An estimated 350,000 to 600,000 American patients experience venous thromboembolism (VTE) annually, which includes deep-vein thrombosis and pulmonary embolism.1-5 Patients undergoing total hip replacement (THR) or total knee replacement (TKR) surgery are at high risk for VTE, because the large veins in the legs carrying blood back to the heart are significantly injured during these procedures.6,7 As a result of the aging and increasingly obese US population, an estimated 500,000 THR operations and 3.5 million TKR operations are expected to be performed by 2030.8
Much is known about how to prevent and minimize the consequences of VTE. A number of established guidelines emphasize the need to provide appropriate thromboprophylaxis to patients at risk for VTE.9-12 However, the approach to the management and prevention of VTE varies, as is evident by the different clinical practice guidelines for VTE that have been developed by several organizations—most notably the American College of Chest Physicians (ACCP) and the American Academy of Orthopaedic Surgeons—according to their medical specialty and specific concerns and needs.9,13
At the heart of this diversity are differing interpretations of evidence supporting anticoagulant therapy in a number of clinical settings, the contrasting opinions regarding the potential risk of bleeding that all anticoagulants carry, and the historical precedents that may unduly influence management decisions.13,14 As a result, many clinicians often find it difficult to know which VTE guidelines to follow and how to adhere to them in everyday practice.
The ACCP’s evidence-based guidelines recommend that anticoagulation be used for at least 10 days and extended for up to 35 days after THR or TKR surgery.9 Bridging therapy is central to the recommendations for perioperative management for patients receiving longterm anticoagulant therapy. In general, bridging therapy may be defined as postoperative anticoagulant overlapping of at least 2 therapies that are administered together during the period of switching from an injectable anticoagulation agent (most often intravenous unfractionated heparin or subcutaneous low-molecular-weight heparin [LMWH]) to oral anticoagulation (most often a vitamin K antagonist, such as warfarin).
Because warfarin requires a minimum of 3 to 4 days to reach therapeutic concentration,15 it is necessary to continue the use of injectable anticoagulants during the transitional (switching) period. When determining whether to use bridging therapy, the risk of bleeding should be balanced against the risk of thromboembolism. Although a consensus exists among the various guidelines on the need for bridging therapy, treatment patterns in clinical practice have yet to be examined.
The aim of this retrospective study was to examine thromboprophylaxis patterns—particularly involving the use of anticoagulant bridging and/or nonbridged strategies— for patients undergoing either THR or TKR surgery. Because a main goal was to make the results of this analysis relevant to practicing clinicians, inpatient data from a large hospital database were linked to outpatient claims for accurate tracking of anticoagulant use.
Data Sources and Study Population
The data used in this study were derived from a subset of data from the MarketScan Hospital Drug Database and its linked outpatient files from the MarketScan Commercial and Medicare Supplemental database from Thomson Reuters covering the period from January 1, 2005, to December 31, 2007. The hospital data consisted of claims linked to detailed service-level hospital bills for the same admissions. The linkage was conducted for 172 hospitals identifiable in hospital and in claims databases.
Data for nearly 22,189 patient-level hospital records were successfully linked to longitudinal claims histories. For each contributing hospital, the hospital database contained the full census annual admissions. Claims data were fully validated and were from covered medical and pharmacy services and geographically dispersed private and public health plans.
The study population included patients admitted for either THR surgery (defined by International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9- CM] codes 81.42-81.47, 81.54-81.55) or TKR surgery (ICD-9-CM codes 81.40, 81.51-81.53), who were aged ≥18 years on the date of surgery (defined as the index date).
Patients were required to have commercial or Medicare insurance and be continuously enrolled in their health plan for at least 180 days before and 90 days after the index date.
Claims and inpatient databases were matched according to each patient’s date of admission, date of discharge, age, sex, and principal diagnosis. For the analysis, patients were categorized to 1 of 2 groups according to their prescribed anticoagulant therapy:
- Bridged anticoagulant therapy (ie, overlapping use of 2 anticoagulants for at least 2 days in the hospital just before discharge or an overlap of at least 2 days in the outpatient setting) or
- Switched anticoagulant therapy (ie, less than 2 days treatment of overlapping anticoagulants in the inpatient or outpatient setting).
The study investigators made an empirical decision to use a minimum of 2 days of overlap as the criterion for bridged therapy, because, based on warfarin’s requirement for 3 to 4 days to reach therapeutic concentration, 15 none of the practice guidelines recommend overlapping anticoagulation therapy for less than that period, whether in the inpatient or the outpatient setting.
Patient and Provider Characteristics
Demographic characteristics (age, gender, geographic location) were available in the enrollment data of the claims database. Using the Deyo adaptation of the Charlson Comorbidity Index,16-18 we assessed each pa - tient’s comorbidities at baseline and used ICD-9-CM codes to identify hypertension, renal disease, and selected cancers—the common diagnoses that affect the outcomes in this patient population. Hospital characteristics (ie, number of beds and teaching status) were available in the hospital drug database.
Cost data were available for patients with fee-for-service health plans. However, cost data were unavailable for partially or fully capitated plans. Therefore, the value of patients’ service utilization under the capitated plans was priced and imputed using the average payments from the MarketScan fee-for-service inpatient and outpatient services, by region, year, and procedure. These were also adjusted for inflation using the medical component of the Consumer Price Index, as defined by the US Bureau of Labor Statistics.19
Diagnostic codes. VTE and major bleeding events were assessed for up to 90 days from the hospitalization index date. VTE events were identified when a patient claim had an ICD-9-CM code for deep-vein thrombosis (code 451.1x-451.81, 451.83-451.9x, 452.xx, or 453.2-453.9x) or pulmonary embolism (code 415.1x). Major bleeding events were identified by ICD-9-CM codes (see Appendix, page 247). Minor bleeding was identified if a patient claim had any of the following ICD-9-CM codes: 784.7, 532.8, 599.7, 791.2, 531.00-531.31, 531.9, 532.00-532.31, 532.9, 533.00-533.31, 533.9, 534.00-534.31, 534.9, 535.01, 535.4, 578.9, 530.82, 569.3, 786.3, 729.92, 432.1.
Tests. Baseline characteristics were compared between the cohorts, and descriptive statistics were calculated as means (±SD) and percentages. Differences between the cohorts were analyzed using the t-test, Mann-Whitney U-test, and the chi-square test.
Propensity score. A key problem that often plagues retrospective cohort studies is the lack of randomization in assigning participants to either a bridged or nonbridged therapy group. Differences in patient and provider characteristics that influence the choice of treatment can confound the outcome measures. One method used to adjust for differences in patient profiles is propensity score analysis.
The propensity score is the conditional probability of a patient receiving treatment, given the patient’s covariates, such as demographic and clinical factors. Previous research has shown that if patients are matched according to their propensity score, a large percentage of the bias resulting from unequal distributions in patient characteristics can be removed.20 For example, if there are 2 patients, 1 in the treatment (ie, bridged therapy) group and 1 in the control (ie, switched therapy) group, with the same or similar propensity scores, then these patients can be considered randomly assigned to each group, and therefore, as equivalently treated or untreated.
Propensity score analysis can be implemented in a variety of ways. For the bridged and nonbridged therapy groups in this present study, we used a logistic regression model to predict the probability that patients belonged in each group on the basis of their observed characteristics. After several matching techniques (1:1, kernel, radius, and Mahalanobis distance) were applied, we used the one that provided the best balance between patients receiving bridged therapy and those receiving switched therapy, according to the guidelines provided by Baser.21 The 1:1 matching technique that was used yielded a total of 174 patients (87 in each group) of 264 patients who were matched.
Outcome measures for the matched group are presented in the article (see Results section), because these values are not confounded by the baseline differences and are therefore risk-adjusted.
Statistical terms. Interaction terms were used, but these were all nonsignificant. The C-statistic of the logistic regression used to create propensity scores was 0.87, justifying the value of propensity score matching.22 Statistical analyses were performed using SAS v9.2 (SAS Institute, Cary, NC) and Stata v10 (StataCorp, College Station, TX). Bootstrapping techniques were used to estimate the standard errors, because the sample size was small.23
Standard difference, which represents the percentage difference between the means after adjusting for differences in the standard deviation between the groups, was used. Values >10% represent significant practical differences.
Of the 1770 patients who were eligible for the analysis, 219 (12%) did not receive anticoagulants. Of the remaining 1551 (88%) patients who received anticoagulants, 264 received combination therapy with intravenous or subcutaneous LMWH and oral warfarin, 75 (28%) of whom underwent THR surgery and 189 (72%), TKR surgery (Table 1).
Of these 264 patients, 163 (62%) received overlapping (≥2 days) anticoagulant therapy; however, this was not considered bridged therapy unless both days of therapy were administered either in the hospital just before discharge or in the outpatient setting. Of these 163 patients, 160 (98%) received overlapping therapy in the hospital and 3 (2%) received it in the outpatient setting. One of the 3 outpatients received only 2 days of overlapping therapy, the second received 5 days, and the third, 14 days. Only 105 (approximately 40%) patients using combination therapy received bridged prophylaxis as defined in this study (ie, overlapping therapy for ≥2 days in either the hospital, just before discharge, or the outpatient setting); 101 patients (38%) were switched to anticoagulant therapy (ie, <2 days of overlapping treatment).
Among the 75 patients who underwent THR surgery and received combination anticoagulation therapy, 48 (64%) received overlapping therapy (46 inpatients, 2 outpatients) and 27 (36%) received switched therapy, all in the hospital. For the 2 outpatients, the overlapping therapy lasted 5 days and 14 days, respectively (Table 1).
Compared with patients who underwent THR surgery, a slightly smaller percentage of those who underwent TKR surgery received bridged therapy. Among patients receiving combination therapy after TKR surgery, 114 (60%) received overlapping therapy, only 1 of whom received it as an outpatient (2 days of overlap; Table 1).
Switched therapy was used in 74 patients (39%) who underwent TKR surgery, all before hospital discharge.
Demographic and Baseline Characteristics
Table 2 outlines baseline characteristics stratified by use of bridging therapy. Overall, patients who received bridged therapy were younger and more likely to have commercial insurance than patients who received nonbridged therapy (ie, switched therapy or overlapping therapy for <2 days in either the hospital, just before discharge, or the outpatient setting), although these differences were not significant.
The frequency of certain baseline comorbidities, such as renal disease, was lower among patients who received bridged therapy than in patients who received nonbridged therapy, but the frequency of selected cancers was higher (not a significant difference) in those receiving bridging therapy. Although there were no differences between the 2 groups in demographic characteristics such as age, sex, and region, patients who received bridging therapy were more likely to be admitted to low-capacity hospitals with 200 to 299 beds and less likely to be admitted to highcapacity hospitals with 300 to 499 beds than patients who received nonbridged therapy (Table 2).
Among patients who underwent THR surgery, the frequency of comorbidities was similar to that in the overall sample. Patients who received bridged therapy tended to live in the north central United States and were less likely to be admitted to high-capacity (300-499 bed) hospitals. Patients who received bridged therapy after TKR surgery were younger, less likely to have renal disease at baseline, and less likely to be admitted to highcapacity hospitals than those who received nonbridged therapy. There were no differences in hospital characteristics (Table 2).
Because cost data will be analyzed separately, we did not consider them to be within the scope of this paper, which focuses on clinical outcomes. It should be noted, however, that cost data presented in Table 2 are in 2007 dollars.
Table 3 provides a comparison of the overall (THR and TKR combined), as well as THR- and TKR-specific, rates of VTE, major bleeding, and minor bleeding in the propensity score–matched subset of patients receiving bridged or nonbridged therapy. Event rates were low in all treatment groups, but the incidence of VTE and minor bleeding events was lower in patients receiving anticoagulant bridging therapy than in those receiving nonbridged therapy; the event rate was significantly lower in the overall analysis (Table 3).
In particular, patients receiving bridged therapy overall were >3 times less likely to experience a VTE event (5.75% vs 18.39%, respectively), 50% less likely to have major bleeding (4.6% vs 2.3%, respectively), and >7 times less likely to experience minor bleeding events (8.05% vs 1.15%, respectively).
In this study population, none of the matched 13 patients who underwent THR and received bridged therapy experienced VTE or any bleeding events. The rates of VTE and of all bleeding were 50% to 80% lower in patients who received bridged therapy after undergoing TKR surgery compared with those who received nonbridged therapy.
Because the number of events was small, standard differences are reported in the rightmost column (Table 3) for all patient comparisons. These values ranged from approximately 10 to 39, indicating significant differences in the comparisons reported.
One of the provisions in the Patient Protection and Affordable Care Act passed in May 2010 is the need to reduce the frequency of preventable complications in healthcare, thereby increasing the quality of care while lowering costs.24 Results from this study showed that after risk adjustment, patients receiving bridged therapy were >3 times less likely to experience VTE events, 50% less likely to have minor bleeding, and >7 times less likely to experience major bleeding.
Because VTE is increasingly understood to have a far reaching impact on patients, payers, and providers, the US Surgeon General issued a call to action in 2008, emphasizing the importance of preventing VTE and pulmonary embolism.5 Similarly, the Centers for Medicare & Medicaid Services’ never event list underscores the urgency to adopt stringent guidelines to prevent VTE in patients who undergo THR/TKR surgery.25 Therefore, stringent guidelines must be adapted for bridged therapy, which is an option that could be adapted for accountable care organizations, organizations aimed at reducing the cost and improving the quality and overall care of Medicare beneficiaries enrolled in the traditional fee-forservice program.
Although postoperative thromboprophylaxis can reduce the incidence of thrombosis dramatically, postdischarge anticoagulation therapy of appropriate duration is essential in the setting of THR/TKR surgery.9 Properly implemented, anticoagulant bridging can be a critical component of such thromboprophylaxis strategies.26,27
Our results, however, show that patients who developed VTE after either THR or TKR while still in the hospital were rarely discharged with appropriate anticoagulant bridge therapy.
Using our criterion of including patients in the bridged therapy group who underwent overlapping therapy for as few as 2 days, we found that among 1770 patients who underwent THR/TKR surgery, only 3 (approximately 0.2%) received bridged anticoagulation therapy as outpatients; all these were in the combination therapy group (Table 1). One of these patients had only 2 days of overlap, the second 5 days, and the third 14 days. Had we adhered to the guideline recommendation of 3 days,9 an even smaller percentage of patients would have received appropriate treatment.
Using linked in-hospital and outpatient claims data, this study’s findings confirmed the observation that anticoagulation use does not conform to established guidelines. 28 Only 264 of 1170 patients received combination anticoagulation therapy. Moreover, as shown in Table 1, only 96 patients (8%) who received it had at least 3 days of overlapping therapy, as recommended in published guidelines.9
Lack of adherence to guidelines may be an effect of the inherent complexities of currently available agents that make proper prophylaxis difficult to achieve for physicians and for patients. Potential barriers to good compliance may include confusion arising out of differences in practice guidelines from different specialists; real or perceived fear of increased risk of bleeding; and problems resulting from the complex issues related to the administration and monitoring that characterize often used available therapies.
This study has several limitations, which is typical of any retrospective claims database.29 Although claims data are extremely valuable for treatment patterns, healthcare resource utilization, and costs, these data are collected for the purpose of payment rather than research. The presence of diagnoses codes on medical claims is not necessarily proof of actual disease, as such codes may have been assigned incorrectly or included as rule-out criteria rather than as documentation of actual disease.
To mitigate such problems, we applied detailed quality checks for the data set before starting the analysis presented here. Medications purchased over the coun - ter (eg, aspirin) or provided as samples by physicians were not measurable in the claims data. Nevertheless, because our analysis is descriptive and based on comparing combination therapy with monotherapy, we do not believe that this lack of information significantly affected our results.
Additional study limitations include the small sample size and the low event rates.
Existing data on anticoagulant bridging practice patterns are limited. Our study shows that few patients undergoing THR or TKR surgery receive appropriate anticoagulant prophylactic therapy. Based on our findings, we suggest that increased use of bridging therapy after THR and TKR surgery can help to improve adherence to guidelines, and ultimately, patient outcomes. Newer anticoagulants may eventually prove useful in this setting, especially if they do not require routine anticoagulation monitoring. The results of the present study may help to clarify the problem of appropriate thromboprophylaxis after THR or TKR surgery, but further research is warranted.
Ruth Sussman, PhD, provided editorial support with funding from Ortho-McNeil Janssen Scientific Affairs, LLC.
This research was funded by Ortho-McNeil Janssen Scientific Affairs, LLC.
Author Disclosure Statement
STATinMED Research is a consultant to Johnson & Johnson; Dr Baser is an employee of STATinMED Research; Dr Supina was an employee of Johnson & Johnson at the time of this study; Dr Sengupta is an employee of Johnson & Johnson; and Dr Wang is an employee of STATinMED Research.
- Bulger CM, Jacobs C, Patel NH. Epidemiology of acute deep vein thrombosis. Tech Vasc Interv Radiol. 2004;7:50-54.
- Fowkes FJ, Price JF, Fowkes FG. Incidence of diagnosed deep vein thrombosis in the general population: systematic review. Eur J Vasc Endovasc Surg. 2003;25:1-5.
- Michota F. Venous thromboembolism: epidemiology, characteristics, and consequences. Clin Cornerstone. 2005;7:8-15.
- Silverstein MD, Heit JA, Mohr DN, et al. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study. Arch Intern Med. 1998;158:585-593.
- US Department of Health and Human Services. The Surgeon General’s Call to Action to Prevent Deep Vein Thrombosis and Pulmonary Embolism 2008. Washington, DC: US Department of Health and Human Services. http://www.surgeongeneral.gov/topics/ deepvein/calltoaction/call-to-action-on-dvt-2008.pdf. Accessed January 18, 2011.
- Warwick D, Dahl OE, Fisher WD. Orthopaedic thromboprophylaxis: limitations of current guidelines. J Bone Joint Surg Br. 2008;90:127-132.
- Furie B, Furie BC. Mechanisms of thrombus formation. N Engl J Med. 2008;359: 938-949.
- Kurtz S, Ong K, Lau E, Mowat F, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007;89:780-785.
- 9. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines, 8th Edition. Chest. 2008;133:381S-453S.
- Johanson NA, Lachiewicz PF, Lieberman JR, et al. American Academy of Orthopaedic Surgeons clinical practice guideline on prevention of symptomatic pulmonary embolism in patients undergoing total hip or knee arthroplasty. J Bone Joint Surg Am. 2009;91:1756-1757.
- National Institute for Health and Clinical Excellence. Reducing the risk of venous thromboembolism (deep vein thrombosis and pulmonary embolism) in inpatients undergoing surgery. NICE clinical guideline No. 92; http://www.nice.org.uk/guidance/index.jsp?action=byID&o=12695. Accessed July 4, 2010.
- Parvizi J, Azzam K, Rothman RH. Deep venous thrombosis prophylaxis for total joint arthroplasty: American Academy of Orthopaedic Surgeons guidelines. J Arthroplasty. 2008;23:2-5.
- American Academy of Orthopaedic Surgeons Clinical Guideline on Prevention of Symptomatic Pulmonary Embolism in Patients Undergoing Total Hip or Knee Arthroplasty. Adopted by the American Academy of Orthopedic Surgeons Board of Directors May 2007. http://www.aaos.org/research/guidelines/PE_guideline.pdf. Accessed January 18, 2011.
- Caprini JA, Tapson VF, Hyers TM, et al. Treatment of venous thromboembolism: adherence to guidelines and impact of physician knowledge, attitudes, and beliefs. J Vasc Surg. 2005;42:726-733.
- Hirsh J. Oral anticoagulant drugs. N Engl J Med. 1991;324:1865-1875.
- Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47:1245-1251.
- D’Hoore W, Bouckaert A, Tilquin C. Practical considerations on the use of the Charlson comorbidity index with administrative data bases. J Clin Epidemiol. 1996; 49:1429-1433.
- Baser O, Palmer L, Stephenson J. The estimation power of alternative comorbidity indices. Value Health. 2008;11:946-955.
- US Department of Labor. Bureau of Labor Statistics. Consumer price index. Measuring price change for medical care in the CPI. http://data.bls.gov/cpi/cpi fact4.htm. Accessed January 18, 2011.
- Heckman JJ. Econometric causality. Int Stat Rev. 2008;76:1-27.
- Baser O. Choosing propensity score matching over regression adjustment for causal inference: when, why and how it makes sense. J Med Econ. 2007;10:379-391.
- Baser O. Too much ado about propensity score models? Comparing methods of propensity score matching. Value Health. 2006;9:377-385.
- Baser O, Crown WH, Pollicino C. Guidelines for selecting among different types of bootstraps. Curr Med Res Opin. 2006;22:799-808.
- Office of the Legislative Counsel for the US House of Representatives. Compilation of Patient Protection and Affordable Care Act. May 2010. http://docs.house.gov/energycommerce/ppacacon.pdf. Accessed January 18, 2011.
- Centers for Medicare & Medicaid Services, US Department of Health and Human Services. CMS improves patient safety for Medicare and Medicaid by addressing never events. http://www.cms.gov/apps/media/press/factsheet.asp?Counter=3224 &intNumPerPage=10&checkDate=&checkKey=&srchType=1&numDays=0&srch Opt=0&srchData=&keywordType=All&chkNewsType=6&intPage=&showAll=1& pYear=1&year=2008&desc=false&cboOrder=date. Accessed January 18, 2011.
- Hyers TM, Agnelli G, Hull RD, et al. Antithrombotic therapy for venous thromboembolic disease. Chest. 2001;119(1 suppl):176S-193S.
- Kearon C. Duration of venous thromboembolism prophylaxis after surgery. Chest. 2003;124:386S-392S.
- Caprini JA, Hyers TM. Compliance with antithrombotic guidelines. Manag Care. 2006;15:49-50,53-60,66. 29. Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med. 2000;342:1878-1886.