Correspondence to: Patrizia Ferroni, Department of Advanced Biotechnologies and Bioimaging, IRCCS San Raffaele Pisana, Via della Pisana 235, 00163, Rome, Italy, Tel.: +39-06-6613-0425, Fax: +39-06-6613-0407, E-mail: email@example.com
Chemotherapy has been associated with an increased risk of venous thromboembolism (VTE). However, the prevalence of coagulation abnormalities or VTE occurrence as a consequence of different anti-cancer agents or treatment schemes is largely uncharacterized. Thus, this study was aimed at analyzing the impact of different anticancer drugs on the prothrombotic status of cancer out-patients scheduled for chemotherapy. To this purpose, a mono-institutional study was prospectively conducted to monitor serial changes of activated protein C (APC) function in 505 consecutive cancer out-patients with primary or relapsing solid cancer at the start of a new chemotherapy regimen. The results obtained showed that age >65 years (p = 0.01), ECOG performance status (p = 0.01), platinum-based (p = 0.035) and fluoropyrimidine-based regimens (p = 0.008) were independent predictors of an acquired APC resistance during the first chemotherapy cycle. Multivariate model of Cox proportional hazards survival analysis demonstrated that a decline in APC functionality (HR = 2.4; p = 0.013) and platinum-based regimens (HR = 2.2; p = 0.042) were both capable of predicting the occurrence of a first VTE episode during chemotherapy. Indeed, 14% of patients with platinum-associated APC impairment had VTE over a 1-year follow-up, compared to 3% of patients treated with other regimens and in whom APC functionality remained stable (HR = 1.5; p = 0.003). We may, thus, conclude that use of platinum-based regimens is responsible for induction of an acquired thrombophilic condition and represents a predictor for VTE even after adjustment for other risk factors.
Eastern Cooperative Oncology Group performance status
factor V leiden
protac induced coagulation inhibition
Cancer patients have an increased risk of venous thromboembolism (VTE), which is at least 4-fold higher than that of the general population and is often represented by sub-clinical or undiagnosed non-fatal VTE. This risk differs in all cancer patients, or even in the same patient over the course of cancer natural history, being highest in the first 3 months after initial diagnosis, declining during the first year and again after 3 years. In this regard, the high rates of VTE observed during the first year after diagnosis have been possibly related to the administration of combined anticancer therapies.
Chemotherapy, indeed, has been associated with a 2- to 6-fold increased VTE risk, especially in the first 3–6 months of treatment. Among the various mechanisms invoked to explain chemotherapy-induced prothrombotic state [see Ref.  for review], dysfunctional alterations of the protein C (PC) pathway may account for an altered balance between procoagulant factors and endogenous anticoagulants. Acquired activated PC (APC) resistance is a common finding among patients with solid tumors, representing a more important risk factor for VTE in cancer than in nonmalignant conditions.[5-7] A significant decline in functional PC activity has been, in fact, demonstrated at mid-therapy,[8, 9] leading to acquired APC resistance,[10, 11] which was predictive of VTE and completely reversed at the end of chemotherapy.[11, 13] The role of chemotherapy as a trigger for coagulation activation and VTE onset was further confirmed in unselected cancer out-patients by monitoring early changes of APC function during the first chemotherapy cycle rather than over a single baseline determination.
Despite these findings, the prevalence of an acquired APC resistance as a consequence of different anticancer agents or treatment schemes is largely uncharacterized. Thus, the aim of this study was to analyze the impact of anticancer drugs on the prothrombotic status of cancer out-patients scheduled for chemotherapy. To this purpose, a mono-institutional study was prospectively conducted to monitor serial changes of APC function in a population of 505 unselected cancer patients representative of a general practice cohort.
Material and Methods
A cohort of 505 consecutive patients with primary or relapsing/recurrent solid cancers, receiving clinical care (6% neoadjuvant, 31% adjuvant and 63% metastatic treatments) in the outpatient department setting of the Medical Oncology Unit of “Tor Vergata” Clinical Center was enrolled between January 2007 and December 2010. Patients were treated for adenocarcinoma (90%), sarcoma (3%), melanoma (2%), small cell lung cancer (3%) or undifferentiated cancer (2%). Patients were required to be at the start of a new chemotherapy regimen. Eligibility criteria were: histologically confirmed diagnosis of malignancy; age >18 years; no previous VTE; absolute neutrophil count ≥2,000 mm−3, platelet count ≥100,000 mm−3, haemoglobin level ≥9.5 g/dl; bilirubin level ≤1.5× upper normal limit (UNL), alanine-aminotransferase and aspartate-aminotransferase ≤2.5× UNL in the absence, or ≤5× UNL in the presence of liver metastasis; normal renal function. Patients' characteristics are summarized in Table 1.
All patients were regularly seen at their scheduled chemotherapy visits or at the occurrence of clinically suspected VTE. All patients were followed-up for a median period of 11.2 months. Deep venous thrombosis (DVT) was confirmed by venography or color-coded duplex sonography (in proximal DVT only). Pulmonary embolism (PE) was established by spiral computed tomography. Additional exclusion criteria were therapeutic doses of any heparin before enrollment or concomitant treatment with anticoagulant or antiplatelet drugs. No patient received prophylactic treatment with any anticoagulant after chemotherapy start. No patient underwent surgery during follow-up, nor was admitted to clinic for acute medical illness requiring thromboprophylaxis. The study was performed in accordance with the Declaration of Helsinki. All patients gave written informed consent, previously approved by our Institutional Ethics Committees.
Blood samples were obtained in all patients prior to chemotherapy start (T0) and before the second cycle (T1). More than 90% of the patients received chemotherapy on a monthly basis. Accordingly, the elapsed time between T0 and T1 was 28 days. Additional blood samples were withdrawn in 51 consenting patients before the start of the third (T3) and the sixth (T6) cycle, i.e. after ∼3 and 6 months from chemotherapy start, to confirm in our setting the pre-existing evidence of a decline in functional PC activity at mid-therapy.[8, 9]
Routine hematology, chemistry and coagulation studies were performed on fresh samples at each time point. Blood was processed, aliquoted and stored at −80°C in the facilities of the Inter-Institutional Multidisciplinary Biobank (BioBIM) of the IRCCS San Raffaele Pisana, Rome. APC functionality was evaluated using the HemosIL ThromboPath assay (kindly provided by Instrumentation Laboratory (IL), Orangeburg, NY) on an ACL-TOP coagulometer (IL, Lexington, MA) as previously reported.[14, 16] A locally-defined cutoff was set at 77% Protac Induced Coagulation Inhibition (PICI%) and all values below were defined as pathologic.[14, 16] For repeated ThromboPath measures during the first chemotherapy cycle, percent changes were calculated as previously reported.
Data are presented as percentages, mean ± SD, or median and IQR. Differences between percentages were assessed by chi-square test. Student's t test, ANOVA test and Pearson correlation analysis were used for normally distributed variables. Appropriate nonparametric tests (Mann–Whitney test, Kruskal–Wallis test and Spearman correlation test) were employed for all other variables. Differences during chemotherapy were analyzed by the Wilcoxon signed-rank test and Friedman ANOVA test. Logistic regression analysis was performed to quantify the relationship between clinical and biochemical variables. Survival curves were calculated by the Kaplan-Meier and log-rank methods. Cox-proportional hazards analysis was used to evaluate the association between clinical variables and time-to-event (TTE). TTE was calculated from the date of enrollment until the event date or the study end. For patients receiving neoadjuvant chemotherapy (all breast cancer) follow-up was stopped at completion of an entire antiblastic treatment and before surgery. Calculations were performed using a computer software package (Statistica 8.0, StatSoft, Tulsa, OK) or free web-based applications (http://statpages.org/).
In agreement with previous findings,[14, 16] global APC functionality was impaired in ∼27% of cases with a mean pretreatment ThromboPath value of 79.9 ± 11.6 PICI%. Of interest, the distribution of APC abnormalities among patients who were chemotherapy naïve at time of study entry (82% of the whole population) or who received prior chemotherapy (18%) were 27% and 29%, respectively, suggesting that impaired APC functionality was related to cancer itself and not to previous treatment. Conversely, a significant decline of APC functionality was observed in the overall population after one cycle of chemotherapy (mean ThromboPath value: 78.0 ± 11.8 PICI%, p = 0.0001). Furthermore, subgroup analysis of the 51 patients who consented to repeated blood withdrawal showed that APC functionality was progressively impaired during the first 3 months of chemotherapy (82.3 ± 8.9 PICI% at T0 vs. 78.4 ± 12.2 PICI% at T1 vs. 78.7 ± 11.5 PICI% at T3), but reverted to baseline levels by the sixth month (84.9 ± 8.6 PICI% at T6; Friedman's ANOVA among the four study points: p = 0.008), thus confirming previous observations.[8, 9]
Median percent change of APC functionality after one cycle of chemotherapy in the overall population was −0.91 (IQR ranging from −5.9 to 3.4), the decrease being more pronounced in those patients who developed VTE during chemotherapy (T0 vs. T1: 79.2 ± 9.1 vs. 74.3 ± 10.1 PICI%, p < 0.0001), with a median percent change of −6.0 (IQR ranging from −10.2 to −3.4). On the basis of value distribution (median, IQR, 10th and 90th percentiles), patients were then categorized into six categories, with category six representing the strongest evidence of positivity (e.g., that impairment of APC functionality was present). Resulting analysis demonstrated that the −6% cut-off had a sensitivity of 72%, a specificity of 50% and a fitted ROC area of 0.66 (SE = 0.04). Hence, patients were categorized as having stable or decreasing ThromboPath values based on this arbitrary cutoff.
To assess the possible determinants of decreasing ThromboPath values among clinical variables and type of chemotherapy/supportive drug used, a logistic regression analysis was performed including gender, age, tumor site, presence of metastasis, ECOG performance status and the type of chemotherapy/supportive drug used as the predictor variables. The results showed that advanced age (>65 years, p = 0.010), ECOG performance status (p = 0.010), platinum-based (p = 0.035) and fluoropyrimidine-based regimens (p = 0.008) were independent predictors of an acquired APC resistance during the first chemotherapy cycle (Table 2).
Table 2. Logistic regression analysis of the predictive value of clinical-pathological variables and type of chemotherapy on the decline of APC function after the first cycle
OR (95% C.I.)
Site of primary tumor
Erythropoietin stimulating agents
Antityrosine Kinase Inhibitors
VTE occurred during chemotherapy in 7% of patients (12 PE and 23 DVT; median TTE: 3.4 months), in agreement with previous reports.[1, 2] Fourteen of 35 patients were incidentally diagnosed with asymptomatic VTE (9 PE) at time of restaging. No patient had asymptomatic VTE on outset as confirmed by restaging procedures. Multivariate model of Cox proportional hazards survival analysis demonstrated that a decline in APC functionality (HR = 2.4; 95%CI: 1.2–4.8; p = 0.013), ECOG performance status (HR = 4.3 95%CI: 2.2–8.4; p < 0.0001) and platinum-based regimens (HR = 2.2; 95%CI: 1.0–4.7; p = 0.042) were capable of predicting a first VTE episode, whereas fluoropyrimidine-based chemotherapy did not show any association (p = 0.284) (Table 3). Figure 1 reports the Kaplan–Meier curves for patients stratified on the basis of ThromboPath percent change during the first cycle of chemotherapy employing or not platinum compounds. As shown, 14% of patients with platinum-associated APC impairment had VTE over a 1-year follow-up, as compared to 3% of patients treated with different regimens and in whom APC functionality remained stable. Univariate Cox proportional hazards survival analysis of the combination of impaired APC function and platinum administration showed an increase of VTE risk with a HR = 1.5 (95%CI: 1.2–2.1, p = 0.003) for increasing category.
Measurement of thrombin generation is a promising approach to detect an individual's coagulation potential, being able to quantify the composite effect of multiple thrombotic risk factors and to predict a prothrombotic state. Here, a cohort of 505 unselected cancer patients was prospectively studied to monitor changes of PC function during chemotherapy using a recently developed APC-dependent thrombin-generation assay. As expected, we observed an impairment of APC function prior to therapy, which is consistent with the occurrence of a cancer-associated prothrombotic condition, most likely arising from tumor cell expression and release of procoagulant activities. This possibility was also suggested by the lack of difference in APC abnormalities between patients who were chemotherapy naïve at time of study entry or who received prior chemotherapy.
The finding of a pretreatment APC resistance, raises the question of whether this represents an acquired condition (associated with clinical-pathological features of cancer) or the result of an interaction with genetic factors in determining coagulation abnormalities and resultant VTE under the trigger of chemotherapy. Data on inherited thrombophilia were available for 314 patients (data not shown). As expected, impaired APC function was found in carriers of both Factor V Leiden (4.8%) or Prothrombin mutation (2.3%), but neither was associated with VTE risk, in agreement with previous observations.[16, 18]
The major result of this study, however, was the finding of an impairment of APC function during the first cycle of chemotherapy, which was more pronounced in patients who developed VTE. Of interest, the result obtained in a small subgroup of patients showed that these changes peaked during the first 3 months and reverted to normal by the sixth month of chemotherapy, which is consistent with data from Refs. [8, 9]. Moreover, we demonstrated that both platinum-based and fluoropyrimidine-based regimens were capable of independently predicting a decline of APC functionality and, thus, the occurrence of an acquired thrombophilic condition possibly responsible for VTE development during chemotherapy. The mechanisms by which platinum compounds may promote thrombosis are generally related to increased production of reactive oxygen species leading to oxidative stress that will be ultimately responsible for endothelial dysfunction, increased TF activity on monocytes and platelet activation [reviewed in Ref. ]. Conversely, fluorouracil infusion has been associated with a reduction in protein C levels, although coronary spasm represents the leading mechanisms for its cardiotoxicity. These considerations might explain why only platinum-based regimens were predictors of a first VTE episode during chemotherapy in this study. Accordingly, we might speculate that the effects of platinum-related oxidant stress are responsible for a more generalized impairment of vascular homeostasis as compared to fluoropyrimidine, involving all the components of the Virchow's triad leading to VTE.
Our results are in partial agreement with those obtained in the PROTECHT study on a population of 391 cancer out-patients receiving chemotherapy for metastatic or locally advanced solid tumors, but no thromboprophylaxis (placebo group), in which the highest rate of VTE was associated with gemcitabine (8.1%) or cisplatin (7.0%), but only in 3.3% of patients treated with 5-fluorouracil The addition of gemcitabine to platinum compounds increased the VTE rate to 10%. Consistent with these observations, a modified Khorana risk assessment score (the Protecht score), which added platinum or gemcitabine-based chemotherapy to the predictive variables, was subsequently proposed.
In the present study, VTE occurred in 9.4 and 8.5% of patients receiving platinum- or fluoropyrimidine-containing regimens, respectively. Gemcitabine combination with platinum compounds did not show any additive effect on VTE incidence as compared to gemcitabine alone (3.9% vs. 5.5%, p = 0.35). However, the low number of patients included in this subgroup analysis must be regarded as a possible limitation when discussing differences with previous studies. The prominent role of platinum as an independent predictor for VTE was evidenced in a multivariate analysis after adjustment for other risk factors including functional impairment of APC. As stated above, the wide spectrum of vascular complications associated with platinum-based chemotherapy is well documented. Moreover, an unacceptable VTE incidence in cancer patients receiving cisplatin-based chemotherapy has been recently demonstrated. Here, we provide evidence on the occurrence of an acquired APC resistance during administration of platinum compounds, which represents and independent predictor of VTE development during chemotherapy.
One possible concern that could be raised is whether VTE occurrence in certain treatment schedules might be related not only to the drug itself but also to the tumor type that is being treated with that specific drug. Nonetheless, in this study 90% of patients had a histological diagnosis of adenocarcinoma, while other histotypes were represented in negligible percentages. Moreover, the site of primary tumor, more than histotype, has been associated with VTE rates, to the point that the Khorana score assigns different points to different cancer sites. This is in agreement with previous findings demonstrating that the extent of thrombin generation and the formation of thrombin/antithrombin complexes did not differ among histological subtypes of lung cancer (squamous or adenocarcinoma).
Although this study may be limited by a relatively small sample size, the results here reported demonstrate that platinum-based regimens are associated with an acquired thrombophilic condition and represent a predictor for VTE even after adjustment for other risk factors (including a decreased APC functionality). All current consensus guidelines by multiple cancer organizations do not recommend routine prophylaxis for the primary prevention of VTE in ambulatory cancer patients receiving chemotherapy (grade 1C). However, further characterization of platinum-associated prothrombotic changes could be useful for stratifying VTE risk in cancer out-patients. Future multicenter prospective studies specifically designed to address this issue will help to improve our knowledge on treatment-associated risks and their impact on patient's quality of life.
The authors wish to thank Dr. Antonia Melino for her excellent technical assistance. This work has been performed within the PhD Programs XXVI and XXVII Ciclo. All authors report no potential conflicts of interest with any company/organization whose products or services may be discussed in this article.