To investigated the incidence and risk factors of venous thromboembolism (VTE) in patients with advanced gastric cancer (AGC) receiving chemotherapy.

All consecutive chemotherapy-naïve patients with AGC who would receive palliative chemotherapy between November 2009 and April 2012 in our hospital were recruited. Their pretreatment clinical and laboratory variables, including D-dimer, were recorded. The frequency of VTE development and survival rates during each chemotherapy cycle and regularly thereafter were assessed.

A total of 241 patients enrolled between November 2009 and April 2012 were analyzed. During a median follow-up duration of 10.8 mo (95%CI: 9.9-11.7), 27 patients developed VTE and the incidence of VTE was 17.5% (95%CI: 10.5-24.0, 12.0 events/100 person-years). The 6-mo and 1-year cumulative incidences were 7.8% (95%CI: 4.2%-11.4%) and 12.4% (95%CI: 7.3-17.2), respectively. Thirteen (48.1%) patients were symptomatic and the other 14 (51.9%) patients were asymptomatic. In multivariate analysis, pretreatment D-dimer level was the only marginally significant risk factor associated with VTE development (hazard ratio = 1.32; 95%CI: 1.00-1.75, P = 0.051).

The incidence of VTE is relatively high in patients with AGC receiving chemotherapy, and pretreatment D-dimer level might be a biomarker for risk stratification of VTE.

In general, patients with cancer carry a significantly higher risk of venous thromboembolism (VTE) than case-control subjects without cancer[1,2]. Gastrointestinal cancers including gastric cancer were ranked third in incidence of VTE, following hematological malignancies and lung cancer[1]. Lee et al[3] reported that patients with advanced gastric cancer (AGC) had a much higher likelihood of developing VTE (24.4%) than patients with lower-stage gastric cancer (0.5%-3.5%), and Blom et al[1] observed that chemotherapy increased the VTE risk 2.2-fold. These findings suggest that VTE might be more common in patients with AGC receiving chemotherapy, who have several potential risk factors of VTE development including cancer, especially highly vulnerable gastric cancer, advanced stage, and chemotherapy. On the other hand, patients with cancer who develop VTE also have shorter survival durations than those who do not develop VTE[4-6]. Activation of hemostasis, as indicated by development of VTE reflects more aggressive tumor biology[7], additionally, VTE development might hinder the continuation of chemotherapy, resulting in poor outcomes. Considering the relatively high incidence of VTE and its impact on survival, information about the VTE is important in patients with AGC receiving chemotherapy. While, the information about VTE in this cohort is not yet clear. Because most of the previous results were retrospectively analyzed with heterogeneous population and included this cohort as a part of small fraction among heterogeneous various groups[8-11].

It has been challenging for oncologists to conduct a practical use of thromboprophylaxis effectively to prevent VTE. The major problem is that the increased rates of complications such as bleeding outweigh its efficacy[12]. Thus, we must more precisely target thromboprophylaxis, especially in gastric cancer, since it has a high risk of bleeding at the endothelial lesion. Therefore, we need to confirm the exact incidence and identify the predictive factors of future VTE in this cohort. To address these issues, we conducted this prospective cohort study to determine the exact incidence and risk factors of VTE in patients with AGC who are undergoing palliative chemotherapy and assess whether VTE development correlates with survival.

The current study assessed the incidence and risk factors of VTE as well as the influence of VTE on survival in AGC patients undergoing palliative chemotherapy. To our knowledge, this is the first prospective study to evaluate VTE in a homogeneous cohort. This study showed a relatively high cumulative incidence of VTE and the role of pretreatment D-dimer as a risk factor for development of VTE. Although the development of VTE was not correlated with poor survival, patients with symptomatic VTE had more interruptions or delays of chemotherapy than those with incidental VTE.

The 6-mo, 1-year, and 2-year cumulative incidences of VTE were 7.8%, 12.3%, and 17.3%, respectively. Our results showed a relatively high incidence of VTE despite AGC being are presentative frequent site in cancer-associated VTE[13]. In our previous report of retrospective analysis of a similar patient population, the 1- and 2-year cumulative incidences of venous VTE were 3.5% and 4.9%, respectively[8]. Other prior studies also reported VTE incidences of 5.3%-11.4%[14-16], which was somewhat lower than our results. The reason for this discrepancy should be preferentially presumed that this study was conducted for homogeneous patients who were high tumor burden of advanced state and receiving palliative chemotherapy. Regarding tumor burden in cancer-associated VTE, Lee et al[3] reported that the VTE incidence increased with stage in gastric cancer patients. Considering that stage is correlated with tumor burden in cancer patients, we can give careful consideration to our results that a history of previous gastrectomy and multiple metastases were slightly related with the development of VTE, but the correlation was not statistically significant. Regarding chemotherapy of cancer-associated VTE, prior studies reported that chemotherapy is another important risk factor for VTE development in cancer patients[1,17]. In the current study, 13 (46.4%) and 20 (71.4%) patients developed VTE within 3 and 6 mo, respectively, which suggests that starting chemotherapy is an important risk factor for VTE development. Blom et al[1] also reported that the risk of thrombosis was the highest in the first 3 mo after the diagnosis of cancer and declined thereafter, which supports the finding of the current study.

In multivariate analysis, pretreatment D-dimer level was the only marginally significant risk factor for VTE development. D-dimer, a general biomarker that globally indicates hemostasis and fibrinolysis activation, is a well-studied biomarker in the diagnosis and management of VTE in non-cancer patients[18]. Khorana et al[17] proposed a predictive model for chemotherapy-associated VTE that included the primary cancer site, pretreatment platelet count, hemoglobin, leukocyte count, and body mass index. However, D-dimer level was not included in this model, which might be because the pretreatment D-dimer level was not routinely checked in the target population. In the subsequent studies, Ay et al[19] reported that a high D-dimer level predicted occurrence of VTE in a variety of cancer patients, and Arpaia et al[20] demonstrated that pretreatment D-dimer was correlated with chemotherapy-associated VTE. However, these studies only included a small number of gastric cancer patients, 35 of 821 (4.3%) and 10 of 124 (8.1%). Moreover, subcohorts of these patients were heterogeneous since the patients were treated variably with surgery or radiotherapy (or even untreated). Thus, the role of D-dimer as a risk factor of VTE in AGC patients receiving chemotherapy remains to be clarified. We demonstrated here that pretreatment D-dimer was indeed an independent risk factor of VTE in this homogeneous cohort. This means that pretreatment D-dimer might be used to more precisely target patients for thromboprophylaxis and lead to enhanced use of thromboprophylaxis in AGC patients treated with palliative chemotherapy. On the other hand, prior established risk factors including the aforementioned Khorana score were not associated with VTE in this study (data not shown).A possible explanation might be the small sample size; however, the unique characteristics of stomach cancer or even ethical characteristics might also contribute.

The occurrence of VTE has been reported to have a significant adverse effect on survival in many studies[3,5,6]. However, we found no statistical difference in survival according to VTE development. The results did not change after adjustment for the presence of symptoms and signs, VTE type, or detection time. The reason for the negative result in the current study may be too small a sample size to detect a difference. On the other hand, due to the short survival of patients with advanced-stage disease, VTE may have a greatly reduced impact on survival. Previous studies reported that an impact of VTE on survival might be somewhat different from that in our study because they mainly included localized stage of cancer[3] or showed a prominent impact of VTE on survival in subgroups of patients with localized cancer[5] or in those with breast cancer, which has a characteristic longer survival duration[6]. Our previous retrospective study with a larger cohort of similar patients also reported that VTE was not a statistically significant factor for survival[8]. As such, whether VTE really has adverse effect on survival in this cohort should be further clarified. Meanwhile, more patients with symptomatic VTE than asymptomatic VTE had chemotherapy interruptions or delays. Although symptomatic VTE did not show poor survival in the subgroup analysis, we should make an effort to detect VTE before symptoms develop.

This study has several limitations. First, we did not evaluate serial measurements of D-dimer; thus, we could not identify the usefulness of the serial changes as a predictive marker for VTE. Considering that the current study showed that D-dimer level at the time of VTE development is significantly increased compared with that at baseline in patients with VTE, serial measurements of D-dimer might detect early changes and predict the development of VTE. Another issue is that the current study did not calculate the proper number of patients, so it might not have been adequately powered to detect statistically significant differences in other characteristics such as risk factors or survival. For these reasons, it is obvious that the present study might not be a confirmative, rather preliminary study for hypothesis generation.Despite these limitations, the current study showed the incidence of VTE and role of pretreatment D-dimer as risk factors in a homogeneous group of AGC patients receiving palliative chemotherapy. Considering the usefulness of D-dimer as a biomarker such as its ease of use and low cost, pretreatment D-dimer might be used as a risk stratification factor for VTE and in patient selection for thromboprophylaxis.

In conclusion, the incidence of VTE is relatively high in patients with AGC receiving chemotherapy, and pretreatment D-dimer level is a risk factor for VTE. Considering the usefulness of D-dimer as a biomarker given its ease of use and low cost, pretreatment D-dimer might be a risk stratification factor for VTE development and patient selection for thromboprophylaxis.