Bahl V, Shuman AG, Hu HM, Jackson CR, Pannucci CJ, Alaniz C, Chepeha DB, Bradford CR. Chemoprophylaxis for Venous Thromboembolism in Otolaryngology. JAMA Otolaryngol Head Neck Surg. 2014;140(11):999-1005. doi:10.1001/jamaoto.2014.2254
Venous thromboembolism (VTE) causes significant morbidity and mortality in surgical patients. Despite strong evidence that thromboprophylaxis reduces the incidence VTE, guidelines for prophylaxis in otolaryngology are not well established. Key to the development of VTE prophylaxis recommendations are effective VTE risk stratification and evaluation of the benefits and harms of prophylaxis.
To evaluate the effectiveness and safety of VTE chemoprophylaxis among a population of otolaryngology patients stratified by risk.
Design, Setting, and Participants
Retrospective cohort study of 3498 adult patients admitted for otolaryngologic surgery at a single-institution academic tertiary care medical center between September 1, 2003, and June 30, 2010.
Patients were stratified into 2 groups based on whether they received VTE chemoprophylaxis.
Main Outcomes and Measures
Incidence of VTE and bleeding-related complications within 30 days after surgery.
Of 1482 patients receiving VTE chemoprophylaxis, 18 (1.2%) developed a VTE compared with 27 of 2016 patients (1.3%) who did not receive prophylaxis (P = .75). Patients with Caprini VTE risk scores greater than 7 were less likely to have a VTE with perioperative chemoprophylaxis (5.3% vs 10.4%; P = .06). Of patients with VTE chemoprophylaxis, 3.5% developed a bleeding complication compared with 1.2% of patients without prophylaxis (P < .001). Bleeding complications were associated with concomitant use of antiplatelet medications and chemoprophylaxis. Among patients undergoing free tissue transfer, chemoprophylaxis significantly decreased the incidence of VTE (2.1% vs 7.7%; P = .002) and increased bleeding complications (11.9% vs 4.5%; P = .01). In all other patients, VTE chemoprophylaxis did not significantly influence the likelihood of VTE (1.0% vs 0.6%; P = .12) or bleeding (1.5% vs 0.9%; P = .15).
Conclusions and Relevance
Effectiveness and safety of VTE chemoprophylaxis differed between patient subgroups, defined by Caprini risk score and by procedure. Effectiveness was most evident in patients with high Caprini risk scores and microvascular free tissue reconstruction. Bleeding complications were associated with VTE chemoprophylaxis administered in close proximity to potent antiplatelet therapy. The Caprini risk assessment model appears to be an effective tool to stratify otolaryngology patients by risk for VTE. Patients undergoing free tissue reconstruction merit further study before developing recommendations for VTE prophylaxis because of their higher risk of both VTE and bleeding.
Venous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE), is a common complication in surgical patients and causes significant morbidity and mortality. The incidence of DVT ranges from 15% to 40% in general surgery patients in the absence of thromboprophylaxis.1 Approximately one-third of 150 000 to 200 000 VTE-related deaths in the United States per year occur after surgery,2 and patients who survive have an increased risk of postthrombotic syndrome, pulmonary hypertension, and recurrent thrombosis.3
Primary thromboprophylaxis with anticoagulant medications (chemoprophylaxis) has been proven to reduce the incidence of VTE in surgical populations. In a seminal meta-analysis of more than 70 randomized trials in 16 000 general, orthopedic, and urologic surgery patients, perioperative subcutaneous unfractionated heparin prevented approximately half of all PE and approximately two-thirds of all DVT while significantly reducing perioperative mortality compared with placebo.4 Another meta-analysis assessing the impact of low-molecular-weight heparin in general surgery patients estimated a 70% reduction in VTE compared with placebo and a trend toward improved survival.5 The benefits of VTE chemoprophylaxis are high but should be considered in light of the risk of VTE and of bleeding complications, for which data are conflicting.6,7
Quiz Ref IDStudies estimating the incidence of VTE in diverse populations of otolaryngology patients reported rates between 0.1% and 0.3% for DVT and 0.05% and 0.2% for PE, which are much lower than in other surgical populations.8,9 The reported incidence was 6 times higher among patients undergoing head and neck procedures compared with those undergoing general otolaryngology procedures (0.6% vs 0.1%) because of higher risks associated with older age and malignancy.8
The literature describes several methods of stratifying otolaryngology patients by risk for VTE.8,10,11 One method is the Caprini risk assessment model.10- 12 This model is used to determine each patient’s individual risk for VTE. The Caprini model uses a point scoring system; the relative scores for each patient’s risk factors are summed to produce a cumulative risk score that defines the patient’s risk level and associated prophylaxis regimen.13 Results using the Caprini model are mixed,10,11 but one study11 demonstrated its effectiveness by identifying patients with low, moderate, and high risk for VTE within 30 days after surgery. The overall incidence of VTE for the study population was 1.3% (0.74% DVT and 0.66% PE), which is comparable to previously published rates in head and neck surgery patients.8 But for high-risk patients, incidence was 18.3%.11
American College of Chest Physician evidence-based guidelines for thromboprophylaxis do not specifically apply to otolaryngology.1 Recommendations for thromboprophylaxis in otolaryngology are not well established because few studies of the incidence of VTE, methods of identifying at-risk patients, and benefits and harms of thromboprophylaxis have been conducted. Otolaryngologists have attempted to develop recommendations by extrapolating from the guidelines for other surgical populations.8,12,14,15 This approach produced recommendations for chemoprophylaxis in otolaryngology patients at high risk for VTE using either subcutaneous unfractionated heparin or low-molecular-weight heparin.
Recommendations for VTE prophylaxis are even more challenging for head and neck patients undergoing microvascular reconstruction with free tissue transfer, where anticoagulation and/or antiplatelet therapy are often used to prevent microvascular thrombosis.16,17 Although patients undergoing free tissue transfer are among the highest risk for VTE within the otolaryngology population, chemoprophylaxis must be considered in the context of the existing anticoagulation regimen to avoid increasing the risk of postoperative hematoma and flap failure.
The purpose of this study was to examine the effectiveness and safety of VTE chemoprophylaxis for an otolaryngology population stratified by risk of VTE.
The study protocol was approved by the institutional review board of the University of Michigan Medical School. A retrospective cohort study was designed to evaluate the incidence of VTE and bleeding outcomes for a population of otolaryngology patients. The study population included adult patients (age ≥18 years) treated at the University of Michigan undergoing otolaryngologic operations requiring hospital admission between September 1, 2003, and June 30, 2010. We provide a descriptive analysis of this population based on the application of the American Board of Otolaryngology procedure groupings.
The study population was divided into 2 groups; one received no VTE chemoprophylaxis and the other received prophylaxis with either unfractionated or low-molecular-weight heparin. The frequencies of the type of VTE prophylaxis each group received, pharmacologic and/or mechanical, were reported.
Venous thromboembolism and bleeding outcomes within 30 days after surgery were identified from several different sources. A DVT was noted for patients with clinical signs and/or symptoms and a positive finding from venous duplex ultrasonography or computed tomographic (CT) scan. A PE was confirmed by ventilation-perfusion scan or PE-protocol CT scan. Venous thromboembolism in patients who experienced sudden death were included if confirmed either pre or post mortem.
Bleeding outcomes were identified from morbidity and mortality conference records and medical records, including operative notes. We included major bleeding complications only, defined as bleeding that required reoperation or caused an airway obstruction, prolonged hospitalization, blood product transfusion, threatened flap, or permanent morbidity or mortality.18 Minor bleeding complications involving bedside hemostatic control or evacuation of hematoma were excluded. Bleeding due to postoperative infection, recurrent or residual tumor, immediate postoperative hypertension, or major hemorrhage from a named vessel were also excluded owing to the possibility of causes other than perioperative chemoprophylaxis. In addition, patients who received therapeutic heparin at any time were excluded from the assessment of bleeding outcomes.
Thirty-day VTE and bleeding outcomes were reported for the overall study population and for the 2 groups. Venous thromboembolism outcomes were also reported by patient risk for VTE, based on the Caprini risk score.
The Caprini risk assessment model consists of approximately 40 risk factors, including predisposing risk factors (genetic and clinical characteristics, such as family history of VTE, obesity, malignancy) and exposing risk factors (presenting illness or procedure, such as hip fracture or major surgery [operations longer than 45 minutes]), each with an assigned relative risk score from 1 to 5. For example, family history of VTE has a relative weight of 3 points, obesity 1 point, and hip fracture 5 points. The scores are summed to produce a cumulative score, which is used to classify the patient to a risk level and determine the type of prophylaxis.13
The Caprini risk score was calculated retrospectively for each patient by identifying risk factors from administrative and clinical data sources, applying relative weights and calculating a cumulative risk score. This risk scoring method has been tested and validated in a previous study of surgery populations.19 The χ2 test was used to test the significance of differences in rates of VTE and bleeding complications between groups at the α = .05 level.
During the initial data analysis, we observed a larger-than-anticipated increase in bleeding complications among patients receiving chemoprophylaxis. Some patients in the study population received antiplatelet medications intraoperatively and/or postoperatively. Logistic regression was used to estimate the odds ratio associated with administration of intraoperative and/or postoperative antiplatelet medications on postoperative bleeding compared with a reference population that received perioperative heparin prophylaxis and no antiplatelet medications. Length of stay was included in the model as a proxy to control for the total number of doses of postoperative antiplatelet and heparin. We explored other confounding factors, such as risk for bleeding due to coagulopathies and thrombocytopenia based on abnormal preoperative laboratory values (international normalized ratio >1.2, partial thromboplastin time >40.5 seconds, or platelet count <50 000/μL). Because none of the patients with abnormal laboratory values had a bleeding outcome, these risk factors were removed from the final model. The C statistic and the Hosmer-Lemeshow goodness-of-fit test were used to assess the fit of the model.
We also observed that patients undergoing microvascular free tissue transfer had a higher incidence of both VTE and bleeding complications compared with all other patients in the study population. Consequently, we conducted separate analyses of 30-day VTE and bleeding outcomes with and without chemoprophylaxis for patients who underwent free tissue transfer vs all other patients.
A total of 3498 hospitalized patients treated by surgeons from the Department of Otolaryngology–Head and Neck Surgery were included in the study population. The population included head and neck (69%), otology/neurotology (12%), plastic and reconstructive (11%), and general otolaryngology (8%) patients. Of these patients, 522 (15%) underwent microvascular free tissue reconstruction, and 1483 (42%) received chemoprophylaxis. Of the patients who received chemoprophylaxis, 97% received subcutaneous unfractionated heparin (5000 IU twice or 3 times daily), enoxaparin, 3% (30-40 mg daily), and fondaparinux, 0.1% (2.5 mg daily). Patients who received chemoprophylaxis had a higher mean Caprini risk score and were more likely to have orders for sequential compression devices (SCDs) (P < .001) (Table 1).
Quiz Ref IDThe incidence of VTE among all patients in the study population was 1.3% (0.74% DVT and 0.66% PE). The incidence of VTE was 1.2% (18 of 1482) for patients who received VTE chemoprophylaxis (0.88% DVT and 0.40% PE) and 1.3% (27 of 2016) for patients who did not receive chemoprophylaxis (0.65% DVT and 0.84% PE). This difference was not statistically significant (P = .75). However, when stratified by Caprini risk score, larger differences between patients with and without chemoprophylaxis were observed. The greatest difference was among patients with a risk score greater than 8, where the incidence of VTE was 10.7% for patients with and 18.3% for patients without chemoprophylaxis (Figure).
The overall incidence of bleeding complications in the study population was 2.2%. Of the patients who received VTE chemoprophylaxis, 3.5% developed a bleeding complication compared with 1.2% of patients without prophylaxis (P < .001).
Nineteen percent of the study population received intraoperative and/or postoperative antiplatelet medications; 25% of patients in the chemoprophylaxis group received antiplatelet medications compared with 15% in the group without chemoprophylaxis. Intraoperatively, ketorolac tromethamine (30 mg) was the most commonly used antiplatelet medication and postoperatively, aspirin (325 mg daily). Quiz Ref IDKetorolac was used for its antiplatelet effects in free tissue transfer and for pain control in other cases. Logistic regression identified the combinations of (1) intraoperative ketorolac, (2) perioperative chemoprophylaxis with heparin, and (3) postoperative aspirin (odds ratio [OR], 6.6; 95% CI, 3.2-13.4 [P < .001]) and of (1) intraoperative ketorolac and (2) perioperative chemoprophylaxis with heparin (OR, 5.0; 95% CI, 2.1-12.3 [P < .001]) as statistically significant predictors of postoperative bleeding compared with a reference population that received perioperative heparin and not ketorolac or aspirin (Table 2).
The difference in incidence of VTE between the groups that did and did not receive chemoprophylaxis was greater among those undergoing free tissue transfer (2.1% vs 7.7%; P = .002) compared with all other patients (1.0% vs 0.6%; P = .12). The difference in bleeding outcomes was also significant among these patients (Table 3). Sixty-eight percent of bleeding complications occurred by postoperative day 2; early bleeding was significantly more likely when intraoperative ketorolac and prophylactic heparin were used in combination than when either (but not both) was used (8.0% vs 1.4%; P < .001). Ketorolac was typically administered near the end of the operation, and heparin, within 12 hours thereafter.
High-level evidence confirms the efficacy of VTE chemoprophylaxis in multiple surgical populations. However, there is a paucity of information in the literature about optimal prophylaxis to prevent VTE in otolaryngology.14,15 Given the lack of specialty-specific guidelines, a perceived lower risk of VTE among otolaryngology patients and a fear of bleeding-related complications, VTE prophylaxis is not systematically used by otolaryngologists.20 Key to the development of recommendations for VTE prophylaxis is effective risk stratification and evaluation of the benefits and harms of prophylaxis.
Quiz Ref IDThe purpose of this study was to build on our recent analysis of risk stratification using the Caprini risk assessment model11 by evaluating the effectiveness and complications associated with chemoprophylaxis in a population of otolaryngology patients. The results of this study show that VTE chemoprophylaxis reduces the incidence of VTE outcomes for patients in an otolaryngology population with high Caprini risk scores. The impact of chemoprophylaxis on bleeding outcomes is less clear because a significant percentage of bleeding events occurred in patients who were also receiving antiplatelet medications.
The Caprini risk assessment model was tested by Shuman et al11 for a subset of the present study population that did not receive chemoprophylaxis.11 The overall incidence of VTE within 30 days for this group was 1.3%. However, this overall rate masked important differences in incidence between patient subgroups defined by Caprini risk score; incidence was 0.5% for low-risk (score <7), 2.4% for medium-risk (score 7-8), and 18.3% for high-risk (score >8) subgroups (Figure).
The difference in overall VTE rates between patients with and without chemoprophylaxis in our study was unremarkable but was more pronounced in patient subgroups with higher Caprini risk scores. Though not statistically significant, the differences grew with each increase in risk level (Figure).
Almost all patients (97%) in the chemoprophylaxis group were given subcutaneous unfractionated heparin (5000 IU twice or 3 times daily), which has been evaluated in many randomized clinical trials of general surgery and other surgical populations and shown to significantly reduce the odds of VTE.1,2,4,5
The decrease in the likelihood off VTE for patients with chemoprophylaxis in our study population, in contrast to the decrease reported for general surgery patients, was not statistically significant. The low incidence of VTE in our population yielded an underpowered analysis liable to a type II error. A study with sufficient power to detect significant differences in VTE rates between subgroups would require 4 times more participants.
Most patients in both groups had orders for SCDs, but the percentage for patients in the chemoprophylaxis group was significantly higher (P < .001). Randomized clinical trials have shown that SCDs can enhance the effectiveness of chemoprophylaxis in preventing VTE in some surgical populations.21- 23 However, in actual clinical practice, SCDs have been less effective in preventing VTE because they are often applied infrequently or improperly.24,25 We did not measure the use or proper application of SCDs and therefore cannot determine the extent to which they contributed to the decrease in VTE outcomes in the chemoprophylaxis group.
Recent guidelines estimated that the baseline risk of major bleeding in general surgery is approximately 1% in the absence of VTE chemoprophylaxis, which matches our findings.2 In a pooled meta-analysis of bleeding risk in patients receiving mechanical VTE prophylaxis vs chemoprophylaxis in general surgery and its subspecialties, the risk of major bleeding complications was 57% lower in those who received mechanical prophylaxis.7 The association between VTE chemoprophylaxis and perioperative bleeding is inconsistent in other surgical fields.
In patients undergoing head and neck procedures and particularly free tissue transfer, the threshold for returns to the operating room to treat postoperative bleeding tends to be lower because of concerns over microvascular thrombosis and flap failure.16,17,26 As a consequence, the rate of bleeding complications is higher.
While our study suggests that chemoprophylaxis reduces the incidence of VTE in high-risk patients, the effect of chemoprophylaxis on bleeding outcomes is less clear because other risk factors for bleeding confound the analysis. Concomitant use of anticoagulants, antiplatelet therapy or thrombolytic drugs engenders a risk of major postoperative bleeding.2 In the entire study population, 19% of patients received both VTE chemoprophylaxis and antiplatelet medications. Other risk factors,2 including thrombocytopenia and coagulopathies, were not prevalent.
Bleeding outcomes were significantly higher in the chemoprophylaxis group; 3.5% developed a bleeding complication compared with 1.2% in the group without chemoprophylaxis (P < .001). The logistic regression showed that the odds of postoperative bleeding increased significantly when VTE chemoprophylaxis was given to patients who had also received intraoperative ketorolac with our without aspirin.
A large percentage (72%) of patients receiving intraoperative ketorolac underwent free tissue transfer. An analysis of free tissue transfer patients produced similar results; bleeding outcomes were significantly higher in the chemoprophylaxis group and were more likely to occur when intraoperative ketorolac and prophylactic heparin were both administered. The 11.9% rate of postoperative bleeding with chemoprophylaxis was much higher than 5.6% reported by Chien et al26 for 216 free tissue transfer patients who received subcutaneous unfractionated heparin (5000 IU twice daily) and aspirin (325 mg daily).26
Ketorolac, like aspirin, is a potent antiplatelet medication. Studies of the effect of aspirin and VTE chemoprophylaxis are mixed, with some demonstrating increases in postoperative bleeding.27,28 The combination of ketorolac tromethamine (10 mg 3 times daily) and subcutaneous heparin (5000 IU twice daily) administered to healthy volunteers did not produce a clinically significant increase in bleeding time,29 but a subsequent study raised the possibility that the increase could have been greater had higher doses of ketorolac (30 mg 3 times daily) been used.30 A randomized multicenter trial demonstrated a significantly higher risk of bleeding complications when postoperative anticoagulants were given to surgical patients receiving ketorolac tromethamine (30 mg 3 times daily).31 Despite conflicting information in the literature, it appears that the initiation of VTE chemoprophylaxis in close proximity to the administration of ketorolac tromethamine (30 mg) was associated with increased bleeding among free tissue transfer patients.
Patients undergoing free tissue transfer are among the highest risk for VTE within otolaryngology because they have many risk factors, including older age, malignant conditions, and long operative and postoperative recovery periods.8,10,12 The higher risk profile of these patients is reflected in our study population where 30% had a Caprini risk score of 7 or higher compared with 16% for all other patients.
Quiz Ref IDInterestingly, when stratified by risk level, the incidence of VTE for patients who did not receive chemoprophylaxis is higher in cases involving free tissue transfer than for other patients in all three risk levels. The difference in incidence in the low-risk level was significant: 4.6% for free tissue transfer patients vs 0.3% for all other patients. The 4.6% rate is much higher than the 0.5% to 1.5% typically considered low risk (Table 4).2 This finding raises questions about whether risk scores for free tissue transfer patients were underestimated. Some risk factors may not have been identified, or alternatively, the Caprini model, which includes a risk factor for operations longer than 45 minutes, may understate the risk associated with free tissue transfer operations that are often many hours in duration.
The overall incidence of VTE in patients who did not receive chemoprophylaxis was 7.7% in free tissue transfer patients. Thai and colleagues10 reported lower rates of between 1.4% and 5.8% for a similar population and, based on these findings, did not recommend routine VTE chemoprophylaxis without additional investigation into the benefit-to-risk ratio.
The difference in incidence of VTE between groups that did and did not receive chemoprophylaxis was significant for free tissue transfer patients; chemoprophylaxis decreased the relative risk of VTE by 73%. The benefit of VTE chemoprophylaxis is clearly positive, but the benefit-to-harm ratio is difficult to assess. The perioperative anticoagulation and/or antiplatelet regimen used in these patients is variable,16,17,26,32 and, as we discovered, the addition of routine VTE chemoprophylaxis must be considered in the context of the existing regimen to avoid a significant increase in bleeding. Furthermore, other clinical outcomes such as flap failure must be evaluated.
To our knowledge, this study is the largest to date assessing the effectiveness and complications of VTE chemoprophylaxis in otolaryngology patients, stratified by VTE risk.
The study has several limitations. First, our efforts to describe the benefits and harms of VTE chemoprophylaxis were constrained by the confounding effects of concomitant use of antiplatelet therapy and chemoprophylaxis on bleeding outcomes. Also, the statistical power of our analysis was limited by the size of the study population. Our method of calculating Caprini risk scores cannot identify all patient risk factors from available data sources and, based on a previous analysis, may understate the risk level for roughly 5% of patients.20 Moreover, VTE and bleeding outcomes may have been underestimated owing to incomplete records for patients who were lost to follow-up or died. Venous thromboembolism outcomes might also have been underestimated because they were reported for 30 days after surgery, whereas studies have shown that VTE can occur up to 90 days after surgery.33 We also did not account for whether patients received chemoprophylaxis at the appropriate time, dose, and duration. Finally, we were not able to evaluate the impact of SCDs on VTE outcomes.
Results from this study provide the basis for future research. The Caprini risk assessment model appears to be an effective tool to stratify otolaryngology patients but may need to be refined to account for higher risks associated with very long operations. An examination of additional benefits and harms of VTE prophylaxis is warranted, including its impact on mortality due to PE. Free tissue transfer patients merit special analysis when developing recommendations for VTE prophylaxis because of the high risk of both VTE and bleeding; a wide range of outcomes should be evaluated, including microvascular thrombosis and flap survival. Finally, further tests of the incidence of VTE by risk level and of the effectiveness and safety of chemoprophylaxis should be conducted for other otolaryngology patients, in populations large enough to produce sufficiently powered analyses.
Findings from this study of the effectiveness and safety of VTE chemoprophylaxis in an otolaryngology population differed between patient subgroups, defined by Caprini risk score and by procedure. Effectiveness of VTE chemoprophylaxis was most evident in patients with high Caprini risk scores. Among patients who underwent microvascular free tissue reconstruction, VTE chemoprophylaxis was associated with a significant decrease in incidence of VTE and a significant increase in bleeding complications. In these patients, bleeding complications were likely due to the initiation of VTE chemoprophylaxis in close proximity to administration of potent antiplatelet therapy.
Submitted for Publication: February 26, 2014; final revision received June 22, 2014; accepted July 30, 2014.
Corresponding Author: Vinita Bahl, DMD, MPP, Director, Performance Assessment and Clinical Effectiveness, Office of Clinical Affairs, University of Michigan Health System, 300 N Ingalls, Ste 7A10, Ann Arbor, MI 48109-5485 (firstname.lastname@example.org).
Published Online: October 2, 2014. doi:10.1001/jamaoto.2014.2254.
Author Contributions: Drs Bahl and Hu had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Bahl, Shuman, Jackson, Pannucci, Chepeha, Bradford.
Acquisition, analysis, or interpretation of data: Bahl, Shuman, Hu, Jackson, Pannucci, Alaniz, Bradford.
Drafting of the manuscript: Bahl, Shuman, Jackson.
Critical revision of the manuscript for important intellectual content: Bahl, Shuman, Hu, Pannucci, Alaniz, Chepeha, Bradford.
Statistical analysis: Shuman, Hu.
Administrative, technical, or material support:Bahl, Shuman, Hu, Jackson, Pannucci, Alaniz.
Study supervision: Bahl, Chepeha, Bradford.
Conflict of Interest Disclosures: None reported.