Published in:
01-12-2018 | Original Research
Venous Thromboembolism Is an Independent Predictor of Mortality Among Patients with Gastric Cancer
Published in: Journal of Gastrointestinal Cancer | Issue 4/2018
Login to get accessAbstract
Background
Venous thromboembolism (VTE) is an independent predictor of death among patients with cancer. Patients with gastric cancer (GC) are at higher risk for VTE when compared to other solid tumors, and if one considers its prevalence, GC may be responsible for one of the highest incidences of cancer-associated thrombosis. The impact of VTE on mortality is not well defined among patients with GC.
Aim
The aim of this study is to measure the impact of VTE as independent predictor of GC mortality.
Methods
Chart review of patients with GC treated in the Department of Oncology at John Stroger Hospital between the years of 2010 and 2015. VTE events were objectively confirmed with imaging in all cases. Active GC was defined as biopsy-proven metastatic disease or on active chemotherapy. Along with cancer-specific data, we abstracted risk assessments tools, non-GC-specific, validated for VTE and mortality prediction cancer, including the Khorana score (KRS), platelet lymphocyte ratio (PLR), and neutrophil lymphocyte ratio (NLR). Continuous variables are expressed by the median as appropriate according to normality. Categorical variables are expressed as percentages. SPSS version 22 was used and chi-square, Mann-Whitney U, Kaplan-Meier curve, and Cox proportional hazard with forward modeling were applied.
Results
We included 112 patients in the analysis. The patients were predominantly men (66%), 58-year-old, with adenocarcinoma (84%) and advanced disease (59%). The median follow-up was 21.3 months (IQR 8.9–42.4). Cumulative incidence of VTE at 1 year was 9%. The median time from diagnosis to VTE occurrence was 59 days (IQR 36 to 258). Patients with VTE had worse OS when compared to the non-VTE group (medians 11.87 vs 29.97 months, p = 0.02). Patients stratified as high risk by the PLR had worse OS (medians 22.6 vs 42.77 months, p = 0.02). There was no statistical difference in OS among patients stratified as high risk by the KRS (medians 23.7 vs 42.5, p = 0.16) and NLR (medians 24.1 vs 42.7 months, p = 0.21). In multivariate analysis, the independent predictors of mortality were VTE (hazard ratio (HR), 2.9; 95% CI, 1.4 to 6.6; p < 0.01), adenocarcinoma (HR, 3.1; 95% CI, 1.1 to 9.0; p = 0.03), advanced disease (HR, 2.8; 95% CI, 1.4 to 5.8; p < 0.01), and PLR (HR, 2.2; 95% CI, 1.3 to 3.8; p < 0.01).
Conclusion
VTE is associated with worse survival among patients with GC along with adenocarcinoma, advanced disease, and PLR. Moreover, these findings were independent of other cancer- and treatment-specific variables. Although potentially predictive in other cancer types, NLR and KRS were not associated with worse survival in this cohort.