- Email: [email protected]
Venous thromboembolism predicts poor prognosis in irresectable pancreatic cancer patients
original article
Annals of Oncology 18: 1660–1665, 2007 doi:10.1093/annonc/mdm284 Published online 28 July 2007
Venous thromboembolism predicts poor prognosis in irresectable pancreatic cancer patients M. Mandala`1*, M. Reni2, S. Cascinu3, S. Barni4, I. Floriani5, S. Cereda2, R. Berardi3, S. Mosconi1, V. Torri5 & R. Labianca1 1 Unit of Medical Oncology, Ospedali Riuniti, Bergamo; 2Division of Medical Oncology, San Raffaele Hospital, Milan; 3Clinica di Oncologia Medica, Universita` Politecnica delle Marche; Azienda Ospedaliero-Universitaria, Ospedali Riuniti di Ancona, Umberto I, Lancisi, Salesi, Ancona; 4 Division of Medical Oncology, Treviglio Hospital, Treviglio; 5Department of Oncology, Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy
Received 15 February 2007; revised 10 May 2007; accepted 14 May 2007
Background: The aim was to investigate the outcomes associated with venous thromboembolism (VTE)
original article
clinical trials between December 2001 and December 2004 in order to evaluate the prognostic impact of symptomatic VTE on clinical outcomes, such as response to treatment, progression-free survival (PFS) and overall survival (OS). Results: Among 227 irresectable pancreatic cancer patients, with Eastern Cooperative Oncology Group performance status (ECOG-PS) £ 2, 59 (26.0%) patients developed a VTE. A synchronous VTE occurred in 28 (12.3%) patients, while a VTE during chemotherapy was observed in 15 (6.6%) patients, and 16 (7.0%) patients experienced both events. Presence of synchronous VTE was associated with a higher probability of not responding to treatment (odds ratio 2.98, 95% CI 1.42–6.27, P = 0.004), but showed no effect on both PFS and OS at least at multivariate analysis. Occurrence of a VTE during chemotherapy showed a statistically significant effect on PFS (hazard ratio [HR] 2.59, 95% CI 1.69–3.97, P < 0.0001) and OS (HR 1.64, 95%CI 1.04–2.58, P = 0.032). Conclusions: Our data suggest that the occurrence of VTE may be associated with a reduced response rate and a shorter PFS and OS among patients with irresectable pancreatic cancer. In these patients the development of VTE may reflect the presence of a biologically more aggressive cancer that in turn leads to a worse prognosis. Key words: venous thromboembolism, prognosis, pancreatic cancer
introduction The strict two-way clinical relationship between cancer and venous thromboembolism (VTE) has been evaluated since Trousseau’s time. While subclinical coagulation activation is encountered in up to 90% of cancer patients, only 4% to 15% of them develop VTE [1]. The relationship among cancer, chemotherapy and thrombosis has been more extensively investigated in breast cancer patients. Few data are available for gastrointestinal (GI) cancer patients receiving chemotherapy, in spite of the general perception of a high incidence of thromboembolic complications in patients with mucinous carcinoma of the pancreas and GI tract. A recent retrospective study on patients with different solid tumors found a remarkably high incidence (7%) of VTE in a group of 39 patients with colon cancer during chemotherapy [2]. *Correspondence to: Dr M. Mandala`, Unit of Medical Oncology, Ospedali Riuniti Bergamo, Largo Barozzi 1, 24100 Bergamo, Italy. Tel: +39-035-269858; Fax: +39-035-266849; E-mail: [email protected]
ª 2007 European Society for Medical Oncology
Pancreatic carcinoma is currently the fourth biggest leading cause of cancer-related death in the USA. About 10%–15% of patients undergo complete tumor resection and of this proportion fewer than one-fifth are alive 5 years after surgery. In locally advanced or metastatic disease, systemic chemotherapy induces responses in only a few patients. The main clinical benefit of chemotherapy is generally for the palliation of pain and cancer-related symptoms [3]. Experimental studies in vitro or in animal models support the hypothesis that the inhibition of blood coagulation may interfere with the progression of malignancy. Experimental and clinical data to support the hypothesis that anticoagulation may influence prognosis of cancer patients have been recently reviewed [4, 5]. There is some evidence that patients with cancer who develop thromboembolism have poorer survival [6, 7]. However most of these studies used population-based data to estimate the incidence and the prognosis of cancer patients, lacking information about performance status and type of treatment, which could influence survival. The objective
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
among irresectable pancreatic cancer patients.
Methods: This is a follow-up study of consecutive irresectable cancer patients, treated and followed up in
original article
Annals of Oncology
of this study was to describe the prognostic role of VTE in a cohort of irresectable pancreatic cancer patients.
materials and methods patients and study design
statistical methods All patients satisfying eligibility criteria and treated with at least one cycle of chemotherapy were considered for analysis. Progression-free survival was defined as the time from the beginning of chemotherapy to first appearance of progressive disease or death for any cause; patients known to be alive and without progressive disease at the time of analysis were censored at their last available follow-up assessment. Overall survival was defined as the time from the beginning of chemotherapy to the date of death from any cause or the date of the last follow-up. Survival curves were estimated using the Kaplan–Meier method. Cox proportional hazards models as implemented in the PHREG program (Statistical Analysis System [SAS] Institute, Cary, NC, USA) were used to estimate and test demographic characteristics and clinical features for their association with PFS and OS. Venous thromboembolism was included both irrespective of the time of occurrence and by distinguishing
Volume 18 | No. 10 | October 2007
results A total of 227 consecutive patients were identified and included in the analysis. Demographic and clinical characteristics of the cohort are summarized in Table 1. Just over half (121, 53.3%) were males and the median age was 66 years, ranging from 38 to 82 years. Forty-four (19.4%) patients had undergone previous surgery, either for curative or palliative intent. Most (63.4%) patients had metastatic disease, while the remaining were classified with locally advanced irresectable disease. Chemotherapy treatment was PEFG for 79 (37.4%) patients; the remaining patients received chemotherapy with gemcitabine or cisplatin and gemcitabine. We excluded five patients who had not received at least one dose of chemotherapy: four patients because of inadequate liver function and one patient because of a concomitant other cancer. Table 1. Characteristics of the study cohort (n = 227) Parameter Agea (years) Gender Male Female Stage of disease Locally advanced Metastatic Previous surgery No Yes Chemotherapy PEFG Other VTE None Synchronous only During chemotherapy only Both
No. of patients 63 (38–82) 121 (53.3%) 106 (46.7%) 83 (36.6%) 144 (63.4%) 183 (80.6%) 44 (19.4%) 79 (34.8%) 148 (65.2%) 168 28 15 16
(74.0%) (12.3%) (6.6%) (7.0%)
a
Median and range.
doi:10.1093/annonc/mdm284 | 1661
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
The cohort was identified by inspecting the electronic database of all patients attending four Italian institutions from December 2001 to December 2004. An additional search was performed using the International Classification of Diseases, 9th revision, clinical modification (ICD9CM) registration systems. We retrieved data concerning clinical outcomes and VTE from patients enrolled in two prospective trials, one sponsored by the Italian Group for the Study of Digestive Tract Cancer (GISCAD) and the other conducted by the San Raffaele Hospital, Milan [8–9]. In addition we retrieved data from an ongoing GISCAD-sponsored Phase III randomized study comparing gemcitabine with cisplatin and gemcitabine in irresectable pancreatic cancer patients. Also included were 27 patients treated with a Cisplatin, epirubicin, 5-fluorouracil, gemcitabine (PEFG) polychemotherapy at the San Raffaele Hospital and not enrolled in the prospective trial previously reported [9]. The study protocols were approved by the local ethical committee and all patients gave informed consent. Eligibility criteria for inclusion in this study included histological diagnosis of locally advanced unresectable and/or metastatic pancreatic carcinoma, treatment with a gemcitabine-based chemotherapy, an Eastern Cooperative Oncology Group performance status (ECOG-PS) £ 2 and normal hepatic (serum bilirubin < 1.5 mg/dL), renal (serum creatinine < 1.5 mg/dL) and bone marrow (leukocyte count > 4000/lL, platelet count > 100000/lL) functions. Exclusion criteria included prophylaxis with warfarin or low molecular weight heparin (LMWH). Information on age, gender, histopathology and surgical and medical treatment were retrieved for each patient, as well as on tumor overall response, progression-free survival (PFS) and overall survival (OS). Stage was assessed according to the American Joint Committee on Cancer (Tumour, Node, Metastasis) classification. Tumor response was classified according to World Health Organization criteria. The medical charts and the radiological history of all patients were checked with regard to ultrasonography of the extremities, chest and abdominal computed tomography (CT) scan and perfusion/ventilation lung scan. In all patients with proven VTE, we actively searched the records and radiological reports for indicators of tumor load at the time of the event. The criterion to diagnose VTE by compression ultrasonography was non-compressibility of a proximal vein. When symptoms suggestive for pulmonary embolism developed, a radionuclide lung scan, a CT scan or both was performed.
between occurrence before or during chemotherapy. Overall VTE occurrence was not included in the multivariate models due to multicolinearity problems and because the other two VTE variables were considered more informative. A proportional hazards model was used for univariate and multivariate analysis. Variables found to be associated with PFS and OS in the univariate analysis (P < 0.10) were considered for the multivariate analysis. A stepwise strategy was then adopted, with a statistical significance cut-off for variable screening of 0.05, and VTE was forced into the model. Results are expressed as hazard ratios (HR) with 95% confidence intervals (CI) reported for each factor. A univariate and multivariate logistic regression model using the same approach was used to assess the effect of the same variables on treatment response, expressing results as odds ratios (OR) and relative 95% CI. With a total of 210 events in our series, this study had a 95% power of detecting a hazard ratio of at least 2.0 associated with the group with VTE (expected to have a prevalence of about 20% [10]) with a type I error of 5%, for a bilateral test. The statistically significant level was set at P £ 0.05. Analysis was performed using SAS (Version 9.0) software.
original article
Annals of Oncology
occurrence of VTE during chemotherapy (HR 3.04, 95% CI 2.12–4.36, P < 0.0001) and overall VTE (HR 2.78, 95% CI 2.07–3.72, P < 0.0001) were associated with a shorter PFS, whereas any 5-year increase of age resulted in a lower risk of progression (HR 0.92, 95% CI 0.85–0.99, P = 0.038). Chemotherapy other than PEFG was not significantly associated with PFS (HR 1.30, 95%CI 0.98–1.72, P = 0.064) but was nevertheless included in the multivariate model since the criterion of P < 0.10 for remaining in the model was met. At multivariate assessment, stage IV (HR 1.78, 95% CI 1.32–2.39, P = 0.0002), 5-year age increase (HR 0.92, 95% CI 0.86–0.99, P = 0.044), chemotherapy (HR 1.39, 95% CI 1.03–1.86, P = 0.030) and VTE during chemotherapy (HR 2.62, 95% CI 1.73–3.97, P < 0.0001) reached statistical significance. As for OS, stage IV (HR 1.95, 95% CI 1.46–2.61, P < 0.0001), synchronous VTE (HR 1.45, 95% CI 1.03–2.04, P = 0.031), occurrence of VTE during chemotherapy (HR 1.95, 95% CI 1.32–2.87, P = 0.0008) and overall VTE (HR 1.76, 95% CI 1.30–2.40, P = 0.0003) showed a prognostic negative effect at univariate analysis. In the multivariate model, only the effects of stage (HR 1.95, 95% CI 1.46–2.61, P < 0.0001) and VTE during chemotherapy (HR 1.64, 95% CI 1.04–2.58, P = 0.032) were confirmed to have a prognostic role on survival Table 3. Figure 1 shows the Kaplan–Meier curves for PFS and OS according to the presence of synchronous VTE and the occurrence of VTE during chemotherapy.
discussion The most striking finding of this study is that the development of VTE seems to predict a poor prognosis in irresectable pancreatic cancer patients. To our best knowledge this is
Table 2. Univariate and multivariate logistic regression models assessing the association between VTE and response to treatment (Progressive Disease vs other) Parameter
Gender Female Male Age (for each 5-year increase) Stage III IV Chemotherapy PEFG Other Synchronous VTE No Yes VTE during chemotherapy No Yes Overall VTEa No Yes
Analysis Univariate OR (95% CI)
P value
Multivariate OR (95% CI)
P value
Reference 0.70 (0.41–1.20) 0.98 (0.84–1.14)
0.194 0.773
Reference 1.96 (1.10–3.50)
0.023
Reference 6.30 (3.13–12.68)
<0.0001
Reference 6.38 (3.12–13.03)
<0.0001
Reference 2.65 (1.34–5.26)
0.005
Reference 2.98 (1.42–6.27)
0.004
Reference 0.94 (0.43–2.06)
0.878
Reference 0.54 (0.22–1.32)
0.179
Reference 2.90 (1.56–5.40)
Not included Not included Not retained
Not includeda <0.0001
a
Included as time-dependent variables.
1662 | Mandala` et al.
Volume 18 | No. 10 | October 2007
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Venous thromboembolism developed in 59 (26.0%) patients, synchronously in 28 (12.3%) patients and during chemotherapy in 15 (6.6%) patients, while 16 (7.0%) patients experienced both events. Response to treatment, available for 218 patients, was as follows: 1 (0.5%) complete response, 54 (24.8%) partial responses, 73 (33.5%) stable diseases and 90 (41.3%) progressive diseases. Results of univariate and multivariate logistic models assessing the prognostic effect of the presence of synchronous VTE, the occurrence of VTE during chemotherapy and the presence of VTE irrespective of the time of occurrence (these latter two included as timedependent variable) are reported in Table 2. At univariate analysis, a significantly higher probability of not responding to treatment was associated with chemotherapy other than PEFG (OR 6.30, 95% CI 3.13–12.68, P < 0.0001), stage IV (OR 1.96, 95% CI 1.10–3.50, P = 0.023), presence of synchronous VTE (OR 2.65, 95% CI 1.34–5.26, P = 0.005) and overall presence of VTE (OR 2.90, 95% CI 1.56–5.40, P < 0.0001). In the stepwise selection model, only chemotherapy treatment other than PEFG (OR 6.38, 95% CI 3.12–13.03, P < 0.0001) and presence of synchronous VTE (OR 3.09, 95% CI 1.40–6.84, P = 0.005) proved to be predictors of progressive disease. No statistically significant effect of occurrence of VTE during chemotherapy on response to treatment was observed either at univariate or at multivariate analysis. At a median follow-up of 35 months, 211 (93.0%) patients progressed and 210 (92.5%) died. Overall 215 (94.7%) patients progressed or died. In the whole group, the median PFS and OS were 5 and 9.6 months, respectively. At univariate analysis, stage IV (HR 2.05, 95% CI 1.53–2.74, P < 0.0001), presence of synchronous VTE (HR 1.82, 95% CI 1.30–2.54, P = 0.0005),
original article
Annals of Oncology
Table 3. Results of Cox univariate and multivariate analyses for progression-free and overall survival Parameter
Progression-free survival Univariate OR (95% CI) P value
P value
Overall survival Univariate OR (95% CI)
P value
0.669 0.976
Not included
Multivariate OR (95% CI)
P value
0.780 0.038
0.92 (0.86–0.99)
0.044
Reference 1.06 (0.81–1.39) 1.00 (0.93–1.08)
<0.0001
Reference 1.78 (1.32–2.39)
0.0002
Reference 2.01 (1.51–2.69)
<0.0001
0.0641
Reference 1.39 (1.03–1.86)
0.030
Reference 1.09 (0.82–1.44)
0.541
0.0005
Reference 1.30 (0.87–1.95)
0.201
Reference 1.45 (1.03–2.04)
0.031
Reference 1.15 (0.77–1.70)
0.496
<0.0001
Reference 2.62 (1.73–3.97)
<0.0001
Reference 1.95 (1.32–2.87)
0.0008
Reference 1.64 (1.04–2.58)
0.032
Reference 1.76 (1.30–2.40)
0.0003
Not includeda <0.0001
Not included Not included
Reference 1.95 (1.46–2.61)
<0.0001
Not included
Not includeda
a
Included as time-dependent variables.
the first trial investigating this issue in a homogeneous cohort of patients. Two large population-based studies investigated the prognostic relevance of VTE in cancer patients [6, 7]. However, these studies may have several limitations because they do not give information about chemotherapy or performance status, which may have an impact on overall survival. Chew et al. demonstrated in a population-based study that a diagnosis of thromboembolism at the time of or within 1 year of cancer diagnosis is a significant predictor of death within that year for several cancer types [7]. In our trial patients developing VTE during chemotherapy had a worse prognosis as compared to patients without any thromboembolic event. Our findings in this particular setting extend the results of Sorensen et al., who reported that patients diagnosed concurrently with thromboembolism and cancer had decreased survival compared with a control group of patients (unmatched for stage) who were diagnosed with cancer but without thromboembolism [6]. In our study, patients developing VTE had a higher probability of developing progression during treatment and a higher progression and mortality rate. There are several biological explanations which support our clinical data. In the last 30 years basic research studies have indicated: (i) that tumor cells possess the capacity to activate blood coagulation by various mechanisms; (ii) that this capacity parallels tumor cell malignant transformation; and (iii) that fibrin formation in tumor tissues, as a final product of the clotting cascade activation, is involved in tumor growth and dissemination. A number of experimental studies in vitro and in animal models support the hypothesis that inhibiting blood coagulation may interfere with the progression of malignancy [11–13].
Volume 18 | No. 10 | October 2007
Recently Nitori et al. investigated, in an elegant study, the prognostic significance of tissue factor (TF) in pancreatic ductal carcinoma. In this study increased TF expression was correlated with the extent of the primary tumor, lymph node metastasis, lymphatic distant metastasis, advanced tumor-node metastasis stage and high tumor grade. More importantly, a high TF expression was an independent negative predictor for survival [14]. In addition to these experimental studies, one may argue that patients with VTE may have more biologically aggressive cancer, have greater underlying comorbidity or simply die earlier because of complications associated with thromboembolism and/or its treatment. However it seems unlikely that complications of venous thromboembolism can account entirely for the increased mortality among the patients in our study who had thromboembolic events. In our retrospective analysis we did not retrieve from medical charts acute death due to pulmonary embolism. Furthermore, at univariate analysis a significantly higher probability of not responding to treatment was associated with the presence of synchronous VTE (OR 2.65, 95% CI 1.34–5.26, P = 0.005). In the stepwise selection model, this variable proved to be a predictor of progressive disease. Several recent reports have demonstrated that activation of coagulation is strictly related to cell proliferation and resistance to apoptosis, which is one of the most common pathways of action for cytotoxic drugs [15, 16]. Further studies are required to determine the reason that VTE is so strongly associated with a poor prognosis. The strength of this study is a relatively large cohort of patients, all treated with a gemcitabine-based chemotherapy,
doi:10.1093/annonc/mdm284 | 1663
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Gender Female Reference Male 0.96 (0.74–1.26) Age (for each 0.92 (0.85–0.99) 5-year increase) Stage III Reference IV 2.05 (1.53–2.74) Chemotherapy PEFG Reference Other 1.30 (0.98–1.72) Synchronous VTE No Reference Yes 1.82 (1.30–2.54) VTE during chemotherapya No Reference Yes 3.04 (2.12–4.36) Overall VTEa No Reference Yes 2.78 (2.07–3.72)
Multivariate OR (95% CI)
original article A
Annals of Oncology
Events No 171 Yes 44
0.8
B
Totals 183 44
1.0
0.6 0.4
6
183 44
12
18
24
89 12
35 4
17 0
5 0
1.0
0
6
Patients at Risk 2 0
No Yes
No Yes
183 44
D
12
18
24
30
Months from randomisation 134 21
78 13
35 5
13 3
1.0
6 1
No Yes
0.8
Overall Survival
0.8 0.6 0.4
0.6 0.4 0.2
0.2
0
6
12
18
24
30
Months from randomisation
0.0
0
6
12
18
24
30
Months from randomisation
Figure 1. Kaplan–Meier estimates of progression-free and overall survival by presence of synchronous VTE and occurrence of VTE during chemotherapy. (A) Patients with synchronous VTE had significantly worse PFS compared to patients without (HR 1.82, 95% CI 1.30–2.54, P = 0.0005). (B) Patients with synchronous VTE had significantly worse OS compared to patients without (median OS 5.8 vs 10.3 months; HR 1.45, 95% CI 1.03–2.04, P = 0.031). (C) Patients with occurrence of VTE during chemotherapy had significantly worse PFS compared to patients without (median PFS 2.6 vs 5.1 months; HR 3.04, 95% CI 2.12–4.36, P < 0.0001). (D) Patients with occurrence of VTE during chemotherapy had significantly worse OS compared to patients without (median OS 4.4 vs 9.9 months; HR 1.95, 95% CI 1.32–2.87, P = 0.0008).
currently the most commonly used standard of care. In addition response to treatment, time to progression and overall survival have been collected prospectively. The limitation of our study is that the clinical research for VTE was performed retrospectively. Indeed, only symptomatic VTE could be considered and therefore a number of asymptomatic VTE may have been missed, with a consequent underestimation of the VTE risk. However, the incidence of VTE found in our study did not differ significantly from that reported by other studies. The high incidence of VTE in locally advanced or metastatic pancreatic cancer patients has been reported recently by Blom et al. [17]. The authors calculated incidence rates of venous thrombosis and compared this with population rates using a standardized morbidity ratio. The risk of venous thrombosis was increased 60-fold compared with the general population, at a cumulative risk of nearly 10%. In this study metastatic cancer patients receiving chemotherapy and with a tumor of the corpus or cauda were at highest risk of developing VTE. Appropriately, Blom et al. suggest that for locally advanced or metastatic pancreatic cancer patients receiving chemotherapy the role of primary prophylaxis should
1664 | Mandala` et al.
be considered, because anticoagulant treatment may prevent serious comorbidity and associated suffering. In previous reports, the incidence of bleeding for cancer patients receiving anticoagulant treatment is markedly lower than the incidence of venous thrombosis in patients with distant metastases [18, 19]. Oral anticoagulation with Vitamin K antagonists (VKA) may be problematic in patients with cancer. Drug interactions, malnutrition and liver dysfunction can lead to wide fluctuations of International Normalized Ratio. In addition cancer patients, compared to non-cancer patients, have a higher rate of VTE recurrence during oral anticoagulant therapy with VKAs; they also show a higher anticoagulation-associated hemorrhagic risk. Results from a recent randomized clinical trial demonstrate that in these patients long-term treatment with LMWH is safe and more effective than treatment with VKAs [20]. It is difficult to compare the incidence of VTE in our study to that in the other large trials evaluating the role of chemotherapy in metastatic pancreatic cancer patients. Indeed, many clinical trials in pancreatic cancer published in recent years have not reported toxic effects related to VTE.
Volume 18 | No. 10 | October 2007
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Progression Free Survival
0.4
30
Months from randomisation
Patients at Risk
0.0
0.6
0.0 0
C
Totals 183 44
0.2
0.2 0.0
No Yes
Events No 168 Yes 42
0.8
Overall Survival
Progression Free Survival
1.0
Annals of Oncology
references 1. Rickels FR, Levine M, Edwards RL. Hemostatic alterations in cancer patients. Cancer Metastasis Rev 1992; 1: 237–248. 2. Otten HM. Mathijssen J, ten Cate H et al. Symptomatic venous thromboembolism in cancer patients treated with chemotherapy, an underestimated phenomenon Arch Intern Med 2004; 164: 190–194. 3. Kosuri K, Muscarella P, BeKaii-Saab TS. Updates and controversies in the treatment of pancreatic cancer. Clin Adv Hematol Oncol 2006; 4: 47–54. 4. Zacharski LR, Ornstein DL. Heparin and cancer. Thromb Haemost 1998; 80: 10–23.
Volume 18 | No. 10 | October 2007
5. Falanga A. The effect of anticoagulant drugs on cancer. J Thromb Haemost 2004; 2: 1–3. 6. Sorensen HT, Mellemkjaer L, Olsen JH et al. Prognosis of cancers associated with venous thromboembolism. N Engl J Med 2000; 343: 1846–1850. 7. Chew HK, Wun T, Harvey D et al. Incidence of venous thromboembolism and its effect on survival among patients with common cancers. Arch Intern Med 2006; 166: 458–464. 8. Cascinu S, Labianca R, Catalano V et al. Weekly gemcitabine and cisplatin chemotherapy: a well-tolerated but ineffective chemotherapeutic regimen in advanced pancreatic cancer patients. A report from the Italian Group for the Study of Digestive Tract Cancer (GISCAD). Ann Oncol 2003; 14: 205–208. 9. Reni M, Cordio S, Milandri C et al. Gemcitabine versus cisplatin, epirubicin, fluorouracil, and gemcitabine in advanced pancreatic cancer: a randomised controlled multicentre phase III trial. Lancet Oncol 2005; 6: 369–376. 10. Khorana AA, Fine RL. Pancreatic cancer and thromboembolic disease. Lancet Oncol 2004; 5: 655–663. 11. Green KB, Silverstein RL. Hypercoagulability in cancer. Hematol Oncol Clin North Am 1996; 10: 499–530. 12. Falanga A, Rickels FR. Pathophysiology of the thrombophilic state in the cancer patient. Semin Thromb Hemost 1999; 25: 173–182. 13. Gale AJ, Gordon SG. Update on tumor cell procoagulant factors. Acta Haematol 2001; 106: 25–32. 14. Nitori N, Yoshinori I, Nakanishi Y et al. Prognostic significance of tissue factor in pancreatic ductal adenocarcinoma. Clin Cancer Res 2005; 11: 2531–2539. 15. Versteeg HH, Ruf W. Emerging insights in tissue factor-dependent signaling events. Semin Thromb Haemost 2006; 32: 24–32. 16. Jiang X, Guo YL, Bromberg ME. Formation of tissue factor-factor VIIa-factor Xa complex prevents apoptosis in human breast cancer cells. Thromb Haemost 2006; 96: 196–201. 17. Blom JW, Osanto S, Rosendaal FR. High risk of venous thrombosis in patients with pancreatic cancer: a cohort study of 202 patients. Eur J Cancer 2006; 42: 410–414. 18. Palareti G, Legnani C, Lee A et al. A comparison of the safety and efficacy of oral anticoagulation for the treatment of venous thromboembolic disease in patients with or without malignancy. Thromb Haemost 2000; 84: 805–810. 19. Hutten BA, Prins MH, Gent M et al. Incidence of recurrent thromboembolic and bleeding complications among patients with venous thromboembolism in relation to both malignancy and achieved international normalised ratio: a retrospective analysis. J Clin Oncol 2000; 18: 3078–3083. 20. Lee A, Levine MN, Baker RI et al. Low-molecular-weight heparin versus a coumarin for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 2003; 349: 146–153. 21. Khorana AA, Francis CW, Culakova E et al. Thromboembolism in hospitalized neutropenic cancer patients. J Clin Oncol 2006; 24: 484–490. 22. Alcalay A, Wun T, Khatri V et al. Venous thromboembolism in patients with colorectal cancer: incidence and effect on survival. J Clin Oncol 2006; 24: 1112–1118. 23. Chew HK, Wun T, Harvey DJ et al. Incidence of venous thromboembolism and the impact on survival in breast cancer patients. J Clin Oncol 2007; 25: 70–76. 24. Kakkar AK, Levine MN, Kadziola Z et al. Low molecular weight heparin, therapy with dalteparin, and survival in advanced cancer: the Fragmin Advanced Malignancy Outcome Study (FAMOUS). J Clin Oncol 2004; 15: 1944–1948. 25. Klerk CPW, Smorenburg SM, Otten HM et al. The effect of low molecular weight heparin on survival in patients with advanced malignancy. J Clin Oncol 2005; 1: 2130–2135.
doi:10.1093/annonc/mdm284 | 1665
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Furthermore, in its initial version the US National Cancer Institute’s criteria for common toxicity did not include VTE, suggesting that this type of complication had been underestimated. In a recent retrospective cohort study using the discharge database of the University Health System Consortium, Khorana et al. were able to demonstrate not only that VTE is frequent in hospitalized neutropenic cancer patients, including perceived low-risk subgroups such as patients with hematologic malignancies and non-metastatic disease, but also that it is associated with increased in-hospital mortality [21]. Since venous thrombosis without embolism is not a life-threatening condition, the authors suggest that the presence of venous thrombosis could signify an aggressive tumor biology and therefore a worse short-term prognosis. Although data from other cancers suggest that the development of thrombosis adversely affects cancer-related survival [22, 23], our study is the first to evaluate this issue in pancreatic cancer patients. Our patients were treated within clinical trials. As a corollary these patients had a ECOG-PS grade of 2 or less and a relatively basal good prognosis. Recently, two prospective randomized, placebo-controlled trials, specifically evaluating LMWH for survival in cancer patients, have been reported. In both studies the survival of patients with a good prognosis was significantly in favor of LMWH compared with placebo [24, 25]. It may be intriguing that from another point of view our data seem to support these results. Finally, the University of Rochester (NY, USA) Community Clinical Oncology Program is currently conducting a Phase III study sponsored by the US National Cancer Institute, which randomly assigns patients with locally advanced or metastatic pancreatic cancer to gemcitabine with or without dalteparin once per day (Protocol URCC-U2200, NCT00031837, URCC 5012, NCI-5012, NCI-CCC-99-45, NCI-P02-0212). A second trial assessing the effects of nadroparin on survival and disease progression in patients with hormone-refractory prostate cancer, locally advanced pancreatic cancer or non–small-cell lung carcinoma is ongoing (Protocol FRX106365 NCT00312013). These trials will provide important information regarding improvements in survival and incidence of thromboembolic disease from the addition of a LMWH to standard therapy for pancreatic cancer.
original article
Annals of Oncology 18: 1660–1665, 2007 doi:10.1093/annonc/mdm284 Published online 28 July 2007
Venous thromboembolism predicts poor prognosis in irresectable pancreatic cancer patients M. Mandala`1*, M. Reni2, S. Cascinu3, S. Barni4, I. Floriani5, S. Cereda2, R. Berardi3, S. Mosconi1, V. Torri5 & R. Labianca1 1 Unit of Medical Oncology, Ospedali Riuniti, Bergamo; 2Division of Medical Oncology, San Raffaele Hospital, Milan; 3Clinica di Oncologia Medica, Universita` Politecnica delle Marche; Azienda Ospedaliero-Universitaria, Ospedali Riuniti di Ancona, Umberto I, Lancisi, Salesi, Ancona; 4 Division of Medical Oncology, Treviglio Hospital, Treviglio; 5Department of Oncology, Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy
Received 15 February 2007; revised 10 May 2007; accepted 14 May 2007
Background: The aim was to investigate the outcomes associated with venous thromboembolism (VTE)
original article
clinical trials between December 2001 and December 2004 in order to evaluate the prognostic impact of symptomatic VTE on clinical outcomes, such as response to treatment, progression-free survival (PFS) and overall survival (OS). Results: Among 227 irresectable pancreatic cancer patients, with Eastern Cooperative Oncology Group performance status (ECOG-PS) £ 2, 59 (26.0%) patients developed a VTE. A synchronous VTE occurred in 28 (12.3%) patients, while a VTE during chemotherapy was observed in 15 (6.6%) patients, and 16 (7.0%) patients experienced both events. Presence of synchronous VTE was associated with a higher probability of not responding to treatment (odds ratio 2.98, 95% CI 1.42–6.27, P = 0.004), but showed no effect on both PFS and OS at least at multivariate analysis. Occurrence of a VTE during chemotherapy showed a statistically significant effect on PFS (hazard ratio [HR] 2.59, 95% CI 1.69–3.97, P < 0.0001) and OS (HR 1.64, 95%CI 1.04–2.58, P = 0.032). Conclusions: Our data suggest that the occurrence of VTE may be associated with a reduced response rate and a shorter PFS and OS among patients with irresectable pancreatic cancer. In these patients the development of VTE may reflect the presence of a biologically more aggressive cancer that in turn leads to a worse prognosis. Key words: venous thromboembolism, prognosis, pancreatic cancer
introduction The strict two-way clinical relationship between cancer and venous thromboembolism (VTE) has been evaluated since Trousseau’s time. While subclinical coagulation activation is encountered in up to 90% of cancer patients, only 4% to 15% of them develop VTE [1]. The relationship among cancer, chemotherapy and thrombosis has been more extensively investigated in breast cancer patients. Few data are available for gastrointestinal (GI) cancer patients receiving chemotherapy, in spite of the general perception of a high incidence of thromboembolic complications in patients with mucinous carcinoma of the pancreas and GI tract. A recent retrospective study on patients with different solid tumors found a remarkably high incidence (7%) of VTE in a group of 39 patients with colon cancer during chemotherapy [2]. *Correspondence to: Dr M. Mandala`, Unit of Medical Oncology, Ospedali Riuniti Bergamo, Largo Barozzi 1, 24100 Bergamo, Italy. Tel: +39-035-269858; Fax: +39-035-266849; E-mail: [email protected]
ª 2007 European Society for Medical Oncology
Pancreatic carcinoma is currently the fourth biggest leading cause of cancer-related death in the USA. About 10%–15% of patients undergo complete tumor resection and of this proportion fewer than one-fifth are alive 5 years after surgery. In locally advanced or metastatic disease, systemic chemotherapy induces responses in only a few patients. The main clinical benefit of chemotherapy is generally for the palliation of pain and cancer-related symptoms [3]. Experimental studies in vitro or in animal models support the hypothesis that the inhibition of blood coagulation may interfere with the progression of malignancy. Experimental and clinical data to support the hypothesis that anticoagulation may influence prognosis of cancer patients have been recently reviewed [4, 5]. There is some evidence that patients with cancer who develop thromboembolism have poorer survival [6, 7]. However most of these studies used population-based data to estimate the incidence and the prognosis of cancer patients, lacking information about performance status and type of treatment, which could influence survival. The objective
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
among irresectable pancreatic cancer patients.
Methods: This is a follow-up study of consecutive irresectable cancer patients, treated and followed up in
original article
Annals of Oncology
of this study was to describe the prognostic role of VTE in a cohort of irresectable pancreatic cancer patients.
materials and methods patients and study design
statistical methods All patients satisfying eligibility criteria and treated with at least one cycle of chemotherapy were considered for analysis. Progression-free survival was defined as the time from the beginning of chemotherapy to first appearance of progressive disease or death for any cause; patients known to be alive and without progressive disease at the time of analysis were censored at their last available follow-up assessment. Overall survival was defined as the time from the beginning of chemotherapy to the date of death from any cause or the date of the last follow-up. Survival curves were estimated using the Kaplan–Meier method. Cox proportional hazards models as implemented in the PHREG program (Statistical Analysis System [SAS] Institute, Cary, NC, USA) were used to estimate and test demographic characteristics and clinical features for their association with PFS and OS. Venous thromboembolism was included both irrespective of the time of occurrence and by distinguishing
Volume 18 | No. 10 | October 2007
results A total of 227 consecutive patients were identified and included in the analysis. Demographic and clinical characteristics of the cohort are summarized in Table 1. Just over half (121, 53.3%) were males and the median age was 66 years, ranging from 38 to 82 years. Forty-four (19.4%) patients had undergone previous surgery, either for curative or palliative intent. Most (63.4%) patients had metastatic disease, while the remaining were classified with locally advanced irresectable disease. Chemotherapy treatment was PEFG for 79 (37.4%) patients; the remaining patients received chemotherapy with gemcitabine or cisplatin and gemcitabine. We excluded five patients who had not received at least one dose of chemotherapy: four patients because of inadequate liver function and one patient because of a concomitant other cancer. Table 1. Characteristics of the study cohort (n = 227) Parameter Agea (years) Gender Male Female Stage of disease Locally advanced Metastatic Previous surgery No Yes Chemotherapy PEFG Other VTE None Synchronous only During chemotherapy only Both
No. of patients 63 (38–82) 121 (53.3%) 106 (46.7%) 83 (36.6%) 144 (63.4%) 183 (80.6%) 44 (19.4%) 79 (34.8%) 148 (65.2%) 168 28 15 16
(74.0%) (12.3%) (6.6%) (7.0%)
a
Median and range.
doi:10.1093/annonc/mdm284 | 1661
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
The cohort was identified by inspecting the electronic database of all patients attending four Italian institutions from December 2001 to December 2004. An additional search was performed using the International Classification of Diseases, 9th revision, clinical modification (ICD9CM) registration systems. We retrieved data concerning clinical outcomes and VTE from patients enrolled in two prospective trials, one sponsored by the Italian Group for the Study of Digestive Tract Cancer (GISCAD) and the other conducted by the San Raffaele Hospital, Milan [8–9]. In addition we retrieved data from an ongoing GISCAD-sponsored Phase III randomized study comparing gemcitabine with cisplatin and gemcitabine in irresectable pancreatic cancer patients. Also included were 27 patients treated with a Cisplatin, epirubicin, 5-fluorouracil, gemcitabine (PEFG) polychemotherapy at the San Raffaele Hospital and not enrolled in the prospective trial previously reported [9]. The study protocols were approved by the local ethical committee and all patients gave informed consent. Eligibility criteria for inclusion in this study included histological diagnosis of locally advanced unresectable and/or metastatic pancreatic carcinoma, treatment with a gemcitabine-based chemotherapy, an Eastern Cooperative Oncology Group performance status (ECOG-PS) £ 2 and normal hepatic (serum bilirubin < 1.5 mg/dL), renal (serum creatinine < 1.5 mg/dL) and bone marrow (leukocyte count > 4000/lL, platelet count > 100000/lL) functions. Exclusion criteria included prophylaxis with warfarin or low molecular weight heparin (LMWH). Information on age, gender, histopathology and surgical and medical treatment were retrieved for each patient, as well as on tumor overall response, progression-free survival (PFS) and overall survival (OS). Stage was assessed according to the American Joint Committee on Cancer (Tumour, Node, Metastasis) classification. Tumor response was classified according to World Health Organization criteria. The medical charts and the radiological history of all patients were checked with regard to ultrasonography of the extremities, chest and abdominal computed tomography (CT) scan and perfusion/ventilation lung scan. In all patients with proven VTE, we actively searched the records and radiological reports for indicators of tumor load at the time of the event. The criterion to diagnose VTE by compression ultrasonography was non-compressibility of a proximal vein. When symptoms suggestive for pulmonary embolism developed, a radionuclide lung scan, a CT scan or both was performed.
between occurrence before or during chemotherapy. Overall VTE occurrence was not included in the multivariate models due to multicolinearity problems and because the other two VTE variables were considered more informative. A proportional hazards model was used for univariate and multivariate analysis. Variables found to be associated with PFS and OS in the univariate analysis (P < 0.10) were considered for the multivariate analysis. A stepwise strategy was then adopted, with a statistical significance cut-off for variable screening of 0.05, and VTE was forced into the model. Results are expressed as hazard ratios (HR) with 95% confidence intervals (CI) reported for each factor. A univariate and multivariate logistic regression model using the same approach was used to assess the effect of the same variables on treatment response, expressing results as odds ratios (OR) and relative 95% CI. With a total of 210 events in our series, this study had a 95% power of detecting a hazard ratio of at least 2.0 associated with the group with VTE (expected to have a prevalence of about 20% [10]) with a type I error of 5%, for a bilateral test. The statistically significant level was set at P £ 0.05. Analysis was performed using SAS (Version 9.0) software.
original article
Annals of Oncology
occurrence of VTE during chemotherapy (HR 3.04, 95% CI 2.12–4.36, P < 0.0001) and overall VTE (HR 2.78, 95% CI 2.07–3.72, P < 0.0001) were associated with a shorter PFS, whereas any 5-year increase of age resulted in a lower risk of progression (HR 0.92, 95% CI 0.85–0.99, P = 0.038). Chemotherapy other than PEFG was not significantly associated with PFS (HR 1.30, 95%CI 0.98–1.72, P = 0.064) but was nevertheless included in the multivariate model since the criterion of P < 0.10 for remaining in the model was met. At multivariate assessment, stage IV (HR 1.78, 95% CI 1.32–2.39, P = 0.0002), 5-year age increase (HR 0.92, 95% CI 0.86–0.99, P = 0.044), chemotherapy (HR 1.39, 95% CI 1.03–1.86, P = 0.030) and VTE during chemotherapy (HR 2.62, 95% CI 1.73–3.97, P < 0.0001) reached statistical significance. As for OS, stage IV (HR 1.95, 95% CI 1.46–2.61, P < 0.0001), synchronous VTE (HR 1.45, 95% CI 1.03–2.04, P = 0.031), occurrence of VTE during chemotherapy (HR 1.95, 95% CI 1.32–2.87, P = 0.0008) and overall VTE (HR 1.76, 95% CI 1.30–2.40, P = 0.0003) showed a prognostic negative effect at univariate analysis. In the multivariate model, only the effects of stage (HR 1.95, 95% CI 1.46–2.61, P < 0.0001) and VTE during chemotherapy (HR 1.64, 95% CI 1.04–2.58, P = 0.032) were confirmed to have a prognostic role on survival Table 3. Figure 1 shows the Kaplan–Meier curves for PFS and OS according to the presence of synchronous VTE and the occurrence of VTE during chemotherapy.
discussion The most striking finding of this study is that the development of VTE seems to predict a poor prognosis in irresectable pancreatic cancer patients. To our best knowledge this is
Table 2. Univariate and multivariate logistic regression models assessing the association between VTE and response to treatment (Progressive Disease vs other) Parameter
Gender Female Male Age (for each 5-year increase) Stage III IV Chemotherapy PEFG Other Synchronous VTE No Yes VTE during chemotherapy No Yes Overall VTEa No Yes
Analysis Univariate OR (95% CI)
P value
Multivariate OR (95% CI)
P value
Reference 0.70 (0.41–1.20) 0.98 (0.84–1.14)
0.194 0.773
Reference 1.96 (1.10–3.50)
0.023
Reference 6.30 (3.13–12.68)
<0.0001
Reference 6.38 (3.12–13.03)
<0.0001
Reference 2.65 (1.34–5.26)
0.005
Reference 2.98 (1.42–6.27)
0.004
Reference 0.94 (0.43–2.06)
0.878
Reference 0.54 (0.22–1.32)
0.179
Reference 2.90 (1.56–5.40)
Not included Not included Not retained
Not includeda <0.0001
a
Included as time-dependent variables.
1662 | Mandala` et al.
Volume 18 | No. 10 | October 2007
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Venous thromboembolism developed in 59 (26.0%) patients, synchronously in 28 (12.3%) patients and during chemotherapy in 15 (6.6%) patients, while 16 (7.0%) patients experienced both events. Response to treatment, available for 218 patients, was as follows: 1 (0.5%) complete response, 54 (24.8%) partial responses, 73 (33.5%) stable diseases and 90 (41.3%) progressive diseases. Results of univariate and multivariate logistic models assessing the prognostic effect of the presence of synchronous VTE, the occurrence of VTE during chemotherapy and the presence of VTE irrespective of the time of occurrence (these latter two included as timedependent variable) are reported in Table 2. At univariate analysis, a significantly higher probability of not responding to treatment was associated with chemotherapy other than PEFG (OR 6.30, 95% CI 3.13–12.68, P < 0.0001), stage IV (OR 1.96, 95% CI 1.10–3.50, P = 0.023), presence of synchronous VTE (OR 2.65, 95% CI 1.34–5.26, P = 0.005) and overall presence of VTE (OR 2.90, 95% CI 1.56–5.40, P < 0.0001). In the stepwise selection model, only chemotherapy treatment other than PEFG (OR 6.38, 95% CI 3.12–13.03, P < 0.0001) and presence of synchronous VTE (OR 3.09, 95% CI 1.40–6.84, P = 0.005) proved to be predictors of progressive disease. No statistically significant effect of occurrence of VTE during chemotherapy on response to treatment was observed either at univariate or at multivariate analysis. At a median follow-up of 35 months, 211 (93.0%) patients progressed and 210 (92.5%) died. Overall 215 (94.7%) patients progressed or died. In the whole group, the median PFS and OS were 5 and 9.6 months, respectively. At univariate analysis, stage IV (HR 2.05, 95% CI 1.53–2.74, P < 0.0001), presence of synchronous VTE (HR 1.82, 95% CI 1.30–2.54, P = 0.0005),
original article
Annals of Oncology
Table 3. Results of Cox univariate and multivariate analyses for progression-free and overall survival Parameter
Progression-free survival Univariate OR (95% CI) P value
P value
Overall survival Univariate OR (95% CI)
P value
0.669 0.976
Not included
Multivariate OR (95% CI)
P value
0.780 0.038
0.92 (0.86–0.99)
0.044
Reference 1.06 (0.81–1.39) 1.00 (0.93–1.08)
<0.0001
Reference 1.78 (1.32–2.39)
0.0002
Reference 2.01 (1.51–2.69)
<0.0001
0.0641
Reference 1.39 (1.03–1.86)
0.030
Reference 1.09 (0.82–1.44)
0.541
0.0005
Reference 1.30 (0.87–1.95)
0.201
Reference 1.45 (1.03–2.04)
0.031
Reference 1.15 (0.77–1.70)
0.496
<0.0001
Reference 2.62 (1.73–3.97)
<0.0001
Reference 1.95 (1.32–2.87)
0.0008
Reference 1.64 (1.04–2.58)
0.032
Reference 1.76 (1.30–2.40)
0.0003
Not includeda <0.0001
Not included Not included
Reference 1.95 (1.46–2.61)
<0.0001
Not included
Not includeda
a
Included as time-dependent variables.
the first trial investigating this issue in a homogeneous cohort of patients. Two large population-based studies investigated the prognostic relevance of VTE in cancer patients [6, 7]. However, these studies may have several limitations because they do not give information about chemotherapy or performance status, which may have an impact on overall survival. Chew et al. demonstrated in a population-based study that a diagnosis of thromboembolism at the time of or within 1 year of cancer diagnosis is a significant predictor of death within that year for several cancer types [7]. In our trial patients developing VTE during chemotherapy had a worse prognosis as compared to patients without any thromboembolic event. Our findings in this particular setting extend the results of Sorensen et al., who reported that patients diagnosed concurrently with thromboembolism and cancer had decreased survival compared with a control group of patients (unmatched for stage) who were diagnosed with cancer but without thromboembolism [6]. In our study, patients developing VTE had a higher probability of developing progression during treatment and a higher progression and mortality rate. There are several biological explanations which support our clinical data. In the last 30 years basic research studies have indicated: (i) that tumor cells possess the capacity to activate blood coagulation by various mechanisms; (ii) that this capacity parallels tumor cell malignant transformation; and (iii) that fibrin formation in tumor tissues, as a final product of the clotting cascade activation, is involved in tumor growth and dissemination. A number of experimental studies in vitro and in animal models support the hypothesis that inhibiting blood coagulation may interfere with the progression of malignancy [11–13].
Volume 18 | No. 10 | October 2007
Recently Nitori et al. investigated, in an elegant study, the prognostic significance of tissue factor (TF) in pancreatic ductal carcinoma. In this study increased TF expression was correlated with the extent of the primary tumor, lymph node metastasis, lymphatic distant metastasis, advanced tumor-node metastasis stage and high tumor grade. More importantly, a high TF expression was an independent negative predictor for survival [14]. In addition to these experimental studies, one may argue that patients with VTE may have more biologically aggressive cancer, have greater underlying comorbidity or simply die earlier because of complications associated with thromboembolism and/or its treatment. However it seems unlikely that complications of venous thromboembolism can account entirely for the increased mortality among the patients in our study who had thromboembolic events. In our retrospective analysis we did not retrieve from medical charts acute death due to pulmonary embolism. Furthermore, at univariate analysis a significantly higher probability of not responding to treatment was associated with the presence of synchronous VTE (OR 2.65, 95% CI 1.34–5.26, P = 0.005). In the stepwise selection model, this variable proved to be a predictor of progressive disease. Several recent reports have demonstrated that activation of coagulation is strictly related to cell proliferation and resistance to apoptosis, which is one of the most common pathways of action for cytotoxic drugs [15, 16]. Further studies are required to determine the reason that VTE is so strongly associated with a poor prognosis. The strength of this study is a relatively large cohort of patients, all treated with a gemcitabine-based chemotherapy,
doi:10.1093/annonc/mdm284 | 1663
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Gender Female Reference Male 0.96 (0.74–1.26) Age (for each 0.92 (0.85–0.99) 5-year increase) Stage III Reference IV 2.05 (1.53–2.74) Chemotherapy PEFG Reference Other 1.30 (0.98–1.72) Synchronous VTE No Reference Yes 1.82 (1.30–2.54) VTE during chemotherapya No Reference Yes 3.04 (2.12–4.36) Overall VTEa No Reference Yes 2.78 (2.07–3.72)
Multivariate OR (95% CI)
original article A
Annals of Oncology
Events No 171 Yes 44
0.8
B
Totals 183 44
1.0
0.6 0.4
6
183 44
12
18
24
89 12
35 4
17 0
5 0
1.0
0
6
Patients at Risk 2 0
No Yes
No Yes
183 44
D
12
18
24
30
Months from randomisation 134 21
78 13
35 5
13 3
1.0
6 1
No Yes
0.8
Overall Survival
0.8 0.6 0.4
0.6 0.4 0.2
0.2
0
6
12
18
24
30
Months from randomisation
0.0
0
6
12
18
24
30
Months from randomisation
Figure 1. Kaplan–Meier estimates of progression-free and overall survival by presence of synchronous VTE and occurrence of VTE during chemotherapy. (A) Patients with synchronous VTE had significantly worse PFS compared to patients without (HR 1.82, 95% CI 1.30–2.54, P = 0.0005). (B) Patients with synchronous VTE had significantly worse OS compared to patients without (median OS 5.8 vs 10.3 months; HR 1.45, 95% CI 1.03–2.04, P = 0.031). (C) Patients with occurrence of VTE during chemotherapy had significantly worse PFS compared to patients without (median PFS 2.6 vs 5.1 months; HR 3.04, 95% CI 2.12–4.36, P < 0.0001). (D) Patients with occurrence of VTE during chemotherapy had significantly worse OS compared to patients without (median OS 4.4 vs 9.9 months; HR 1.95, 95% CI 1.32–2.87, P = 0.0008).
currently the most commonly used standard of care. In addition response to treatment, time to progression and overall survival have been collected prospectively. The limitation of our study is that the clinical research for VTE was performed retrospectively. Indeed, only symptomatic VTE could be considered and therefore a number of asymptomatic VTE may have been missed, with a consequent underestimation of the VTE risk. However, the incidence of VTE found in our study did not differ significantly from that reported by other studies. The high incidence of VTE in locally advanced or metastatic pancreatic cancer patients has been reported recently by Blom et al. [17]. The authors calculated incidence rates of venous thrombosis and compared this with population rates using a standardized morbidity ratio. The risk of venous thrombosis was increased 60-fold compared with the general population, at a cumulative risk of nearly 10%. In this study metastatic cancer patients receiving chemotherapy and with a tumor of the corpus or cauda were at highest risk of developing VTE. Appropriately, Blom et al. suggest that for locally advanced or metastatic pancreatic cancer patients receiving chemotherapy the role of primary prophylaxis should
1664 | Mandala` et al.
be considered, because anticoagulant treatment may prevent serious comorbidity and associated suffering. In previous reports, the incidence of bleeding for cancer patients receiving anticoagulant treatment is markedly lower than the incidence of venous thrombosis in patients with distant metastases [18, 19]. Oral anticoagulation with Vitamin K antagonists (VKA) may be problematic in patients with cancer. Drug interactions, malnutrition and liver dysfunction can lead to wide fluctuations of International Normalized Ratio. In addition cancer patients, compared to non-cancer patients, have a higher rate of VTE recurrence during oral anticoagulant therapy with VKAs; they also show a higher anticoagulation-associated hemorrhagic risk. Results from a recent randomized clinical trial demonstrate that in these patients long-term treatment with LMWH is safe and more effective than treatment with VKAs [20]. It is difficult to compare the incidence of VTE in our study to that in the other large trials evaluating the role of chemotherapy in metastatic pancreatic cancer patients. Indeed, many clinical trials in pancreatic cancer published in recent years have not reported toxic effects related to VTE.
Volume 18 | No. 10 | October 2007
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Progression Free Survival
0.4
30
Months from randomisation
Patients at Risk
0.0
0.6
0.0 0
C
Totals 183 44
0.2
0.2 0.0
No Yes
Events No 168 Yes 42
0.8
Overall Survival
Progression Free Survival
1.0
Annals of Oncology
references 1. Rickels FR, Levine M, Edwards RL. Hemostatic alterations in cancer patients. Cancer Metastasis Rev 1992; 1: 237–248. 2. Otten HM. Mathijssen J, ten Cate H et al. Symptomatic venous thromboembolism in cancer patients treated with chemotherapy, an underestimated phenomenon Arch Intern Med 2004; 164: 190–194. 3. Kosuri K, Muscarella P, BeKaii-Saab TS. Updates and controversies in the treatment of pancreatic cancer. Clin Adv Hematol Oncol 2006; 4: 47–54. 4. Zacharski LR, Ornstein DL. Heparin and cancer. Thromb Haemost 1998; 80: 10–23.
Volume 18 | No. 10 | October 2007
5. Falanga A. The effect of anticoagulant drugs on cancer. J Thromb Haemost 2004; 2: 1–3. 6. Sorensen HT, Mellemkjaer L, Olsen JH et al. Prognosis of cancers associated with venous thromboembolism. N Engl J Med 2000; 343: 1846–1850. 7. Chew HK, Wun T, Harvey D et al. Incidence of venous thromboembolism and its effect on survival among patients with common cancers. Arch Intern Med 2006; 166: 458–464. 8. Cascinu S, Labianca R, Catalano V et al. Weekly gemcitabine and cisplatin chemotherapy: a well-tolerated but ineffective chemotherapeutic regimen in advanced pancreatic cancer patients. A report from the Italian Group for the Study of Digestive Tract Cancer (GISCAD). Ann Oncol 2003; 14: 205–208. 9. Reni M, Cordio S, Milandri C et al. Gemcitabine versus cisplatin, epirubicin, fluorouracil, and gemcitabine in advanced pancreatic cancer: a randomised controlled multicentre phase III trial. Lancet Oncol 2005; 6: 369–376. 10. Khorana AA, Fine RL. Pancreatic cancer and thromboembolic disease. Lancet Oncol 2004; 5: 655–663. 11. Green KB, Silverstein RL. Hypercoagulability in cancer. Hematol Oncol Clin North Am 1996; 10: 499–530. 12. Falanga A, Rickels FR. Pathophysiology of the thrombophilic state in the cancer patient. Semin Thromb Hemost 1999; 25: 173–182. 13. Gale AJ, Gordon SG. Update on tumor cell procoagulant factors. Acta Haematol 2001; 106: 25–32. 14. Nitori N, Yoshinori I, Nakanishi Y et al. Prognostic significance of tissue factor in pancreatic ductal adenocarcinoma. Clin Cancer Res 2005; 11: 2531–2539. 15. Versteeg HH, Ruf W. Emerging insights in tissue factor-dependent signaling events. Semin Thromb Haemost 2006; 32: 24–32. 16. Jiang X, Guo YL, Bromberg ME. Formation of tissue factor-factor VIIa-factor Xa complex prevents apoptosis in human breast cancer cells. Thromb Haemost 2006; 96: 196–201. 17. Blom JW, Osanto S, Rosendaal FR. High risk of venous thrombosis in patients with pancreatic cancer: a cohort study of 202 patients. Eur J Cancer 2006; 42: 410–414. 18. Palareti G, Legnani C, Lee A et al. A comparison of the safety and efficacy of oral anticoagulation for the treatment of venous thromboembolic disease in patients with or without malignancy. Thromb Haemost 2000; 84: 805–810. 19. Hutten BA, Prins MH, Gent M et al. Incidence of recurrent thromboembolic and bleeding complications among patients with venous thromboembolism in relation to both malignancy and achieved international normalised ratio: a retrospective analysis. J Clin Oncol 2000; 18: 3078–3083. 20. Lee A, Levine MN, Baker RI et al. Low-molecular-weight heparin versus a coumarin for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 2003; 349: 146–153. 21. Khorana AA, Francis CW, Culakova E et al. Thromboembolism in hospitalized neutropenic cancer patients. J Clin Oncol 2006; 24: 484–490. 22. Alcalay A, Wun T, Khatri V et al. Venous thromboembolism in patients with colorectal cancer: incidence and effect on survival. J Clin Oncol 2006; 24: 1112–1118. 23. Chew HK, Wun T, Harvey DJ et al. Incidence of venous thromboembolism and the impact on survival in breast cancer patients. J Clin Oncol 2007; 25: 70–76. 24. Kakkar AK, Levine MN, Kadziola Z et al. Low molecular weight heparin, therapy with dalteparin, and survival in advanced cancer: the Fragmin Advanced Malignancy Outcome Study (FAMOUS). J Clin Oncol 2004; 15: 1944–1948. 25. Klerk CPW, Smorenburg SM, Otten HM et al. The effect of low molecular weight heparin on survival in patients with advanced malignancy. J Clin Oncol 2005; 1: 2130–2135.
doi:10.1093/annonc/mdm284 | 1665
Downloaded from http://annonc.oxfordjournals.org/ by guest on July 13, 2015
Furthermore, in its initial version the US National Cancer Institute’s criteria for common toxicity did not include VTE, suggesting that this type of complication had been underestimated. In a recent retrospective cohort study using the discharge database of the University Health System Consortium, Khorana et al. were able to demonstrate not only that VTE is frequent in hospitalized neutropenic cancer patients, including perceived low-risk subgroups such as patients with hematologic malignancies and non-metastatic disease, but also that it is associated with increased in-hospital mortality [21]. Since venous thrombosis without embolism is not a life-threatening condition, the authors suggest that the presence of venous thrombosis could signify an aggressive tumor biology and therefore a worse short-term prognosis. Although data from other cancers suggest that the development of thrombosis adversely affects cancer-related survival [22, 23], our study is the first to evaluate this issue in pancreatic cancer patients. Our patients were treated within clinical trials. As a corollary these patients had a ECOG-PS grade of 2 or less and a relatively basal good prognosis. Recently, two prospective randomized, placebo-controlled trials, specifically evaluating LMWH for survival in cancer patients, have been reported. In both studies the survival of patients with a good prognosis was significantly in favor of LMWH compared with placebo [24, 25]. It may be intriguing that from another point of view our data seem to support these results. Finally, the University of Rochester (NY, USA) Community Clinical Oncology Program is currently conducting a Phase III study sponsored by the US National Cancer Institute, which randomly assigns patients with locally advanced or metastatic pancreatic cancer to gemcitabine with or without dalteparin once per day (Protocol URCC-U2200, NCT00031837, URCC 5012, NCI-5012, NCI-CCC-99-45, NCI-P02-0212). A second trial assessing the effects of nadroparin on survival and disease progression in patients with hormone-refractory prostate cancer, locally advanced pancreatic cancer or non–small-cell lung carcinoma is ongoing (Protocol FRX106365 NCT00312013). These trials will provide important information regarding improvements in survival and incidence of thromboembolic disease from the addition of a LMWH to standard therapy for pancreatic cancer.
original article