Role of acquired and hereditary thrombotic risk factors in colon ischemia of ambulatory patients

GASTROENTEROLOGY 2001;121:561–565

Role of Acquired and Hereditary Thrombotic Risk Factors in Colon Ischemia of Ambulatory Patients I. E. KOUTROUBAKIS,* A. SFIRIDAKI,‡ A. THEODOROPOULOU,* and E. A. KOUROUMALIS* *Department of Gastroenterology, University Hospital Heraklion and ‡Regional Blood Bank Center, Venizelion Hospital Heraklion, Crete, Greece

See editorial on page 724. Background & Aims: Hypercoagulable states may play an important role in the pathogenesis of colon ischemia. Aim of this study was to assess this hypothesis investigating the role of acquired and hereditary thrombotic risk factors in patients with definite diagnosis of colon ischemia. Methods: We compared the frequency of antiphospholipid antibodies, protein C, protein S, and antithrombin deficiencies, factor V Leiden, prothrombin gene mutation G20210GA, and methylenetetrahydrofolate reductase C677T in 36 patients (23 men, 13 women; mean age, 64.8 years) with colon ischemia, 18 patients with diverticulitis, and 52 healthy controls. Results: The prevalence of antiphospholipid antibodies was significantly higher in patients with colon ischemia compared with inflammatory and healthy controls (19.4% vs. 0% and 1.9%). Among genetic factors, only factor V Leiden was significantly associated with colon ischemia (22.2% vs. 0% and 3.8%). A combination of thrombophilic disorders was found in 25% of the cases. Overall, one or several prothrombotic abnormalities were present in 26 patients (72%). Conclusions: A comprehensive thrombophilic screening in colon ischemia reveals a congenital or acquired thrombophilic state in 72% of patients. Hereditary and acquired thrombotic risk factors may play an important role in the disease pathogenesis.

eversible colon ischemia (CI) as described by Boley et al. in 19631 represents the most common form of gastrointestinal ischemia.2 Its presentation is nonspecific, and numerous conditions may favor its development. As a result, frequent misdiagnosis of CI and the underestimation of its incidence may occur.3,4 Several distinct clinicopathological subtypes of CI have been proposed.5 CI may occur in association with serious illnesses such as shock, colon cancer, or the period after a surgical intervention on the aorta or the mesenteric vessels. It may also spontaneously appear in apparently healthy individuals.6 Several risk factors have been proposed associated with CI, often through isolated observations,7–11 and rarely as

R

clinical studies.12,13 An increasing number of young people have been identified in whom CI is associated with various medications, oral contraceptives, cocaine use, long-distance running, sickle cell disease, vasculitis, and the newly described coagulopathies.14 The involvement of hereditary or acquired thrombotic risk factors is established in the pathogenesis of splanchnic vein thrombosis. The role of these factors in CI has not been extensively investigated. Cases of CI associated with genetic defects such as deficiencies of protein C, protein S, and antithrombin (AT),10,15,16 factor V Leiden (FVL) mutation,17,18 and prothrombin 20210G/A mutation,19 as well as acquired factors such as antiphospholipid antibodies,20 have been reported. The present study was designed to study the prevalence of newly recognized inherited prothrombotic disorders in definite CI of ambulatory patients, and also to assess the implication of both inherited and acquired prothrombotic disorders in the development of CI.

Materials and Methods Patients We undertook a retrospective review of all patients with nonocclusive colon ischemia who were nonsurgically treated at the University Hospital of Heraklion (Heraklion, Crete, Greece) from January 1995 to December 2000. Case records were retrieved, and the accuracy of the diagnosis was assessed from clinical, endoscopic, and histologic findings. Clinically, patients with mild-to-moderate abdominal pain, diarrhea, or lower intestinal bleeding with minimal-to-moderate abdominal tenderness were investigated for possible CI. Patients with a previous history of inflammatory bowel disease, or with pathogenic bacteria in their stools, or with recent use of antibiotics were excluded. Endoscopic findings suggestive of Abbreviations used in this paper: ␣␤2GPI, anti-beta2-glycoprotein I; aCL, anti-cardiolipin; ANA, antinuclear antibody; APCR, activated protein C resistance; AT, antithrombin; CI, colon ischemia; FVL, factor V Leiden; MTHFR, methylenetetrahydrofolate reductase; SMA, smooth muscle antibody. © 2001 by the American Gastroenterological Association 0016-5085/01/$35.00 doi:10.1053/gast.2001.27227

562

KOUTROUBAKIS ET AL.

CI were the existence of mucosal edema with mucosal hemorrhagic infiltration and/or scattered large superficial ulcers or a bluish discoloration of the mucosa with deep ulceration. Histologic findings suggestive of CI were the loss of epithelium, edema of mucosa and muscularis mucosa, the intravascular thrombosis with dilated capillaries, the submucosal hemosiderosis, and the presence of fibrosis in the deeper layers. Clinical features compatible with CI, endoscopic findings suggestive of CI, and histologic findings definite for or suggestive of CI were all required for confirmation of diagnosis. Moreover, reexamination with colonoscopy and biopsies (when necessary) 2 months after the acute episode was done in all cases, to exclude other causes of intestinal inflammation. We identified 63 patients from discharge summary codes, gastrointestinal endoscopy reports, and pathology reports with the diagnosis of CI. From all these patients, cases with mesenteric arterial thrombosis, mesenteric arterial embolic disease, mesenteric venous thrombosis, or ischemia secondary to conditions such as colon carcinoma, diverticulitis, shock, low cardiac output, aortic surgery, and prolonged pelvic surgery were excluded (15 cases). Five patients had already died. The remaining patients who had a definite diagnosis of CI (43 cases) were invited to attend a special follow-up clinic, where a full medical history was taken, clinical examination was performed, and blood samples were collected. Finally, 36 patients (23 men, 13 women; mean age, 64.8 years) with CI were included in the study. In all these cases, the diagnosis was definite based on the above-mentioned criteria. These were matched for age and sex and compared with 18 inflammatory controls (cases with diverticulitis; mean age, 62.6 years) and 52 healthy controls with a mean age of 61.7 years. Healthy controls were recruited from visitors of hospital wards.

Laboratory Studies Venous blood was collected from all patients and controls, and plasma, serum, and EDTA-blood from each patient were stored in ⫺80°C until assay. All laboratory studies were performed at least 1 month after the acute episode of the disease. Protein C was determined by enzyme-linked immunosorbent assay according to the manufacturer’s instructions (Biochem Diagnostics, Asnie`res, France; normal range, 65%– 140%). Protein S (total and free) was also assayed by enzymelinked immunosorbent assay (Biochem Diagnostics; normal range of free protein S, 70%–140%). AT activity was measured using a chromogenic substrate kit (Organon Technica, Durham, NC; normal range, 85%–111%). Activated protein C resistance (APCR) was measured by determining the activated partial thromboplastin time in the absence and presence of activated protein C (APC; Coatest activated protein C resistance kit, Chromogenix, Moendal, Sweden). APCR was diagnosed if the APC ratio was ⬍1.9. Commercial kits were used to assay anti-cardiolipin (aCL) and anti-beta2-glycoprotein I (␣␤2GPI) antibodies of immunoglobulin (Ig) G and IgM types (Callestad Pasteur, Marnes la Coquette, France and Quanta Lite, INOVA Diagnostics, San

GASTROENTEROLOGY Vol. 121, No. 3

Diego, CA, respectively). Lupus anticoagulant was investigated as previously described.21 Antinuclear antibody (ANA), smooth muscle antibody (SMA), and perinuclear antineutrophil cytoplasmic autoantibody were assayed by indirect immunofluorescence and the rheumatoid factor by the RoseWaler sheep red cell agglutination test. Genomic DNA isolation from EDTA blood, polymerase chain reaction, and detection of the FVL mutation, prothrombin 20210A mutation, and C677T methylenetetrahydrofolate reductase (MTHFR) polymorphism were done using Factor V Gene Mutation Assay (VIENNA LAB, Vienna, Austria), Prothombin Gene Mutation Assay (VIENNA LAB), and MTHFR Mutation Assay (VIENNA LAB). FVL mutation was assayed in patients and controls with abnormal APCR, whereas the other 2 mutations were assayed in all cases.

Statistical Analysis Comparisons between the 3 diagnostic groups in terms of continuous measurements were made by one-way analysis of variance. The difference in prevalence of acquired and genetic thrombophilic defects between CI patients and controls was studied by means of 2 ⫻ 2 table analysis using the Fisher exact test. A 5% significance threshold was adopted.

Results The clinical characteristics of the patients included in the study are presented in Table 1. Twentynine patients (80%) were taking medications. The use of nonsteroidal anti-inflammatory drugs was established in 10 cases, neuroleptics in 2 cases, antidepressants in 7 cases, and antihypertensives in 18 cases, 7 of them diuretics.

Table 1. Clinical Details of the Patients Included in the Study Number Mean age Male Female Smokers Nonsmokers Under medications Disease type Reversible ischemic colopathy Transient CI Chronic ulcerative CI Ischemic colonic stricture Colonic gangrene Fulminant universal CI Disease localization Right-sided Splenic flexure Rectosigmoid Extensivea aExtensive

36 64.8 23 13 24 12 29 8 17 7 3 1 0 7 18 8 3

CI was defined as a form involving more than 1 of the previous 3 locations.

September 2001

THROMBOTIC RISK FACTORS IN ISCHEMIC COLITIS

563

Table 2. Serum Levels of Antiphospholipid Antibodies and Plasma Levels of Other Thrombophilic Parameters in Patients With CI, Compared With Patients With Diverticulitis and Healthy Controls Parameter IgG aCL antibodies (units/mL; median, range) IgG ␣␤2GPI antibodies (units/mL; median, range) Protein C (%, mean ⫾ SD) Protein S (%, mean ⫾ SD) AT (%, mean ⫾ SD) APC ratio (mean ⫾ SD)

CI (0.1–39.6)a

5.9 5.2 (0–40)b 110.1 ⫾ 30.1 84.2 ⫾ 27.7 91.5 ⫾ 18.9 2.2 ⫾ 0.6c

DIV

HC

Normal range

2.9 (0.1–9.6) 2.2 (0.1–9.6) 102.5 ⫾ 23.6 94.5 ⫾ 19.0 94.9 ⫾ 15.1 2.8 ⫾ 0.7

2.6 (0.1–16) 1.8 (0–9.6) 108.8 ⫾ 36.3 95.5 ⫾ 29.1 98.3 ⫾ 18.8 3.1 ⫾ 0.8

⬍10 ⬍10 65–140 70–140 85–111 1.8–5

DIV, diverticulitis; HC, healthy controls. ⫽ 0.07. bP ⫽ 0.03. cP ⫽ 0.02. aP

The serum levels of antiphospholipid antibodies in patients with CI compared with patients with diverticulitis and healthy controls are presented in Table 2. Four CI cases had elevated aCL antibodies. Two of them were also found to have elevated ␣␤2GPI antibodies. Two patients were aCL-positive and ␣␤2GPI-negative, and the other 2 were aCL-negative and ␣␤2GPI-positive. The titers of these antibodies were moderately positive in 2 cases (26 – 40 units) and low positive (10 –25 units) in the rest. Median levels of IgG aCL (5.9 units/mL; range, 0.1–39.6) and of IgG ␣␤2GPI (5.2 units/mL; range, 0 – 40) were higher in CI than in inflammatory and healthy controls (Table 2). However, the differences among groups reached statistical significance only for IgG ␣␤2GPI (P ⫽ 0.03). The Lupus anticoagulant was positive in 2 CI patients (5.6%) and none of the inflammatory and healthy controls. One patient was both aCL– and Lupus anticoagulant–positive. Overall, 7 cases (4 men, 3 women; 19.4%) had elevated antiphospholipid antibodies compared with none of the inflammatory controls and only 1 healthy control (1.9%; P ⫽ 0.047 and P ⫽ 0.007, respectively). Three patients (8.8%) with CI had raised levels of rheumatoid factor; 1 of them was a patient with an established diagnosis of rheumatoid arthritis. ANA and SMA titers were elevated in 2 and 1 case, respectively, whereas no patient was perinuclear antineutrophil cytoplasmic autoantibody–positive. None of the CI patients had vasculitis, and therefore, it could be suggested that the presence of rheumatoid factor, ANA, and SMA have no thrombophilic importance. Only 1 of the controls had both rheumatoid factor and ANA. Table 3 shows the prevalence of prothrombotic genetic risk factors in patients with CI, patients with diverticulitis, and in healthy controls. Free protein S, protein C, and AT deficiencies were found in 5 (13.9%), 2 (5.6%), and 4 (11.1%) CI cases, 1 (5.6%), 1 (5.6%), and 0 (0%) inflammatory controls, and in 4 (7.7%), 1 (1.9%), and 2 (3.8%) healthy controls, respectively. Mean values of

these parameters did not differ significantly among groups (Table 2). In 7 CI cases, the values of these proteins were only slightly below the normal levels, whereas a profound deficiency (⬍50%), such as observed in the genetically deficient groups, was found in 4 cases (3 free protein S, 1 protein C). Only 1 healthy control had a profound deficiency of free protein S. APCR was found in 10 CI patients, 1 inflammatory, and 2 healthy controls. The prevalence of APCR in CI (27.8%) was significantly higher than both groups of controls (5.6% and 3.8%; P ⫽ 0.05 and P ⫽ 0.003, respectively). FVL mutation was tested in all patients and controls with abnormal APCR and was present in 8 of the CI cases (5 men, 3 women), none of the inflammatory, and 2 of the healthy controls. All patients and controls with FVL were heterozygous for the mutation. Prevalence of FVL in CI patients was 22.2%, significantly higher than that in both groups of controls (0% and 3.8%; P ⫽ 0.03 and P ⫽ 0.01). Prothrombin 20210A mutation was found in only 1 patient (homozygous for the mutation) with CI (2.8%) Table 3. Specific Prothrombotic Risk Factors in Patients With CI Compared With Patients With Diverticulitis and Healthy Controls Parameter

CI N (%)

DIV N (%)

HC N (%)

Antiphospholipid antibodies Protein S deficiency Protein C deficiency AT deficiency APCR FVL mutation Prothrombin 20210A mutation C677T MTHFR

7 (19.4)a,b 5 (13.9) 2 (5.6) 4 (11.1) 10 (27.8)c,d 8 (22.2)e,f 1 (2.8) 8 (22.2)

0 (0) 1 (5.6) 1 (5.6) 0 (0) 1 (5.6) 0 (0) 0 (0) 3 (16.7)

1 (1.9) 4 (7.7) 1 (1.9) 2 (3.8) 2 (3.8) 2 (3.8) 0 (0) 9 (17.3)

DIV, diverticulitis; HC, healthy controls. ⫽ 0.007 vs. HC. bP ⫽ 0.047 vs. DIV. cP ⫽ 0.003 vs. HC. dP ⫽ 0.05 vs. DIV. eP ⫽ 0.01 vs. HC. fP ⫽ 0.03 vs. DIV. aP

564

KOUTROUBAKIS ET AL.

Table 4. Combination of Prothrombotic Disorders and Risk Factors for Thrombosis in Patients With CI Patient CI-2 CI-6 CI-9 CI-12 CI-16 CI-18 CI-21 CI-27 CI-32

Protein S deficiency, APCR, heterozygous FVL Protein S deficiency, aCL, ␣␤2GPI AT deficiency, aCL, ␣␤2GPI APCR, heterozygous FVL, aCL Protein S deficiency, APCR, heterozygous FVL AT deficiency, homozygous C677T MTHFR Prothrombin 20210A mutation, ␣␤2GPI Protein C deficiency, APCR, heterozygous FVL, lupus anticoagulant, rheumatoid factor Protein C deficiency, protein S deficiency, rheumatoid factor

and none of the controls. The C677T MTHFR polymorphism was identified in 8 CI cases (22.2%; 4 heterozygous and 4 homozygous), 3 inflammatory controls (16.7%; 2 heterozygous and 1 homozygous), and 9 healthy controls (17.3%; 5 heterozygous and 4 homozygous) for the C677T variant. The prevalence of homozygous C677T MTHFR was not statistically different between CI and controls. Combinations of multiple prothrombotic risk factors for thrombosis were found in 9 CI patients (Table 4). In 10 patients, no thrombophilic risk factor was identified. Overall, one or several prothrombotic disorders were demonstrated in 26 of the 36 patients (72%) with CI. No significant differences in coagulation disturbances among the different clinical stages of CI were found. However, the number of patients in some groups (ischemic colonic stricture, colonic gangrene, and fulminant universal CI) was too small to draw conclusions. Although in our series there was a male predominance, no gender difference among the thrombophilic factors was found. In relation to other cofactors associated with ischemia, there were only 2 cases with associated cancers (1 breast cancer and 1 lung cancer). No other factors (estrogen use, long distance running, sickle cell disease, vasculitis, intravascular volume depletion, etc.) were found.

Discussion Various factors have been reported to be associated with colonic ischemia,2 but the pathogenetic mechanisms remain obscure. An etiologic role of infection with Escherichia coli 0157:H7 has been proposed, but this association has not been conclusively demonstrated.22 On the other hand, it could be suggested that hypercoagulable states may play an important role in the disease, leading to the development of thrombotic occlusion of small vessels supplying the colon. However, the role of acquired or hereditary thrombotic risk factors in CI is

GASTROENTEROLOGY Vol. 121, No. 3

poorly documented, and only case reports with such associations have appeared.10,11,15–19 In a recent study of 24 patients with ischemic colitis,13 clotting factor abnormalities were found in 3 of 9 patients who were checked. One of these patients had raised aCL antibodies, ANAs, and rheumatoid factor; the other 2 had aCL antibodies alone. No other thrombophilic risk factors (deficiency of proteins S and C, AT, and FVL) were found. In our study, which is the first with a comprehensive thrombophilic screening in CI patients, the prevalence of acquired and hereditary thrombotic risk factors was found significantly higher compared with the prevalence of these factors in matched inflammatory and healthy controls. A thrombophilic tendency was demonstrated in many CI patients, and the most significant associations were with antiphospholipid antibodies and with FVL mutation. Elevated levels of antiphospholipid antibodies were found in a significantly higher number of patients with CI (19.4%) than that in controls (0% and 1.9%). Based on the known association of antiphospholipid antibodies and deep vein thrombosis,23 it could be suggested that antiphospholipid antibodies may be implicated in the disease manifestation of at least some CI patients. The present study has shown that FVL was present in 22.2% of patients with CI significantly higher compared with inflammatory (0%) and healthy controls (3.8%). In a population of apparently healthy subjects from various geographic areas in Europe, the reported prevalence of FVL was 0.45%–7%, with a mean prevalence of 5%.24 FVL has been identified in 20%–25% of patients with venous thromboembolism,24 which makes this anomaly the most common prothrombotic state currently identified. In Greece, FVL has been found in 4.8% of the healthy population25,26 and in 28.1% of patients with a documented thrombophilic tendency.25 The finding of a similarly high prevalence of FVL in CI patients suggests that thrombotic mechanisms with the participation of FVL may play an important role in the pathogenesis of CI. In the present study, the prevalence of the C677T MTHFR gene mutation in CI patients was similar to that in controls, and the prothrombin 20210A mutation was also found in only 1 patient. It could therefore be suggested that these mutations do not play an important role in CI. Because C677T MTHFR mutation is associated with the presence of hyperhomocysteinemia,27 it seems that homocysteine may not be involved in the pathogenesis of this disease.

September 2001

Based on our data and the review of the literature, we could suggest that the mechanism of CI is multifactorial. Acquired and genetic factors may interact, leading to the disease manifestation. A high frequency of arrhythmia and embolic conditions has been found in CI patients.6 Oral contraceptives, cocaine, and several other medications have been proposed to play a possible role in the disease pathogenesis. Antiphospholipid antibodies commonly present in CI patients, as shown by this study, could also interfere in the disease pathogenesis. On the other hand, an inherited predisposition owing to one or more genes may also play an important role. Our results of increased prevalence of FVL support this hypothesis. The occurrence of multiple prothrombotic factors in 25% of our patients also supports the hypothesis of an interaction between acquired and genetic factors. Our findings suggest that prothrombotic disorders are important in the development of CI. The interactions between acquired and hereditary thrombotic risk factors are possibly the key to understanding why certain persons develop CI at a specific time point. Questions raised by these findings should address the possible benefits from an extensive thrombophilic screening for the patients with CI and the subsequent administration of anticoagulants in patients with prothombotic abnormalities.

THROMBOTIC RISK FACTORS IN ISCHEMIC COLITIS

11.

12.

13. 14. 15. 16.

17.

18.

19.

20.

21.

22.

References 1. Boley SJ, Lash J, Schwartz S, Sternhill V. Reversible vascular occlusion of the colon. Surg Gynecol Obstet 1963;113:53– 60. 2. Gandhi SK, Hanson MM, Vernava AM, Kaminski DL, Longo WE. Ischemic colitis. Dis Colon Rectum 1996;39:88 –100. 3. Robert JH, Mentha G, Rohner A. Ischaemic colitis: two distinct patterns of severity. Gut 1993;34:4 – 6. 4. Alapati SV, Mihas AA. When to suspect ischemic colitis. Why is this condition so often missed or misdiagnosed? Postgrad Med 1999;105:177–180. 5. Brandt LJ, Boley SJ. Colonic ischemia. Surg Clin North Am 1992; 72:203–229. 6. Collet T, Even C, Bouin M, Lecluse E, Piquet MA, Crampon D, Grollier G, Dao T, Verwaerde JC. Prevalence of electrocardiographic and echocardiographic abnormalities in ambulatory ischemic colitis. Dig Dis Sci 2000;45:23–25. 7. Yang RD, Han MW, McCarthy JH. Ischemic colitis in a crack abuser. Dig Dis Sci 1991;36:238. 8. Boutros HH, Pautler S, Chakrabarti S. Cocaine-induced ischemic colitis with small-vessel thrombosis of colon and gallbladder. J Clin Gastroenterol 1997;24:49 –53. 9. Dowd J, Bailey D, Moussa K, Nair S, Doyle R, Culpepper-Morgan JA. Ischemic colitis associated with pseudoephedrine: four cases. Am J Gastroenterol 1999;94:2430 –2434. 10. Blanc P, Bories P, Donadio D, Parelon G, Rouanet C, Paleirac G, Michel H. Ischemic colitis and recurrent venous thrombosis

23.

24. 25.

26.

27.

565

caused by familial protein S deficiency. Gastroenterol Clin Biol 1989;13:945. Mann DE Jr, Kessel ER, Mullins DL, Lottenberg R. Ischemic colitis and acquired resistance to activated protein C in a woman using oral contraceptives. Am J Gastroenterol 1998;93:1960 –1962. Deana DG, Dean PJ. Reversible ischemic colitis in young women. Association with oral contraceptive use. Am J Surg Pathol 1995; 19:454 – 462. Arnott ID, Ghosh S, Ferguson A. The spectrum of ischaemic colitis. Eur J Gastroenterol Hepatol 1999;11:295–303. Brandt LJ, Boley SJ. AGA technical review on intestinal ischemia. Gastroenterology 2000;118:954 –968. Verger P, Blanc C, Feydy P, Boey S. Ischemic colitis caused by protein S deficiency. Presse Med 1996;25:1350. Knot EA, ten Cate JW, Bruin T, Iburg AH, Tytgat GN. Antithrombin III metabolism in two colitis patients with acquired antithrombin III deficiency. Gastroenterology 1985;89:421– 425. Ludwig D, Stahl M, David-Walek T, Bruning A, Siemens A, Zwaan M, Schmucker G, Stange EF. Ischemic colitis, pulmonary embolism, and right atrial thrombosis in a patient with inherited resistance to activated protein C. Dig Dis Sci 1998;43:1362–1367. Yee NS, Guerry D 4th, Lichtenstein GR. Ischemic colitis associated with factor V Leiden mutation. Ann Intern Med 2000;132: 595–596. Balian A, Veyradier A, Naveau S, Wolf M, Montembault S, Giraud V, Borotto E, Henry C, Meyer D, Chaput JC. Prothrombin 20210G/A mutation in two patients with mesenteric ischemia. Dig Dis Sci 1999;44:1910 –1913. Cappell MS. Esophageal necrosis and perforation associated with the anticardiolipin antibody syndrome. Am J Gastroenterol 1994;89:1241–1245. Rosner E, Pauzner R, Lusky A, Modan M, Many A. Detection and quantitative evaluation of lupus circulating anticoagulant activity. Thromb Haemost 1987;57:144 –147. Su C, Brandt LJ, Sigal SH, Alt E, Steinberg JJ, Patterson K, Tarr PI. The immunohistological diagnosis of E. coli O157:H7 colitis: possible association with colonic ischemia. Am J Gastroenterol 1998;93:1055–1059. Bick RL. Antiphospholipid thrombosis syndromes: etiology, pathophysiology, diagnosis and management. Int J Hematol 1997;65:193–213. Rosendaal FR. Venous thrombosis: a multicausal disease. Lancet 1999;353:1167–1173. Lambropoulos AF, Foka Z, Makris M, Daly M, Kotsis A, Makris P. Factor V Leiden in Greek thrombophilic patients: relationship with activated protein C resistance test and levels of thrombin-antithrombin complex and prothrombin fragment 1 ⫹ 2. Blood Coagul Fibrinolysis 1997;8:485– 489. Koutroubakis IE, Sfiridaki A, Mouzas IA, Maladaki A, Kapsoritakis A, Roussomoustakaki M, Kouroumalis EA, Manousos ON. Resistance to activated protein C and low levels of free Protein S in Greek patients with inflammatory bowel disease. Am J Gastroenterol 2000;95:190 –194. Huisman MV, Rosendaal F. Thrombophilia. Curr Opin Hematol 1999;6:291–297.

Received January 25, 2001. Accepted May 16, 2001. Address requests for reprints to: Ioannis E. Koutroubakis, M.D., Department of Gastroenterology, University Hospital Heraklion, P.O. Box 1352, 71110 Heraklion, Crete, Greece. e-mail: ktjohn@her. forthnet.gr; fax: 003081542085.