Correspondence
Venous Thrombosis and Cancer
N Engl J Med 1993; 328:885-887March 25, 1993
- Article
To the Editor:
Armand Trousseau died in 1867. Therefore, he was incapable of making observations in 1868 or lecturing in 1872, as suggested by Prandoni et al.1 and Silverstein and Nachman2 (Oct. 15 issue). These investigators cited an English translation of Trousseau's work without realizing that it had been published posthumously.
Trousseau published his initial observations of cancer associated with venous thrombosis in 18653. In 1866 he had the frightening experience of diagnosing his own syndrome in himself, when unilateral leg swelling developed4. He died the following year of gastric carcinoma5.
5 ReferencesEvan S. Siegelman, M.D.
Laurence Needleman, M.D.
Jefferson Medical College, Philadelphia, PA 191071
Prandoni P ,Lensing AWA ,Buller HR , et al. Deep-vein thrombosis and the incidence of subsequent symptomatic cancer. N Engl J Med 1992;327:1128-1133
Full Text | Web of Science | Medline2
Silverstein RL ,Nachman RL . Cancer and clotting -- Trousseau's warning. N Engl J Med 1992;327:1163-1164
Full Text | Web of Science | Medline3
Trousseau A. Phlegmasia alba dolens. In: Clinique Medicale de l'Hotel-Dieu de Paris. 2nd ed. Vol. 3. Paris: JB Bailliere, 1865:654-712.
4
Pioneer clinicianMD Medical Newsmagazine 1967;11:247-253
5
Lourie J. Medical eponyms: who was Coude? London: Pitman Publishing, 1982:184.
To the Editor:
Prandoni and colleagues make statistical errors that may very well affect their conclusions. They compare the incidence of cancer among three study groups: patients with secondary thrombosis, patients with idiopathic thrombosis, and patients with recurrent idiopathic thrombosis.
The comparisons of the length of time until the development of cancer between the group with recurrent idiopathic thrombosis and each of the other two groups are invalid. The comparison between the group with recurrent idiopathic thrombosis and the group with idiopathic thrombosis is flawed because the group of 35 patients with recurrent idiopathic thrombosis is a subgroup of the 145 patients with idiopathic thrombosis. The Mantel-Haenszel (log-rank) method assumes that independent groups of patients are being compared.
Moreover, although follow-up for the patients with idiopathic thrombosis begins at the initial diagnosis of thrombosis, follow-up for the patients with recurrent idiopathic thrombosis is calculated from the subsequent thrombotic episode that defines this group. Thus, the six events occurring in the 35 patients contained in both the idiopathic-thrombosis and recurrent-idiopathic-thrombosis groups do not match in time, because the time from the initial to the subsequent thrombosis has been subtracted from the follow-up of the patients in the group with recurrent idiopathic thrombosis. This statistical comparison is therefore also wrong, because the time scale for the comparison differs for the two groups. The comparison between the group with recurrent idiopathic thrombosis and the group with secondary thrombosis is also flawed, for the same reason. Appropriate comparison of the recurrent-idiopathic-thrombosis group with the other groups would require statistical methods designed to compare groups whose membership is determined by events occurring during follow-up: a time-dependent Cox model or the Mantel-Byar procedure.1
Nevertheless, the primary comparison between patients with idiopathic thrombosis and patients with secondary thrombosis is valid and suggests that unrecognized cancer may be a cause of a subgroup of these idiopathic thromboses.
1 ReferencesJames R. Anderson, Ph.D.
Paula K. Roberson, Ph.D.
University of Nebraska Medical Center, Omaha, NE 68198To the Editor:
Regarding the recommendation by Silverstein and Nachman of carcinoembryonic antigen as a screening test for cancer, we believe that there is no proof of its value and that it will lead to many unnecessary endoscopic procedures. Only 30 to 40 percent of patients with localized colon cancer have elevated titers of carcinoembryonic antigen1,2. Considering that these patients will almost all be taking anticoagulants, it would seem likely that tests of the stool for occult blood would be a much more useful way to screen for cancer. It is unlikely that a patient with a colonic tumor would have a stool negative for blood if the test was done properly on three separate days.
Our greater concern is the number of unnecessary and unrewarding colonoscopic procedures that a borderline elevation in carcinoembryonic antigen would lead to.
2 ReferencesBarry H. Epstein, M.D., F.A.C.P., F.A.C.G.
Anil Kankaria, M.D.
Prince George's Hospital Center, Cheverly, MD 207851
Woolfson K . Tumor markers in cancer of the colon and rectum. Dis Colon Rectum 1991;34:506-511
CrossRef | Web of Science | Medline2
Go VL . Carcinoembryonic antigen: clinical application. Cancer 1976;37:562-566
CrossRef | Web of Science | Medline
To the Editor:
I challenge the recommendation in the editorial by Silverstein and Nachman that all patients with idiopathic deep venous thrombosis should be examined for occult neoplasms. A malignant tumor should be suspected if thrombosis is recurrent, occurs at an odd site, or is associated with thrombophlebitis, although in patients with such tumors the presence of an underlying cancer is usually obvious1.
The careful study by Prandoni et al. suggests that idiopathic deep venous thrombosis may precede the diagnosis of an underlying cancer,2 but such a link is not adequately supported by the other three articles cited in the accompanying editorial3-5. A major concern is how the term “idiopathic” is interpreted. I understand it to mean that no precipitating factor for deep venous thrombosis has been identified, yet in two of the four studies cited the incidence of “idiopathic” deep venous thrombosis was 42 percent (145 of 3422 and 35 of 833), whereas in other series deep venous thrombosis was denoted idiopathic in less than 20 percent of the instances4,6,7. In another study, four of seven patients with “idiopathic” deep venous thrombosis in whom overt cancer later developed had either lost weight or had an abnormal chest film at presentation4. A critical reading of these reports shows that patients with recurrent or atypical-site thrombosis are included,3,5 that venography was not always used to establish the diagnosis of deep venous thrombosis,3,5 and that it is unclear whether any of the patients with idiopathic deep venous thrombosis had coexisting thrombophlebitis, which would raise the possibility of an underlying neoplasm2-5. Finally, the editorial ignores work that contradicts the purported linkage between truly idiopathic deep venous thrombosis and the later development of cancer6-8. Although these three series were retrospective, they studied all patients given diagnoses of lower-limb deep venous thrombosis over a designated period, and follow-up data were provided on 98 to 100 percent of the cases.
Even if a link between idiopathic deep venous thrombosis and malignant disease of later onset had been satisfactorily established, it would be a far cry from concluding that such patients would benefit from early diagnosis of an underlying cancer2. A solitary idiopathic lower-limb deep venous thrombosis without thrombophlebitis merits a hunt for cancer only if there are abnormalities on physical examination, blood count, erythrocyte sedimentation rate, or chest film6-8.
8 ReferencesNigel O'Connor, M.D., M.R.C.P.
Royal Shrewsbury Hospital, Shrewsbury SY3 8QR, United Kingdom1
Sack GH Jr ,Levin J ,Bell WR . Trousseau's syndrome and other manifestations of chronic disseminated coagulopathy in patients with neoplasms. Medicine (Baltimore) 1977;56:1-37
Web of Science | Medline2
Prandoni P ,Lensing AWA ,Buller HR , et al. Deep-vein thrombosis and the incidence of subsequent symptomatic cancer. N Engl J Med 1992;327:1128-1133
Full Text | Web of Science | Medline3
Aderka D ,Brown A ,Zelikovski A ,Pinkhas J . Idiopathic deep vein thrombosis in an apparently healthy patient as a premonitory sign of occult cancer. Cancer 1986;57:1846-1849
CrossRef | Web of Science | Medline4
Monreal M ,Lafoz E ,Casals A , et al. Occult cancer in patients with deep venous thrombosis: a systematic approach. Cancer 1991;67:541-545
CrossRef | Web of Science | Medline5
Goldberg RJ ,Seneff M ,Gore JM , et al. Occult malignant neoplasm in patients with deep venous thrombosis. Arch Intern Med 1987;147:251-253
CrossRef | Web of Science | Medline6
Anlyan WG ,Shingleton WW ,DeLaughter GD Jr . Significance of idiopathic venous thrombosis and hidden cancer. JAMA 1956;161:964-966
Web of Science | Medline7
O'Connor NT ,Cederholm-Williams SA ,Fletcher EW ,Allington M ,Sharp AA . Significance of idiopathic deep venous thrombosis. Postgrad Med J 1984;60:275-277
CrossRef | Web of Science | Medline8
Griffin MR ,Stanson AW ,Brown ML , et al. Deep venous thrombosis and pulmonary embolism: risk of subsequent malignant neoplasms. Arch Intern Med 1987;147:1907-1911
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: We thank Drs. Siegelman and Needleman for their historical footnote. Drs. Anderson and Roberson state that statistical errors may affect the validity of the increased risk of malignant disease that we observed in patients with recurrent idiopathic deep-vein thrombosis. As we indicated in our article, with regard to the incidence of malignant disease we compared patients with recurrent idiopathic venous thromboembolism with patients with idiopathic deep-vein thrombosis that did not recur and patients with secondary thrombosis, using logistic-regression analysis. This type of statistical analysis is appropriate and sufficient for this purpose. The results were both statistically significant and clinically important (P = 0.024 for the comparison with patients with idiopathic thrombosis, and P = 0.008 for the comparison with patients with secondary thrombosis; odds ratio, 4.3 and 9.8, respectively), and they form the basis for our conclusions.
In addition, we presented curves of the cumulative incidences of malignant disease to show the occurrence of such disease in relation to the time elapsed since the thrombotic episode. For the subgroup of patients with recurrent idiopathic venous thromboembolism, this curve was calculated as a function of the time elapsed since the recurrent episode (i.e., they were regarded as a cohort assembled at the moment of their recurrence). This latter curve is of particular interest to the clinician, since patients in whom recurrent venous thromboembolism will later develop cannot be identified at the moment of their first event.
Anthonie W.A. Lensing, M.D., Ph.D.
Martin H. Prins, M.D.
Academic Medical Center, 1105 AZ Amsterdam, the NetherlandsPaolo Prandoni, M.D., Ph.D.
University Hospital, 35100 Padua, ItalyAuthor/Editor Response
We thank Drs. Siegelman and Needleman for correcting our historical revisionism. We share the concern of Drs. Epstein and Kankaria that routine population screening with carcinoembryonic antigen might result in a large number of negative (and expensive) colonoscopic examinations. The reason for this concern is not as clear, however, in patients with idiopathic deep venous thrombosis who have a serious risk of underlying cancer.
We do not agree that the studies reported by Aderka et al.,1 Monreal et al.,2 and Goldberg et al.,3 referred to in our editorial, fail to support the conclusion that idiopathic deep venous thrombosis may precede the diagnosis of an underlying cancer. We agree that the concern with the incidence of “idiopathic” thrombosis is important, but in fact all three published prospective studies report similar incidences of idiopathic deep venous thrombosis (27 percent,2 42 percent,1 and 42 percent4). The two studies noted by Dr. O'Connor that report incidences of less than 20 percent were both retrospective5,6. One of these, the one by Anlyan et al.,5 is seriously flawed, in that only patients seen on the surgical and gynecology services were included, and no information was provided about how the diagnoses of deep venous thrombosis were made. In 1956, when this study was published, diagnostic methods were much less sophisticated than at present and were likely to include a substantial number of false positive errors. Although it is true that the study by Aderka et al.1 did not rely on phlebography and included patients with recurrent and unusual thrombotic events, these presentations did not have a statistically significant effect on the risk of underlying cancer. Impedance plethysmography is reported in numerous large studies to have sensitivities and specificities of more than 95 percent, so it is unlikely that the failure to perform contrast phlebography influenced this study.
The three retrospective studies referred to by Dr. O'Connor were not mentioned in our editorial because their nature did not allow conclusions to be drawn. The study by Griffin et al.7 did in fact show a twofold increase in the relative risk of cancer in patients with deep venous thrombosis as compared with all residents of Olmsted County. The design of the study, however, did not distinguish between idiopathic and secondary deep venous thrombosis. In addition, there were significant differences between patients and controls in such important variables as age, sex, and underlying diseases. The retrospective study by Anlyan et al.5 was seriously flawed, as mentioned above, whereas the study by O'Connor et al.6 included only 15 patients with idiopathic deep venous thrombosis (as compared with 153 in the study by Prandoni et al4). Finally, although none of the prospective studies reported the incidence of “coexisting thrombophlebitis” in patients with deep venous thrombosis, we know of no studies showing that this might increase the risk of underlying cancer.
7 ReferencesRoy L. Silverstein, M.D.
Ralph L. Nachman, M.D.
Cornell University Medical College, New York, NY 100211
Aderka D ,Brown A ,Zelikovski A ,Pinkhas J . Idiopathic deep vein thrombosis in an apparently healthy patient as a premonitory sign of occult cancer. Cancer 1986;57:1846-1849
CrossRef | Web of Science | Medline2
Monreal M ,Lafoz E ,Casals A , et al. Occult cancer in patients with deep venous thrombosis: a systematic approach. Cancer 1991;67:541-545
CrossRef | Web of Science | Medline3
Goldberg RJ ,Seneff M ,Gore JM , et al. Occult malignant neoplasm in patients with deep venous thrombosis. Arch Intern Med 1987;147:251-253
CrossRef | Web of Science | Medline4
Prandoni P ,Lensing AWA ,Buller HR , et al. Deep-vein thrombosis and the incidence of subsequent symptomatic cancer. N Engl J Med 1992;327:1128-1133
Full Text | Web of Science | Medline5
Anlyan WG ,Shingleton WW ,DeLaughter GD Jr . Significance of idiopathic venous thrombosis and hidden cancer. JAMA 1956;161:964-966
Web of Science | Medline6
O'Connor NT ,Cederholm-Williams SA ,Fletcher EW ,Allington M ,Sharp AA . Significance of idiopathic deep venous thrombosis. Postgrad Med J 1984;60:275-277
CrossRef | Web of Science | Medline7
Griffin MR ,Stanson AW ,Brown ML , et al. Deep venous thrombosis and pulmonary embolism: risk of subsequent malignant neoplasms. Arch Intern Med 1987;147:1907-1911
CrossRef | Web of Science | Medline
- Citing Articles (4)
Citing Articles
1
B. Psaila, D. Lyden, I. Roberts. (2011) Megakaryocytes, Malignancy and Bone Marrow Vascular Niches. Journal of Thrombosis and Haemostasisno-no
CrossRef2
Christina Hart, Reinhard Andreesen, Joachim R. Hahn. (2009) Thromboembolische Erkrankungen in der Hämatologie und Onkologie / Thrombotic disease in cancer patients. LaboratoriumsMedizin 33:3, 133-139
CrossRef3
Christina Hart, Reinhard Andreesen, Joachim R. Hahn. (2009) Thrombotic disease in cancer patients 1. LaboratoriumsMedizin 33:3, ---
CrossRef4
Rohan J. K. Hettiarachchi, Judith Lok, Martin H. Prins, Harry R. B
ller, Paolo Prandoni. (1998) Undiagnosed malignancy in patients with deep vein thrombosis. Cancer 83:1, 180-185
CrossRef
