Key Points
Question
Is keratinocyte carcinoma an independent risk factor for venous thromboembolism?
Findings
In this population-based analysis of data from 740 246 individuals in a nationwide insurance claims database, there was no statistically significant difference in the risk of venous thromboembolism in the cohort of patients with keratinocyte carcinoma compared with the cohort not previously diagnosed with cancer. Both cohorts were matched across patient characteristics and known risk factors for venous thromboembolism.
Meaning
Prophylactic anticoagulation should not be administered perioperatively in patients with keratinocyte carcinoma and no other risk factors for venous thromboembolism.
Abstract
Importance
Although malignancy is an established risk factor for venous thromboembolism (VTE), the risk of VTE specifically in patients with keratinocyte carcinoma (KC) has not been previously studied.
Objective
To determine the risk of VTE in patients with KC compared with patients not diagnosed with cancer and with patients diagnosed with common malignant neoplasms associated with VTE.
Design, Setting, and Participants
Population-based retrospective analysis of patient insurance claims made between January 1, 2007, and December 31, 2014, from the Truven MarketScan Commercial and Medicare Supplemental Databases. Patients treated across the United States were divided into 3 cohorts: patients with KC, patients with pancreatic cancer or acute myelogenous leukemia who are thus at high risk for VTE, and patients without a history of common malignant neoplasms. Patients were excluded from the KC cohort if they had a history of another type of cancer. Data were analyzed between April 1, 2017, and January 15, 2018.
Main Outcomes and Measures
Diagnosis of VTE within 1 year following the index date (for the KC and high-risk cohorts, the date of the initial diagnosis of cancer; for the control cohort, the date following 365 days of continuous insurance enrollment). Logistic regression was used to assess the risk of VTE in the KC cohort compared with the high-risk and control cohorts before and after matching across patient characteristics and known risk factors for VTE.
Results
Of 5 753 613 potentially eligible patients, the final sample consisted of 740 246 patients (12.8%) across 3 cohorts. Of the 740 246 study participants, 417 839 were in the KC cohort (223 986 [53.6%] men, mean [SD] age, 64.2 [13.6] years); 314 736 were in the control cohort (135 203 [43.0%] men, 42.9 [15.2] years); and 7671 were in the high-risk cohort (3502 [45.7%] men, 59.4 [14.4] years) The risk of VTE in the KC cohort was lower compared with the high-risk cohort in univariable analysis (odds ratio [OR], 0.22; 95% CI, 0.20-0.23; P < .001), multivariable analysis (OR, 0.29; 95% CI, 0.26-0.32; P < .001), and after matching across patient characteristics and known risk factors (OR, 0.52; 95% CI, 0.35-0.78; P = .001). The risk of VTE in the KC cohort was higher in the univariable analysis (OR, 2.31; 95% CI, 2.23-2.41; P < .001), lower in the multivariable analysis (OR, 0.85; 95% CI, 0.80-0.90; P < .001), and not different after matching of patient characteristics and risk factors (OR, 0.95; 95% CI, 0.89-1.01; P = .08) than that of the control cohort.
Conclusions and Relevance
The results of this study provided no evidence supporting the increased risk of VTE in the KC cohort compared with the control cohort. Given the inherent risks of chemoprophylaxis, the need for prophylactic anticoagulation in patients with KC who are scheduled for surgery should be carefully considered.
Level of Evidence
NA.
This population-based analysis uses data from a US nationwide insurance claims database to assess the risk of venous thromboembolism associated with keratinocyte carcinoma within 1 year following the date of diagnosis and compares rates of venous thromboembolism among patients with or without cancer.
Introduction
Venous thromboembolism (VTE), a condition that includes both deep venous thrombosis and pulmonary embolism, is associated with high morbidity and mortality in patients, resulting in 200 000 to 300 000 hospitalizations and 50 000 to 100 000 deaths annually in the United States.1,2
There is a well-known association between VTE and malignant neoplasms. Venous thromboembolism has been found to occur in up to 20% of patients with cancer, who may have a 4- to 7-fold increased risk of VTE compared with the general population.3,4 The risk of VTE is not uniform across different types of cancer. Cancers associated with the highest risk of VTE include pancreatic, brain, ovarian, and bone as well as leukemia and lymphoma.4,5 Risk of VTE in other types of cancer may be much lower; for example, Blom et al4 found a cumulative incidence of VTE to be 22.7 per 100 000 patients for those with pancreatic cancer and 32.6 per 100 000 patients for those with ovarian cancer compared with 2.7 per 100 000 patients with melanoma.
Despite the large variability in the incidence of VTE across various malignant neoplasms, a diagnosis of cancer—regardless of type—is typically treated as a uniform risk factor in risk assessment models for VTE prophylaxis. One of the most commonly used risk stratification models for VTE is the Caprini score,6 a validated model that assigns points based on numerous risk factors for VTE and subsequently stratifies patients into risk categories. Based on a patient’s score, various prophylaxis regimens may be used for patients undergoing surgery, ranging from early ambulation to mechanical prophylaxis with sequential compressive devices, or chemoprophylaxis with fondaparinux sodium, unfractionated heparin, or low-molecular-weight heparin.7 Based on the Caprini score, a diagnosis of cancer—either present or previous—is worth 2 points, which alone equates to the upper end of the “low risk” category for which mechanical prophylaxis is recommended. Patients with a diagnosis of cancer and just 1 additional risk factor (age 41-60 years or planned minor surgery) are classified as “moderate risk,” for which both chemoprophylaxis and mechanical prophylaxis are recommended.6,8,9 Recently, Caprini10 has begun to exclude basal cell carcinoma from this calculation, but no reference to evidence is given.
Although rates of VTE in numerous malignant neoplasms have been previously reported, the association between keratinocyte carcinoma (KC)—which includes both basal cell carcinoma and squamous cell carcinoma—and VTE has not been previously studied. The additional risk of VTE associated with KC remains unknown, but patients with a diagnosis of squamous cell carcinoma or basal cell carcinoma are routinely classified as being at heightened risk for a thromboembolic event because of their cancer diagnosis, consequently affecting treatment considerations such as perioperative anticoagulation. This study aims to determine the risk of VTE associated with KC by comparing rates of VTE in these patients with rates in patients without cancer, or in patients with malignant neoplasms known to incur a high risk of VTE.
Methods
Data Source
A retrospective analysis was performed using data from the Truven Health MarketScan Commercial and Medicare Supplemental Databases (Truven Health Analytics), which contains deidentified insurance claims data corresponding to more than 200 million individuals across the United States who receive private health insurance from self-insured employers and other health plans. Claims reflect services provided to enrollees, their spouses, and their dependents in inpatient and outpatient settings and include outpatient pharmacy claims. The database was queried for diagnoses made between January 1, 2007, and December 31, 2014, using International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes for patient selection. Analysis of these data occurred between April 1, 2017, and January 15, 2018. The Research and Compliance Office at the Stanford University School of Medicine, Stanford, California, deemed the study exempt from human participants review. The Stanford University School of Medicine waived the need for patient informed consent; an exemption was provided given the lack of identifiable patient information.
Study Cohort
Of 3 cohorts identified, a cohort of patients with KC was identified by querying the database for all patients with a diagnosis (ICD-9 diagnosis code) of cutaneous squamous cell carcinoma (173.x2) or basal cell carcinoma (173.x1); ICD-9 diagnosis codes used to identify diagnoses of basal cell carcinoma are inclusive of KC with basosquamous morphologic features. Patients were excluded if they were younger than 18 years, did not have 12 months of continuous insurance enrollment before or after the initial diagnosis of KC, or had a medical history of the 14 most commonly diagnosed types of cancers in the United States that together account for more than 75% of all new cancer cases (Table 1).11 These 14 types of cancer were excluded to minimize the risk of confounding resulting from the increased risk of VTE associated with these malignant neoplasms. Patients were also excluded if they were treated prior to January 1, 2011, to ensure that the KC cohort was specific to the squamous cell carcinoma and basal cell carcinoma diagnoses. Prior to January 1, 2011, all types of nonmelanoma skin cancer (squamous cell carcinoma, basal cell carcinoma, unspecified, and other specified) were represented by 1 ICD-9 diagnosis code for each skin subsite; it was not until January 1, 2011, that each type of cancer received its own code.12 Finally, patients were excluded if demographic data were missing or ambiguous (eg, different geographic regions for treatment were recorded for services provided on the same diagnosis date). The index date for the KC cohort was defined as the date of the initial diagnosis of cancer as recorded in the data set.
Table 1. Types of Cancer Accounting for at Least 75% of All New Cases of Cancer in the United States, January 1, 2007, to December 31, 2014.
| Type of Cancer | Cancer Cases, % of All New Cases |
|---|---|
| Men | |
| Prostate | 19 |
| Lung and bronchus | 14 |
| Colon and rectum | 9 |
| Urinary bladder | 7 |
| Melanoma of the skin | 6 |
| Kidney and renal pelvis | 5 |
| Non-Hodgkin lymphoma | 5 |
| Leukemia | 4 |
| Oral cavity and pharynx | 4 |
| Liver and intrahepatic bile duct | 3 |
| Total | 76 |
| Women | |
| Breast | 30 |
| Lung and bronchus | 12 |
| Colon and rectum | 8 |
| Uterine corpus | 7 |
| Thyroid | 5 |
| Melanoma of the skin | 4 |
| Non-Hodgkin lymphoma | 4 |
| Leukemia | 3 |
| Pancreas | 3 |
| Kidney and renal pelvis | 3 |
| Total | 79 |
A second cohort was created with patients diagnosed (ICD-9 diagnosis code) with pancreatic cancer (157.x) and acute myelogenous leukemia (AML) (205.xx), 2 malignant neoplasms previously shown to have a high risk of VTE.5,13,14 Exclusion criteria for the high-risk cohort included those criteria used to create the KC cohort. Patients were also excluded if they had a history of KC or any of the 12 remaining most common types of cancer other than AML and pancreatic cancer. The index date for the high-risk cohort, as with the KC cohort, was defined as the date of the initial diagnosis of cancer as recorded in the data set.
A third cohort was created to approximate a noncancer or control cohort to represent individuals with an average risk of VTE. We selected a random sample of all patients in the database who did not have a history of KC or the 14 most common types of cancer. Similarly, patients were excluded if they were younger than 18 years, were treated before January 1, 2011, and did not have 12 months of continuous insurance enrollment prior to and after the index date, defined as the initial date of 365 days of continuous insurance enrollment. Patients in this control cohort were also excluded if they were diagnosed with any form of metastatic disease after the index date.
Explanatory Variables
Patient characteristics included age, sex, type of insurance, geographic region, and year of treatment, which were captured on the index date for patients with KC and high-risk patients and on the latest date of service for the control patients in the 1-year look-back period. Covariates also included risk factors for VTE specified in the Caprini score. These covariates (medications and comorbidities within 30 days prior to the index date) included warfarin sodium and oral contraceptive use and sepsis, stroke, congestive heart failure, and bone fracture of the lower extremity. Comorbidities within 1 year prior to the index date included obesity, pregnancy, pedal edema, prior VTE, varicose veins, myocardial infarction, inflammatory bowel disease, chronic obstructive pulmonary disease, and other hypercoagulable states (Table 2). In addition, we captured whether a patient was diagnosed (ICD-9 diagnosis code) with any form of metastatic disease (196.x, 197.x, 198.xx, and 199.x) in the year following the index date, as metastases have been previously associated with an increased risk of VTE.15 A patient’s overall comorbidity burden was further assessed using the van Walraven index, a formulation of the Elixhauser index, which represents 30 comorbidities as a single score that reflects association with mortality.16
Table 2. Initial Cohort Characteristics.
| Characteristic | Type of Cohorta | ||
|---|---|---|---|
| Noncancer Control (n = 314 736) | KC (n = 417 839) | High-Risk (n = 7671)b | |
| Age, mean (SD), y | 42.9 (15.2) | 64.2 (13.6) | 59.4 (14.4) |
| Male, No. (%) | 135 203 (43.0) | 223 986 (53.6) | 3502 (45.7) |
| Insurance plan, No. (%) | |||
| Comprehensive | 7483 (2.4) | 85 999 (20.6) | 1199 (15.6) |
| EPO | 7838 (2.5) | 5815 (1.4) | 158 (2.1) |
| HMO | 30 249 (9.6) | 41 119 (9.8) | 970 (12.6) |
| POS | 23 290 (7.4) | 24 320 (5.8) | 541 (7.1) |
| PPO | 219 890 (69.9) | 234 310 (56.1) | 4397 (57.3) |
| POS with capitation | 1746 (0.6) | 1525 (0.4) | 20 (0.3) |
| CDHP | 12 405 (3.9) | 14 609 (3.5) | 232 (3.0) |
| HDHP | 11 835 (3.8) | 10 142 (2.4) | 154 (2.0) |
| Region, No. (%) | |||
| Northeast | 72 011 (22.9) | 75 385 (18.0) | 2032 (26.5) |
| North Central | 61 698 (19.6) | 90 987 (21.8) | 1667 (21.7) |
| South | 112 747 (35.8) | 155 019 (37.1) | 2357 (30.7) |
| West | 60 438 (19.2) | 86 779 (20.8) | 1443 (18.8) |
| Unknown | 7842 (2.5) | 9669 (2.3) | 172 (2.2) |
| Year, No. (%) | |||
| 2011 | 173 536 (55.1) | 72 545 (17.4) | 2524 (32.9) |
| 2012 | 103 810 (33.0) | 195 104 (46.7) | 2690 (35.1) |
| 2013 | 37 390 (11.9) | 150 190 (35.9) | 2457 (32.0) |
| Prescription medication use within 30 d prior to index date, No. (%)c | |||
| Oral contraceptives | 11 383 (3.6) | 11 522 (2.8) | 210 (2.7) |
| Warfarin | 1088 (0.3) | 28 501 (6.8) | 451 (5.9) |
| Comorbidities within 30 d prior to index date, No. (%)c | |||
| Sepsis | 46 (<0.1) | 2948 (0.7) | 187 (2.4) |
| Congestive heart failure | 564 (0.2) | 27 374 (6.6) | 713 (9.3) |
| Stroke | 180 (0.1) | 12 373 (3.0) | 307 (4.0) |
| Bone fracture of the lower extremity | 321 (0.1) | 10 914 (2.6) | 208 (2.7) |
| Comorbidities, within 1 y prior to index date, No. (%)c | |||
| Pedal edema | 6121 (1.9) | 18 013 (4.3) | 510 (6.6) |
| Varicose veins | 2526 (0.8) | 7716 (1.8) | 126 (1.6) |
| Obesity | 15 678 (5.0) | 15 038 (3.6) | 769 (10.0) |
| Pregnancy | 10 258 (3.3) | 1265 (0.3) | 43 (0.6) |
| Myocardial infarction | 834 (0.3) | 3155 (0.8) | 145 (1.9) |
| Inflammatory bowel disease | 2008 (0.6) | 4309 (1.0) | 121 (1.6) |
| Chronic obstructive pulmonary disease | 2408 (0.8) | 9148 (2.2) | 284 (3.7) |
| Prior VTE | 1750 (0.6) | 11 326 (2.7) | 358 (4.7) |
| Other hypercoagulable state | 484 (0.2) | 906 (0.2) | 69 (0.9) |
| Comorbidities, within 1 y after index date, No. (%)c | |||
| Metastatic disease | 0 | 2366 (0.6) | 1490 (19.4) |
| van Walraven index, mean (SD) | 0.56 (3.27) | 2.08 (4.88) | 9.03 (8.64) |
| VTE after index date, No. (%)c | 3347 (1.1) | 10 142 (2.4) | 788 (10.3) |
Abbreviations: AML, acute myelogenous leukemia; CDHP, consumer-driven health plan; EPO, exclusive provider organization; HDHP, high-deductible health plan; HMO, health maintenance organization; KC, keratinocyte carcinoma; POS, point of service; PPO, preferred provider organization; VTE, venous thromboembolism.
Comparison among cohorts was statistically significant for all characteristics with P < .001.
High-risk cohort includes patients with AML and pancreatic cancer.
The index date for the KC and high-risk cohorts was defined as the date of initial diagnosis of cancer as recorded in the data set. The index date for the control cohort was defined as the date following 365 days of continuous insurance enrollment.
Outcome Variables
The primary outcome of this study was the risk of VTE within 1 year following a patient’s index date, defined as the date of initial diagnosis of cancer; for the control cohort, index date is defined as the date following 365 days of continuous insurance enrollment. For patients in each of the 3 cohorts, diagnoses of VTE were identified using ICD-9 codes 453.4x, 453.8x, and 415.1x.
Statistical Analysis
The association between cancer status and risk of VTE was explored by independently comparing the KC cohort with the high-risk cohort and the control cohort. The risk of VTE was represented as an odds ratio (OR) and was assessed for each comparison using univariable and multivariable logistic regression, which included patient characteristics and risk factors for VTE as covariates. Risk of VTE was also assessed after matching (through an exact matching procedure) the 2 cohorts in a given comparison across patient characteristics and risk factors for VTE. Odds ratios were represented with 95% CIs and a 2-sided significance level of P < .05. Summary statistics were produced using analysis of variance tests for quantitative variables and χ2 analyses for categorical variables.
Data extraction and manipulation was performed using SAS, version 9.4 (SAS Institute Inc), and further analysis was performed using R, version 3.4.2 (R Foundation for Statistical Computing). We used the MatchIt R package to perform exact matching.17
Results
We identified a total of 5 753 613 patients with KC, patients with high-risk cancers (AML or pancreatic cancer), or as part of a potential noncancer control cohort. Of all potentially eligible patients, 2 303 472 patients (40.0%) did not meet the enrollment criteria, 864 763 (15.0%) were younger than 18 years, 2 445 871 (42.5%) were treated prior to 2011, and 1 725 650 (30.0%) had missing or ambiguous demographic data (Figure 1). Of the 4 659 945 patients potentially eligible for the control cohort, 1750 (0.04%) were diagnosed with metastatic disease in the year following the index date. Patients excluded from the sample may have failed multiple criteria. After applying exclusion criteria, the final study sample consisted of 740 246 patients (12.8% of all patients assessed).
Figure 1. Flowchart Illustrating Selection Criteria of the Study Population.
Of the 740 246 study participants, 417 839 were in the KC cohort (223 986 [53.6%] of these were men); 314 736 were in the control cohort (135 203 [43.0%] of these were men); and 7671 were in the high-risk cohort (3502 [45.7%] of these were men) (Table 2). The patients in the control cohort were younger than the patients in the other 2 cohorts, with a mean (SD) age of 42.9 (15.2) years compared with 64.2 (13.6) years for the KC cohort and 59.4 (14.4) years for the high-risk cohort (P < .001). The control cohort also had lower prevalence of nearly all of the assessed comorbidities, including heart and lung disease (congestive heart failure, myocardial infarction, and chronic obstructive pulmonary disease), history of stroke, and history of VTE as well as a lower mean (SD) van Walraven index (0.56 [3.27] compared with 2.08 [4.88] for the KC cohort and 9.03 [8.64] for the high-risk cohort; P < .001). Risk factors for VTE that were more prevalent in the control group than in either of the 2 cancer cohorts were obesity, pregnancy, and oral contraceptive use.
Univariable analysis showed that the KC cohort had a lower risk for VTE compared with the high-risk cohort (OR, 0.22; 95% CI, 0.20-0.23; P < .001) (Figure 2). This association was also observed following multivariable analysis (OR, 0.29; 95% CI, 0.26-0.32; P < .001). Table 3 shows the results of exact matching across patient characteristics and known risk factors for VTE, including the number of matched and unmatched participants. Logistic regression following matching also showed lower odds of VTE in the KC cohort compared with the high-risk cohort (OR, 0.52; 95% CI, 0.35-0.78; P = .001) (Figure 2). Thus, all 3 models suggested that the risk for VTE in the KC cohort population were lower compared with the high-risk (AML or pancreatic cancer) cohort population.
Figure 2. Risk of Venous Thromboembolism (VTE) in the Keratinocyte Carcinoma (KC) Cohort vs the High-Risk or Noncancer Control Cohorts.
The high-risk cohort includes patients with acute myeloid leukemia and those with pancreatic cancer. Matched is defined as matching patient characteristics and known risk factors for VTE between 2 cohorts. AML indicates acute myelogenous leukemia.
Table 3. Exact Matching Results of Cohorts Across Patient Characteristics and Known Risk Factors for VTE.
| Patient Subset | KC vs Noncancer Control, No. | KC vs High-Risk, No.a | ||
|---|---|---|---|---|
| Noncancer Control Cohort, No. | KC Cohort, No. | High-Risk Cohort, No. | KC Cohort, No. | |
| All | 314 736 | 417 839 | 7671 | 417 839 |
| Matched | 214 883 | 237 851 | 2191 | 54 487 |
| Unmatched | 99 853 | 179 988 | 5480 | 363 352 |
Abbreviations: KC, keratinocyte carcinoma; VTE, venous thromboembolism.
High-risk cohort includes patients with acute myelogenous leukemia and pancreatic cancer.
The KC cohort showed higher risk for VTE compared with the control cohort in univariable analysis (OR, 2.31; 95% CI, 2.23-2.41; P < .001) (Figure 2). Multivariable analysis showed the risk for VTE in the KC cohort as slightly lower (OR, 0.85; 95% CI, 0.80-0.90; P < .001) than for those in the control cohort. A statistically significant association was not observed in the analysis following matching across patient characteristics and known risk factors for VTE (OR, 0.95; 95% CI, 0.89-1.01; P = .08).
When assessed in multivariable regression, we found that a diagnosis of metastasis within 1 year following the index date was a strong risk factor for VTE in comparisons of both the KC vs the control cohort (OR, 1.76; 95% CI, 1.45-2.14; P < .001) and the KC vs the high-risk cohort (OR, 2.82; 95% CI, 2.50-3.18; P < .001).
Discussion
This study assessed the risk of VTE in patients with KC, which has not been examined previously to our knowledge. Compared with the high-risk cohort of patients diagnosed with either AML or pancreatic cancer, patients with KC showed significantly lower risk of VTE following the cancer diagnosis in univariable analysis, multivariable analysis, and after matching of patient characteristics and known risk factors of VTE. Although the KC cohort also had higher risk of VTE compared with the control cohort in univariable regression, analysis after matching of the cohorts across all covariates—including known risk factors for VTE, which were more prevalent in the KC cohort—did not find a statistically significant difference between the 2 cohorts. Of note, risk for VTE in the KC cohort was slightly lower than that of the control cohort in multivariable analysis prior to matching. This result may be because the control cohort did not completely represent a typical healthy adult but rather reflected a sicker population that was at higher risk for VTE. One possible explanation for this increased baseline risk for VTE was that we only excluded the most common cancers, representing 75% of all cancer diagnoses, from the control cohort for simplicity; a subset of the control cohort may have been diagnosed with other cancers that could have increased the risk for VTE. Given that the effect size estimated in multivariable regression was much smaller than that observed in the KC vs high-risk cohort comparison, and that analysis following the matching of cohorts did not find a statistically significant difference in the rates of VTE between the KC vs control cohort, it is possible that the association observed in multivariable regression was not clinically significant.
Use of prophylactic anticoagulants in patients at increased risk for VTE can be lifesaving but may also introduce its own set of risks. In the perioperative setting, VTE chemoprophylaxis has been associated with an increased rate of bleeding complications, including wound hematoma, bleeding of the injection, and the need for postoperative transfusion.18,19,20 These complications can be particularly troublesome when dealing with resection and reconstruction, which often involves a variety of local skin flaps and skin grafts that can be impaired by localized bleeding. Furthermore, even prophylactic dosing of heparin can increase the risk of heparin-induced thrombocytopenia, an adverse drug reaction in which antibodies to heparin result in thrombocytopenia and an increased risk of thrombotic events.18 While rates of this uncommon but life-threatening complication vary widely in different populations, surgical patients and those who receive unfractionated heparin rather than low-molecular-weight heparin are at greatest risk for heparin-induced thrombocytopenia, with a rate as high as 5% reported in patients receiving unfractionated heparin while undergoing orthopedic surgical procedures.21
A recent systematic review by 2 of us (S. P. Moubayed and S. P. Most) found that rates of VTE vary significantly by surgical specialty; thus, interpretation of Caprini scores varies by specialty.22 Because most current risk assessment models, such as the Caprini score, do not specifically include or exclude the diagnosis of KC; however, providers who strictly adhere to such models may use preoperative chemoprophylaxis solely on the basis of a diagnosis of KC and in the absence of other risk factors for VTE. Given the potential for adverse events associated with VTE chemoprophylaxis, it is paramount that such prophylaxis is not given reflexively to all patients who undergo surgery, but rather only to patients with a truly elevated risk of a thromboembolic event. The findings from our study argue that, in the absence of other risk factors, patients with nonmetastatic KC undergoing a minor surgical procedure (defined in the Caprini scoring system as any surgical procedure taking <45 minutes) should not receive VTE chemoprophylaxis.
Limitations
This study has several limitations. The MarketScan database includes only patients with private health insurance and may not be generalizable to uninsured patients or those insured patients covered by Medicaid or Medicare. While we were able to identify general diagnoses of metastatic disease, billing codes do not provide information of the corresponding primary tumor or of the tumor characteristics, precluding further examination of stage- or morphologic feature–specific associations and limiting our conclusions to patients with primary KC. Furthermore, multivariable regression and matching can only balance cohorts based on measured covariates; potential confounders not captured in billing codes, such as the duration of the surgical procedure and the use of sequential compressive devices during hospitalization, could bias results. Finally, this study is limited to follow-up of 1 to 3 years given that ICD-9 diagnosis codes specific to cutaneous KC were introduced only in late 2011.
Conclusions
Venous thromboembolism is a significant problem because of its morbidity, mortality, and associated health care costs. While chemoprophylaxis is important when treating patients with an increased risk of VTE, it is equally important that such agents are not administered inappropriately because they can lead to perioperative complications. In this population-based study, we found no evidence of increased risk of VTE in patients with KC compared with a control cohort. These results argue for careful consideration of risk assessment models, such as the Caprini score, when a surgical procedure is planned for a patient with KC and no other risk factors for VTE in order to limit unnecessary exposure to the potential risk of VTE chemoprophylaxis.
References
- 1.Horlander KT, Mannino DM, Leeper KV. Pulmonary embolism mortality in the United States, 1979-1998: an analysis using multiple-cause mortality data. Arch Intern Med. 2003;163(14):1711-1717. doi: 10.1001/archinte.163.14.1711 [DOI] [PubMed] [Google Scholar]
- 2.Goldhaber SZ. Pulmonary embolism. N Engl J Med. 1998;339(2):93-104. doi: 10.1056/NEJM199807093390207 [DOI] [PubMed] [Google Scholar]
- 3.Imberti D, Benedetti R. Primary prophylaxis of VTE in cancer outpatients. Thromb Res. 2016;140(suppl 1):S103-S108. doi: 10.1016/S0049-3848(16)30108-6 [DOI] [PubMed] [Google Scholar]
- 4.Blom JW, Vanderschoot JPM, Oostindiër MJ, Osanto S, van der Meer FJM, Rosendaal FR. Incidence of venous thrombosis in a large cohort of 66,329 cancer patients: results of a record linkage study. J Thromb Haemost. 2006;4(3):529-535. doi: 10.1111/j.1538-7836.2006.01804.x [DOI] [PubMed] [Google Scholar]
- 5.Blom JW, Osanto S, Rosendaal FR. High risk of venous thrombosis in patients with pancreatic cancer: a cohort study of 202 patients. Eur J Cancer. 2006;42(3):410-414. doi: 10.1016/j.ejca.2005.09.013 [DOI] [PubMed] [Google Scholar]
- 6.Caprini JA, Arcelus JI, Hasty JH, Tamhane AC, Fabrega F. Clinical assessment of venous thromboembolic risk in surgical patients. Semin Thromb Hemost. 1991;17(suppl 3):304-312. [PubMed] [Google Scholar]
- 7.Gould MK, Garcia DA, Wren SM, et al. Prevention of VTE in nonorthopedic surgical patients: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest. 2012;141(2)(suppl):e227S-e277S. doi: 10.1378/chest.11-2297 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Obi AT, Pannucci CJ, Nackashi A, et al. Validation of the Caprini venous thromboembolism risk assessment model in critically ill surgical patients. JAMA Surg. 2015;150(10):941-948. doi: 10.1001/jamasurg.2015.1841 [DOI] [PubMed] [Google Scholar]
- 9.Gharaibeh L, Albsoul-Younes A, Younes N. Evaluation of VTE prophylaxis in an educational hospital: comparison between the institutional guideline (Caprini 2006) and the ACCP guideline (ninth edition). Clin Appl Thromb/Hemost. 2016;22(7):627-632. doi: 10.1177/1076029615575344 [DOI] [PubMed] [Google Scholar]
- 10.Caprini DVT risk assessment. Venous Resource Center. https://venousdisease.com/caprini-dvt-risk-assessment/. Accessed August 20, 2017.
- 11.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin. 2017;67(1):7-30. doi: 10.3322/caac.21387 [DOI] [PubMed] [Google Scholar]
- 12.ICD-9-CM index to diseases and injuries addenda effective October 1, 2011 (FY12). https://www.cdc.gov/nchs/data/icd/icd-9-cmindexaddendafy12.pdf. Accessed June 1, 2017.
- 13.Habib M, Saif MW. Thromboembolism and anticoagulation in pancreatic cancer. JOP. 2013;14(2):135-137. doi: 10.6092/1590-8577/1480 [DOI] [PubMed] [Google Scholar]
- 14.Ku GH, White RH, Chew HK, Harvey DJ, Zhou H, Wun T. Venous thromboembolism in patients with acute leukemia: incidence, risk factors, and effect on survival. Blood. 2009;113(17):3911-3917. doi: 10.1182/blood-2008-08-175745 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Falanga A, Russo L, Milesi V, Vignoli A. Mechanisms and risk factors of thrombosis in cancer. Crit Rev Oncol Hematol. 2017;118:79-83. doi: 10.1016/j.critrevonc.2017.08.003 [DOI] [PubMed] [Google Scholar]
- 16.van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care. 2009;47(6):626-633. doi: 10.1097/MLR.0b013e31819432e5 [DOI] [PubMed] [Google Scholar]
- 17.Ho D, Imai K, King G, Stuart EA. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw. 2011;42(8):1-28. [Google Scholar]
- 18.Ho Y-H, Seow-Choen F, Leong A, Eu KW, Nyam D, Teoh MK. Randomized, controlled trial of low molecular weight heparin vs. no deep vein thrombosis prophylaxis for major colon and rectal surgery in Asian patients. Dis Colon Rectum. 1999;42(2):196-202. doi: 10.1007/BF02237127 [DOI] [PubMed] [Google Scholar]
- 19.Leonardi MJ, McGory ML, Ko CY. The rate of bleeding complications after pharmacologic deep venous thrombosis prophylaxis: a systematic review of 33 randomized controlled trials. Arch Surg. 2006;141(8):790-799. doi: 10.1001/archsurg.141.8.790 [DOI] [PubMed] [Google Scholar]
- 20.Keith JN, Chong TW, Davar D, Moore AG, Morris A, Gimbel ML. The timing of preoperative prophylactic low-molecular-weight heparin administration in breast reconstruction. Plast Reconstr Surg. 2013;132(2):279-284. doi: 10.1097/PRS.0b013e318295870e [DOI] [PubMed] [Google Scholar]
- 21.Keeling D, Davidson S, Watson H; Haemostasis and Thrombosis Task Force of the British Committee for Standards in Haematology . The management of heparin-induced thrombocytopenia. Br J Haematol. 2006;133(3):259-269. doi: 10.1111/j.1365-2141.2006.06018.x [DOI] [PubMed] [Google Scholar]
- 22.Moubayed SP, Eskander A, Mourad MW, Most SP. Systematic review and meta-analysis of venous thromboembolism in otolaryngology-head and neck surgery. Head Neck. 2017;39(6):1249-1258. doi: 10.1002/hed.24758 [DOI] [PubMed] [Google Scholar]