Patterns and predictors of thromboembolic events among patients with gastric cancer

Patients with gastric cancer are at higher risk for venous thromboembolic events (VTE). Majority of such patients are treated in ambulatory settings where thromboprophylaxis is not routinely offered. In this study, we report on VTE rates and search for predictors that may help identify patients at higher risk to justify VTE-prophylaxis in ambulatory settings. Patients with pathologically-confirmed gastric adenocarcinoma were retrospectively reviewed for VTE detected by imaging studies. Clinical and pathological features known to increase the risk of VTE were studied. Khorana risk assessment model was applied on patients receiving chemotherapy. A total of 671 patients; median age 55 years, were recruited. VTE were diagnosed in 150 (22.4%) patients, including 42 (28.0%) pulmonary embolism and 18 (12.0%) upper extremity deep vein thrombosis (DVT). Majority (> 80%) developed VTE while in ambulatory settings and none had been on thromboprophylaxis. Rate was higher (27.1%) among 365 patients with metastatic compared to 16.7% among 306 patients with nonmetastatic disease, p = 0.001. Patients with metastatic disease who received multiple lines of chemotherapy (n = 85) had significantly higher rate of VTE compared to those who received a single line; 48.2% versus 19.4%, p < 0.001. Among the whole group, Khorana risk score, age, gender, smoking and obesity had no impact on VTE rates. Patients with metastatic gastric cancer, especially when treated with multiple lines of chemotherapy, are at a significantly higher risk of VTE. Khorana risk score had no impact on VTE rates. Thromboprophylaxis in ambulatory patients with metastatic gastric cancer worth studying.


Scientific Reports
| (2020) 10:18516 | https://doi.org/10.1038/s41598-020-75719-w www.nature.com/scientificreports/ cancer, are treated and managed in ambulatory settings where VTE prophylaxis is not routinely practiced 13 . Several risk assessment models (RAM) had tried to link tumor-, patient-and treatment-related factors to the risk of VTE 14 . None of these models, including the Khorana 15 , and COMPASS 16 , were gastric cancer specific. Lack of strong evidence to routinely offer VTE prophylaxis for ambulatory cancer patients, and the fear of bleeding complications had resulted in higher rate of thromboembolic events. In this study, we review our experience of VTE among patients with gastric cancer, mostly ambulatory, trying to identify factors that may increase the risk of VTE to levels high-enough to recommend VTE prophylaxis in ambulatory settings.

Patients and methods
Medical records and hospital database of adult patients (≥ 18 years old) with pathologically-confirmed gastric adenocarcinoma; diagnosed, treated and followed at our institution between January 2007 and December 2017 were retrospectively reviewed. Only VTE confirmed by imaging studies were included. Clinical and pathological features known to increase the risk of VTE, including risk factors used to calculate the Khorana risk score were collected; such risk factors include the presence of central venous catheter (CVC), hemoglobin (Hb) level, platelet and white blood cell (WBC) counts, disease stage, and body mass index (BMI). Pharmacy database was also searched for any anticoagulant therapy for all patients enrolled.
Thrombosis was considered chemotherapy-related if diagnosed any time after the first dose and up to 4 weeks after the last. Ambulatory VTE was defined as those occurring in outpatient settings or diagnosed more than 4 weeks of the most recent hospitalization. Deep venous thrombosis was diagnosed by doppler ultrasound while PE was diagnosed by CT angiogram. Major bleeding, as a complication of anticoagulation, was defined as fatal bleeding and/or symptomatic bleeding in a critical organ and/or bleeding causing a fall in hemoglobin level of ≥ 2.0 g/dL, or leading to transfusion of two or more units of red blood cells. The study was approved by the King Hussein Cancer Center's institutional review board (IRB).

Statistical analysis.
The primary objectives were to determine the incidence and characteristics of VTE in adult patients diagnosed with gastric cancer. The secondary objective was to analyze the importance of the tumor, patients and treatment'-related characteristics in predicting the occurrence of thrombosis. Descriptive statistics were performed for patients' clinical and pathological characteristics. Data was tabulated and described by ranges, medians and categorial variables were expressed as percentages (%). Variables and VTE incidences were examined for correlation.
Univariate analysis was done for patients' data using chi-square test. Factors were categorized into risk groups and included in multivariate analysis using logistic regression analysis. A value of P < 0.05 was considered statistically significant. In result, the association between studied variables and occurrence of VTE was evaluated. Following this analysis, it was hypothesized that particular group(s) of patients with specific pathological or clinical features can be identified as high risk for VTE.

Results
During the study period, a total of 671 patients (59.6% males), median age (range) 55 (18-95) years fulfilled our inclusion criteria and were enrolled. More than half (n = 365, 54.4%) of the patients had metastatic disease at diagnosis and were treated with palliative intent. Among the 306 patients with early-stage disease, 291 (95.1%) underwent gastric surgery including 99 (34.0%) with total gastrectomy. Neoadjuvant chemotherapy was given to 84 (27.5%) while 182 (59.5%) patients were treated with adjuvant chemotherapy. A central venous access (CVC) was inserted to deliver chemotherapy for patients with early-and advanced-stage disease (n = 86, 12.8%), details of patients' demographics and characteristics are presented in Table 1.
During the first 12 months of follow up, 150 (22.4%) patients had a confirmed diagnosis of VTE, including 42 (28.0%) PE and 18 (12.0%) upper extremity DVT. All patients were treated with enoxaparin, a low molecular weight heparins (LMWH), at the usual dose of 1.0 mg/kg subcutaneously twice daily. None of the patients received warfarin or direct oral anticoagulants (DOAC). Anticoagulation was complicated by bleeding in 14 (9.3%) patients; two were major but not fatal, while 6 (4.0%) had thrombocytopenia and 33 (22.0%) had recurrent thrombosis.
Majority of the patients (n = 132, 88.0%) had their VTE while having active disease; 94 (62.7%) within 4 weeks from chemotherapy. Only 25 (16.7%) patients had their clots while in-patient or within 4 weeks of recent hospitalization, while majority (83.3%) developed VTE while in ambulatory settings and none were on thromboprophylaxis,  Table 4.
We also applied a univariate analysis to further address the risk of VTE. The effect of age, gender, BMI, smoking history, disease stage, the need for multiple lines of chemotherapy and surgery were studied. Only disease stage and the need for multiple lines of chemotherapy were significantly associated with VTE. Both factors remained significant in multivariate analysis with Odd Ratio (OR) of 2.24 [95% Confidence Interval (CI) 1.423, 3.528, P < 0.001] and 2.51 [95% CI 1.428, 4.420, P = 0.001], respectively. BMI and smoking history were not significant, Table 5.

Discussion
Venous thromboprophylaxis in patients with gastric cancer is always a challenge. Balancing the known high-risk of thrombosis with the high-risk of bleeding associated with anticoagulation is not easy. Patients with gastric cancer are at higher risk of bleeding compared to other patients with other solid tumors. Efforts pushing for more prophylaxis are always pulled back when faced with higher rates of bleeding. However, the risk of bleeding following full-dose anticoagulation is known to be significantly higher than the risk associated with prophylactic dose. Such fact might not be in the mind of most practicing physicians when making such decisions. Therefore, we need to factor-in and may accept the few episodes of bleeding, mostly minor, encountered in most of VTE prophylaxis trials. Such bleeding, in quantity and frequency, is significantly lower than those associated with full-dose therapeutic anticoagulation, if we wait for the thrombus to happen. Risk assessment models, in high risk tumors like gastric cancer, should calculate the gain of preventing VTE against the net loss due to bleeding in both occasions; therapeutic versus prophylactic dose of anticoagulation.
Our VTE rate of 22% is higher than several previously published studies and can be explained, at least in-part, by the higher percentage of patients with metastatic disease in our cohort (54%), but other contributing factors cannot be ruled out. Several other findings in our study worth highlighting and may help us make decisions on VTE prophylaxis for ambulatory patients: First, 87% of our confirmed VTE episodes were among patients with active disease at time of diagnosis. This is expected given the hypercoagulable state that is known to be induced by adenocarcinomas in general, and gastric cancer in particular 17 . Additionally, patients with active disease are likely to be treated with chemotherapy which is by itself can increase the risk of VTE. Additional factors that may increase such risk include poor performance status and frequent hospitalizations associated with active disease state.
Second, the fact that those with early stage disease who never had any form of chemotherapy had a relatively low rate of VTE (9.9%) might indicate that such patients, without the need to apply any risk assessment model, be considered not at high-enough risk to offer VTE prophylaxis while ambulatory.
Third, patients who had multiple lines of chemotherapy had significantly higher VTE rate than those who only had one. This could be a surrogate marker of disease activity and might also reflect poor performance status and ambulatory level.
Fourth, chemotherapeutic agents used in gastric cancer are also thought to affect the risk of VTE. A prospective controlled trial of 964 patients treated with epirubicin, platinum and fluoropyrimidine combination chemotherapy in stage 3 and 4 gastric cancer found more thromboembolic events in the cisplatin group compared with the oxaliplatin groups (15.1% versus 7.6%; p = 0.0003) 18 . We have previously reported our experience with VTE associated with cisplatin-based chemotherapy in a group of 1677 patients with various solid tumors including 191 patients with gastric cancer. Among the whole group, thromboembolic events were confirmed in 6.6% but was highest (20.9%) in patients with gastric cancer. In multivariate analysis, significantly higher rates of thrombosis were associated with gastric, as the primary tumor, and advanced-stage disease 19 . More recently, the FLOT4 (fluorouracil plus leucovorin, oxaliplatin and docetaxel) trial also showed lower risk of thromboembolic events in the cisplatin-free regimen (FLOT4) compared to other cisplatin-containing regimens 20 . An increase in the incidence of VTE was also observed in patients with gastric cancer receiving other agents like irinotecan or bevacizumab in locally advanced and metastatic settings, reaching up to 25% of patients in the study 21 . The high VTE rate among our cohort may explain the lack of difference between cisplatin-based regimen and others in our study.
Last, despite having gastric cancer among the highest risk tumors in Khorana RAM, the model couldn't separate patients in our study into different risk levels. The high VTE rates among our cohort may explain such finding. Additionally, as defined by Khorana RAM and just because of type of their cancer, patients with gastric cancer are either at intermediate or high risk but not low risk (Table 6).
Our study is not without limitations; we have not addressed underlying illnesses and comorbidities which may increase the rate of VTE. Though our data is derived from a single institution, the large number of patients included minimize selection bias. Our findings should highlight the need for a special RAM for patients with gastric cancer. Such model should take into consideration the stage of the disease, number of lines of chemotherapy and specific chemotherapeutic agents used.