Correspondence
Venous Thromboembolism and Cancer
N Engl J Med 2000; 343:1337-1338November 2, 2000
- Article
To the Editor:
In their report concerning newly diagnosed cancer in patients with venous thromboembolism, Schulman and Lindmarker, for the Duration of Anticoagulation Trial (June 29 issue),1 state that their findings strongly support the impression that warfarin has an antineoplastic effect. However, the study has a number of important limitations.
First, the number of patients who were lost to follow-up is not reported. Second, an attempt was made to identify all cases of newly diagnosed cancer through the Swedish Cancer Registry, but more than 20 percent of cancers are not reported to this registry. This high rate of underreporting is a cause for concern because the difference between the two groups in the reported rate of events was modest. The possibility of incomplete case ascertainment is further suggested by the low incidence of breast cancer and of colorectal cancer reported at eight years of follow-up. Third, mortality data are not provided. The apparent lack of an effect of warfarin on the incidence of death from cancer may be due to insufficient follow-up or to inaccuracies in the reported cause of death, which are extremely common,2 even in Sweden.3 The results of this study might have been more convincing if data had been presented to demonstrate that there were no differences in cause-specific mortality and all-cause mortality. Finally, the data on which the authors' conclusions are based represent an a posteriori, secondary analysis that should be interpreted with great caution and in the light of all available information about the disease and potential pharmacologic mechanisms of its treatment.4 In this context, the excess incidence of urogenital cancers in the patients treated for six weeks with warfarin, which accounts entirely for the observed difference in the incidence of cancer between the two groups, is inconsistent with findings in previous reports5 and lacks known biologic mechanisms. This highlights the possibility that the higher incidence of newly diagnosed cancer in patients treated for six weeks with warfarin than in those treated for six months with warfarin may simply be due to chance.
5 ReferencesJohn W. Eikelboom, M.B., B.S.
Shamir R. Mehta, M.D.
McMaster University, Hamilton, ON L8L 2X2, Canada1
Schulman S ,Lindmarker P . Incidence of cancer after prophylaxis with warfarin against recurrent venous thromboembolism. N Engl J Med 2000;342:1953-1958
Full Text | Web of Science | Medline2
Kircher T ,Nelson J ,Burdo H . The autopsy as a measure of accuracy of the death certificate. N Engl J Med 1985;313:1263-1269
Full Text | Web of Science | Medline3
Johansson LA ,Westerling R . Comparing Swedish hospital discharge records with death certificates: implications for mortality statistics. Int J Epidemiol 2000;29:495-502
CrossRef | Web of Science | Medline4
Yusuf S ,Wittes J ,Probstfield J ,Tyroler HA . Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. JAMA 1991;266:93-98
CrossRef | Web of Science | Medline5
Zacharski LR ,Henderson WG ,Rickles FR , et al. Effect of warfarin anticoagulation on survival in carcinoma of the lung, colon, head and neck, and prostate: final report of VA Cooperative Study #75. Cancer 1984;53:2046-2052
CrossRef | Web of Science | Medline
To the Editor:
After six to eight years, Schulman and Lindmarker found that the rate of prostate cancer, though not of death or other cancers, was higher in patients given warfarin for six weeks after thrombosis than in patients given warfarin for six months. They dismiss the possibility of a difference in screening because most cancers were found after the second year. However, borderline levels of prostate-specific antigen found initially might lead to the diagnosis, years later, in men without symptoms. Did the groups differ in terms of medical supervision during the first six months, and thus possibly in the number tested for prostate-specific antigen? Patients with a probable cause for their thrombosis are less likely to have been tested, and this might affect the difference in the risk of cancer (Figure 3 of the article).
It would be interesting to know the number of elderly men in each group and the criteria for a diagnosis of prostate cancer.
T.H. Hughes-Davies, F.R.C.P.
Breamore Marsh, Fordingbridge, Hampshire SP6 2EJ, United KingdomTo the Editor:
Schulman and Lindmarker made an important contribution to the treatment of venous thromboembolism by demonstrating that anticoagulation for six months is more effective than anticoagulation for six weeks.1 However, in their recent article they may have incorrectly concluded that six months of oral anticoagulation has an antineoplastic effect. The principal question is whether the difference in the incidence of cancer they noted was due to a lower incidence in the six-month group or a higher incidence in the six-week group. The answer may lie in the standardized incidence ratios for cancer in their two groups as compared with that in patients with deep-vein thrombosis as estimated from retrospective reviews of national registries.2 Figure 3 of the article by Schulman and Lindmarker and Figure 1 of the article by Sørensen et al.2 show that the standardized incidence ratio in the six-month group was similar to, and that the standardized incidence ratio in the six-week group was higher than, that in patients with thrombosis as reported by Sørensen et al. Thus, the suggestion that six months of oral anticoagulation has anticancer properties may be erroneous. Could a higher incidence of cancer in the six-week group explain the difference noted by Schulman and Lindmarker? The higher rate of recurrent thrombosis in the six-week group could have led to more sequential screening and a higher frequency of newly diagnosed cancers.
2 ReferencesDebasish Roychowdhury, M.D.
Lilly Research Laboratories, Indianapolis, IN 462851
Schulman S ,Rhedin A-S ,Lindmarker P , et al. A comparison of six weeks with six months of oral anticoagulant therapy after a first episodeof venous thromboembolism. N Engl J Med 1995;332:1661-1665
Full Text | Web of Science | Medline2
Sorensen HT ,Mellemkjaer L ,Steffensen FH ,Olsen JH ,Nielsen GL . The risk of a diagnosis of cancer after primary deep venous thrombosis or pulmonary embolism. N Engl J Med 1998;338:1169-1173
Full Text | Web of Science | Medline
To the Editor:
Schulman and Lindmarker state that the earlier demonstration of improved patient outcome with warfarin treatment for small-cell carcinoma of the lung has never been confirmed. On the contrary, two prospective, randomized trials of warfarin and one of anticoagulation with unfractionated heparin showed improved outcome in patients with small-cell carcinoma of the lung.1 The lack of effect in a third study of warfarin in patients with small-cell carcinoma of the lung was explained by the fact that the patients had disease of an earlier stage than the patients in the previous studies and received more intense radiation and chemotherapy, which probably resulted in increased thrombin generation and negated the modest anticoagulant effect of warfarin.2 Small-cell carcinoma of the lung is one of a small number of tumors that express thrombin in situ.3 Tumors that do not express thrombin, including prostate cancer, are unaffected by treatment with warfarin.2
Therefore, we do not believe that the anticoagulant activity of warfarin accounts for the findings reported by Schulman and Lindmarker. The fact that a reduced risk of cancer was observed years after exposure suggests an effect of warfarin at the stage of tumor initiation or promotion rather than an effect on an established tumor. Supporting this idea is evidence that coumarin derivatives may influence carcinogenesis, the growth of prostate tissue, and the immunologic status of the host.2
3 ReferencesLeo R. Zacharski, M.D.
Veterans Affairs Medical Center, White River Junction, VT 05009Deborah L. Ornstein, M.D.
Wilford Hall Medical Center, Lackland AFB, TX 782361
Zacharski LR, Rickles FR. Warfarin in the treatment of cancer. In: Poller L, Hirsh J, eds. Oral anticoagulants. London: Arnold, 1996:228-38.
2
Maurer LH ,Herndon JE II ,Hollis DR , et al. Randomized trial of chemotherapy and radiation therapy with or without warfarin for limited-stage small-cell lung cancer: a Cancer and Leukemia Group B Study. J Clin Oncol 1997;15:3378-3387
Web of Science | Medline3
Zacharski LR ,Memoli VA ,Morain WD ,Schlaeppi J-M ,Rousseau SM . Cellular localization of enzymatically active thrombin in intact human tissues by hirudin binding. Thromb Haemost 1995;73:793-797
Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Eikelboom and Mehta as well as Roychowdhury ignore the fact that our data are based on information from prospective clinical follow-up in combination with information from the Swedish Cancer Registry. This probably explains why we found a higher incidence of cancer in the six-week group than did Sørensen et al.1 A direct comparison with that study or with the study by Baron et al.2 is also hampered by other differences; we included only patients with objectively verified venous thromboembolism and excluded all patients with previous cancer. Loss from clinical follow-up at any time before 72 months was 4.6 percent in our trial and was equal in the two treatment groups. This loss does not affect the reporting to the Swedish Cancer Registry. Cancers unknown to the clinician and the patient may have been missed in instances of death without autopsy. We do not agree with Eikelboom and Mehta that 47 percent more cancers in the six-week group is only a modest increase.
The incidence of breast or colorectal cancer is not presented in our article, but the numbers in most of the subgroups of cancer are low and therefore cannot be used as evidence of incomplete case ascertainment. The article cited about Swedish mortality statistics does not show that we have a problem with inaccuracies but rather that the diagnoses in the hospital-discharge summary and in the death certificate do not always match.3 We are planning to publish data on cause-specific and all-cause mortality elsewhere and are unable to provide details here, but the results will not change our previous conclusions.
In our discussion, we stated that a weakness of our study was that it was not designed to demonstrate a difference between the treatment groups in the incidence of cancer. However, the study by Zacharski et al.4 cannot be used to negate our results. They tested the effect of warfarin mainly on disseminated cancers, such as prostate cancer with bone metastases and disseminated or recurrent colorectal cancer. We agree with Zacharski and Ornstein that the possible antineoplastic effect is probably on tumor initiation and promotion.
In response to Hughes-Davies: There was no difference between the treatment groups in medical supervision at any time, and as stated in the Results section, screening for prostate cancer contributed to the diagnosis in only one case. The numbers of men 80 years old or older at enrollment were 12 and 17 in the six-week and six-month groups, respectively.
4 ReferencesSam Schulman, M.D.
Per Lindmarker, M.D.
Karolinska Hospital, SE-171 76 Stockholm, Sweden1
Sorensen HT ,Mellemkjaer L ,Steffensen FH ,Olsen JH ,Nielsen GL . The risk of a diagnosis of cancer after primary deep venous thrombosis or pulmonary embolism. N Engl J Med 1998;338:1169-1173
Full Text | Web of Science | Medline2
Baron JA ,Gridley G ,Weiderpass E ,Nyren O ,Linet M . Venous thromboembolism and cancer. Lancet 1998;351:1077-1080[Erratum, Lancet 2000;355:758.]
CrossRef | Web of Science | Medline3
Johansson LA ,Westerling R . Comparing Swedish hospital discharge records with death certificates: implications for mortality statistics. Int J Epidemiol 2000;29:495-502
CrossRef | Web of Science | Medline4
Zacharski LR ,Henderson WG ,Rickles FR , et al. Effect of warfarin anticoagulation on survival in carcinoma of the lung, colon, head and neck, and prostate: final report of VA Cooperative Study #75. Cancer 1984;53:2046-2052
CrossRef | Web of Science | Medline
- Citing Articles (3)
Citing Articles
1
T. P. Ahern, L. Pedersen, C. Svaerke, K. J. Rothman, H. T. Sorensen, T. L. Lash. (2011) The Association Between Vitamin K Antagonist Therapy and Site-specific Cancer Incidence Estimated by Using Heart Valve Replacement as an Instrumental Variable. American Journal of Epidemiology 174:12, 1382-1390
CrossRef2
M. R. Taliani, G. Agnelli, P. Prandoni, C. Becattini, M. Moia, M. Bazzan, W. Ageno, C. Tomasi, G. Guazzaloca, G. B. Ambrosio, A. Bertoldi, R. Salvi, R. Poggio, M. Silingardi, . (2003) Incidence of cancer after a first episode of idiopathic venous thromboembolism treated with 3 months or 1 year of oral anticoagulation. Journal of Thrombosis and Haemostasis 1:8, 1730-1733
CrossRef3
Leo R Zacharski. (2002) Anticoagulants in cancer treatment: malignancy as a solid phase coagulopathy. Cancer Letters 186:1, 1-9
CrossRef
