SEARCH

SEARCH BY CITATION

Keywords:

  • gastroesophageal junction;
  • overall survival;
  • protocol chemotherapy;
  • thromboembolic events

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND.

Thromboembolic events (TEEs) are considered common in patients with gastroesophageal carcinoma, but their frequency at baseline and during chemotherapy is not known. Because prophylactic anticoagulation results in improved overall survival (OS) of solid tumor patients, the authors hypothesized that TEEs at baseline and during chemotherapy would have an adverse effect on OS.

METHODS.

The authors analyzed patients with advanced gastroesophageal carcinoma who were treated on 4 prospective chemotherapy Phase II/III trials. Baseline and subsequent TEEs were documented and correlated with OS.

RESULTS.

On the 4 trials, 191 patients received single-agent or a combination of a taxane, camptothecin, platinum, or fluoropyrimidine. At baseline, TEEs occurred in 5.3% of untreated patients compared with 8.5% of previously treated patients (who had received prior treatment for metastatic disease). The median OS was only 3.9 months for patients who had a TEE at any time versus 8.7 months for patients who never developed a TEE (P = .007). TEEs at baseline were correlated with poor median OS in untreated patients (4.9 months vs 8.9 months for patients without a TEE; P = .014). There was no associated between TEEs and the type of chemotherapy used.

CONCLUSIONS.

The current results established that TEEs at baseline and/or during chemotherapy are frequent and result in poor OS for patients with advanced gastroesophageal carcinoma. Aggressive methods to treat or prevent TEEs are warranted. Cancer 2007. © 2007 American Cancer Society.

The association between thrombosis and malignancy was reported first in 1865 by Trousseau1 and now is documented well in the medical literature.2–4 Gastric or gastroesophageal junction cancer often is associated with a high incidence of thromboembolic events (TEEs). In a review of >1 million Medicare patients who were admitted with a diagnosis of malignancy, gastric cancer had the fifth highest rate of deep vein thrombosis (DVT) or pulmonary embolism (PE): only lymphoma and cancers of the pancreas, brain, and ovaries with higher rates.2 It is known that, if diagnoses of cancer and TEE are made simultaneously or close together, then patients often have more advanced-stage cancer and a poor prognosis.4 It also has been demonstrated that chemotherapy can increase the risk of TEEs in cancer patients; however, those data generally were limited to patients with breast cancer who received biologic and cytotoxic agents.5–7 The rate of baseline TEEs and their development during systemic chemotherapy in patients with advanced gastroesophageal carcinoma is not known. In addition, to our knowledge, the impact of TEEs on overall survival (OS) has not been documented in this group of cancers. A closer focus on TEEs is even more important, because a recent randomized study of 302 patients with solid tumors demonstrated a survival advantage for patients who received treatment with prophylactic low-molecular-weight heparin.8

Limited information on the association of TEEs (venous and arterial) has been reported in patients with advanced gastroesophageal cancer who were receiving biochemotherapy (irinotecan, cisplatin, and bevacizumab).9 In 24 patients, the rate of development of TEEs was 25% (95% confidence interval, 11–45%) during therapy. The baseline rate of TEEs was not an issue in that trial, because those patients were ineligible, and the numbers were too small to assess the impact of TEEs on OS.

We chose to study a large number of patients with advanced gastroesophageal carcinoma who were entering systemic chemotherapy clinical trials. Therefore, patients were staged thoroughly and followed carefully. We hypothesized that, because prophylactic anticoagulation prolongs survival, the presence TEEs at baseline or their development during protocol chemotherapy would have an adverse impact on OS. To our knowledge, this is the first analysis of TEEs at baseline and during chemotherapy and their relation to OS in patients with gastroesophageal cancer.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

We evaluated 191 patients with advanced gastric or gastroesophageal junction adenocarcinoma who participated in approved chemotherapy protocols at the University of Texas M. D. Anderson Cancer Center between August 1997 and September 2003. Patients had a complete history and physical examination, complete blood count, serum chemistry, and computed tomography (CT) scans of the abdomen and pelvis. A CT scan of the chest was obtained when it was indicated clinically (examples include involvement of the lower esophagus, an abnormal chest radiograph, or pulmonary symptoms). Patients did not have any uncontrolled comorbid conditions (uncontrolled diabetes, active angina or congestive heart failure, uncontrolled hypertension, or an active psychiatric condition that would prevent participation and compliance). All patients signed a written informed-consent form and were enrolled into 1 of 4 consecutive clinical trials. Three trials evaluated camptothecin-based chemotherapy (irinotecan, pegylated camptothecin, and rubitecan), and 1 trial investigated noncamptothecin-based chemotherapy. One of the camptothecin-based trials was a combination chemotherapy regimen with cisplatin and irinotecan, and the other 2 trials examined single-agent pegylated-camptothecin and single-agent rubitecan. In the noncamptothecin-based trial, patients received combinations of cisplatin, docetaxel, and fluorouracil as part of a Phase II/III trial.

All data and clinical events were collected prospectively and reported while patients were enrolled on the clinical trial. During treatment, TEEs of the extremities were confirmed by duplex ultrasonography or venography. PEs were confirmed by CT scans, ventilation perfusion studies, or angiography. CT scanning or angiography was used to confirm the presence of intraabdominal TEEs and coronary artery thrombi. Events were categorized as DVT of the extremities, PE, cerebrovascular accidents, myocardial infarction, or intraabdominal TEE. The anatomic location of thrombi or emboli was recorded, and all events were graded based on the National Cancer Institute Common Toxicity Criteria for Adverse Events (version 3.0).10 Events that occurred in patients who had not received any treatment for metastatic disease were used to determine the baseline incidence of TEEs in untreated patients. In treated patients (who received 1 prior chemotherapy regimen for metastatic disease), TEEs prior to the protocol chemotherapy were used as the baseline. TEEs that occurred after the initiation of the protocol chemotherapy and within 60 days of the last dose of chemotherapy were classified as treatment-related TEEs. TEEs that occurred within the immediate postoperative period (within 60 days of surgery) or thrombosis of central venous catheters were excluded from the study, because they could be attributable to causes other than malignancy or chemotherapy. Superficial thrombophlebitis, transient ischemic attacks, and thrombosis that occurred in the setting of disseminated intravascular coagulation were not included in the study. OS was calculated from the date protocol chemotherapy was initiated. Time points until death or last follow-up were determined by chart review or by using the Social Security Death Index. This analysis was approved by the University of Texas M. D. Anderson Cancer Center Institutional Review Board.

Statistical Methods

Descriptive statistics were used to describe patient characteristics. Methods such as the 2-sided chi-square test and the Fisher exact test were used to determine differences in the incidence of TEEs between groups. The Kaplan-Meier method was used to estimate survival, and Mantel-Cox log-rank tests were used to test differences in survival distribution. Data from patients who remained alive at the time of the analysis were censored. All statistical analyses were 2-sided and performed at a significance level of .05.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient Characteristics

Prospectively collected data on 191 patients was rereviewed for the purpose of this analysis. Table 1 presents the age, sex, primary tumor location, tumor histology, and history of prior chemotherapy for all patients at the time of enrollment. With the exception of prior treatment, all 4 protocols had similar criteria for inclusion and exclusion. Age, sex, histology, and location of the primary tumor did not differ significantly between patients with and without TEEs diagnosed at baseline or during protocol chemotherapy.

Table 1. Patient Characteristics
CharacteristicNo. of patients (%)
  1. GE indicates gastroesophageal; DCF, docetaxel, cisplatin, and fluorouracil; DC, docetaxel and cisplatin; CF, cisplatin and fluorouracil.

All patients191 (100)
Sex
 Men151 (79.06)
 Women40 (20.94)
Median age [range], y56 [20–80]
Histology
 Poorly differentiated139 (72.77)
 Moderately differentiated46 (24.08)
 Well differentiated2 (1.05)
 Unknown differentiation4 (2.09)
Location of primary tumor
 Proximal (including GE junction)111 (58.12)
 Medial44 (23.04)
 Distal36 (18.85)
Prior chemotherapy
 No132 (69.11)
 Yes59 (30.89)
Camptothecin-based protocol treatment136 (71.20)
 Irinotecan/cisplatin67 (35.08)
 Pegylated campthecin19 (9.95)
 Rubitecan50 (26.18)
Noncamptothecin-based protocol treatment (DCF, DC, or CF)55 (28.80)

TEEs

The incidence of TEEs at baseline for all patients was 6.3%. In untreated patients, the baseline incidence of TEEs was 5.3% compared with 8.5% in treated patients (P = .519). During the receipt of protocol chemotherapy, TEEs were observed in 7.9% of patients overall. TEEs were observed in 6.8% of untreated patients during protocol chemotherapy and in 10.2% of treated patients during protocol chemotherapy (P = .561). When untreated patients received a camptothecin-based chemotherapy regimen, 6.5% had a treatment-related TEE. TEEs were observed in 7.3% of untreated patients who received noncamptothecin-based chemotherapy. In untreated patients, the chemotherapy regimen with the lowest treated-related incidence of TEEs was the combination chemotherapy regimen of cisplatin and irinotecan (2.6%). In treated patients, the percentage of TEEs increased to 13.8% in patients who received cisplatin and irinotecan (P = .158). The highest incidence of treatment-related TEEs in untreated patients was observed in patients who received rubitecan (15.0%). The incidence of TEEs at baseline or during treatment was not statistically different between any of the different chemotherapy groups (Table 2).

Table 2. Thromboembolic Events at Baseline and During Treatment
Treatment groupNo./Total patients (%)
  1. TEE indicates thromboembolic event;±, with or without.

TEE at baseline
 Overall12/191 (6.28)
 Chemotherapy naive7 /132 (5.30)
  Irinotecan/cisplatin1/38 (2.63)
  Pegylated camptothecin1/19 (5.26)
  Rubitecan3/20 (15.0)
  Cisplatin± fluorouracil± docetaxel2/55 (3.64)
  Camptothecin-based therapy5/77 (6.49)
  Noncamptothecin-based therapy2/55 (3.64)
 Prior chemotherapy5/59 (8.47)
  Irinotecan/cisplatin3/29 (10.34)
  Rubitecan2/30 (6.67)
TEE during treatment
 Overall15/191 (7.85)
 Chemotherapy naive9/132 (6.82)
  Irinotecan/cisplatin1/38 (2.63)
  Pegylated camptothecin1/19 (5.26)
  Rubitecan3/20 (15.0)
  Cisplatin± fluorouracil± docetaxel4/55 (7.27)
  Camptothecin-based therapy5/77 (6.49)
  Noncamptothecin-based therapy4/55 (7.27)
 Prior chemotherapy6/59 (10.17)
  Irinotecan/cisplatin4/29 (13.79)
  Rubitecan2/30 (6.67)

Twenty-six of 191 patients (13.6%) had a TEE documented either at baseline or during protocol chemotherapy. Twelve patients (6.3%) had a TEE diagnosed prior to the start of protocol chemotherapy; and, of those 12 patients, only 1 patient had a recurrent TEE diagnosed during chemotherapy. That patient had an asymptomatic intraabdominal TEE diagnosed prior to chemotherapy and later developed a DVT. DVT of the extremities was the most frequent and accounted for 47.1% of TEEs during chemotherapy. None of the patients who had DVT had a PE diagnosed concurrently. However, 1 patient with DVT had a concurrent portal vein thrombosis. PE was observed in 4 patients on protocol chemotherapy. One event resulted in death, and 1 patient had a portal and splenic vein thrombosis diagnosed concurrently (Table 3).

Table 3. Distribution of Thromboembolic Events During Chemotherapy
LocationNo. of TEEs (%)
All TEEs17 (100)
Deep vein thrombosis8 (47.1)
Pulmonary embolism4 (23.5)
Myocardial infarction1 (5.9)
Cerebral vascular accident0 (0)
Intraabdominal thrombosis/embolism4 (23.5)

TEE and OS

The median OS for all 191 patients was 8.3 months. Two patients remained alive >24 months after the completion of protocol chemotherapy. In patients who had a TEE diagnosed at baseline or during treatment, the median OS was only 3.9 months compared with 8.7 months for patients who did not have a TEE (P = .007) (Fig. 1). In a subgroup analysis, the difference in median OS remained significant for untreated patients who had a TEE at baseline or during chemotherapy compared with patients who did not have a TEE (6.6 months vs 9 months; P = .025). The survival difference did not reach statistical significance for treated patients (most likely because of the small number of patients in this group).

thumbnail image

Figure 1. This Kaplan-Meier plot compares the overall survival of patients who had a thromboembolic event (TEE) at baseline or during chemotherapy with the overall survival of patients who did not have a TEE.

Download figure to PowerPoint

At baseline, patients who had a TEE had a median OS of only 3 months compared with a median OS of 8.4 months for patients who did not have a TEE (P = .042). In untreated patients who had a TEE at baseline, the median OS was 5 months compared with 8.9 months in patients who did not have a TEE (P = .014) (Fig. 2). In previously treated patients, a difference in OS was not observed between patients who had a TEE prior to protocol chemotherapy and patients who did not have a TEE prior to protocol chemotherapy. Again, the lack of significance may have been caused by the small number of patients in this group.

thumbnail image

Figure 2. This Kaplan-Meier plot compares the overall survival of untreated patients who had a thromboembolic event (TEE) at baseline with the overall survival of patients who did not have a TEE.

Download figure to PowerPoint

Patients who had a TEE diagnosed during chemotherapy had a median survival of 4.8 months compared with 8.5 months in patients who did not have a TEE during chemotherapy (P = .078) (Fig. 3). Untreated patients who developed a TEE during protocol chemotherapy had a median OS of 6.7 months compared with a median OS of 9 months in patients who did not develop a TEE during chemotherapy (P = .269). In treated patients who had a TEE during protocol chemotherapy, the median OS was 3.7 months compared with 6.3 months in patients who did not have a TEE during chemotherapy (P = .168). The median survival for all groups is presented in Table4. OS was not affected by histologic differentiation (P = .871), primary tumor location (P = .215), sex (P = 751), or age (P = .083).

thumbnail image

Figure 3. This Kaplan-Meier plot compares the overall survival of patients who had a thromboembolic event (TEE) during treatment (Tx) (green line) with the overall survival of patients who did not have a TEE (blue line).

Download figure to PowerPoint

Table 4. Kaplan-Meier Survival Estimates
Treatment groupMedian survival, moP
TEENo TEE
  1. TEE indicates thromboembolic event.

TEE at baseline
 Overall38.4.042
  Chemotherapy naive4.98.9.014
  Prior chemotherapy35.8.721
TEE during treatment
 Overall4.88.5.078
  Chemotherapy naive6.79.269
  Prior chemotherapy3.76.3.168
TEE at baseline or during treatment
 Overall3.98.7.007
  Chemotherapy naive6.69.025
  Prior chemotherapy3.76.5.216

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

It is understood generally that gastroesophageal carcinomas are associated with a high rate of thrombosis. However, the largest studies of cancer and thrombosis are limited to Medicare database reviews of patients who were hospitalized with a diagnosis of cancer and thrombosis.2, 3 The clinical relevance of this information is unclear. Furthermore, there have been scarce data regarding the incidence of TEEs in patients with advanced gastroesophageal cancer undergoing chemotherapy. In addition, to our knowledge, the impact of TEEs diagnosed at baseline or during chemotherapy has not been reported in this group of patients. Klerk et al.8 studied the effect of low-molecular-weight heparin or placebo as prophylaxis in a randomized fashion in 302 patients with solid tumors and excluded the patients who had a baseline TEE. That intent-to-treat analysis demonstrated a statistically significant improvement in OS (P = .021) in favor of the group that received low-molecular-weight heparin compared with the group that received placebo. With this observation, our hypothesis that TEEs (at baseline or during chemotherapy) would have an adverse impact on OS is strengthened further. Our data support the finding TEEs at baseline or during chemotherapy result in poorer OS for patients with advanced gastroesophageal cancer who are receiving protocol chemotherapy (P = .007) (Fig. 1). In addition, our data demonstrate the adverse impact of TEEs at baseline in untreated patients (P = .014) (Fig. 2) and of TEEs that develop during protocol chemotherapy in both patient populations studied, although the impact was not statistically significant but did show a trend (P = .078). The adverse impact of TEEs at baseline in treated patients did not reach statistical significance; however, this most likely was because of the small number of patients in this group.

Another issue related to the development of TEE is that of chemotherapy agents. Irinotecan, which is a semisynthetic analogue, has been associated with fatal TEEs.11 In addition, it has been demonstrated that biotherapy, such as bevacizumab, has a higher incidence of arterial TEEs. In a pooled analysis of 5 randomized, controlled trials, arterial TEEs were more common in the 963 patients who received bevacizumab and chemotherapy compared with the 782 patients who received chemotherapy alone (4.4% vs 1.9%; P < .01).12 In the current analysis, untreated patients who received cisplatin and irinotecan had rates of TEE similar to those in patients who received other agents. Our data on all patients who received any of the camptothecins also suggest that the incidence of TEE was not increased compared with the incidence in patients who received other classes of agents. Because agents such as bevacizumab are more likely to be used in future investigations, our data on the development of TEEs with chemotherapy alone provide a baseline that has not been available previously.

Why TEEs result in poor survival needs to be pondered. There is no clear answer available at the moment. It has been reported that patients with cancer diagnosed at the time of venous thromboembolism or within 1 year more often have metastatic disease and a lower survival rate than patients without venous thromboembolism.4 TEEs may be related to or manifestations of aggressive cancer biology (metastatic potential) while imparting morbidity of their own. However, further research in this area to include consideration for prophylactic anticoagulation for protocol patients with advanced gastroesophageal cancer is needed. A well-designed clinical trial would randomize patients to treatment with low-molecular-weight heparin or placebo for a defined period. Survival would be the primary endpoint, and the study would be powered to detect a clinically relevant difference in survival. Correlative and translational studies would be completed to investigate the impact of biochemotherapy, tumor biology, and patient genetics. The concerns regarding morbidity from prophylactic anticoagulation are legitimate; because, in the study by Klerk et al., a higher rate of bleeding was reported in the treatment group than in the placebo group.8 Secondary endpoints would have to include the incidence of TEE, bleeding complications, patient compliance, and quality of life. Currently, prophylactic anticoagulation remains controversial, and further studies are needed to direct the development of treatment guidelines.

To decrease the probability of overtreating patients with anticoagulation, it is possible that, before starting chemotherapy, patients with a TEE could be selected for prolonged anticoagulation therapy based on D-dimer assays. Palareti et al. conducted a study of 627 patients with a history of idiopathic venous thromboembolism who had been treated previously for at least 3 months with a vitamin K antagonist.13 Patients with abnormal D-dimer assays were assigned randomly to resume or discontinue treatment. Patients with normal D-dimer levels did not resume anticoagulation. In patients who had abnormal D-dimer levels, the incidence of recurrent thromboembolism was reduced by resuming anticoagulation (hazard ratio, 4.26; 95% confidence interval, 1.23–14.6; P = .02). The utility of D-dimer assays needs to be studied in a population of cancer patients to assess the need and potential benefit of prolonged anticoagulation with low-molecular-weight heparin.

We studied only those patients who were participating on approved chemotherapy protocols; thus, the data reported here were collected prospectively. The patients were staged thoroughly and observed closely, and no patients were excluded from the analysis. We acknowledge that a limitation to this study was the lack of a suitable control group. However, it would be difficult to find a control group of patients with advanced gastroesophageal cancer who are eligible for protocol chemotherapy and could be assigned to observation. Another consideration is the evolving imaging technology, which may detect an unsuspected TEE. Most of the TEEs in our report were DVTs, and very few were PEs. Nevertheless, the impact of improved imaging could not be addressed in our study or in any study with an observation/recruitment span >2 years. In patients with gastroesophageal cancers, CT scans of the chest are obtained only when they are indicated clinically; and, in our study, chest CT scans were obtained in 56.5% of patients. If all patients had routine CT scans of the chest, then the results may have been different; however, considering the low frequency of confirmed PEs in our study, we do not believe that the difference would have been substantial.

In conclusion, to our knowledge, this report presents the first data that demonstrate the frequency of TEEs at baseline and during chemotherapy in patients with advanced gastroesophageal cancer who were enrolled on protocol chemotherapy. In addition, the diagnosis of a TEE (at baseline or during chemotherapy) had an adverse impact on OS. Further research is warranted on methods to include aggressive therapy for TEEs and their prevention to improve the survival of patients with advanced gastroesophageal carcinoma.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Trousseau A. Phlegmasia alba dolens. In: BaillierJB, ed. Clinique Medicale de l'Hotel-Dieu de Paris.2nd ed. London: New Sydenham Society; 1865: 3: 9495.
  • 2
    Levitan N,Dowlati A,Remick S, et al. Rates of initial and recurrent thromboembolic disease among patients with malignancy versus those without malignancy: the risk analysis using Medicare claims data. Medicine. 1999; 78: 285291.
  • 3
    Sorensen H,Mellemkjaer L,Steffensen F, et al. The risk of a diagnosis of cancer after primary deep venous thrombosis or pulmonary embolis. N Engl J Med. 1998; 338: 11691173.
  • 4
    Sorensen H,Mellemkjaer L,Olsen J, et al. Prognosis of cancers associated with venous thromboembolism. N Engl J Med. 200; 343: 18461850.
  • 5
    Fisher B,Castantino J,Redmond C, et al. A randomized clinical trial evaluating tamoxifen in the treatment of patients with node-negate breast cancer who have estrogen-receptor-positive tumors. N Engl J Med. 1989; 320: 479484.
  • 6
    Fisher B,Digman J,Wolmark N, et al. Tamoxifen and chemotherapy for lymph node-negative, estrogen receptor-positive breast cancer. J Natl Cancer Inst. 1997; 89: 16731682.
  • 7
    Lee A,Levine M. The thrombophiliic state induced by therapeutic agents in cancer patients. Semin Thromb Hemost. 1999; 25: 137145.
  • 8
    Klerk CP,Smorenburg SM,Otten HM, et al. The effect of low molecular weight heparin on survival in patients with advanced malignancy. J Clin Oncol. 2005; 23: 21302135.
  • 9
    Shah M,Ilson D,Kelson D. Thromboembolic events in gastric cancer: high incidence in patients receiving irinotecan- and bevacizumab-based therapy [letter]. J Clin Oncol. 2005; 23: 25742526.
  • 10
    National Cancer Institute Website. Available at URL: http://ctep.cancer.gov/reporting-/ctc.html Accessed August 29, 2005.
  • 11
    Rothenberg ML,Meropol NJ,Poplin EA, et al. Mortality associated with irinotecan plus bolus fluorouracil/leucovorin: summary findings of an independent panel. J Clin Oncol. 2001; 19: 38013807.
  • 12
    Skillings JR,Johnson DH,Miller K, et al. Arterial thromboembolic events (ATEs) in a pooled analysis of 5 randomized, controlled trials (RCTs) of bevacizumab (BV) with chemotherapy. J Clin Oncol. 2005; 23( suppl): 196s. Abstract 3019.
  • 13
    Palareti G,Cosmi B,Legnani C, et al. D-dimer testing to determine duration of anticoagulation therapy. N Engl J Med. 2006; 355: 1781790.