Anticardiolipin (acl) antibodies and lupus anticoagulant

Transcription

1 Anticardiolipin Antibodies and Recurrent Coronary Events A Prospective Study of 1150 Patients Androniki Bili, MD; Arthur J. Moss, MD; Charles W. Francis, MD; Wojciech Zareba, MD, PhD; Luc F. Miller Watelet, PhD; Ignacio Sanz, MD; for the Thrombogenic Factors and Recurrent Coronary Events Investigators* Background The association of anticardiolipin (acl) antibodies with coronary artery disease has been shown in several studies but remains controversial. We evaluated the association of acl and anti 2 -glycoprotein I (a 2 GPI) antibodies with the risk of recurrent cardiac events in postinfarction patients. Methods and Results The study population consisted of 1150 patients with acute myocardial infarction. Levels of IgG and IgM acl and a 2 GPI antibodies were determined on sera collected before hospital discharge. There were 131 recurrent cardiac events (nonfatal myocardial infarctions or cardiac deaths) over a mean follow-up period of 24.6 months. Patients with elevated IgG acl antibodies had a higher event rate than patients with low levels (P 0.05). Multivariate Cox analysis after adjustment for relevant clinical covariates showed that elevated levels of IgG acl (hazard ratio 1.63; P 0.01) and low levels of IgM acl (hazard ratio of 1.76; P 0.02) antibodies contribute independent risks for recurrent cardiac events. Patients with elevated IgG acl and low IgM acl antibody levels had a 3-fold higher risk of recurrent cardiac events than patients with low IgG acl and elevated IgM acl antibody levels (P 0.001). There was no significant association of the a 2 GPI antibodies with recurrent cardiac events. Conclusions In postinfarction patients, elevated IgG acl and low IgM acl antibodies are independent risk factors for recurrent cardiac events. Patients with both elevated IgG acl and low IgM acl antibodies have the highest risk. These findings shed additional light on the mechanistic role of acl antibodies in coronary artery disease in patients without autoimmune diseases. (Circulation. 2000;102: ) Key Words: antibodies thrombosis coronary disease myocardial infarction Anticardiolipin (acl) antibodies and lupus anticoagulant are the main antibodies associated with the antiphospholipid syndrome, which manifests clinically as vascular thrombosis and pregnancy morbidity. 1 acl antibodies are strongly associated with venous and arterial thrombosis, both in patients with systemic lupus erythematosus (SLE) and in patients without any apparent autoimmune diseases. 2,3 It is known that acl antibodies require a serum cofactor, 2 -glycoprotein I ( 2 GPI) or apolipoprotein H, for binding to cardiolipin in vitro. 4 Antibodies to 2 GPI can be measured directly and were shown to be more specific than acl antibodies for the clinical manifestations of antiphospholipid syndrome. 5 7 However, these are mostly retrospective studies and mainly involve patients with SLE. An association of acl antibodies with coronary artery disease has been shown in several 8 13 but not all studies. 14,15 In the present study, we evaluated the association of acl and anti 2 -GPI (a 2 GPI) antibodies with recurrent myocardial infarction (MI) or cardiac death in a large cohort of patients without autoimmune diseases. The patients were participants in the Thrombogenic Factors and Recurrent Coronary Events (Thrombo) study, a multicenter prospective cohort study that examined the risk of different thrombogenic factors for recurrent coronary events. 16 Methods Patients The Thrombo study recruited 1161 patients hospitalized for acute MI (index MI). Serum samples were collected before hospital discharge and at 2 months after the acute event. We conducted our study on the predischarge sera (average time from index MI to discharge 10.9 days, with range of 1 to 31 days). The study included men or women 21 years of age with enzyme confirmation of the index MI with symptoms and/or ECG changes consistent with an acute MI. The index MI could be Q-wave or non Q-wave in type and either a first or a recurrent infarction. Exclusion criteria were coronary artery bypass grafting surgery during the hospital phase of the index MI, malignancy, or severe hepatic, renal, or cerebral disease. 16 Received January 4, 2000; revision received April 6, 2000; accepted April 7, From the Allergy, Immunology, and Rheumatology Unit (A.B., I.S.), the Cardiology Unit (A.J.M., W.Z.), and the Vascular Medicine Unit (C.W.F.) of the Department of Medicine and the Department of Biostatistics (L.F.M.W.), University of Rochester Medical Center, Rochester, NY. *The Thrombo Investigators are listed in Reference 16. Correspondence to Arthur J. Moss, MD, University of Rochester Medical Center, 601 Elmwood Ave, BOX 653, Rochester, NY heartajm@heart.rochester.edu 2000 American Heart Association, Inc. Circulation is available at

2 Bili et al Anticardiolipin Antibodies and Coronary Events 1259 TABLE 1. Baseline Patient Characteristics According to acl Antibody Type and Quartile Level acl Antibody IgG acl Antibody IgM Characteristic All Patients (n 1150) Q 4 (n 288) Q 1 3 (n 862) Q 4 (n 257) Q 1 3 (n 893) Age 60 y Male sex Nonwhite race 26 35* 23 19* 28 History Prior MI Treated hypertension Diabetes mellitus Stroke Cigarette smoking status Never smoked Ex-smoker Current smoker Index MI New Q wave Pulmonary congestion Ejection fraction Medications at enrollment Acetylsalicylic acid ACE inhibitors Blockers Calcium-channel blockers Diuretics Oral anticoagulants Lipid-lowering agents acl antibodies IgG, GPL IgM, MPL Antibodies to 2 GPI IgG, SGU IgM, SMU Q 4 indicates highest quartile; Q 1 3, lower 3 quartiles. Values are percentage; plus-minus values are mean SD. *P 0.01 between highest quartile and lower 3 quartiles. Dicumarol or warfarin. Includes statin-type medications. Study Protocol Baseline demographic data regarding known risk factors for coronary artery disease and recurrent MI were collected in all patients. The patients were followed up by phone and with regular clinic visits. Eighteen patients were lost to follow-up, and they were included in the analysis up to the time of their last contact. The human investigation committees of all the participating hospitals approved the study protocol. Written informed consent was obtained from all study participants. Measurement of Antibodies All the tests were done in duplicate with sera stored at 70 C, and the mean value was obtained. One of the authors performed all the tests in the same laboratory environment. Determination of acl antibodies was performed by standardized ELISA for both IgG and IgM isotypes (REAADS Medical Products, Inc) with bovine calf serum in the sample diluent as the source of 2 GPI. Results are expressed as GPL units for the IgG (positive 23 GPL) and MPL units for the IgM (positive 11 MPL) acl antibodies, with 1 GPL or MPL unit being equivalent to 1 g/ml of an affinity-purified standard IgG or IgM acl antibody sample. 17 Determination of IgM and IgG a 2 GPI antibodies was performed by ELISA with irradiated and chemically activated plastic microwell plates containing purified human 2 GPI (INOVA, Inc). Results are expressed in standardized units, 18 SGU for the IgG (positive 20 SGU) and SMU for the IgM (positive 20 SMU) a 2 GPI antibodies. All tests met the quality control standards as determined by the manufacturer. Statistical Analysis The end points were cardiac mortality and cardiac morbidity with rehospitalization as the result of recurrent nonfatal MI (cardiac

3 1260 Circulation September 12, 2000 events). The same diagnostic criteria as for the index MI were used for the recurrent nonfatal MI. The first recurrent cardiac event (cardiac death or recurrent nonfatal MI) that occurred was used as the end point for each patient. The baseline characteristics of patients with versus patients without cardiac events were compared by 2 test. Univariate analysis of the IgG and IgM acl and a 2 GPI antibodies in relation to the outcome was performed by Wilcoxon s rank-sum test and 2 test where appropriate. For the survival analyses, the IgG and IgM acl and IgM a 2 GPI antibodies were dichotomized at a cutoff point closest to the 75th percentile, and comparisons of the highest quartile (Q 4 ) versus the lower 3 quartiles (Q 1 3 ) were performed. This dichotomization parallels the analysis of the original Thrombo study and resolves the problem of skewness. The cutoff values for the highest quartile were 12.5 GPL for IgG acl, 4.1 MPL for IgM acl, and 2.0 SMU for IgM a 2 GPI antibodies. For IgG a 2 GPI antibodies, values of the 95th percentile were used for analysis rather than the quartile approach because there were only 47 patients with any detectable level of these antibodies. Kaplan-Meier survival analysis was used for estimation of event rates in association with the IgG and IgM acl and a 2 GPI antibodies. The Cox proportional-hazards survivorship model was used to evaluate the independent contribution of risk variables to outcome events. In the construction of the Cox model, a clinical model was first obtained by including any clinical variable that entered the model at a level of significance 0.1 after a forward stepwise regression scheme. This approach was used to ensure inclusion of any added contribution to this model from important covariates with borderline levels of significance. Individual medications (as listed in Table 1) were tested in the model one at a time. The SAS statistical program (SAS Institute, Inc) was used for data analysis. Results Of the 1161 patients recruited, serum samples were available for testing on The baseline characteristics for the entire study population and for patients in the highest and lower 3 IgG and IgM acl antibody quartiles are shown in Table 1. Race was the only covariate with significantly different frequencies in the high and low quartile distributions of both IgG and IgM acl antibodies. During the mean follow-up period of 24.6 months (range 0 to 42 months), a total of 131 recurrent cardiac events (84 cardiac deaths and 47 nonfatal MI) occurred. The group with events was older, with a higher percentage having had a prior MI before the index MI and having hypertension, diabetes mellitus, pulmonary congestion, and ejection fraction 30%. The mean levels and number of patients in the highest quartiles of IgG and IgM acl and a 2 GPI antibodies in those with and without coronary events are shown in Table 2. There is a trend for higher IgG acl and lower IgM acl antibody levels in patients with than without events (P 0.06 for both). The highest quartiles of IgG acl and IgM acl antibodies show a similar trend (Table 2). The IgG and IgM a 2 GPI antibodies did not show a significant association or a meaningful trend with recurrent cardiac events. Kaplan-Meier survivorship analysis for the IgG and IgM acl antibodies is shown in the Figure. Patients in the highest quartile of IgG acl antibodies had a higher coronary event rate than patients in the lower 3 quartiles (P 0.05). A reverse trend was observed for the IgM acl antibodies (P 0.06). The risk contributed by acl antibodies to the Cox model after adjustment for relevant clinical covariates is shown in Table 3. The variables that entered the clinical model at the 0.1 level of significance were age 60 years, prior MI before TABLE 2. Distribution of acl and a 2 GPI Antibodies in Patients With and Without Recurrent Coronary Events Antibodies Patients With Events (n 131) Patients Without Events (n 1019) P * IgG acl Level, GPL Q 4, 12.5 GPL 41 (31) 245 (24) 0.07 Positive, 23 GPL 14 (11) 64 (6) 0.06 IgM acl Level, MPL Q 4, 4.1 MPL 23 (18) 258 (25) 0.05 Positive, 11 MPL 4 (3) 40 (4) 0.6 IgG a 2 GPI Level, SGU NS Detectable, any 7 (5) 40 (4) NS Positive, 20 SGU 2 (2) 7 (1) NS IgM a 2 GPI Level, SMU NS Q 4, 2.0 SMU 25 (19) 263 (26) NS Positive, 20 SMU 5 (4) 41 (4) NS Values are numbers of patients (percentage); plus-minus values are mean SD. Q 4 indicates highest quartile; Q 1 3, lower 3 quartiles. * 2 test. Only 47 patients had any detectable IgG a 2 GPI (95th percentile). P the index MI, diabetes mellitus, smoking, pulmonary congestion, and ejection fraction 30%. In this model, IgG acl antibodies in the highest quartile of the distribution contributed an independent hazard ratio (HR) of 1.63 (95% CI 1.11 to 2.38; P 0.01) for recurrent coronary events. In contrast, IgM acl antibodies in the lower 3 quartiles of the distribution contributed an independent HR of 1.76 (95% CI 1.10 to 2.82; P 0.02) for recurrent coronary events. Because the IgG and IgM acl antibodies contributed comparable opposite effects, we explored the joint risk for patients in the high- and low-risk partitions of both acl antibody isotypes. The group of patients with IgM acl antibodies in the highest quartile and IgG acl antibodies in the lower 3 quartiles had the lowest risk and was chosen as the reference group. Patients with IgG acl antibodies in the highest quartile and IgM acl antibodies in the lower 3 quartiles had higher risks than any other acl antibody group and 3-fold higher risk than the reference group (P 0.001) (Table 4). The cutoff values for the upper quartiles of the IgG ( 12.5 GPL) and IgM ( 4.1 MPL) acl antibodies are lower than the values that are considered positive in the acl antibody assay that was used in our study ( 23 GPL for IgG and 11 MPL for IgM acl antibodies). When the data were analyzed according to the positive values, positive IgG and IgM acl antibodies were associated with hazard ratios of 2.0 (P 0.01) and 0.8 (P 0.6), respectively. The low number of positive IgG and IgM acl antibodies in the group of patients with events (Table 2) was the main obstacle to the interpretation of these results.

4 Bili et al Anticardiolipin Antibodies and Coronary Events 1261 Figure 1. Kaplan-Meier cumulative cardiac event rates for highest quartile (Q 4 ) vs lower 3 quartiles (Q 1 3 ) for IgG (A) and IgM (B) acl antibodies. The correlation coefficients for acl and a 2 GPI antibodies by isotype are shown in Table 5. A moderately strong correlation (r 0.50) was present only between IgM acl and a 2 GPI IgM antibodies. Discussion This report represents the largest study to date evaluating the association between acl and a 2 GPI antibodies and recurrent cardiac events in patients after MI. The data presented herein show that IgG acl antibodies in the highest quartile represent an independent risk factor for recurrent MI and cardiac death, with an HR of 1.63 after adjusting for other relevant clinical variables. This risk is comparable to other known risk factors for recurrent coronary events in our study: prior MI (HR 1.77), non insulin-dependent diabetes (HR 1.87), ejection fraction 30% (HR 1.69), and smoking (HR 1.81). Our data also revealed an inverse association of IgM acl antibody levels with recurrent cardiac events. Thus, patients with elevated IgG acl and low IgM acl antibodies had the highest risk compared with all other acl antibody groups. The association between myocardial infarction and IgG acl antibodies has been suggested in several smaller studies With regard to noncardiac events, IgG acl antibod-

5 1262 Circulation September 12, 2000 TABLE 3. Hazard Ratios of acl Antibodies in Multivariate Cox Model acl Antibodies HR* 95% CI P IgG (Q 4 :Q 1 3 ) IgM (Q 1 3 :Q 4 ) Q 1 3 indicates lower 3 quartiles; Q 4 : highest quartile. Cutoff values for Q 4 : IgG acl 12.5 GPL; IgM acl 4.1 MPL. Cox model includes adjustment for 6 clinical variables (P 0.1 to enter): age 60 y, diabetes mellitus, smoking, prior MI, pulmonary congestion, and ejection fraction *Ratio of the risk of having a cardiac event per unit time for patients in the top-risk partition vs patients in the lower-risk partition. ies at levels above the 95th percentile have been implicated as a significant risk factor for venous thrombosis or pulmonary embolus in a case-control study from the Physicians Health Study 2 and for first stroke in a large case-control study. 3 In other studies that found a positive association of the IgG acl antibodies with arterial thrombosis, information about the IgM acl antibodies was not reported 19 or the association was not conclusive. 8 Our study, because of its large size, had the power to detect such an association. Before our study, minimal information was available regarding the association between a 2 GPI antibodies and coronary events. We found no association between these antibodies (either IgG or IgM) with recurrent cardiac events despite a meaningful correlation (r 0.50) between acl IgM and a 2 GPI IgM antibodies. Our findings confirm and extend the only relevant study available in which only the IgG isotype was measured. 20 This result is consistent with Vaarala s 21 observation that a 2 GPI antibodies are infrequently detected in nonautoimmune patients with MI. The question of whether acl antibodies can be induced in response to tissue necrosis that occurs in MI is still unresolved. 22 However, there is clinical evidence that acl antibodies precede the development of first MI 12 and that acl antibody titers are stable for up to 3 months after MI. 11 These studies indicate that acl antibodies are not generated by tissue necrosis but rather they participate in the pathogenesis of MI. This hypothesis is biologically plausible because of the proinflammatory and procoagulant properties of these antibodies. The acl auto- TABLE 4. Joint Risks for Patients in High- and Low-Risk Partitions of IgG and IgM acl Antibodies IgG acl Antibodies IgM acl Antibodies Events/Patients, % HR P Q 1 3 Q 4 12/ * (6.8) Q 1 3 Q / (11) Q 4 Q 4 11/ (14) Q 4 Q / (14) Q 1 3 indicates lower 3 quartiles; Q 4, highest quartile. Cutoff values for Q 4 : IgG acl 12.5 GPL; IgM acl 4.1 MPL. *Group with the lowest risk (reference group). TABLE 5. Correlation Between acl and a 2 GPI Antibodies by Genotype Antibodies IgM acl IgG a 2 GPI IgM a 2 GPI IgG acl r 0.16 r 0.14 r 0.13 Igm acl r 0.08 r 0.50 IgG a 2 GPI r 0.12 P for all r correlation values. antibodies detected in SLE are primarily directed against the so-called cofactor proteins, including 2 GPI (the predominant cofactor), prothrombin, annexin V, thrombomodulin, and proteins C and S, among others. 26,30 Such cofactors share the property of being phospholipid-binding proteins with intrinsic anticoagulant activity. Hence, binding of acl antibodies may diminish the function of these anticoagulant systems. In addition, IgG acl antibodies enhance platelet activation and thromboxane production 27 and activate endothelial cells in the presence of 2 GPI. 28,29 The interplay among acl antibodies, the anticoagulant system, and the endothelial cells could contribute to the prothrombotic state that is central to the pathogenesis of recurrent MI. 16,31 A unique finding of our study is the reverse association of the IgM acl antibodies with recurrent cardiac events, even in the presence of IgG acl antibodies. Several non mutually exclusive mechanisms could account for this observation. IgM acl antibodies might represent protective natural autoantibodies 32 or they could have rheumatoid factor activity and thus play a beneficial role in immune homeostasis. 33 Alternatively, it could be postulated that the acl antibody response in our patients starts with a primary IgM antibody response against nonpathogenic epitopes that subsequently, through epitope spreading, results in the recognition of previously ignored pathogenic epitopes by antigen-selected IgG antibodies. 34 A hypothetical example of a protective initial IgM antibody response might be the binding and inhibition of the colonization of the vascular endothelium by infectious agents such as Chlamydia pneumoniae that may play a role in the development of atherosclerosis. 35,36 The ability of acl antibodies to recognize chlamydial antigens, or conversely, the ability of chlamydial infection to elicit acl antibodies, remains to be tested. This model is appealing because it is also consistent with the lack of association of recurrent coronary events with a 2 GPI antibodies that was observed in our study. Indeed, acl antibodies induced by infections do not recognize 2 GPI, 37 and these antibodies may be prevalent in our cohort of nonautoimmune patients. It should be noted that acl antibodies cross-react with oxidized LDL 25 and that antibodies against oxidized LDL have been implicated in the development of atherosclerosis. 22 When considered that both C pneumoniae and chlamydial hsp60, an inflammatory antigen that was recently localized in atheromas, can induce cellular oxidation of LDL, 38 a reasonable link between chlamydial infection, acl antibodies, and atherosclerosis emerges. In summary, IgG acl antibodies are an independent risk factor for recurrent cardiac events in this large population of nonautoimmune patients. Our findings contribute new data in the ongoing research for identification of additional coronary risk factors. 39 At present, the use of IgG acl antibodies to

6 Bili et al Anticardiolipin Antibodies and Coronary Events 1263 stratify patients for coronary risk is limited by the heterogeneity of these antibodies, 40 the available commercial kits, 41 and the lack of standardization of the acl antibody assays between the different laboratories. 42 As we learn more about the different specificities of the acl antibodies and more standardized assays develop, acl antibodies may become useful for identifying patients at risk for coronary events. Although high levels IgG and low levels of IgM acl antibodies convey independent risk, the predictive value of these antibodies is too low for use in risk-stratification application at this time. Acknowledgments This study was supported by National Institutes of Health grants HL and HL and by the University of Rochester School of Medicine and Dentistry. We thank Elsa Welch for technical support and Dr Elena Rey for advice with the ELISA tests. We thank Dr John Condemi for reviewing the manuscript. References 1. Wilson WA, Gharavi AE, Koike T, et al. International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome. Arthritis Rheum. 1999;42: Ginsburg KS, Liang MH, Newcomer L, et al. Anticardiolipin antibodies and the risk for ischemic stroke and venous thrombosis. Ann Intern Med. 1992;117: Antiphospholipid Antibodies in Stroke Study (APASS) Group. Anticardiolipin antibodies are an independent risk factor for first ischemic stroke. Neurology. 1993;43: Sheng Y, Kandiah DA, Krilis SA. 2 -Glycoprotein I: target antigen for antiphospholipid antibodies : immunological and molecular aspects. Lupus. 1998;2:S5 S9. 5. Tsutsumi A, Matsuura E, Ichikawa K, et al. Antibodies to 2 -glycoprotein I and clinical manifestations in patients with systemic lupus erythematosus. Arthritis Rheum. 1996;39: Roubey RAS, Maldonado MA, Byrd SN. Comparison of an enzymelinked immunosorbent assay for antibodies to 2 -glycoprotein I and a conventional anticardiolipin immunoassay. Arthritis Rheum. 1996;39: Cabiedes J, Cabral AR, Alarcon-Segovia D. Clinical manifestations of the antiphospholipid syndrome in patients with systemic lupus erythematosus associate more strongly with anti- 2 -glycoprotein I than with antiphospholipid antibodies. J Rheumatol. 1995;22: Morton KE, Krilis SA, Baron DW, et al. Coronary artery bypass graft failure: an autoimmune phenomenon? Lancet. 1986;2: Hamsten A, Bjorkholm M, Norberg R, et al. Antibodies to cardiolipin in young survivors of myocardial infarction: an association with recurrent cardiovascular events. Lancet. 1986;1: Klemp P, Cooper RC, Strauss FJ, et al. Anticardiolipin antibodies and ischaemic heart disease. Clin Exp Immunol. 1988;74: Zuckerman E, Toubi E, Shiran A, et al. Anticardiolipin antibodies and acute myocardial infarction in non-systemic lupus erythematosus patients: a controlled prospective study. Am J Med. 1996;171: Vaarala O, Manttari M, Manninen V, et al. Anticardiolipin antibodies and risk of myocardial infarction in a prospective cohort of middle-aged men. Circulation. 1995;91: Wu R, Nityanand S, Berglund L, et al. Antibodies against cardiolipin and oxidatively modified LDL in 50-year-old men predict myocardial infarction. Arterioscler Thromb Vasc Biol. 1997;17: Sletnes KE, Smith P, Abdelnoor M, et al. Antiphospholipid antibodies after myocardial infarction and their relation to mortality, reinfarction, and non-haemorrhagic stroke. Lancet. 1992;339: Phadke KV, Phillips RA, Clarke DTR, et al. Anticardiolipin antibodies in ischaemic heart disease: marker or myth? Br Heart J. 1993;69: Moss AJ, Goldstein RE, Marder VJ, et al. Thrombogenic factors and recurrent coronary events. Circulation. 1999;99: Harris EN, Gharavi AE, Patel SP, et al. Evaluation of the anti-cardiolipin antibody test: report of an international workshop held 4 April Clin Exp Immunol. 1987;68: Erickson EN, Najmey SS, Keil LB, et al. Reference calibrators for IgG antibodies to 2 -glycoprotein I: preparation, properties and availability to investigators. Clin Chem. 1996;42: Levine SR, Salowich-Palm L, Sawaya KL, et al. IgG anticardiolipin antibody titer 40 GPL and the risk of subsequent thrombo-occlusive events and death. Stroke. 1997;28: Vaarala O, Puurunen M, Manttari M, et al. Antibodies to prothrombin imply a risk of myocardial infarction in middle-aged men. Thromb Haemost. 1996;75: Vaarala O. Antiphospholipid antibodies and myocardial infarction. Lupus. 1998;7(suppl 2):S132 S Vaarala O. Antiphospholipid antibodies and atherosclerosis. Lupus. 1996; 5: Shoenfeld Y, Harats D, George J. Atherosclerosis and the antiphospholipid syndrome: a link unraveled? Lupus. 1998;7(suppl 2):S140 S Hasunuma Y, Matsuura E, Katahira MT, et al. Involvement of 2 -glycoprotein I and anticardiolipin antibodies in oxidatively modified lowdensity lipoprotein uptake by macrophages. Clin Exp Immunol. 1997;107: Vaarala O, Alftan G, Jauhiainen M, et al. Crossreaction between antibodies to oxidised low-density lipoprotein and to cardiolipin in systemic lupus erythematosus. Lancet. 1993;341: Roubey RAS. Mechanisms of autoantibody-mediated thrombosis. Lupus. 1998;7(suppl 2):S114 S Martinuzzo ME, Maclouf J, Carreras LO, et al. Antiphospholipid antibodies enhance thrombin-induced platelet activation and thromboxane formation. Thromb Haemost. 1993;70: Cuadrado MJ. Thrombosis in primary antiphospholipid syndrome: a pivotal role for monocyte tissue factor expression. Arthritis Rheum. 1997; 40: Simantov R, LaSala JM, Lo SK, et al. Activation of cultured vascular endothelial cells by antiphospholipid antibodies. J Clin Invest. 1995;96: Sheng Y, Kandiah DA, Krilis SA. 2 -glycoprotein I: target antigen for antiphospholipid antibodies: immunological and molecular aspects. Lupus. 1998;7(suppl 2):S5 S DeWood MA, Spores J, Notske R, et al. Prevalence of total coronary occlusion during the early hours of transmural myocardial infarction. N Engl J Med. 1980;303: Coutinho A, Kazatchkine MD, Avrameas S. Natural autoantibodies. Curr Opin Immunol. 1995;7: Carson DA, Chen PP, Kipps TJ. New roles for rheumatoid factor. J Clin Invest. 1991;87: Vanderlugt CL, Begolka WS, Neville KL, et al. The functional significance of epitope spreading and its regulation by costimulatory molecules. Immunol Rev. 1998;164: Melnick SL, Shahar E, Folsom AR, et al. Past infection by Chlamydia pneumoniae strain TWAR and asymptomatic carotid atherosclerosis: Atherosclerosis Risk in Communities (ARIC) Study Investigators. Am J Med. 1993;95: Laurila A, Bloigu A, Nayha S, et al. Chronic Chlamydia pneumoniae infection is associated with a serum lipid profile known to be a risk factor for atherosclerosis. Arterioscler Thromb Vasc Biol. 1997;17: Hunt JE, McNeil HP, Morgan GJ, et al. A phospholipid- 2 -glycoprotein I complex is an antigen for anticardiolipin antibodies occurring in autoimmune disease but not in infection. Lupus. 1992;1: Kalayoglu MV, Hoerneman B, LaVerda D, et al. Cellular oxidation of low-density lipoprotein by Chlamydia pneumoniae. J Infect Dis. 1999; 180: Ridker PM. Evaluating novel cardiovascular risk factors: can we better predict heart attacks? Ann Intern Med. 1999;130: Roubey RA. Autoantibodies to phospholipid-binding plasma proteins: a new view of lupus anticoagulant and other antiphospholipid autoantibodies. Blood. 1994;84: Fevaloro EJ, Silvestrini R, Mohammed A. Clinical utility of anticardiolipin antibody assays: high inter-laboratory variation and limited consensus by participants of external quality assurance programs signals a cautious approach. Pathology. 1999;31: Reber G, Arvieux J, Comby E, et al. Multicenter evaluation of nine commercial kits for the quantitation of anticardiolipin antibodies. Thromb Haemost. 1995;73: