Venous Thromboembolism in Patients with Sarcoma: A Retrospective Study

Risk factors and management guidelines for venous thromboembolism (VTE) are well established; however, little clinical research has been undertaken on VTE in sarcomas. This retrospective study aimed to identify the clinical characteristics of adult sarcoma patients with VTE.

Abstract

Background.

Little has been published about the association of venous thromboembolism (VTE) and sarcoma. In this study, we sought to identify clinical features of patients with sarcoma presenting at least one VTE episode.

Methods.

Our study was a retrospective case–control study of a single‐institution database with univariate and multivariate analysis using chi‐square and Student's t test. A p value less than .05 was considered significant.

Results.

The overall incidence of VTE in patients with sarcoma was 7.9%. Predictive factors identified by multivariate analysis were metastatic disease and administration of chemotherapy. It was not statistically possible to correlate the risk of VTE with specific sarcoma subtypes, but observations suggested malignant peripheral nerve sheath tumor, osteosarcoma, and liposarcoma as having the highest propension.

Conclusion.

VTE is not infrequent in patients with sarcoma. Adoption of common guidelines for cancer‐associated thrombosis is recommended.

Introduction

Venous thromboembolism (VTE) is a common complication of cancer at both the localized and the metastatic stages. Although risk factors are well established and management guidelines are regularly published, most clinical research on cancer‐associated VTE has been performed in epithelial cancers [1], [2]. Few papers specifically address VTE in sarcomas, perhaps because carcinomas and hematological malignancies are more frequent or considered particularly thrombogenic [3]. Yet differences in epidemiology or natural history of VTE in sarcomas, if found, could lead to specific recommendations. In this retrospective study, we sought to identify the clinical characteristics of adult patients with both sarcoma and VTE.

Methods

The McGill University Health Centre (MUHC) is a supraregional referral center for sarcomas in the province of Quebec, Canada. The MUHC Sarcoma Program prospectively maintains a database of all patients seen at its multidisciplinary clinics. Demographic, clinical, radiological, and laboratory data are regularly entered and validated, and most patients receive treatment at our center. The use of this database for retrospective studies has been authorized by the Research Ethics Board of the MUHC.

Inclusion criteria for our retrospective study were diagnosis of VTE by a recognized imaging method (ultrasound, computed tomography angiography, ventilation/perfusion scan), a concurrent histological diagnosis of bone or soft‐tissue sarcoma made by an expert pathologist, and that the VTE episode could have occurred within 90 days preceding the diagnosis. Exclusion criteria were a diagnosis other than sarcoma, VTE episode more than 90 days prior to diagnosis of sarcoma, and diagnosis of VTE not established by imaging investigations.

Patients meeting the study criteria were matched with a cohort of control subjects from the same database in a 1:4 proportion in regard to age and sarcoma type (bone or soft tissue). Chi‐square and Student's t test were performed to compare groups. Identification of risk factors was done with univariable and multivariable logistic regression models, using the following parameters: sarcoma type (bone or soft‐tissue), histological grade, metastatic stage, administration of chemotherapy, and recent radiation or surgery (within 30 days). A p value less than .05 was considered significant.

Results

Between 2002 and 2015, the data of 957 consecutive patients were entered in the sarcoma database. Primary disease site was bone in 203 patients (21%) and soft tissues in 754 patients (79%). Seventy‐six patients met the inclusion criteria, 52 men and 24 women having been diagnosed with at least one VTE episode (incidence 7.9%, including symptomatic pulmonary embolism in 1.8% of cases; Table 1). One case of pulmonary embolism was fatal. Bone sarcomas were seen in 21 patients, and soft‐tissue sarcomas were seen in 55 patients. Mean age was 52 years (±17).

Table 1. Baseline demographic, tumor, and treatment‐related characteristics among patients with sarcoma.

Abbreviations: AJCC, American Joint Committee on Cancer; STS, soft‐tissue sarcoma; VTE, venous thromboembolism.

The sarcoma subtypes most commonly associated with VTE (Table 2) were, in decreasing frequency, liposarcoma (15 of 76, 20%), undifferentiated pleomorphic sarcoma (formerly malignant fibrous histiocytoma; 13 of 76, 17%), and leiomyosarcoma and osteosarcoma, both in third position (11 of 76, 14.5% each). Another observation was that the risk of VTE seemed to vary according to histology subtype: malignant peripheral nerve sheath tumor (27%), osteosarcoma, and liposarcoma (17% each) were associated with the highest incidence. Sites of first VTE included lower and upper extremity (60%), trunk (20%), lung (17%), and head and neck (3%). Sixty‐two patients with VTE (82%) had metastatic or recurrent disease. Identified or putative risk factors for VTE were administration of chemotherapy (74%), disease progression or recurrence (43%), central line insertion (25%), and postoperative setting (24%). More than one such risk factor could be present in any given patient.

Table 2. Sarcoma subtypes and incidence of venous thromboembolism.

Abbreviations: MPNST, malignant peripheral nerve sheath tumor; UPS, undifferentiated pleomorphic sarcoma; VTE, venous thromboembolism.

Predictive factors identified by multivariate analysis were metastatic disease and administration of chemotherapy. No predictive value was found for the following factors: type of sarcoma (bone vs. soft tissue), histological grade, or anatomic location (Table 3).

Table 3. Factors predictive of venous thromboembolism.

Abbreviations: CI, confidence interval; OR, odds ratio; STS, soft‐tissue sarcoma.

Discussion

This retrospective study shows that VTE is not uncommon in patients with sarcoma, an incidence of 7.9% being comparable to that of many other cancers [1], [2], [3], [4]. Few studies have been published about VTE in patients with sarcoma. For instance, a SEER‐Medicare retrospective study [5] limited to older patients (>65 years) found a higher total VTE rate of 16.7%, likely explained in part by the median age of the cohort (77 years). Another publication found a risk of 9.1% in patients operated on for bone or soft‐tissue sarcomas, whether localized or metastatic, when not given thromboprophylaxis, and a nonstatistically significant trend in risk reduction when prophylaxis was administered [6]. Taken together, these results and ours show that VTE is not uncommon in patients with sarcoma and that both administration of chemotherapy and a postoperative setting are significant risk factors for this complication.

Chemotherapy appears to be a significant contributor to the risk of VTE, but the overall low incidence of chemotherapy‐associated VTE in sarcomas and the low associated mortality do not point to routine prophylactic anticoagulation concomitantly with chemotherapy outside of a clinical trial. Indeed, randomized studies show that prophylactic anticoagulation decreases the risk of cancer‐associated VTE [7], [8], but this has not led to its widespread use because of a perceived unfavorable risk/benefit ratio and cost‐effectiveness [9]. Postoperative prophylaxis after cancer surgery or during hospitalization is, however, standard of care [9]. Our findings justify adoption of commonly used guidelines for cancer‐associated VTE in patients with sarcoma [9].

However, our study has limitations. Its retrospective nature and long time span, straddling 13 years, explain why some clinical and biological data, currently known to have a predictive value for VTE, were not included in the database. Nonetheless, it appears evident that many of the recognized risk factors in well‐known scores such as Khorana and PROTECHT [10], [11] were present in our patients. In the control cohort, patients were not matched by stage and grade, which could have led to some unbalanced baseline features between the two groups that were compared. In addition, patients were not prospectively assessed for VTE, and, therefore, the true incidence of this complication may have been underestimated. Finally, the relatively small number of patients does not allow us to dissect the relative risks proper to every disease subtype or chemotherapy regimen. It is recognized that sarcomas represent a collection of at least 50–100 clinically and/or molecularly defined entities [12], which could all have specific characteristics regarding VTE. For instance, a population‐based study revealed that the risk of VTE in non‐Hodgkin lymphomas varies according to the subtype to a large extent [13]. A population‐based study looking at distinct sarcoma subtypes could provide relevant information in that regard and confirm our findings.

Conclusion

This retrospective study brings specific information relative to the association of sarcomas and VTE but does not suggest a different natural history from VTE associated with other cancers. Therefore, observance of general treatment guidelines for VTE in patients with cancer is recommended for patients with sarcoma as well. However, retrospective analyses of larger databases or prospective assessment of histology‐specific cohorts could show precise risk factors in given disease entities and lead to more tailored recommendations.

Disclosures

The authors indicated no financial relationships.