POD indicates postoperative day; US, ultrasonographic evaluation; VTE, venous thromboembolism.
Clayburgh DR, Stott W, Cordiero T, Park R, Detwiller K, Buniel M, Flint P, Schindler J, Andersen P, Wax MK, Gross N. Prospective Study of Venous Thromboembolism in Patients With Head and Neck Cancer After Surgery. JAMA Otolaryngol Head Neck Surg. 2013;139(11):1143-1150. doi:10.1001/jamaoto.2013.4911
Copyright 2013 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Venous thromboembolism (VTE) is associated with significant morbidity and mortality in surgery patients, but little data exist on the incidence of VTE in head and neck cancer surgical patients.
To determine the incidence of VTE in postoperative patients with head and neck cancer.
Design, Setting, and Participants
A prospective study of 100 consecutive patients hospitalized at a tertiary care academic surgical center who underwent surgery to treat head and neck cancer. Routine chemoprophylaxis was not used. On postoperative day (POD) 2 or 3, participants received clinical examination and duplex ultrasonographic evaluation (US). Participants with negative findings on clinical examination and US were followed up clinically; participants with evidence of deep venous thrombosis (DVT) or pulmonary embolism (PE) were given therapeutic anticoagulation. Participants with superficial VTE underwent repeated US on POD 4, 5, or 6. Participants were monitored for 30 days after surgery.
Main Outcome and Measure
Total number of new cases of VTE (superficial and deep) identified within 30 days of surgery and confirmed on diagnostic imaging.
Of the 111 participants enrolled, 11 withdrew before completing the study; thus, 100 participants were included. The overall incidence of VTE was 13%. Eight participants were identified with clinically significant VTE: 7 DVT and 1 PE. An additional 5 participants had asymptomatic lower extremity superficial VTE detected on US alone. Fourteen percent of patients received some form of postoperative anticoagulation therapy; the rate of bleeding complications in these patients (30.1%) was higher than that in patients without anticoagulation therapy (5.6%) (P = .01).
Conclusions and Relevance
Hospitalized patients with head and neck cancer not routinely receiving anticoagulation therapy after surgery have an increased risk of VTE. Bleeding complications are elevated in patients receiving postoperative anticoagulation.
Venous thromboembolism (VTE) is responsible for 5% to 10% of all hospital deaths, affecting as many as 600 000 patients per year in the United States.1 Surgical oncology patients are considered among the highest risk patients for VTE. The risk of VTE is increased nearly 20-fold following surgery,2 and patients with cancer are twice as likely to develop VTE after surgery as patients without cancer.3 In fact, 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. Thus, chemoprophylaxis with anticoagulants such as low-molecular-weight heparin or fondaparinux is commonly recommended in postsurgical patients.5 However, VTE chemoprophylaxis is not without risks. In patients with head and neck cancer, possible complications include bleeding or hematoma that could cause airway compromise, wound complications, or failure of a microvascular reconstruction. Retrospective studies of general otolaryngology patients have shown a very low risk for VTE, generally between 0.1% and 2.4%.6- 9 Not surprisingly, compliance with VTE chemoprophylaxis guidelines has traditionally been quite low among head and neck surgeons.10
Patients undergoing head and neck cancer surgery are at increased risk for development of VTE compared with general otolaryngology patients for several reasons, including older age, tobacco use, major surgery, decreased mobility, and decreased pulmonary function.5 Our research group previously performed a retrospective study to specifically examine the incidence of VTE in high-risk patients undergoing head and neck cancer surgery11 and found that the incidence of VTE ranged from 1.4% (confirmed) to 5.8% (confirmed and suspected). These results suggest that the incidence of VTE may be underestimated in high-risk patients undergoing head and neck cancer surgery. The true incidence of VTE may be further underestimated given the retrospective nature of all previous studies. Thus, the primary purpose of this study was to prospectively determine the incidence of VTE following head and neck cancer surgery requiring prolonged hospitalization. This information is critical for the design of future studies to properly risk stratify and test prophylaxis in this patient population.
This prospective, observational cohort study was designed to assess the incidence of VTE in patients with head and neck cancer undergoing surgery requiring a prolonged hospitalization. Participants older than 18 years with a diagnosis of head and neck cancer undergoing surgery with curative intent were candidates for this study. Participants were required to have an expected postsurgical hospital stay of at least 4 days and be willing to participate in all study activities. Exclusion criteria included distant metastases or other malignant conditions, preoperative anticoagulation, any investigational medications, known hypercoagulable state or bleeding disorders, or psychiatric illness or social situations that could limit compliance with study activities. The study was designed to evaluate 100 participants for VTE monitoring; an additional 11 participants were enrolled to account for patients who failed to complete the study protocol. A prestudy power calculation determined that with an expected VTE incidence ranging from 1.4% to 5.8%,11 a total of 100 eligible participants would produce a 2-sided 95% confidence interval (CI) with a width equal to 0.079 (0.006-0.085), assuming 3% of the participants would experience a VTE after surgery (1-proportion CI using the Exact [Clopper-Pearson] method [Pass 2008; NCSS Statistical Software]). This study was approved by the Oregon Health and Science University institutional review board. Informed consent was obtained from all participants.
After patients were recruited into the study, their baseline demographic and medical characteristics were recorded. Karnofsky performance status,12 Charlson comorbidity index,13 and Caprini score14 were obtained for each participant. Caprini scoring is detailed in the Box. A serum D-dimer level was also obtained from each participant preoperatively. Participants 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 (SCDs), worn during the procedure and while in bed postoperatively, and early ambulation. Pharmacologic VTE prophylaxis was not routinely prescribed, even in those participants undergoing free-tissue transfer reconstruction. However, since this was an observational study, participants were not restricted from receiving prophylactic doses of heparin sodium or enoxaparin sodium postoperatively if deemed medically necessary by the treating team. Additional clinical information was collected at the time of surgery, including details of the surgical procedure.
Age 41 to 60 years
Minor surgery planned
Recent major surgery (<1 month)
History of inflammatory bowel disease
Swollen legs (currently)
Obesity (BMI >25)
Acute myocardial infarction
Congestive heart failure (<1 month)
Sepsis (<1 month)
Serious lung disease, including pneumonia (<1 month)
Abnormal pulmonary function
Medical patient currently prescribed bed rest
Oral contraceptives or hormone therapy
Pregnancy or postpartum status (<1 month)
History of unexplained stillbirth or recurrent spontaneous abortion
Age 60 to 74 years
Malignant condition (currently or previously)
Major surgery (>45 minutes)
Patient confined to bed (>72 hours)
Immobilizing plaster cast (<1 month)
Central venous access
Age >75 years
History of DVT/PE
Family history of thrombosis
Positive factor V Leiden
Positive prothrombin 20210A
Elevated serum homocysteine level
Positive lupus anticoagulant
Elevated anticardiolipin antibodies
Other congenital or acquired thrombophilia
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); DVT, deep venous thrombosis; PE, pulmonary embolism.
Following surgery, participants received daily clinical examinations, including calculation of a Wells score (Table 1)15 for VTE risk on postoperative day 2 or 3. On the same day, participants also underwent a bilateral lower extremity venous duplex ultrasonographic examination (Figure).16 If no VTE was seen on ultrasonography, then participants were followed up clinically throughout the remainder of their hospitalization without further intervention. Participants with a VTE in the deep venous circulation were recommended for anticoagulation therapy per current clinical guidelines,17 typically of 6 months’ duration. If a VTE was seen in the superficial veins of the lower extremity, another duplex ultrasonographic evaluation was scheduled between postoperative day 4 and 6. If a persistent superficial VTE was seen, the patient was recommended for anticoagulation therapy per current clinical guidelines,17 typically of 6 weeks’ duration. If the repeated ultrasonographic evaluation showed resolution of the VTE, then no anticoagulation was recommended. If the ultrasonographic evaluation showed progression of the VTE to involve the deep venous system, then participants instead were recommended for anticoagulation as described herein. After discharge, additional clinical screening for VTE was performed at all clinic visits up to 30 days after surgery and at least 1 clinic visit 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. All VTEs were further categorized as clinically significant and non–clinically significant. Clinically significant VTEs were those requiring more than 6 weeks of anticoagulation therapy or associated with any negative impact on clinical course. By this definition, any deep venous thrombosis (DVT) or pulmonary embolism (PE) was considered clinically significant, as was symptomatic superficial VTE.
Descriptive statistics were applied to all variables. Comparison was made between patients with VTE and those without VTE using χ2 analysis for categorical variables and the t test for continuous variables. A 2-tailed P value of <.05 was considered significant in all analyses. A Bonferroni correction was applied to the P value when multiple variables were tested at once.
A total of 111 participants were enrolled, including 11 who did not complete the study protocol: 2 participants withdrew consent after surgery; 5 participants had a change in surgical plan that led to shorter hospitalization (<4 days) than anticipated; and 4 participants failed to undergo postoperative ultrasonography within the allotted time frame. Participants who did not complete the study protocol were excluded from data analyses. Demographic and baseline screening data of the study cohort are listed in Table 2. Mean (SD) participant age was 63.5 (12.4) years. Most participants had a history of smoking (n = 73, 73%) and had been diagnosed as having squamous cell carcinoma (n = 78, 78%). Participants underwent a variety of head and neck cancer surgical procedures, including 80% who had some form of microvascular reconstruction (Table 3). As expected, this cohort demonstrated moderate functional impairment found on Karnofsky performance status evaluation and fell in the moderate to high risk for VTE based on the Caprini VTE risk assessment.
A total of 13 VTEs were detected in the study, yielding an overall VTE incidence rate of 13%. Details of the observed VTE are summarized in Table 4. Five VTEs were asymptomatic superficial VTEs and deemed non–clinically significant. The remaining 8 VTEs were judged to be clinically significant (symptomatic superficial VTEs, DVTs, or PEs), yielding a clinically significant VTE rate of 8%. Of note, only 4 of the clinically significant VTEs provoked any signs or symptoms. So it is likely that only these 4 (50%) would have been discovered in the absence of our rigorous prospective VTE screening. At the time of diagnosis, the mean (SD) Wells score was slightly higher in the patients with VTE (1.62 [0.65]) than in those without VTE (1.28 [0.54]) (P = .10), although this was not statistically significant.
Participants were grouped according to the presence or absence of VTE and compared with respect to known risk factors for VTE (Table 5). While this study was not powered to detect differences in risk factors among these groups, there was a trend in participants diagnosed as having VTE toward lower mean (SD) Karnofsky performance status score (72  in patients with VTE vs 79  in patients without VTE) (P = .09) and higher Caprini risk assessment score (7.6 [1.4] in patients with VTE vs 6.9 [1.4] in patients without VTE) (P = .09). However, none of the risk factors examined reached statistical significance.
Postoperative anticoagulation was not routinely administered during the study period. However, since this was an observational trial, participants were not restricted from receiving anticoagulation postoperatively. Fourteen participants (14%) received postoperative anticoagulation (excluding participants treated for a VTE). In 10 of these cases, anticoagulation was given by the intensive care unit (ICU) team per ICU protocols, and in the other 4 cases, no specific reason for anticoagulation was provided by the treating team. One VTE was observed among the 14 patients who received anticoagulation (7% VTE rate among patients undergoing anticoagulation treatment). Eight bleeding complications occurred among the cohort of 100 participants (Table 6): 4 in participants undergoing anticoagulation (4 of 14, 29%) and 4 in participants who did not receive anticoagulation (4 of 86, 5%) (P = .01).
Venous thromboembolism is a significant cause of perioperative morbidity and mortality following a wide range of surgical procedures. It accounts for approximately 10% of hospital deaths annually,5 and patients who survive VTE are at risk for further complications, including recurrent VTE, pulmonary hypertension, venous stasis, and complications of long-term anticoagulation.5,18,19 In addition, VTE substantially adds to length of hospital stay and inpatient costs of affected patients.20 Surgical oncology inpatients are generally considered one of the highest-risk groups for VTE,3- 5 given both the generalized proinflammatory state induced by cancer that serves to activate the clotting cascade, and the increased risk that is seen in postsurgical patients.21 Based on these data, 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.5,22 Such prophylaxis would include the routine use of low-molecular-weight heparin or fondaparinux along with possible use of mechanical compression devices. Numerous other organizations have also advocated the routine use of postoperative anticoagulation, including the Office of the Surgeon General of the United States, the Center for Medicare and Medicaid Services, the Joint Commission, and the National Quality Forum.23,24
It should be noted that the bulk of the data supporting the recommendations for routine postoperative anticoagulation come from general surgery, orthopedic surgery, and gynecologic surgery patients. There is clear benefit to the use of anticoagulation in these patient populations: baseline rates of VTE in general surgery and oncologic surgery patients is 10% to 40%, rising to 40% to 60% in orthopedic patients.5 Multiple studies have shown that the routine use of postoperative anticoagulation may bring VTE rates down to 2% to 13% in these high-risk groups.4,25 Compliance with VTE guidelines has historically been poor among otolaryngologists10 presumably because patients are often able to ambulate soon after surgery, and the potential consequences of airway compromise from bleeding or hematoma are catastrophic. Furthermore, there is relatively little data supporting the use of routine postoperative anticoagulation in head and neck surgery patients.
Previous studies of the incidence of VTE in otolaryngology and head and neck surgery patients have been retrospective in nature (Table 7). Several studies examining general otolaryngology patients6,7,26 report VTE rates lower than 1%.6,7,9,26 These studies included both inpatient and outpatient cases. More recently, Shuman et al8 evaluated general otolaryngology inpatients treated at a single institution. The authors retrospectively used Caprini scores to risk stratify patients and found extremely high rates of VTE (>18%) in the highest-risk patient group (Caprini >8). Most of the high-risk patients were head and neck cancer surgery patients.8 Of note, none of the patients in this study received postoperative anticoagulation.
There are fewer studies that exclusively examine head and neck surgery patients; the largest retrospective study of head and neck surgery patients analyzed discharge data on more than 90 000 patients and found a 2% incidence of VTE.20 Our group previously reported the results of a retrospective study of head and neck cancer patients after surgery that included microvascular reconstruction.11 In this study only 1.4% of patients had a confirmed VTE, although events that were suggestive for VTE occurred in up to 5.8% of patients, highlighting the limitations of retrospective analysis in determining the rate of VTE. It was this finding that prompted the current prospective study in the hopes of establishing a benchmark for future research including possible interventional studies.
We found a 13% overall incidence of VTE using a rigorous screening protocol in high-risk head and neck cancer surgery patients who did not routinely undergo anticoagulation therapy after surgery. Our VTE incidence is significantly greater than previous retrospective studies of otolaryngology and head and neck surgery patients but compares favorably to VTE rates observed in other high-risk surgery groups. For example, general or gynecologic surgical oncology patients have been shown to have a 5% to 20% incidence of VTE even when various forms of pharmacologic prophylaxis are used.22 In our study, 4 (30%) of the documented VTEs were symptomatic. Thus, the overall rate of symptomatic VTE was 4%. The substantial number of asymptomatic VTEs found on routine postoperative duplex ultrasonography suggests that the incidence of VTE may have been underestimated in previous retrospective studies.
The clinical significance of asymptomatic VTEs discovered on routine postoperative duplex ultrasonographic screening has yet to be defined. While most postoperative VTEs originate in the superficial calf veins, half of these resolve within 72 hours, and fewer than 15% progress to become a DVT.28- 30 The difference between asymptomatic superficial vein VTE and DVT is meaningful. In one study, 45% of DVTs led to PEs compared with 0% of asymptomatic superficial vein VTEs.29 However, there is evidence that even asymptomatic superficial VTEs carry an increased risk of mortality.31 Given the uncertain clinical significance of asymptomatic superficial VTEs, we classified all other thrombi (PEs, DVTs, and symptomatic superficial VTEs) as clinically significant, producing a clinically significant VTE rate of 8%.
It is also possible that some participants may have had VTEs prior to surgery, which were then only detected by the screening duplex ultrasonography. Indeed, the ultrasonographic report of a participant with an asymptomatic superficial VTE noted that the thrombus appeared potentially chronic. A preoperative physical examination was performed for all participants. Preoperative ultrasonographic evaluation was only performed for participants scheduled for fibula free-tissue transfer reconstruction. So, it is possible that the overall VTE rate due to surgery in this study is artificially elevated. However, it could be argued that patients with a chronic and or preoperative VTE are at greatest risk of complications. Therefore, it is reasonable to include these participants when considering the use of VTE chemoprophylaxis. Furthermore, it is unlikely that any of the symptomatic VTEs were present preoperatively. Taken together, the clinically significant VTE rate of 8% in this study should be representative despite the lack of preoperative VTE ultrasonographic screening.
Identification of patients at the highest risk of VTE is critical to appropriately directing surveillance and prevention resources. With this in mind, we also assessed known potential risk factors for VTE in our head and neck surgery patients with cancer. We noted a trend toward increased Caprini score and decreased Karnofsky performance status in participants with a VTE, although these results were not statistically significant. Other risk factors examined, including age, BMI, Charlson comorbidity Index, preoperative D-dimer, and time to ambulation failed to predict the presence of VTE. A recent large-scale retrospective study of more than 90 000 patients with head and neck cancer had generally similar findings, wherein only the presence of major comorbidities and major surgery were associated with VTE, while age, obesity, weight loss, and tobacco use showed no association.20 In our study, all participants had a Caprini score of 5 or higher. Five of 15 participants (33%) with Caprini scores greater than 8 developed VTE compared with 8 of 85 participants (9%) with a score of 8 or lower (P = .02). Our study confirms that the Caprini score may be a useful method of risk stratifying patients for possible VTE. However, it is important to note that for our high-risk cohort, the incidence of VTE was substantial even for participants with lower Caprini scores. This suggests that Caprini risk stratification may be less useful in a high-risk head and neck cancer surgery cohort.
Routine postoperative chemoprophylaxis for VTE is rapidly becoming standard practice across many surgical disciplines. Head and neck surgeons have been slower to apply the use of routine postoperative anticoagulation possibly out of fear of bleeding complications. Data on complications after anticoagulation is conflicting. Some retrospective studies have demonstrated an increase in complications with the use of anticoagulation,32 while others have not found a significant difference.33,34 A prospective assessment of the risk of routine anticoagulation after head and neck cancer surgery would be beneficial to help define the risk-benefit ratio of routine postoperative pharmacologic chemoprophylaxis. Although the present study was not designed to assess the risk of anticoagulation, surgical complications were tracked prospectively. We noted a significantly higher rate of bleeding complications in the 14 participants who received anticoagulation. This group included patients treated with prophylactic dose low-molecular-weight heparin as well as patients treated with heparin infusion. So the data should not be extrapolated to all patients with head and neck cancer receiving VTE chemoprophylaxis. Prospective studies examining the role of VTE chemoprophylaxis should carefully monitor for bleeding complications.
There are limitations to this study. This study was designed as an observational study to determine the incidence of VTE in patients with head and neck cancer after surgery. While all patients were given mechanical prophylaxis with SCDs, only a small subset received pharmacologic prophylaxis. The use of anticoagulation in some patients may have slightly decreased the overall rate of VTE observed. This study was not designed to assess the efficacy or complication rates of various forms of thromboprophylaxis, nor was it designed to assess risk factors for the development of VTE. However, most of the participants who received anticoagulation did so because they were admitted to an ICU postoperatively and were treated according to standardized ICU protocols. It could be argued that excluding these participants would exclude the sickest participants who are potentially at greatest risk for VTE. Thus, we elected to include all participants as an intention-to-treat analysis despite the potential confounding effects of the use of VTE chemoprophylaxis. The study population was also highly selected. It is unknown whether the results can be generalized to a broader population. In this study, we included the highest-risk head and neck cancer surgical patients at a tertiary care academic surgical center. Our results may not be extrapolated to the general otolaryngologist or 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 determination of the incidence of VTE in head and neck cancer surgery patients. The results indicate that VTE in this population occurs at a higher rate than that reported in previous retrospective studies. Our results support the use of routine VTE chemoprophylaxis in patients with head and neck cancer admitted for more than 72 hours after surgery. Importantly, these data establish a baseline VTE rate in high-risk head and neck cancer surgery patients that can serve as a benchmark for future prospective trials of VTE chemoprophylaxis and risk stratification.
Corresponding Author: 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 (email@example.com).
Submitted for Publication: February 2, 2013; final revision received April 14, 2013; accepted May 30, 2013.
Published Online: September 26, 2013. doi:10.1001/jamaoto.2013.4911.
Author Contributions: Drs Clayburgh and Gross 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: Clayburgh, Stott, Cordiero, Andersen, Wax, Gross.
Acquisition of data: Clayburgh, Stott, Cordiero, Park, Detwiller, Buniel, Schindler, Wax, Gross.
Analysis and interpretation of data: Clayburgh, Stott, Flint, Andersen, Wax, Gross.
Drafting of the manuscript: Clayburgh, Wax.
Critical revision of the manuscript for important intellectual content: Clayburgh, Stott, Cordiero, Park, Detwiller, Buniel, Flint, Schindler, Andersen, Wax, Gross.
Statistical analysis: Clayburgh.
Obtained funding: Clayburgh, Gross.
Administrative, technical, or material support: Stott, Cordiero, Park, Detwiller, Buniel, Flint, Wax, Gross.
Study supervision: Stott, Schindler, Andersen, Wax, Gross.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented as an abstract (No. 48430) at the American Head and Neck Society 2013 Annual Meeting; April 10-11, 2013; Orlando, Florida.