Total Plasma Antioxidant Capacity Predicts ... - Diabetes Care

2 downloads 0 Views 721KB Size Report
TRAP, but no single other antioxidant, had a strong inverse association with ... antioxidant capacity rather than any single plasma antioxidant that is the most ...
Pathophysiology/Complications N A L

A R T I C L E

Total Plasma Antioxidant Capacity Predicts Thrombosis-Prone Status in NIDDM Patients ANTONIO CERIELLO, MD NADIA BORTOLOTTL, PHD MARIO PIRISI, MD ADRIANA CRESCENTINI LAURA TONUTTI, MD

ENRICO MOTZ, MD ASSUNTA RUSSO, MD ROBERTA GIACOMELLO, BSC GlULIANA STEL, BSC CLAUDIO TABOGA, MD

OBJECTIVE — To explore the hypothesis that a relationship exists between free radical activity and abnormalities in hemostasis in NIDDM. RESEARCH DESIGN A N D METHODS— The use of the total radical-trapping antioxidant parameter (TRAP) has very recently been proposed to explore the antioxidant property of a plasma sample, taking into consideration known and unknown antioxidants present in the plasma and their mutual cooperation. In the present study, TRAP, vitamin E, vitamin C, vitamin A, uric acid, protein-bound SH (thiol) groups, fibrinogen, prothrombin fragments F l + 2 , and D-dimer have been evaluated in 46 NIDDM patients and 47 healthy matched control subjects. RESULTS — In NIDDM patients, TRAP, vitamin A, SH groups, and uric acid were significantly reduced, whereas the level of vitamin E was significantly increased. Vitamin C was similar in the two groups. Fibrinogen, prothrombin fragment 1 + 2, and D-dimer were increased in diabetic patients. TRAP, but no single other antioxidant, had a strong inverse association with fibrinogen, prothrombin fragment 1+2, and D-dimer. CONCLUSIONS — These findings are consistent with the hypothesis that oxidative stress may condition coagulation activation in diabetes. However, the data suggest that it is the total antioxidant capacity rather than any single plasma antioxidant that is the most relevant parameter.

T

he existence of a thrombosis-prone condition has long been recognized in diabetes (1,2). Free radicals are capable of activating coagulation (3). This has recently raised the question whether oxidative stress, which is known to be increased in patients with diabetes (4,5), is associated with changes in hemostatic variables in these patients (6-9). Extracellular fluids lack protection by the antioxidant enzymes, but contain various antioxidants that delay or inhibit the oxidative process (10). It has been suggested that the reduction in the antioxidant state contributes to the development of

oxidative stress in diabetes (4). Several previous studies evaluating plasma antioxidants in diabetic patients have produced conflicting results (11-17), possibly because they focused on a limited number of individual antioxidants (11-17). Indeed, it has been shown that the plasma total antioxidant capacity is not determined by the mere sum of the relative concentration of antioxidants, but is determined also by their synergism (10). Recently, the assay of the total plasma antioxidant radical-trapping activity (total radical-trapping antioxidant parameter [TRAP]) has been proposed to evaluate

From the Department of Medicine and Pathology, Clinical and Experimental (A.Ce., N.B., M.P, A.Cr, E.M., A.R., R.G., G.S.), University of Udine, and the Diabetes Unit (L.T., C.T.), Udine General Hospital, Udine, Italy. Address correspondence and reprint requests to Antonio Ceriello, MD, Chair of Internal Medicine, University of Udine, P. le S. Maria della Misericordia, 33100 Udine, Italy. Received for publication 31 December 1996 and accepted in revised form 17 June 1997. Abbreviations: ABAP, 2,2'-azobis-(2-amidinopropane) dihydrochloride; SH, thiol; TRAP, total radicaltrapping antioxidant parameter.

DIABETES CARE, VOLUME 20, NUMBER 10, OCTOBER

1997

plasma antioxidant capacity (18). TRAP takes into consideration known and unknown antioxidants present in the plasma and their mutual cooperation (18). Reduced TRAP activity has already been reported in both IDDM (17,18) and NIDDM (19) patients. In this study, we measured TRAP and the major individual antioxidants in the plasma of NIDDM patients and correlated them with markers of thrombosis, such as fibrinogen, prothrombin fragments 1+2, and D-dimer. Numerous epidcmiological studies have concurred in recognizing fibrinogen as an important marker of cardiovascular risk (20), and studies in diabetic patients agree with these results (21,22). Prothrombin fragments 1 + 2 represent a reliable marker of the amount of thrombin released in the circulation (23), while Ddimer represents a breakdown fragment of fibrin (24). Increased prothrombin fragment 1 + 2 and D-dimer plasma levels have been already demonstrated in diabetes (25-27).

RESEARCH DESIGN AND M E T H O D S — Forty-six NIDDM patients on diet plus hypoglycemic agent therapy (27 men and 19 women; mean age ± SE, 58.2 ± 1.5 years; duration of diabetes, 10.0 ± 1.3 years; BMI, 26.2 ± 1.2 kg/m 2 ), selected according to the National Diabetes Data Group criteria (28), gave informed consent to participate in this study. Fortyseven healthy normal blood donor subjects, selected to have a similar distribution of sex (30 men and 17 women), age (55.1 ± 1.3 years), and BMI (25.9 ±1.4 kg/m 2 ), served as the control group. All were nonsmokers. Moreover, the diabetic patients and the control subjects were also well-matched for diet, with regard to their consumption of fruit and fresh vegetables. Dietary information was obtained by trained interviewers from the administration of a questionnaire developed according to Block et al. (29). The reliability and validity of this tool for assessing food intake in the population has already been evaluated (30). None of the selected subjects of this study was taking hypolipidemic drugs or

1589

TRAP and thrombosis in N1DDM

was on antioxidant supplementation. Even though it is well known that it is practically impossible to exclude the presence of macroangiopathy in diabetic patients, none had clinically symptomatic macroangiopathy, as judged by pathological changes in the resting electrocardiogram, a previous history of cardiac angina, intermittent claudication, and myocardial or cerebral infarction. None had micro- or macroalbuminuria or suffered from hypertension. This study was approved by the ethical committee of our institution. In all subjects, fasting plasma glucose, HbAlc, cholesterol, triglycerides, fibrinogen, prothrombin fragments 1 + 2, Ddimer, uric acid, vitamin A, vitamin C, vitamin E, protein-bound SH groups, and TRAP levels were evaluated. Plasma glucose was assayed by the glucose-oxidase method. HbAlc levels were measured by aminophenylboronic acid affinity chromatography (31). Cholesterol and triglycerides were measured enzymatically on a clinical chemistry analyzer (MONARCH, Instrumentation Laboratory, Lexington, MA).

The resulting value of x is then multiplied by 2.0 (the stoichiometric factor of trolox) and by the dilution factor of plasma (250), and the values are expressed as micromoles per liter.

immunoenzymatically according to Rylatt et al. (37).

Statistical analysis Statistical analysis of data was performed by means of the BMDP statistical software package. Normal distribution of each parameters Sample preparation According to the procedure of Ghiselli et al. was confirmed by normality test (Shapiro (18), the reaction mixture consisted of 1.5 and Wilk's W test [38]). Data between diaX 10~8 mol/1 R-Phycoerythrin in 75 mmol/1 betic patients and control subjects were then phosphate buffer (pH 7). Eight microliters compared by unpaired Students t test. of plasma or 30 pi of 120-pmol/l Trolox Multiple regression was performed to were added to afinalvolume of 2.0 ml, and determine the way in which the variable of the resulting solution was maintained at interest (dependent variable) was influ37°C for 5 min in 10-mm quartz fluorom- enced by several other variables (predictor eter cells. The oxidation reaction was started variables or covariates). A forward stepby adding ABAP to afinalconcentration of wise approach was used (39,40). With this 4.0 mmol/1, and the decay of R-Phycoery- method, thefirststep is to find, in the set of thrin fluorescence was monitored every 5 covariates chosen, the variable bearing the min on a Perkin-Elmer LS-50 Lumines- strongest association with the dependent cence Spectrometer equipped with a ther- variable. The strength of the association mostatically controlled cell-holder; between the two variables is given by the F monochromators were operating at an exci- ratio. The variable with the highest F is the tation wavelength of 495 nm/5 nm slit first to be entered into the model. The folwidth and an emission wavelength 575 lowing step is to add to the model obtained nm/5 nm slit width. TRAP values were cal- so far the variable that explains the largest culated as above described. Intra- and amount of the remaining variability. This interassay coefficients of variation for this step is repeated until the addition of an method were 10 and 12%, respectively. extra variable does not improve further the TRAP measurement predictive value of the model. The higher 2,2' -azobis-(2-amidinopropane) dihythe F is to enter of a variable, the stronger drochloride (ABAP), R-Phycoerythrin, Plasma antioxidants assay Trolox, and all other chemicals were pur- Vitamin C was evaluated spectrophoto- the contribution it gives to the significance chased from Sigma (St. Louis, MO). metrically according to Tsan et al. (32). of the multiple regression (39,40). TRAP was evaluated according to Ghi- Vitamin E and Vitamin A levels were measselli et al. (18). In this method, the pro- ured by high-performance liquid chro- RESULTS— TRAP, vitamin A, SHgroups, and uric acid were significantly duction of peroxyl radicals obtained by matography (33). thermal decomposition of ABAP leads to a The protein SH groups were measured reduced in diabetic patients, while the level linear decrease in R-phycoerythrin fluores- by the method described by Koster et al. of triglycerides and vitamin E were significence emission over 1 h. When plasma is (34), while the uric acid assay was per- cantly increased, even when vitamin E was added to the reaction mixture, a period of formed by a commercially available enzy- corrected for serum cholesterol (Table 1) complete protection of R-phycoerythrin is matic colorimetric test (Instrumentation (41). Vitamin C was not different between the two groups (Table 1). Fibrinogen, proobserved. The length of this lag-phase (T) is Laboratory). thrombin fragments 1+2, and D-dimer here taken to be directly related to total were increased in diabetic patients (Table 1). plasma antioxidant capacity. To quantify Coagulation the TRAP, the T produced by plasma is To study coagulation parameters, blood Only TRAP was inversely associated with compared with the T produced by a known was collected through a silicone-treated fibrinogen (r = -0.32, P < 0.03), proamount of trolox. The T values are calcu- needle (2-mm diameter) and was allowed thrombin fragments 1+2 (r = — 0.61, P < lated by extrapolating the slope of maximal to flow freely into silicone-treated glass 0.001), and D-dimer (r = - 5 1 , P < 0.001) R-phycoerythrin decay to intersect with the tubes, where it was mixed with l/9th its (Fig. 1), while none of the measured antioxslope of plasma and trolox protection. The volume of 0.1 mol/1 trisodium citrate. The idants was. Prothrombin fragments 1+2 (r projection of these intersection points on blood was immediately centrifuged at = 0.30, P < 0.05) and D-dimer (r = 0.33, P the x-axis gives the T values, which repre- l,700g for 20 min at 4°C and frozen at < 0.02) were significantly correlated with fibrinogen (Fig. 2). Stepwise multivariate sent the time required to achieve the max- -80°C until assayed. imal R-phycoerythrin peroxidation rates. Plasmafibrinogenwas measured func- analysis showed that TRAP (F to enter, 5.1) By comparing the T of plasma with the T of tionally by the method of Rossi et al. (35) in and D-dimer (F to enter, 4.8) were strong trolox, taking into account the concentra- an ACL autoanalyzer (Instrumentation predictors of fibrinogen levels in diabetic tion of trolox, the TRAP value of a plasma Laboratory). Prothrombin fragments 1+2 patients. However, only TRAP was a predicsample is obtained according the following plasma levels were evaluated by enzyme- tor when prothrombin fragments 1+2 (F to proportion: [Trolox] : T Trolox = x : T linked immunosorbent assay according to enter, 26.5) and D-dimer (F to enter, 15.2) Plasma (18). Pelzer et al. (36). D-dimer was measured were considered as independent variables. 1590

DIABETES CARE, VOLUME 20, NUMBER 10, OCTOBER

1997

Ceriello and Associates

No correlations were found in normal subjects. CONCLUSIONS— Patients with diabetes, a known risk factor for developing cardiovascular disease, have an increased cardiovascular morbidity and mortality (42,43). Indeed, thrombosis can be demonstrated at the onset of many cardiovascular events, and the presence of marked thrombophilia has been shown in many patients at risk (44,45). Thus, laboratory indexes of a thrombophilic status are of great clinical interest. Prothrombin fragments 1+2 represent a reliable marker of the amount of thrombin released in the circulation (23), whereas D-dimer represents a breakdown fragment of fibrin (24). It has been shown that prothrombin fragments 1+2 are a predictor of the severity of coronary artery disease (46) and D-dimer of myocardial infarction (47) in nondiabetic subjects. The role of fibrinogen as an independent cardiovascular risk factor in diabetic and nondiabetic subjects cannot be substantiated (20-22). In the present study, we show that in comparison with control subjects, NIDDM patients have high levels offibrinogen,prothrombin fragments 1 + 2, and D-dimer. Thesefindingsare in agreement with those reported by others (21,22,25-27). In addition, the present study confirms previous observations by our group regarding the existence in NIDDM patients of a significant correlation between plasma levels of prothrombin fragments 1+2, D-dimer, and fibrinogen, with independent association between D-dimer andfibrinogenon multivariate analysis (48). Recent epidemiological studies have suggested that decreased levels of circulating antioxidants might favor cardiovascular disease (49-52). Thereafter, several studies have addressed the issue of the antioxidant status in NIDDM, generating conflicting results (11-17). Our data indicate that NIDDM patients have decreased levels of vitamin A, protein-bound SH groups, and uric acid and increased levels of vitamin E. These data are in agreement with several previous reports and in contrast with others reporting decreased levels of vitamin C and vitamin E or increased concentrations of uric acid (11-17). However, we think that these results should not be interpreted as evidence in favor of the greater or lesser importance of individual antioxidants, since plasma total antioxidant capacity is a function not only of the concentrations of DIABETES CARE, VOLUME 20, NUMBER 10, OCTOBER

Table 1—Various parameters in diabetic patients and control subjects Diabetic patients

Control subjects

46 10.5 ±0.5* 6.4 ± 0.2* 5.7 ±0.2 1.7 ±0.2* 346.8 ± 10.4* 1.3 ±0.1* 26.8 ± 3.0* 677 ± 15.8* 490.8 ± 8.0* 24.4 ± 1.2t 5.0±0.1t 17.8 ± 1.8 1.5±0.1t 279.5 ±10.1*

47 4.1 ±0.2 4.3 ±0.1 5.6 ±0.1 1.0 ±0.1 295.5 ± 12.3 0.7 ± 0.05 14.2 ±1.7 950.1 ± 16.0 547.2 ±10.1 20.6 ± 0.5 4.0 ±0.1 19.8 ± 1.1 1.7 ±0.1 309.2 ± 8.3

n Plasma glucose (mmol/1) HbAlc (%) Cholesterol (mmol/1) Triglycerides (mmol/1) Fibrinogen (mg/dl) Prothrombin fragments 1 + 2 (nmol/1) D-dimer (ng/ml) TRAP (pmol/l) SH groups (pmol/1) Vitamin E (umol/1) Vitamin E (umol/l)/cholesterol (mmol/1) Vitamin C (umol/1) Vitamin A (umol/1) Uric acid (pmol/1)

Data are means ± SE. *P < 0.001, t p < 0.005, *P < 0.01 vs. control subjects.

• - ! 4*0

5M

«••

s

7M

TRAP umoM

*••

7M

M0

TRAP uraoM

& 400

g 300 200 100

500

600

700

TRAP urn.*!

Figure 1—Correlation between TRAP andfibrinogen, prothrombin fragments 1 + 2 , and D-dimer in diabetic patients.

1997

1591

TRAP and thrombosis in NIDDM

4.

5.

6.

100

200

300

400

500

600

700

Fibrinogen mg/dJ

7.

8.

100

200

300

400

500

600

9.

700

Ftbrinogen mg/dl

Figure 2—Correlation between fibrinogen and prothrombin fragments 1 + 2 and D-dimer in diabetic patients.

individual antioxidants but also their synergism. It should be kept in mind that the various antioxidants give very different contributions to TRAP, uric acid being the most important (~50%) and vitamin E the least (~5%) (18,53). In this view, it is explained why an increase in vitamin E concentrations in diabetics is accompanied by a decrease of TRAP Low TRAP levels have been reported in IDDM and NIDDM (17-19). In this study, we confirm low levels of TRAP, which represents a more reliable estimation of plasma antioxidant capacity than the measurement of each of the known antioxidants (18), in NIDDM patients, compared with healthy normal control subjects. Moreover, we show a good correlation between TRAP and some markers of thrombosis in diabetes. In our opinion, these data support the hypothesis that oxidative stress may condition coagulation activation in diabetes. Even though it is well known that free radicals are capable of activating coagulation (3), how oxidative stress in diabetes would induce coagulation activation remains hypothetical. It has been shown that endothelial cells in hyperglycemic conditions overproduce tissue factor (54), while

1592

decreasing nitric oxide production (55) and fibrinolysis (56). These dysfunctions may be sufficient to produce a prothrombotic situation. However, some of these alterations are reversed by antioxidants, suggesting that free radicals are involved (55,56). In conclusion, we believe that the major novelty of our study consists in the finding that TRAP, but no single antioxidant, is inversely related tofibrinogen,prothrombin fragments 1 + 2, and D-dimer levels in NIDDM patients. This evidence stresses the concept that in evaluating the impact of oxidative stress on thrombosis in diabetes, a cumulative index of plasma antioxidant power should be considered, with less interest for the role of each single antioxidant.

10.

11. 12.

13.

14.

15.

16.

Oxygen radicals, lipid peroxidation and the coagulation system. Agent Actions 22:347348, 1987 Giugliano D, Ceriello A, Paolisso G: Oxidative stress and diabetic vascular complications. Diabetes Care 19:257-267, 1996 Baynes JW: Role of oxidative stress in development of complications in diabetes. Diabetes 40:405-412, 1991 Jennings PE, McLaren M, Scott NA, Saniabadi AR, Belch JJF: The relationship of oxidative stress to thrombotic tendency in type 1 diabetic patients with retinopathy. DiabetMed 8:860-865, 1991 Collier A, Rumley AG, Paterson JR, Leach JP, Lowe GDO, Small M: Free radical activity and hemostatic factor in NIDDM patients with and without microalbuminuria. Diabetes 41:909-913, 1992 Knobl P, Schernthaner G, Schack C, Pietschnann R Griesmacher A, Prager R, Miiller M: Thrombogenic factors are related to urinary albumin excretion rate in type 1 (insulin-dependent) and type 2 (noninsulin-dependent) diabetic patients. Diabetologia 36:1045-1050, 1993 Ceriello A, Giacomello R, Stel G, Motz E, Taboga C, Tonutti L, Pirisi M, Failed E, Bartoli E: Hyperglycaemia-induced thrombin formation in diabetes: the possible role of oxidative stress. Diabetes 44:924-928,1995 Halliwell B, Gutteridge JMC: The antioxidants of human extracellular fluids. Arch BiochemBiophys 280:1-8, 1990 Oberley LW: Free radicals and diabetes. Free Radic Biol Med 5:113-124,1988 Jones AF, Winkles JW, Jennings PE, Florkowski CM, Lunec J, Barnett AH: Serum antioxidant activity in diabetes mellitus. Diabetes Res 7:89-92,1988 Stankova L, Riddle M, Lamed J, Burry K, Menashe D, Hart J, Bigley R: Plasma ascorbate concentrations and blood cell dehydroascorbate transport in patients with diabetes mellitus. Metabolism 33:347-353, 1984 Karpen CW, Cataland S, O'Dorisio T, Panganamala R: Interrelation of platelet vitamin E and thromboxane synthesis in type I diabetes mellitus. Diabetes 33:239-243, 1984 Strain JJ: Disturbances of micronutrients and antioxidant status in diabetes. Proc NutrSoc 50:591-604, 1991 Asayama K, Uchida N, Nakane T, Hayashibe H, Dobashi K, Amemiya S, Kato K, Nakazawa S: Antioxidants in the serum of children with insulin-dependent dia-

References betes mellitus. Free Radic Biol Med 1. Ostermann H, van de Loo J: Factors of the 15:597-602, 1993 hemostatic system in diabetic patients: a survey of controlled studies. Haemostasis 17. Basu TK, Tze WJ, Leichter J: Serum vitamin A and retinol-binding protein in patients 16:386-416, 1986 with insulin-dependent diabetes mellitus. 2. Ceriello A: Coagulation activation in diaAmJClinNutr 50:329-331, 1989 betes mellitus: the role of hyperglycaemia and therapeutic prospects. Diabetologia 18. Ghiselli A, Serafini M, Maiani G, Azzini E, Ferro-Luzzi A: A fluorescence-based method 36:1119-1125,1993 for measuring total plasma antioxidant capa3. Barrowcliffe TW, Gutteridge JM, Gray E:

DIABETES CARE, VOLUME 20, NUMBER 10, OCTOBER 1997

Ceriello and Associates

bility Free Radio Biol Med 18:29- 36,1995 19. Ceriello A, Bortolotti N, Falleti E, Taboga C, Tonutti L, Crescentini A, Motz E, Lizzio S, Russo A, Bartoli E: Total radical-trapping antioxidant parameter in non-insulin dependent diabetic patients. Diabetes Care 20:194-197, 1997 20. Ernst E, Resch KL: Fibrinogen as a cardiovascular risk factor: a meta-analysis and review of the literature. Ann Intern Med 118:956-963, 1993 21. Kannel WB, D'Agostino RB, Wilson PW, Belanger AJ, Gagnon DR: Diabetes, fibrinogen and risk of cardiovascular disease: the Framingham experience. Am Heart ] 120:672-676, 1990 22. Ceriello A: Fibrinogen and diabetes mellitus: is it time for interventional trials? Diabetologia 40:731-734, 1997 23. Aronson DL, Stevan L, Ball AE Franza BR Jr., Finlayson JS: Generation of the combined prothrombin activation peptide (F1.2) during the clotting of blood and plasma. J Clin Invest 60:1410-1418, 1977 24. Graeff H, Hafter R: Detection and relevance of cross linked fibrin derivatives in blood. Sem Thromb Hemost 8:57-68, 1982 25. Ceriello A, Giacomello R, Colatutto A, Taboga C, Gonano F: Increased prothrombin fragment 1+2 in type 1 diabetic patients. Haemostasis 22:50-51, 1992 26. Myrupp B, Rossing P Jensen T, Gram J, Kluft C, Jespersen J: Procoagulant activity and intimal dysfunction in IDDM. Diabetologia 38:73-78,1995 27. Van Wersch JWJ, Westerhuis LWJJM, Venekamp WJRR: Coagulation activation in diabetes mellitus. Haemostasis 20:263-269, 1990 28. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1037-1057, 1979 29. Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L: A data-based approach to diet questionnaire design and testing. Am ] Epidemiol 124:453-469,1986 30. Mares-PerlmanJA, Klein BEK, Klein R, Ritter LL, Fisher MR, Freudenheim JL: A diet history questionnaire ranks nutrient intakes in middle-aged and older men and women similarly to multiple food records. J Nutr 123:489-501, 1993 31. Mallia AK, Heremenson GT, Krohn RS, Fujimoto EK, Smith PK: Preparation and use of a boronic acid affinity chromatography for separation and quantification of glycosylated hemoglobin. Analyt Lett 14:649-661, 1981

32. Tsan ZL, Chin N, Kiser MD, Bigler WN: Specific spectrophotometry of ascorbic acid in serum or plasma by use of ascorbate oxidase. Clin Chem 28:2225-2228, 1982 33. McCormickDB, Wright LD (Eds.): Vitamins and coenzymes, part F In Methods in Enzymology. Vol. 67. New York, Academic, 1980 34. Koster JF, Biemond P, Swaak AJ: Intracellular and extracellular sulphydryl levels in rheumatoid arthitis. Ann Rheum Dis 45:44-46, 1986 35. Rossi E, Lombardi A, Preda L: Method for determination of functional (clottable) fibrinogen by the new family of ACL coagulometers. Thromb Res 52:453-468, 1988 36. Pelzer H, Schwarz A, Stuber W: Determination of human prothrombin activation fragment F l + 2 in plasma with an antibody against a synthetic peptide. Thromb Haemostas 65:153-159, 1991 37. Rylatt DB, Blake AS, Cottis LE, Massingham DA, Fletcher WA, Masci PP, Whitaker AN, Elms M, Bunce I, Webber AJ, Wyatt D, Bundesden G: An immunoassay for human D-dimer using monoclonal antibodies. Thromb Res 31:767-778, 1983 38. Shapiro SS, Wilk MB: An analysis of variance test for normality (complete samples). Biometrika 52:591-611, 1965

DIABETES CARE, VOLUME 20, NUMBER 10, OCTOBER

1997

47.

48.

49.

50.

51.

39. Altman DG: Practical Statistics for Medical

Research. London, Chapman & Hall, 1991, p. 340-344 52. 40. Dixon WJ: BMDP Statistical Software Man-

41.

42. 43.

44.

45.

46.

ual. Berkeley, CA, University of California Press, 1992, p. 387-425. Blanchard P, Anton B, Larousse C, Auget D, Mainard F, Charbonnel B, Krempf M: Plasma vitamin E, non-high density lipoprotein cholesterol, and apolipoprotein B in diabetic patients. Clin Chem 38:23392340,1992 Colwell JA: Vascular thrombosis in type II diabetes mellitus. Diabetes 42:8-11, 1993 Garcia MJ, McNamara PM, Gordon T, Kannel WB: Morbidity and mortality in diabetes in the Framingham population: sixteen year follow-up study. Diabetes 23:105-111,1974 Theroux P, LatourJG, Leger-Gaudiier C, De Lara J: Fibrinopeptide A and platelet factor levels in unstable angina pectoris. Circulation 75:156-162, 1987 Kruskal JB, Commerford PJ, Franks JJ, Kirsch RE: Fibrin and fibrinogen-related antigens in patients with stable and unstable coronary artery disease. N Engl J Med 317:1361-1365, 1987 Siemens A, Giannitsis E, Brueckner S, Mitusch R, Schmuecker G, Sheikzadeh A,

53.

54.

55. 56.

Diederich KW, Wagner T: Severe coronary artery disease proves to be associated with increased levels of prothrombin fragments 1+2 (Abstract). Thromb Haemostas 73: A215,1995 Ridker PM, Hennekens CH, Cerskus A, Stampfer MJ: Plasma concentration of cross-linked fibrin degradation product (Ddimer) and the risk of future myoeardial infarction among apparently healthy men. Circulation 90:2236-2240, 1994 Ceriello A, Taboga C, Giacomello R, Falleti E, De Stasio G, Motz E, Lizzio S, Gonano F, Bartoli E: Fibrinogen plasma levels as a marker of thrombin activation in diabetes. Diabetes 43:430-432, 1994 Riemersa RA, Wood DA, Macintyre CCA, Elton RA, Gey KE. Oliver MF: Risk of angina pectoris and plasma concentrations of vitamins A, C and E and carotene. Ixincet 337:1-5, 1991 Stampfer MJ, Hennekens CH, Manson JE, Colditz GA, Rosner B, Willett WC: Vitamin E consumption and the risk of coronary disease in women. N Engl J Med 328:1444-1449, 1993 Rimm EB, Stampfer MJ, Ascherio A, Giovannucci E, Colditz GA, Willett WC: Vitamin E consumption and the risk of coronary disease in men. N Engl J Med 328:1450-1456,1993 Hertog MGL, Feskens EJM, Hollman PCH, Katan MB, Kromhout D: Dietary antioxidant flavonoids and risk of coronary heart disease: the Zutphen Elderly Study. Lancet 342:1007-1011, 1993 Wayner DDM, Burton GW, Ingold KU, Barclay LRC, Locke SJ: The relative contributions of vitamin E, urate, ascorbate and proteins to the total peroxyl radical-trapping antioxidant activity of human blood plasma. Biochcm Biophys Ada 924:408-419, 1987 Boeri D, Almus FE, Maiello M, Cagliero E, Rao LVM, Lorenzi M: Modification of tissue factor mRNA and protein response to thrombin and interleukin 1 by high glucose in cultured endothelial cells. Diabetes 38:212-218, 1989 Tesfamariam B: Free radicals and diabetic endothelial dysfunction. Free Radic Biol Med 16:383-391, 1994 Ceriello A, Curcio E Dello Russo P, Pegoraro I, Stel G, Amstad P, Cerutti P: The defense against free radicals protects endothelial cells from hyperglycaemiainduced plasminogen activator inhibitor 1 over-production. Blood Coag Fibrinolysis

6:133-137, 1995

1593