The association between cancer and thrombosis is well known, and the occurrence of thrombotic complications in patients with malignancies commonly correlates with poor prognosis . The protein C pathway is the main natural in vivo anticoagulant mechanism, and abnormalities such as acquired activated protein C (APC) resistance and congenital APC resistance, mostly resulting from the factor V Leiden G1691A mutation, are major risk factors for venous thromboembolism (VTE), with odds ratios (ORs) of 2.5 and 2.0–7.0, respectively [2-4].
Our previous study demonstrated the association of acquired APC resistance (APC resistance without FV Leiden) with VTE in cancer patients . Moreover, high levels of FVIII, FV, FII and FX were found to correlate with acquired APC resistance in cancer patients, which could explain the plasmatic mechanism of thisphenomenon .
In addition to its anticoagulant activity in malignancy, APC induces direct cellular effects that could promote cancer cell migration, invasion, and angiogenesis, and inhibit cancer cell apoptosis [7-9].
Another important component of the protein C pathway is the endothelial protein C receptor (EPCR), which binds protein C and enhances the rate of its activation . A soluble form of EPCR (sEPCR) can be detected in plasma, probably resulting from shedding of membrane EPCR. sEPCR binds protein C and APC with similar affinity, and inhibits APC anticoagulant activity by blocking its interaction with negatively charged phospholipids . High levels of sEPCR were found to be associated with the EPCR A6936G polymorphism, and appeared to constitute a moderate risk factor for VTE, with an OR of 1.8 .
The aims of this prospective study were to determine a possible prognostic value of hypercoagulable parameters, especially those of the anticoagulant protein C pathway, for the survival of patients with non-small cell lung carcinoma (NSCLC), and to compare them with known predictors, such as gender, weight loss, disease stage, performance status, histological subtype, leukocytosis, anemia, and lactate dehydrogenase (LDH) levels .
The study was approved by the Rambam Institutional Review Board (approval number 2092), and by the Israel Ministry of Health (approval number 920051312), and is registered on the NIH ClinicalTrials.gov website (ClinicalTrials.gov ID: NCT00192829).
One hundred and two patients with newly diagnosed NSCLC who were referred to the Rambam Oncology Institute were eligible to participate in the study, according to the following criteria: (i) histologically or cytologically confirmed diagnosis of NSCLC; (ii) presence of active malignancy; (iii) no prior malignancy apart from treated basal cell or squamous cell carcinoma of the skin; (iv) no prior chemotherapy or radiotherapy; (v) no major surgery during the last month before investigation; (vi) no anticoagulant therapy during the last month before investigation; (vii) no evidence of active infectious disease; (viii) normal serum bilirubin; (ix) serum transaminases ≤ 2.0 times the upper limit of normal; (x) serum creatinine level of ≤ 1.5 mg dL−1; and (xi) availability of signed informed consent.
The patients' ages ranged between 43 and 90 years (median: 66 years); 77 patients were male. The majority (61%) had stage IIIB or IV disease. First-line therapy included a carboplatin-based combination in most of these patients (78%).
Before the start of therapy, blood samples were collected into 3.2% sodium citrate tubes. Fibrinogen and D-dimer determination, protein C global assay and APC resistance tests were performed on fresh plasma samples. All other coagulation assays were performed on thawed frozen plasma samples with ELISA. A detailed description of the methods is provided in Data S1. The FV Leiden G1691A mutation was assessed in patients with APC resistance (APC sensitivity ratio [APC-SR] < 2.15). The complete blood count was determined and routine blood chemistry investigation was performed.
Kaplan–Meier survival analysis (log-rank test) was used to identify the covariates that were significantly associated with survival. All parameters were analyzed with the median, high or low cut-off level of the normal range for each parameter. Survival duration was determined as the time from entering the study to the death of the patient. The variables that were found by Kaplan–Meier analysis to be significantly associated with survival (i.e. platelet count > 400 × 103 μL−1, fibrinogen > 535.5 mg dL−1, hemoglobin > 10 g dL−1, white blood cell count > 11 × 103 μL−1, D-dimer > 0.5 mg L−1, sEPCR > 110 ng mL−1, polymorphism EPCR 6936 AG + GG, weight loss ≥ 5%, and male gender) were further assessed in the multivariate Cox regression analysis. At the time of survival analysis, only 17 patients (16.6%) remained alive. The median follow-up periods were 325 days for the 85 patients who died during the study and 1393 days for the 17 survivors.
Within the first 6 months after enrollment, seven patients (6.8%) developed VTE: three developed pulmonary embolism, three had upper extremity deep vein thrombosis, and one was diagnosed with portal vein thrombosis.
The median survival time of these seven patients was significantly shorter than that of patients who did not develop VTE during the first 6 months post-enrollment (198 ± 78 vs. 466 ± 46 days, P = 0.03).
A hypercoagulable state was observed in most patients: 78% had high levels of D-dimer (> 0.5 mg L−1, the upper level of normal), and 70% of patients had abnormal results of the protein C global assay, which were expressed as the protein C activation time normalized ratio (PCAT-NR), below the cut-off level of 0.8. A PCAT-NR of < 0.8 demonstrates low activity of the protein C pathway, and could be considered to be a risk factor for VTE . In addition, 31 patients (30%) had APC resistance; only six of them were carriers of the FV Leiden mutation. APC resistance and abnormal results of the protein C global assay, analyzed with or without FV Leiden, were not found to be associated with survival (Table 1), regardless of whether they were calculated according to a median level, a level below the lower cut-off (APC-SR < 2.15, PCAT-NR < 0.8), or a level above the lower cut-off .
|Parameter||Cut-off||Patients||Survival time (days, Kaplan–Meier test)|
|Platelet count||< 400 × 103 μL−1||77||76||478||26||0.015|
|> 400 × 103 μL−1||24||24||242||48|
|Fibrinogen||≤ 535.5 mg dL−1||51||50||503||35||0.005|
|> 535.5 mg dL−1||51||50||261||19|
|Hemoglobin||≤ 10 g dL−1||5||5||149||106||< 0.0001|
|> 10 g dL−1||96||95||466||42|
|WBC count||≤ 11 × 103 μL−1||70||69||483||25||0.009|
|> 11 × 103 μL−11||31||31||261||36|
|D-dimer||< 0.5 mg L−1||23||22||744||323||0.006|
|> 0.5 mg L−1||79||78||388||65|
|F1+2||≤ 192.0 pmol L−1||51||51||351||94||0.7|
|> 192.0 pmol L−1||49||49||470||66|
|TF||≤ 32 pg mL−1||50||50||459||53||0.7|
|> 32 pg mL−1||49||50||434||96|
|Protein C global assay (PCAT-NR)||< 0.8||71||70||425||66||0.3|
|sEPCR||≤ 110 ng mL−1||52||52.5||287||40||< 0.0001|
|> 110 ng mL−1||47||47.5||520||130|
|Polymorphism EPCR 6936||AA||79||79||351||72||0.008|
|AG + GG||21||21||1464||866|
However, a high sEPCR level was positively related to survival. The median survival was significantly longer in patients with sEPCR > 110 ng mL−1 than in those with sEPCR ≤ 110 ng mL−1 (Table 1; Fig. S1).
sEPCR levels in NSCLC patients did not significantly differ from those in matched controls (146.6 ± 94.6 ng mL−1 vs. 135.3 ± 86.5 ng mL−1, respectively, P = 0.43).
As in the controls, sEPCR levels in NSCLC patients were distributed trimodally according to the EPCR A6936G polymorphism. This is in line with findings of previous studies [12, 15].
Remarkably, polymorphisms EPCR 6936 AG and GG, which are associated with high levels of sEPCR, were found to correlate with prolonged survival in NSCLC patients (Table 1; Fig. S1).
Additionally, high platelet counts and high fibrinogen and D-dimer levels were found, by univariate analysis, to be associated with shorter survival (Table 1). Notably, levels of tissue factor and the coagulation activation marker F1+2 were not related to survival time (Table 1). Moreover, of the evaluated clinical parameters, only male gender and a loss of ≥ 5% of body weight were associated with shorter survival.
The multivariate analysis revealed five significant and independent predictors of survival: weight loss ≥ 5%, male gender, D-dimer > 0.5 mg L−1 and fibrinogen > 535.5 mg dL−1 were found to be associated with shorter survival, whereas sEPCR > 110 ng mL−1 appeared to be related to longer survival.
Certain prognostic factors were previously reported to predict survival duration in patients with NSCLC. Favorable prognostic factors included diagnosis at an early stage of the disease, good performance status, the absence of significant weight loss, and female gender , whereas high levels of LDH, leukocytosis, anemia, the presence of bone and liver metastases and VTE were considered to be poor prognostic factors [16, 17]. The lack of associations of survival with disease stage, performance status or LDH levels, demonstrated in the present study, could be explained by the small size of the patient group. In addition to the known predictors of poor prognosis, the current study has established high platelet count and high fibrinogen and D-dimer levels as unfavorable prognostic factors in NSCLC patients.
Remarkably, our study has shown a high level of sEPCR to be a new independent predictor of longer survival in cancer patients. This finding is intriguing at first sight, given that high sEPCR levels are known to constitute a moderate risk factor for VTE in non-cancer patients .
The results obtained demonstrated no correlation between the APC-SR, protein C global assay, and sEPCR levels. This suggests that the association between sEPCR levels and survival in NSCLC patients is mediated by mechanisms other than the protein C pathway anticoagulant activity. Although the exact underlying mechanisms are still unknown, one plausible explanation for this association could be APC cell signaling, as APC has been reported to promote cancer cell migration and invasion, and to inhibit apoptosis through EPCR and protease-activated receptor-1 . Given that sEPCR in plasma inhibits APC, high levels of sEPCR could protect against the prometastatic effects of APC [11, 18-20]. However, high levels of EPCR on tumor cells, determined by immunohistochemistry, have lately been suggested to be associated with a poor prognosis in stage I lung carcinoma patients . Therefore, the role of the protein C pathway in malignancy, mediated through either cell signaling or anticoagulant function, and its potential impact on disease progression and survival remain to be further explored.
The results of this study should be interpreted with caution, owing to the relatively small group size and the fact that the assays were performed only once during the study. sEPCR in plasma could originate from either vascular endothelium or tumor cells; however, its exact source needs to be investigated.
Assessment of the protein C pathway has shown sEPCR to be a new, independent predictor of survival in lung cancer patients. As high levels of sEPCR are often associated with the congenital EPCR polymorphism A6936G, which has been shown to be related to survival duration, further studies of the role of EPCR polymorphisms and their expression in cancer are warranted.