended chemical VTE prophylaxis post-operatively.
Supplemental material
Demographics of the study population
Cost-effectiveness analysis
In the base case model, no pharmacologic prophylaxis was less costly and more effective than prophylactic apixaban or prophylactic enoxaparin (Table 3). The difference was due to the increased cost of VTE in the no pharmacologic prophylaxis arm. On one-way sensitivity analysis comparing no pharmacologic prophylaxis to prophylactic enoxaparin, variables that had the greatest impact on the model results included the utility of prophylactic enoxaparin, the probability of a major bleed on prophylactic enoxaparin, and the probability of a minor bleed on no pharmacologic prophylaxis. However, no pharmacologic prophylaxis remained a cost saving strategy over all parameters examined.
Cost-effectiveness model results, 28-day horizon
When no pharmacologic prophylaxis was compared with prophylactic apixaban, the variables with the greatest impact on the model results were the utility of prophylactic apixaban, the probability of a major bleed on prophylactic apixaban, and the probability of a minor bleed on no pharmacologic prophylaxis. No pharmacologic prophylaxis remained the favored strategy over all parameters examined at a willingness-to-pay threshold of $100 000 per QALY.
When all patients were assigned to pharmacologic prophylaxis, prophylactic apixaban was both less costly and more effective than prophylactic enoxaparin. Variables with the greatest impact on the model results were the probability of major bleeding on prophylactic apixaban and prophylactic enoxaparin, and the probability of pulmonary embolism on prophylactic apixaban and prophylactic enoxaparin. Using the predetermined willingness-to-pay of $100 000 per QALY, prophylactic apixaban remained favored even with a major bleeding risk of 3.5%. On Monte Carlo probabilistic sensitivity analysis, no pharmacologic prophylaxis was favored in 41.1% of trials, prophylactic apixaban was favored in 33.7% of trials, and prophylactic enoxaparin was favored in 25.2% of trials (Figure 2).
Cost-effectiveness acceptability curve. CE, cost-effectiveness; pApix, prophylactic apixaban; pEnox, prophylactic enoxaparin.
In this model, prophylactic enoxaparin became cost-effective compared with no pharmacologic prophylaxis at a willingness-to-pay threshold of $100 000 per QALY when the risk of pulmonary embolism was at least 13% or the risk of DVT was at least 40%. Similarly, prophylactic apixaban became cost-effective when compared with no pharmacologic prophylaxis when the risk of DVT was at least 14%. When all patients were assigned to prophylaxis, prophylactic enoxaparin became cost-effective compared with prophylactic apixaban when the risk of DVT while on apixaban was equal to or exceeding 13%.
Discussion
Summary of main results
The rate of VTE after minimally invasive surgery for endometrial cancer at our institution is less than 1%. In our cost-effectiveness model, no extended anticoagulation prophylaxis was superior to extended prophylactic enoxaparin and extended prophylactic apixaban in clinically early-stage endometrial cancer patients undergoing minimally invasive surgery.
Results in the context of published literature
At our institution, the rate of VTE following minimally invasive staging surgery for endometrial cancer was 0.36% within 30 days of surgery. Notably, this does not account for risk stratification and adds to the growing body of literature demonstrating very low rates of VTE in this patient population in general.9–14 In our cost-effectiveness analysis, we found that no extended post-operative thromboprophylaxis is a cost-effective strategy in patients with endometrial cancer undergoing minimally invasive surgery. When comparing no pharmacologic prophylaxis with prophylactic enoxaparin, the model was most sensitive to the utility (ie, quality of life) associated with enoxaparin injections. It is well known that patients have a strong dislike of injections, and it would be clinically plausible to assume that oral medications or no medications would be preferred. Also as expected, the probability of bleeding complications were important factors identified on multiple one-way sensitivity analyses. Interestingly, the base case scenario comparing pharmacologic anticoagulation strategies trended towards favoring prophylactic enoxaparin when the risk of major bleeding on prophylactic apixaban increased; however, it did not cross the willingness-to-pay threshold of $100 000 per QALY. On probabilistic sensitivity analysis, no prophylaxis was favored over either mode of anticoagulation in 42% of trials, confirming that no prophylaxis is the favored strategy in this model.
Strengths and weaknesses
The greatest strengths of this cost-effectiveness study is the use of probabilities from large trials after extensive review to maximize accuracy of the model. We recognize the limitations of our study. Namely, this is a single-institution, retrospective report on the rate of VTE which is subject to the bias of the provider decision-making, since clinical guidelines do not exist in this scenario. We did not account for medication adherence in our cost-effectiveness model. However, prior studies have found that post-operative medication adherence for prophylactic enoxaparin is similar to prophylactic apixaban so this likely would not have an impact on our findings.16 29 As with all models, our analysis and results are limited by the inputs used. Although probabilities were estimated from the literature when available, expert opinion was used to define those probabilities that have not been distinctly described. This is particularly relevant to our definition of the low-risk patient. We recognize several scoring systems for VTE risk exist, but these scores may not be directly applicable to every cancer type, and specifically to clinically early-stage endometrial cancer. Moreover, the type of surgery (open vs minimally invasive) is not always accounted for in recommendations for peri-operative thromboprophylaxis, and minimally invasive surgery is known to have a lower risk of VTE compared with laparotomy.9 As such, we have made assumptions for a ‘low-risk’ patient within the limitations of the data. In clinical practice, other individual patient characteristics will influence the decision to use extended thromboprophylaxis such as frailty, mobility status, other medical comorbidities, and patient preference. Lastly, while $100 000 per QALY is a widely accepted willingness-to-pay threshold for cost-effectiveness studies, the true threshold for any system varies based on multiple factors.
Implications for practice and future research
Based on our model, prophylactic apixaban for 28 days would be cost-effective compared with no prophylaxis if the risk of DVT is 14% or higher. The VTE risk required for enoxaparin to be cost-effective was much greater, including a risk of pulmonary embolism≥13% and risk of DVT≥40%. While the definition of a low-risk endometrial cancer patient is not clearly defined, our model attempts to incorporate a large breadth of patients by only excluding those with a history of VTE, long-term anticoagulation use, or known thrombophilia.12 Our range sensitivities likely include the high-risk patients identified by Barber et al of age greater than 60 years, BMI greater than 40 kg/m2, and surgery length greater than 180 min.9 Although available data do not suggest population VTE rates that exceed those which were needed to make anticoagulation cost-effective in this model, it may be useful to clinicians when deciding which individual patients should receive post-operative anticoagulation. In our model, when the risk of DVT is 4.8% or higher, a 7-day course of prophylactic apixaban was cost-effective compared with no prophylaxis. Therefore, it is reasonable to consider a short course of anticoagulation for patients with increased risk of VTE.
Future directions include the incorporation of any prospective data relevant to this patient population and risk of VTE post-operatively. Additionally, recent research has focused on a direct oral anticoagulant compared with aspirin after surgery for VTE prevention with equivalent outcomes.30 ,31 32 While these findings occurred in orthopedic patients without malignancy, it may be useful to consider further research into aspirin for VTE prophylaxis in post-operative gynecologic oncology patients.
Conclusions
In this model, no prophylaxis was superior to prophylactic enoxaparin and prophylactic apixaban in a patient population of clinically early-stage endometrial cancer undergoing minimally invasive surgery. We acknowledge that the model is limited by available data in this clinical arena and patient-specific factors that might affect decision-making. Future directions should include development of a risk stratification calculator to determine which individual patients undergoing minimally invasive surgery for endometrial cancer should be considered for prophylactic anticoagulation.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @sarahgracebell
Contributors SB: made a substantial contribution to conception and hypothesis generation for this project and was responsible for data acquisition, data analysis, table and figure preparation, and manuscript writing. SB is the guarantor of the manuscript. TO: contributed to the conception and hypothesis generation, study design, and critical evaluation of tables, figures, and manuscript writing. AG: contributed the retrospective review of venous thromboembolism rate at our institution. KS: contributed to study design and critical evaluation of manuscript writing. HK: contributed to study design and critical evaluation of manuscript writing. AR: contributed to data acquisition. JB: contributed to study design and critical evaluation of manuscript writing. SR: contributed to critical evaluation of tables, figures, and manuscript writing. JLL: responsible for conception of the project, study design, and critical evaluation of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.