Effect of atorvastatin on hemorheologic-hemostatic parameters and serum fibrinogen levels in hyperlipidemic patients

Effect of Atorvastatin on HemorheologicHemostatic Parameters and Serum Fibrinogen Levels in Hyperlipidemic Patients Carlos A. Dujovne, MD, William S. Harris, PhD, Raul Altman, Ron W. Overhiser, and Donald M. Black, MD

MD, PhD,

Plasma fibrinogen and hemorheologic-hemostatic factors contribute to dyslipidemia-induced morbidity. Some of these parameters can be favorably affected when abnormal serum lipoprotein levels are corrected. Thus, we investigated whether treatment with atorvastatin would result in changes in plasma viscosity and other hemorheologic and hemostatic parameters. Twenty-two hyperlipidemic men at a university lipid clinic were treated single-blinded with atorvastatin 80 mg/day for 12 weeks to determine hemostatic-hemorheologic parameters including blood viscosity, fibrinogen levels, whole blood platelet aggregation, tissue plasminogen activator antigen, hematocrit, plasminogen activator inhibitor activity, factor VII activity, red blood cell (RBC) deformity and lipid ratio, sedimentation rate, and fasting serum lipoprotein levels. Atorvastatin treatment provided significant lowering of serum lipoprotein levels:

low-density lipoprotein ⴚ53% (p ⴝ 0.0001), very low density lipoprotein ⴚ43% (p ⴝ 0.0001), and triglycerides ⴚ35% (p <0.0001). These effects were accompanied by changes in plasma viscosity ⴚ10% (p ⴝ 0.0007), arachidonic acid-induced whole blood platelet aggregation ⴚ11% (p ⴝ 0.006), factor VII ⴚ8% (p ⴝ 0.001), RBC lipid composition ⴙ5% (p ⴝ 0.0003), and RBC sedimentation ⴚ33% (p ⴝ 0.0002). Plasma fibrinogen levels were not affected. Thus, atorvastatin 80 mg/day produced marked reductions in serum lowdensity lipoproteiin cholesterol (ⴚ53%), very low density lipoprotein cholesterol (ⴚ43%), and triglycerides levels (ⴚ35%), and significant changes in plasma viscosity as well as other hemorheologic-hemostatic parameters, but no changes in plasma fibrinogen levels. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;85:350 –353)

bnormal lipoprotein levels affect structure and function of the arterial endothelium and promote A platelet aggregation, thrombosis, and/or thromboem-

ering as seen with atorvastatin may impact blood rheology and hemostatic parameters more favorably and/or consistently.7,8 A recent controversial report described a significant elevation in serum fibrinogen after atorvastatin administration.9 We report the effects of atorvastatin on this and other related parameters in a single-blinded, placebo-controlled pilot study in dyslipidemic patients.

bolic phenomena through changes in hemostatic and hemorheologic factors such as fibrinogen, factor VII, plasma viscosity, hematocrit, and red cell agglutination. Hemostatic-hemorheologic factors contribute to morbidity and mortality in dyslipidemia.1 Lipoprotein-regulating drugs can affect hemostatic and hemorheologic parameters; the mechanisms are unclear and results of clinical studies have been equivocal.2 Pravastatin or lovastatin, but not cholestyramine or simvastatin, decreased plasma viscosity and fibrinogen in dyslipidemic patients.3 Lovastatin and pravastatin can decrease platelet aggregation.4 We published results showing that drugs or diets having moderate effects on reduction of serum lipids provided there are no changes in platelet aggregation and other hemostatic parameters.5,6 However, more pronounced lipid lowFrom the Lipid, Arteriosclerosis and Metabolic Clinic, University of Kansas Medical Center, Kansas City, Kansas; and Parke-Davis Pharmaceutical Research, Ann Arbor, Michigan. This study was supported in part by Parke-Davis Pharmaceuticals, Ann Arbor, Michigan. Manuscript received February 23, 1999; revised manuscript received and accepted August 30, 1999. Address for reprints: Carlos A. Dujovne, MD, Kansas Foundation for Clinical Pharmacology (KFCP), Lipid, Heart Disease and Stroke Prevention Clinic Radiant Research-Kansas City, 10550 Quivira Road, Suite 220, Overland Park, Kansas 66215. E-mail: carlosdujovne @radiantresearch.com.

350

©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 85 February 1, 2000

METHODS

Patients: Twenty-two men, aged 40 to 65 years (mean 53) with primary hyperlipoproteinemia were enrolled; 8 were classified as Fredrickson phenotype IIa (isolated hypercholesterolemia) (low-density lipoprotein [LDL] cholesterol ⬎160 mg/dl and triglycerides ⬍200 mg/dl), 8 type IIb (mixed hyperlipidemia, LDL cholesterol ⬎160 mg/dl and triglycerides ⱖ200 or ⬍650 mg/dl), and 6 type IV (isolated hypertriglyceridemia, LDL cholesterol ⬍160 mg/dl and triglycerides ⬎300 or ⬍650 mg/dl). Women were excluded to avoid possible hemorheologic-hemostatic variations associated with possible effects of female hormones on thrombosis-hemostasis parameters. Patients with endocrine, liver, hepatic, renal, or pulmonary diseases, ⬎30% of ideal body weight, substance abuse, smoking, or intake of ⬎14 oz/week of ethanol equivalents were excluded. Drugs with a known or potential effect on lipid levels or hemorheologic-hemostatic parameters (i.e., antihistamines, aspirin, ␤ blockers, cortico0002-9149/00/$–see front matter PII S0002-9149(99)00745-6

steroids) were prohibited. The protocol was approved by the Human Subjects’ Committee of the University of Kansas and patients provided written informed consent. Dietary and stabilization phase: Diet composition can affect lipid as well as hemostatic-hemorheologic parameters; thus, to maximize uniformity of diets and lifestyle habits, patients had dietary stabilization 6 weeks before placebo baseline period after counseling by a dietitian on the National Cholesterol Education Program Step 1 diet and dietary constancy monitored by analysis of 3 day’s eating diaries by validated methods at every visit (Professional Nutrition Systems, Inc., Overland Park, Kansas).10 Measurements of whole blood-platelet aggregation and hemorheologic parameters: Whole fresh blood

was used for platelet aggregation tests. Plasma was analyzed for fibrinogen, plasminogen activation inhibitor (PAI)-1 activity, tissue plasminogen activator antigen, and factor VII activity. The following precautions were taken to minimize variables known to affect these tests: (1) Subjects were recumbent for at least 15 minutes before the procedure; (2) all blood samples were obtained with a 19-gauge needle attached to a catheter taped to the skin to prevent vein wall irritation by needles and ensure constant blood flow without suction; (3) no tourniquet was applied; (4) all visits were at the same time of day after a 12-hour fast; (5) samples were analyzed fresh (except lipoprotein(a) [Lp(a)]). Baseline, placebo, and atorvastatin treatment phases: After the 6-week baseline phase of placebo treat-

ment, dietary monitoring, and counseling, eligible patients began 12 weeks of treatment with atorvastatin 80 mg/day, visiting every 4 weeks for complete clinical laboratory tests and monitoring of dietary constancy. At the end of the baseline phase and the 12th week of treatment, 2 blood samples 2 to 3 days apart were drawn for measurement of all parameters. Laboratory analyses: All clinical procedures and lipoprotein cholesterol determinations were done at the Lipid, Arteriosclerosis and Metabolic Clinic and Laboratory at the University of Kansas under the Lipid Standardization Program of the Centers for Disease Control and Prevention (National Heart, Lung, and Blood Institute).11 Lp(a) was determined by immunoturbidity assay. Plasma samples were analyzed fresh except for Lp(a), which was frozen at ⫺80°C and analyzed at termination of the study. Plasma viscosity was reported as the ratio of plasma flow time to water (normal range 1.4 to 1.8).12 Fibrinogen was assayed using the method of Clauss13 (normal range 1.46 to 3.90 g/L). The t-PA antigen was measured with a double-antibody enzyme-linked immunosorbent assay kit from Biopool (Grand Island, New York)14 (normal range 2 to 15 ng/ml). Factor VII activity was assayed by the 1-step method using factor VII deficient plasma15 (normal range 50% to 150% of control). Plasminogen activator inhibitor-1 activity was assayed spectrophotometrically (Stachrom PAI kit, Diagnostica Stago, Asnieres, France; normal range 0 to 15 U/ml).16 Red blood cell (RBC) deform-

TABLE I Mean Absolute Levels and Changes (%) from Baseline at 12 Weeks of Atorvastatin Therapy in All Patients (n ⫽ 22) Parameter LDL cholesterol (mg/dl) Baseline Final visit (12 wk) % Change VLDL cholesterol (mg/dl) Baseline Final visit (12 wk) % Change Total triglycerides (mg/dl) Baseline Final visit (12 wk) % Change

Levels

p Value

172 85 ⫺53a*

0.0001

49 23 ⫺43a*

0.0001

262 159 ⫺35a*

0.0001

*Statistically significantly different from baseline (p ⬎0.05). a ⫽ median change; VLDL ⫽ very low density lipoprotein.

ability was assayed by measuring the RBC filtration rate at which a 20% suspension of washed RBCs passed through a 5-␮m filter17 (normal rate 50 to 150 ␮l/s). RBC membrane lipid (cholesterol:phospholipid) ratio was determined by extracting RBC membrane “ghosts.” The cholesterol/phospholipid ratio of the membrane lipid extract was then quantitated enzymatically18 (normal ratio 0.6 to 1.1). Whole blood platelet aggregation was measured using the impedance technique after addition of each of 3 agonists (2 ␮g/ml collagen, 5 ␮mol adenosine diphosphate, or 0.5 mM arachidonic acid) to the test sample.19 Results were reported as the change in impedance after 6 minutes of incubation (normal ranges: collagen, 15 to 27 ohms; adenosine phosphate, 1 to 17 ohms; arachidonic acid, 5 to 17 ohms). Statistical analysis: The primary efficacy parameter was the change from baseline of plasma viscosity, which is known to be correlated to serum fibrinogen levels. For all other secondary efficacy parameters, the t test was performed to determine if the mean percent or absolute change from baseline was significantly different from zero. When the assumption for performing the t test was not met (i.e., when the data were not normally distributed but were symmetric), Wilcoxon signed rank test for medians was used. When the data did not appear to be symmetric, the sign test for medians was used.

RESULTS

Lipid parameters: The effects at 12 weeks of 80 mg/day atorvastatin therapy on the lipoprotein parameters showing a statistically significant difference from baseline are shown in Table I. We chose to study patients of the 3 most common phenotypes because there was no knowledge on which lipoprotein (LDL or very low density lipoprotein, or both) has major impact on the hemorheologic-hemostatic variables investigated. Results are not described separately by Fredrickson phenotype because all types responded with similarly statistically significant reductions from baseline.

PREVENTIVE CARDIOLOGY/HEMORHEOLOGY–HEMOSTATIC CHANGES AFTER ATORVASTATIN

351

TABLE II Changes in Hematologic Hemorheologic and Hemostatic Parameters Before and at the End (3 months) of Atorvastatin Therapy in 22 Patients After Three Months in All Patients (n ⫽ 22) Parameter Plasma viscosity Fibrinogen t-PA antigen ADP-induced aggregation AA-induced aggregation Collagen-induced aggregation Hematocrit PAI-1 activity Factor VII activity RBC deformability RBC lipid ratio RBC sedimentation Lp(a) concentration

Baseline

On Atorvastatin

2 316 20 9 16 20 43 10 114 114 0.8 10 23

1.8 315 19 9 12 21 43 10 104 109 0.84 6 32

⌬ ⫺.2 ⫺1

% Change

p Value ⌬

p Value % ⌬

⫺10*

0.0005

0.0007

0.06

0.07

⫺1

⫺1.7

⫺11*

⫺10

⫺8*

.045* ⫺3.3 9

⫺33a* 36*

95% CI ⫺14, ⫺4 ⫺2.3, ⫺0.04

0.006

⫺18, ⫺4

0.0005

0.001

⫺12, ⫺4

0.0003 0.0002

2.6, 6.8 0.0002 0.008

⫺50, ⫺24 12, 58

*Significantly different from baseline (p ⬍0.05). ⌬ showing p values ⬎0.07 (2-tailed) are not shown. a ⫽ median change; AA ⫽ arachidonic acid; ADP ⫽ adenosine diphospate; CI ⫽ 95% confidence intervals; t-PA ⫽ tissue-type plasminogen activator.

Hemorheologic parameters: Table II shows the changes in hematologic, hemorheologic, fibrinolytic, and hemostatic parameters for all patients combined because there were no apparent or statistically significant correlations found exclusively for any one of the phenotypes above. Atorvastatin significantly reduced plasma viscosity and whole blood platelet aggregation, which is a preferable parameter over aggregation of platelet-rich plasma because it allows for evaluation of interactions among all synergistic blood components that affect platelet reactivity.20 Factor VII activity and RBC sedimentation were also decreased. Atorvastatin and plasma fibrinogen level: Plasma fibrinogen levels were not changed to any extent that could be clinically or statistically significant; baseline values and values after atorvastatin treatment were almost identical. There was no statistically significant correlation between hematologic-hemorheologic changes and changes in lipoprotein parameters.

DISCUSSION Triglycerides, very low density lipoprotein cholesterol, and LDL cholesterol may relate to a different extent to the mechanisms of hemostasis that were measured.21 As in previous trials with atorvastatin,7,8 there were pronounced improvements in LDL and very low density lipoprotein cholesterol levels (reduced by 53% and 43%, respectively). Baseline hemostatic-hemorheologic parameters were similar across lipoprotein phenotypes with the exception of PAI-1 activity in type IV patients, which was 3 times larger than that of type IIa and IIb patients, but nevertheless not changed at all with atorvastatin. Factor VII activity levels, which have been directly correlated with cardiovascular mortality,22 were statistically significantly reduced. Inhibition of arachidonic acid-induced platelet ag352 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 85

gregation indicates a lower reactivity to thromboxane and prostaglandin 2 receptors in platelets or a decreased production of these intermediate products. Lp(a) values were elevated with atorvastatin treatment in this study but barely above the normal range. The exact role for this degree of change in Lp(a) on cardiovascular risk is not well established.23 The risk of coronary disease is increased 2 to 3 times in subjects with fibrinogen levels within the upper third percentile.24 Serum fibrinogen is directly correlated with plasma viscosity,25 which is also a cardiovascular risk factor.26 A 6-year follow-up study showed that for persons with high LDL cholesterol, cardiovascular risk increased over sixfold with high fibrinogen concentrations.27 Pravastatin and lovastatin were shown to lower serum fibrinogen levels in some studies. Yet these results have not been consistently reproduced with this or other statins.28 Wierzbicki et al9 showed an increase in plasma fibrinogen levels measured with an immunoturbidimetric method in patients treated with atorvastatin; plasma viscosity was not measured in that study. In our study, fibrinogen levels, measured by the clotting method of Clauss,13 were not changed with atorvastatin treatment. The difference in results may be related to methodologic factors. Fibrinogen is an acute phase reaction protein and can be influenced in a variety of situations. Strenuous physical activity and various types of stress or myocardial infarction can increase serum fibrinogen up to 37% for several days.23 Twenty-five percent of patients reported by Wierzbicki had coronary heart disease, but details of the clinical conditions of those patients or time of clinical events at testing were not described or controlled. The immunologic assay used by Wierzbicki et al9 is sensitive to fibrinogen as well as to fibrinogen degradation products, and thus may register a higher value under certain conditions.24 The measures in our FEBRUARY 1, 2000

patients by the Clauss method showed no increase in mean fibrinogen levels, but an overall nonsignificant 5% to 15% decrease in PAI-1 antigen in some patients after atorvastatin treatment. If PAI-1 was affected in the patients of Wierzbicki et al, the increase in fibrinogen titers could be due to increased fibrinogen degradation products resulting from decreased PAI-1 antigen levels. Plasma viscosity, the primary end point of our study, was significantly decreased. Elevated fibrinogen levels are always positively correlated with plasma viscosity. Thus, our results on significant reduction in plasma viscosity are discrepant, with the possibility of an elevation in serum fibrinogen levels by atorvastatin. The change in decreased plasma viscosity documented in this study, regardless of the mechanism, was obtained after a brief (3-month) treatment period, and may present a useful additional surrogate end point for documenting short-term benefits of highly effective lipid-lowering therapies.28 Acknowledgment: We thank Parke-Davis Pharmaceutical Research for partial support of this study. We also thank Rodolfo Paoletti, MD, University of Milan, Italy, for reviewing the manuscript and for his suggestions on methods, results, and interpretation of the studies. We are grateful to Marjorie Zucker, MD, for performing fibrinogen, tissue-type plasmingoen activator, factor VII, and PAI-1 activities, to Kathy Harris, RN, for patient recruitment and trial coordination, to Sheryl Windsor, MT (ASCP), and Jeff Lickteig for technical support in the laboratory, to Rachel Laskey, PhD, for assistance with the manuscript, and to Patricia McFarland for editorial assistance. Jolene Held, RD (Professional Nutrition Systems, Inc.) was responsible for analysis of dietary diaries for the documentation of dietary constancy during the trial.

1. Meade TW. Hypercoagulability and ischaemic heart disease. Blood Rev

1987;1:2– 8. 2. Beigel Y, Fuchs J, Snir M, Green P, Lurie Y, Djaldetti M. Lovastatin therapy

in hypercholesterolaemia: effect on fibrinogen, hemorheologic parameters, platelet activity, and red blood cell morphology. J Clin Pharmacol 1991;31:512–517. 3. Jay RH, Rampling MW, Betteridge DJ. Abnormalities of blood rheology in familial hypercholesterolemia: effects of treatment. Atherosclerosis 1990;85: 249 –256. 4. Hochgraf E, Levy Y, Aviram M, Brook J, Cogan U. Lovastatin decreases plasma and platelet cholesterol levels and normalizes elevated platelet fluidity and aggregation in hypercholesterolemic patients. Metabolism 1994;43:11–17. 5. Zucker ML, Woodroof J, Krehbiel P, Jackson B, Chernoff S, Dujovne C. The effect of variable fat diets and cholesterol-lowering drugs on antithrombin III levels in hyperlipoproteinemic and normal subjects. Thromb Res 1983;30:661– 669.

6. Zucker ML, Trowbridge C, Krehbiel P, Jackson B, Chernoff S, Dujovne C.

Platelet function in hypercholesterolemics before and after hypolipidemic drug therapy. Haemostasis 1986;16:57– 64. 7. Davidson M, McKenney J, Stein E, Schrott H, Bakker-Arkema R, Fayyad R, Black D, for the Atorvastatin Study Group I. Long-term efficacy and safety of atorvastatin compared to lovastatin in hypercholesterolemic patients. Am J Cardiol 1997;79:1475–1481. 8. Dart A, Jerums G, Nicholson G, Nicholson G, d’Emden M, Hamilton-Craig I, Tallis G, Best J, West M, Sullivan D, Bracs P, Black D. A multi-center, double-blind, one-year study comparing safety and efficacy of atorvastatin and simvastatin in patients with hypercholesterolemia. Am J Cardiol 1997; 80:39 – 44. 9. Wierzbicki AS, Lumb PJ, Semra YK, Crook MA. Effect of atorvastatin on plasma fibrinogen. Lancet 1998:351:569 –570. 10. Harris W, Held S, Dujovne CA. Comparison of two scoring systems used to monitor diets in outpatient clinical trials. J Cardiovasc Risk 1995;2:359 – 365. 11. Lipid Research Clinics Program. Manual of Laboratory Operations. vol 1. Lipid and lipoprotein analysis. (DHEW publication no. (NIH)75– 628). Washington, DC: U.S. Government Printing Office, 1974. 12. Wolf R, Levin W, Ritzmann S. Serum protein abnormalities, diagnostic and clinical aspects. Thermoproteins 1975;20:494 – 498. 13. Clauss A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haematol 1957;17:237–246. 14. Ranby M, Nguyen G, Scarabin PY, Samama M. Immunoreactivity of tissue plasminogen activator and of its inhibitor complexes. Biochemical multicenter validation of a two site immunosorbent assay. Thromb Haemost 1989;61:409 – 414. 15. Simidge M, Shannon R, eds. Laboratory Evaluation of Hemostasis and Thrombosis. 3rd ed, Philadelphia: Lea & Febiger, 1983:151–160. 16. Eriksson E, Ranby M, Gyzander E, Risberg B. Determination of plasminogen activator inhibitor in plasma using t-PA and a chromogenic single-point poly-Dlysine assay. Thromb Res 1988;50:91–101. 17. Yamaguchi H, Ailers M, Roberts D. The effects of urea on red cell deformability during cardiopulmonary bypass. Scand J Thorac Cardiovasc Surg 1983; 18:119 –122. 18. Rodriguez-Vico F, Martinex-Cayuela M, Flora Z, Garcia-Peregrin E, Ramirez H. A procedure for the simultaneous determination of lipid and protein in biomembranes and other biological samples. Lipids 1991;26:77– 80. 19. Reiss H, Braun G, Brehm G, Hiller E. Critical evaluation of platelet aggregation in whole human blood. Am J Clin Pathol 1986;85:50 –58. 20. Valles J, Santos T, Aznar J, Marcus A, Martinez-Sales V, Portoles M, Broekman M, Safier L. Erythrocytes metabolically enhanced collagen-induced platelets responsiveness via increasing thromboxin production, ADP release and recruitment. Blood 1991;78:154 –162. 21. Hachinski V, Graffagnino C, Beaudry M, Bernier G, Buck C, Donner A, Spence J, Doig G, Wolfe B. Lipids and stroke: a paradox resolved. Arch Neurol 1996;53:303–308. 22. Meade TW, Mellows S, Brozovic M, Miller G, Chakrabarti R, North W, Haines A, Stirling Y, Imeson J, Thompson S. Haemostatic function and ischemic heart disease: principal results of the Northwick Park Heart Study. Lancet 1986;2:533–537. 23. Altman R. Assessment of effects of lipid-lowering drugs on hemostasis. Am J Cardiol 1998;81:73F-74F. 24. Ernst E, Resch KL. Fibrinogen as a cardiovascular risk factor: a meta-analysis and review of the literature. Ann Intern Med 1993;118:956 –963. 25. Sweetnam P, Thomas H, Yarnel J, Beswick A, Baker I, Elwood P. Fibrinogen, viscosity and the 10-year incidence of ischaemic heart disease. Eur Heart J 1996;17:1814 –1820. 26. Ernst E, Resch K. Fibrinogen as a cardiovascular risk factor: a meta-analysis and review of the literature. Ann Intern Med 1993;118:956 –963. 27. Assman G, Cullen P, Heinrich J, Schulte H. Hemostatic variables in the prediction of coronary risk: results of the 8 year follow-up of healthy men in the Munster Heart Study (PROCAM). Prospective Cardiovascular Munster Study. Isr J Med Sci 1996;32:364 –370. 28. Sirtori C, Colli S. Influences of lipid-modifying agents on hemostasis. Cardiovasc Drugs Therap 1993;7:817– 823.

PREVENTIVE CARDIOLOGY/HEMORHEOLOGY–HEMOSTATIC CHANGES AFTER ATORVASTATIN

353