Figure 1. Caprini venous thromboembolism risk assessment. BMI indicates body mass index (calculated as weight in kilograms divided by height in meters squared); DVT, deep venous thrombosis; and PE, pulmonary embolism.
Figure 2. Wells criteria for diagnosis of deep venous thrombosis (DVT).
Figure 3. Study design. POD indicates postoperative day; US, ultrasonography; and VTE, venous thromboembolism.
Figure 4. Study outcomes. POD indicates postoperative day; US, ultrasonography; and VTE, venous thromboembolism.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Clayburgh D, Stott W, Kochanowski T, et al. Prospective Study of Venous Thromboembolism in Patients With Head and Neck Cancer After Surgery: Interim Analysis. JAMA Otolaryngol Head Neck Surg. 2013;139(2):161–167. doi:10.1001/jamaoto.2013.1372
Authors Affiliations: Department of Otolaryngology–Head and Neck Surgery, Oregon Health and Science University, Portland.
Objectives To prospectively determine the incidence of venous thromboembolism (VTE) following major head and neck surgery. At the midpoint of enrollment, an interim analysis was performed to determine if it was ethical to continue this study as an observational study without routine anticoagulation.
Design Prospective, observational cohort study.
Setting Academic surgical center.
Patients The interim analysis comprised 47 subjects.
Main Outcome Measure The total number of new cases of VTE (superficial and deep) identified within 30 days of surgery and confirmed on diagnostic imaging. These cases were further categorized as clinically relevant and nonclinically relevant. Clinically relevant VTEs were those requiring more than 6 weeks of anticoagulation or were associated with any negative impact on clinical course. On postoperative day 2 or 3, subjects were clinically examined and received duplex ultrasonography. Subjects with negative findings from examination and ultrasonography were followed up clinically; subjects with evidence of deep venous thrombus or pulmonary embolism were given therapeutic anticoagulation. Subjects with superficial VTE received repeated ultrasonography on postoperative days 4 to 6. Subjects were monitored for 30 days after surgery.
Results Three subjects (6%) were identified as having clinically significant VTE: 2 cases of deep venous thrombus and 1 case of pulmonary embolism. Two additional subjects had lower extremity superficial VTE without clinical findings, which were detected by ultrasonography alone. No statistically significant differences were seen between patients with VTE and those without VTE.
Conclusions This interim analysis of the first prospective study of the incidence of VTE in patients with head and neck cancer showed a VTE rate slightly higher than previously estimated in retrospective studies. There have been no unexpected serious adverse events and no rationale for early termination of the study.
Venous thromboembolism (VTE) is an important cause of morbidity and mortality in hospitalized patients, with consequences including chronic leg swelling, pulmonary embolism, and death. It is estimated that VTE contributes to 5% to 10% of all hospital deaths, affecting as many as 600 000 patients per year in the United States.1 Surgery is estimated to increase the risk of VTE more than 20-fold, while the presence of cancer confers a 6.5-fold increase in the risk of VTE.2 Patients with cancer are twice as likely to develop VTE after surgery compared with patients without cancer,3 and VTE may be the most common cause of death in the postoperative period in patients with cancer.4
The estimated baseline incidence of VTE in general surgery patients is 10% to 40%, rising to 40% to 60% in orthopedic patients.5 Given the high morbidity associated with VTE, prophylaxis is routinely recommended in these and other surgical patient groups.5 However, the risk-benefit ratio of VTE prophylaxis in head and neck surgery patients is not as clear. Given the potential for bleeding to cause airway compromise, wound complications, or ischemia of free flap reconstruction, the use of anticoagulant medications in the head and neck surgery patient is not to be taken lightly. Furthermore, the true incidence of VTE in head and neck surgery patients is not well defined. Retrospective studies of general otolaryngology patients have placed the risk of VTE between 0.1% and 2.4%.6-8 However, these studies may underestimate the true incidence of VTE for head and neck cancer surgery patients, since lower-risk patients were included.
We previously performed a retrospective study to specifically examine the incidence of VTE in high-risk head and neck cancer surgery patients and found that the incidence of VTE ranged from 1.4% (confirmed) to 5.8% (confirmed and suspicious).9 These results suggest that the incidence of VTE may be underestimated in high-risk head and neck cancer surgery patients. Thus, the primary purpose of this study was to prospectively determine the incidence of VTE following major head and neck surgery. At the midpoint of enrollment, an interim analysis was performed to determine if it was ethical to continue this study as an observational study without routine anticoagulation. Given that this is the first prospective study of VTE incidence in high-risk head and neck surgery patients, we elected to publish these interim data to help guide clinical decision making.
This prospective, observational cohort study was designed to assess the incidence of VTE in patients with head and neck cancer undergoing surgery requiring extensive hospitalization. Subjects older than 18 years with a diagnosis of head and neck cancer who were undergoing surgery with curative intent were candidates for this study. In addition, subjects had to have an expected postsurgical hospital stay of at least 4 days and be willing to participate in all study activities. Exclusion criteria included the following: patients with distant metastases or other malignant conditions, patients treated with anticoagulants preoperatively, subjects receiving any investigational medications, subjects with known hypercoagulable or bleeding disorders, or subjects with psychiatric illness or social situations that could limit compliance with study activities. The study was designed to enroll 100 subjects for VTE monitoring. With the expected incidence of VTE ranging from 1.4% to 5.8%,9 a total of 100 eligible patients will produce a 2-sided 95% confidence interval with a width equal to 0.079 (0.006-0.085), assuming that 3% of the patients will experience a VTE after surgery (1-proportion confidence interval using the exact [Clopper-Pearson] method (Pass 2008 Statistical Software). An interim analysis of the first 50 subjects enrolled was conducted and is presented herein. This study was approved by the Oregon Health and Science University (OHSU) institutional review board.
After recruitment into the study, baseline demographic and medical characteristics were recorded. Karnofsky performance status,10 Charlson comorbidity index,11 and Caprini risk scores (Figure 1)12 were obtained for each patient. A serum D-dimer level was also obtained from each patient preoperatively. Subjects then underwent surgery and were treated with the current standard practice for perioperative VTE prevention in patients with head and neck cancer at our institution. This consisted of bilateral (if possible) sequential compression devices during the procedure and while in bed postoperatively and early ambulation. Because this was designed as an observational study, subjects were not restricted from receiving a prophylactic dose of heparin sodium or enoxaparin sodium if deemed medically necessary by the treating team. Pharmacologic VTE prophylaxis was not routinely prescribed. Additional information was collected at the time of surgery, including details of the surgical procedure.
Following surgery, subjects were followed with daily clinical examination, including calculation of a Wells score (Figure 2)13 for VTE risk on postoperative day (POD) 2 or 3. On the same day, subjects also underwent bilateral lower extremity venous duplex ultrasonography (Figure 3). If no VTE was seen on ultrasonography, subjects were followed clinically throughout the remainder of their hospitalization without further intervention. Subjects with a VTE in the deep venous circulation were recommended anticoagulation per current clinical guidelines,14 typically of 6 months' duration. If a VTE was seen in the superficial veins of the lower extremity, repeated duplex ultrasonography was scheduled between PODs 4 and 6. If a persistent superficial VTE was observed, the patient was recommended anticoagulation per current clinical guidelines,14 typically of 6 weeks' duration. If repeated ultrasonography revealed resolution of the VTE, then no anticoagulation was recommended. If ultrasonography revealed progression of the VTE to involve the deep venous system, then anticoagulation was recommended for subjects, as previously described. After discharge, additional clinical screening for VTE was performed at all clinic visits within 30 days of surgery, and at least 1 clinic visit was made more than 30 days after surgery. Additional imaging was performed as needed, based on clinical findings.
The primary end point of this study was total number of new cases of VTE (superficial and deep) identified within 30 days of surgery and confirmed on diagnostic imaging. Cases of VTE were further categorized as clinically relevant and non–clinically relevant. Clinically relevant VTE cases were those requiring more than 6 weeks of anticoagulation or were associated with any negative impact on clinical course. By this definition, any deep venous thrombosis or pulmonary embolism was considered clinically relevant.
Descriptive statistics were performed on all variables. A comparison was made between patients with VTE and those without VTE using χ2 analysis for categorical variables and the unpaired t test for continuous variables. A 2-tailed P value of <.05 was considered significant in all analyses.
Of the initial 50 subjects enrolled in this study, 3 were excluded from final analysis—1 subject withdrew consent after surgery and 2 subjects had a change in surgical plan that led to shorter hospitalization (<4 days) than anticipated. At the time of the interim analysis, a total of 47 subjects completed the study protocol and were included. Demographic and baseline screening data of these subjects are given in Table 1. The mean (SD) subject age was 62.9 (12.5) years. Most subjects had a history of smoking (n = 35 [74%]) and had been diagnosed as having squamous cell carcinoma (n = 36 [77%]).
Subjects underwent a variety of head and neck procedures, and a majority included simultaneous microvascular free flap reconstruction (n = 35 [74%]) (Table 2). This is not unexpected because patients receiving a free flap at our institution stay in the hospital for at least 5 days and thus meet a major criterion for inclusion in this study. While standard VTE prophylaxis for head and neck surgery patients at our institution includes sequential compression devices and early ambulation without chemoprophylaxis, a total of 6 subjects received prophylactic doses of heparin sodium or enoxaparin sodium at some point during their hospital stay. One subject was placed on a heparin infusion after experiencing free flap ischemia. The mean (SD) baseline Caprini score was 7.1 (1.6).
The VTE outcomes observed in these 47 subjects are shown in Figure 4. In all, 5 subjects were diagnosed as having VTE within 30 days of surgery (Table 3). None of these subjects received chemoprophylaxis prior to the discovery of their VTE. Two cases of VTE were asymptomatic and were only detected on screening duplex ultrasonography. These subjects were treated with anticoagulation for 6 weeks without further sequelae. Three subjects were diagnosed as having clinically significantVTE:1 subject developed shortness of breath on POD 1 and was found to have a pulmonary embolism on chest computed tomography; 1 subject presented with unilateral leg swelling on POD 27 and on ultrasonography was found to have a superficial calf VTE requiring treatment; and 1 subject developed bilateral arm swelling and discomfort on POD 4 and was found to have bilateral upper extremity VTE on ultrasonographic imaging. These data give an overall VTE rate in this study of 11% (5 of 47) and a clinically significant VTE rate of 6% (3 of 47).
Subjects were then grouped based on the presence or absence of VTE and compared with respect to known risk factors for VTE (Table 4). While patients with VTE trended toward a higher mean age (70.4 vs 62.0 years), having a higher mean Caprini score (8.0 vs 7.0), longer period of immobilization after surgery (2.6 vs 1.8 days), and longer hospital stay (7.8 vs 6.9 days), no variables were statistically significant. This may be owing to the limited sample size in this interim analysis. Other variables, including comorbidities, body mass index, length of surgery, and preoperative D-dimer level, did not appear to correlate with VTE.
Venous thromboembolism is a major source of perioperative morbidity and mortality and is the third most common cause of hospital-related death in the United States. Patients who survive VTE are at risk for a range of problems, including recurrent VTE, pulmonary hypertension, venous stasis, and complications of long-term anticoagulation.5,15,16 Surgical patients admitted to the hospital have a nearly 70-fold increase in VTE incidence compared with the general population and have 10 times more VTEs than patients undergoing outpatient surgery.17 Furthermore, oncologic patients are also at a substantially increased risk for VTE, making surgical oncology patients among the highest-risk patients for VTE overall.3-5 The presence of cancer induces a generalized proinflammatory state and activates components of the clotting cascade, putting patients at risk for the development of VTE.3
Given the significant morbidity and mortality that may be caused by VTE, many groups have proposed recommendations and guidelines to decrease the incidence of VTE, including the Office of the Surgeon General of the United States, the Centers for Medicare & Medicaid Services, the Joint Commission on Accreditation of Health Care Organizations, and the National Quality Forum.18,19 In light of the increased risk in oncologic surgical patients, the American College of Chest Physicians and the American Society of Clinical Oncology have proposed that patients with cancer undergoing surgery should receive VTE prophylaxis for 1 month after surgery.5,20 These recommendations suggest for the routine use of low-molecular-weight heparin or fondaparinux, along with possible use of mechanical compression devices. However, there remains no consensus among head and neck surgeons regarding routine pharmacologic thromboprophylaxis. There are several possible reasons for this. First, compared with patients with cancer undergoing general surgery, many patients with head and neck cancer have relatively superficial resections and are often able to ambulate soon after surgery. Second, the risk of anticoagulation in patients after head and neck surgery includes the possibility of hematoma formation that could compromise the airway or, if applicable, the viability of free flap reconstruction. Finally, the data regarding risk of VTE were derived from groups of patients that rarely included patients with head and neck cancer. Given these considerations, it is not surprising that adherence to VTE guidelines has historically been poor among otolaryngologists.21 So it would seem that many head and neck surgeons remained unconvinced that VTE is clinically relevant to their patients. This discrepancy highlights the need for more robust data to allow for better assessment of the risk-benefit ratio of routine anticoagulation in head and neck cancer surgery patients.22
Compared with general, gynecologic, and urologic surgery,4,5 only a handful of studies examining the risk of VTE are available for otolaryngology patients, and even fewer in the head and neck cancer population specifically. Moreano and colleagues23 performed a retrospective study of more than 12 800 otolaryngology patients undergoing surgery and found an overall VTE incidence of 0.3%. Within this group, head and neck surgery patients had the highest VTE incidence (0.6%). Similarly, Lee et al6 retrospectively reviewed all otolaryngology surgical procedures at a single center over a 5-year period and found an overall VTE incidence of 0.1%.6 Innis et al7 retrospectively reviewed more than 6100 cases and found a VTE incidence of 0.1%, with an incidence of 0.6% in patients with head and neck cancer. More recently, Shuman et al8 assessed otolaryngology patients requiring inpatient admission and found a VTE incidence of 1.3%.
Two previous studies have addressed the head and neck cancer population specifically. Chen et al24 examined the incidence of VTE after reconstruction of head and neck defects and found a 0.31% incidence of deep venous thrombosis and a 0.44% incidence of pulmonary embolism. Interestingly, VTE rates were much higher in the subset of patients undergoing free flap reconstruction (0.85% for deep venous thrombosis and 1.5% for pulmonary embolism). Finally, we recently performed a retrospective review of patients with head and neck cancer undergoing simultaneous resection and free tissue transfer microvascular reconstruction and found a confirmed VTE incidence of 1.4%.9 However, we also assessed for possible VTEs based on symptoms of extremity edema, sudden cardiac arrest, or sudden respiratory failure without imaging confirmation of VTE. This yielded a possible VTE incidence as high as 5.8%, suggesting that retrospective review for confirmed VTE may significantly underestimate the true incidence of VTE in the head and neck cancer surgery population. Thus, this study was designed to be a prospective assessment of the incidence of VTE in a high-risk head and neck cancer surgery cohort.
In the present interim analysis, we found the overall incidence of VTE to be 10.6%, which is substantially higher than that reported in the previously mentioned studies. It is important to note that 2 of these VTE cases were asymptomatic superficial calf thrombi that likely would not have been detected if screening duplex ultrasonography was not performed. The clinical significance of asymptomatic superficial VTE is questionable. While isolated superficial VTE is generally considered a benign condition,25 there is some evidence that they may contribute to an increased risk of death.26 The 3 VTE cases that we defined as being clinically relevant (requiring >6 weeks of anticoagulation or was associated with any negative impact on clinical course) yielded a rate of 6%, similar to that seen in our previous retrospective review.9 The interim analysis was performed to determine if continuing the protocol as an observational study (eg, without routine anticoagulation) was ethical. The results were reviewed among the head and neck surgical team at OHSU, and it was determined that the incidence rate did not meet the a priori threshold for discontinuing the study. The results of the interim analysis were also shared with the OHSU institutional review board, which approved continuation of this observational study. While the observed VTE rate from our interim analysis is higher than those previously reported for otolaryngologic surgery, it compares favorably with VTE rates in general or gynecologic surgical oncology patients, which range as high as 40% to 80% in the absence of VTE prophylaxis to approximately 5% to 20% when various forms of VTE prophylaxis are used.20 In our study, all subjects received mechanical thromboprophylaxis with sequential compression devices along with early ambulation postoperatively. Furthermore, participants are not restricted from being anticoagulated. In fact, 13% of subjects received some form of anticoagulation based on the clinical recommendation of the treating team. Combined pharmacologic and mechanical VTE prophylaxis has been shown to be more effective in preventing VTE than single-modality prophylaxis in other colorectal surgery27 and may hold promise for further reducing the VTE rate in head and neck cancer surgery patients as well.
Preoperative risk stratification for VTE may improve patient outcomes by determining which patients would most benefit from aggressive thromboprophylaxis after surgery, while helping to avoid the potential complications of thromboprophylaxis in patients at low risk for VTE. We assessed several potential risk factors for VTE for their role in predicting the incidence of VTE. There were trends toward greater age, higher Caprini score, delay to ambulation, and longer hospital stay as predictors of VTE, although owing to the limited cohort size in this interim analysis, none of these rose to the level of statistical significance. Body mass index, length of surgery, preoperative D-dimer level, Karnofsky performance status, and Charlson comorbidity index did not appear to correlate with the presence of VTE. These findings are broadly in line with the results of Shuman et al,8 who recently analyzed the Caprini risk assessment as a predictor of VTE in otolaryngology patients. They found that patients scoring higher than 8 on the Caprini scale faced an 18.3% incidence of VTE, while those with a score less than 6 had a 0.5% incidence of VTE. When we performed this analysis on our study population, all subjects had a Caprini score greater than 5. Ten subjects had Caprini scores greater than 8, 3 of whom developed VTE (30% incidence), while 37 subjects had score of 8 or less with only 2 VTEs in this group (5.4%). Thus, the Caprini score may be a useful method of risk stratifying head and neck cancer surgery patients for possible VTE. Completion of data collection in our study should allow for a more definitive assessment of risk factors that may be useful for predicting the risk of VTE in these high-risk patients.
Despite the interesting early results of this study, it should be noted that the data presented herein have several limitations. First and foremost is the recognition that this is an interim analysis of this project, and data collection is not yet complete on the entire study population. The true VTE incidence and significance of VTE risk factors may change significantly once the study is completed. Furthermore, this was designed as a purely observational study, thus there is some heterogeneity in the thromboprophylaxis regimen used. While all patients were given mechanical prophylaxis with sequential compression devices, only a small subset received pharmacologic prophylaxis. This study is not designed to assess the efficacy or complication rates of various forms of thromboprophylaxis, but rather to prospectively establish the baseline incidence of VTE in this surgical population. Finally, it is important to recognize that these data represent the treatment of the highest-risk head and neck surgical patients at an academic surgical center, and as such probably represents the upper end of VTE incidence for head and neck surgical patients. The results of this study may not necessarily be extrapolated to the general otolaryngologist and the treatment of early-stage limited head and neck cancer (eg, thyroid cancer).
Despite these limitations, to our knowledge, this study represents the first prospective assessment of the incidence of VTE in head and neck cancer surgery patients. The results of this interim analysis suggest a higher rate of VTE than that seen in previous studies. Once completed, this study will provide an important benchmark for future research to determine the optimal methods for preventing the morbidity and mortality of VTE in patients undergoing head and neck cancer surgery.
Correspondence: Neil D. Gross, MD, Department of Otolaryngology–Head and Neck Surgery, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Mail Code PV01, Portland, OR 97239 (firstname.lastname@example.org).
Submitted for Publication: August 9, 2012; final revision received September 27, 2012; accepted October 30, 2012.
Author Contributions: Drs Clayburgh and Gross had full access to all 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: Clayburgh, Stott, Kochanowski, Flint, Andersen, Wax, and Gross. Acquisition of data: Clayburgh, Stott, Kochanowski, Park, Detwiller, Buniel, Schindler, and Gross. Analysis and interpretation of data: Clayburgh, Stott, Kochanowski, Schindler, Andersen, Wax, and Gross. Drafting of the manuscript: Clayburgh, Stott, and Wax. Critical revision of the manuscript for important intellectual content: Clayburgh, Stott, Kochanowski, Park, Detwiller, Buniel, Flint, Schindler, Andersen, Wax, and Gross. Statistical analysis: Clayburgh and Stott. Obtained funding: Gross. Administrative, technical, and material support: Stott, Kochanowski, Park, Detwiller, Flint, and Gross. Study supervision: Stott, Flint, Schindler, Andersen, Wax, and Gross.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This manuscript was presented at the American Head and Neck Society Eighth International Conference on Head and Neck Cancer; July 22, 2012; Toronto, Ontario, Canada.