- Email: [email protected]
Plasma concentration of soluble intercellular adhesion molecule 1 and risks of future myocardial infarction in apparently healthy men
ARTICLES
Plasma concentration of soluble intercellular adhesion molecule 1 and risks of future myocardial infarction in apparently healthy men
Paul M Ridker, Charles H Hennekens, Beatrice Roitman-Johnson, Meir J Stampfer, Jean Allen
Summary Background The intercellular adhesion molecule ICAM-1 mediates adhesion and transmigration of leucocytes to the vascular endothelial wall, a step proposed to be critical in the initiation and progression of atherosclerosis. Whether concentrations of soluble ICAM-1 (sICAM-1) are raised in apparently healthy individuals who later suffer acute myocardial infarction is unknown. Methods We obtained baseline plasma samples from a prospective cohort of 14 916 healthy men enrolled in the Physicians’ Health Study. With a nested case-control design, we measured sICAM-1 concentrations for 474 participants who developed a first myocardial infarction, and 474 controls (participants who remained healthy throughout the 9-year follow-up). Cases were matched to controls according to age and smoking status at the time of myocardial infarction. Findings We found a significant association between increasing concentration of sICAM-1 and risk of future myocardial infarction (p=0·003), especially among participants with baseline sICAM-1 concentrations in the highest quartile (>260 ng/mL; relative risk 1·6 [95% CI 1·1–2·4], p=0·02). This association was present overall as well as among non-smokers, and persisted after control for lipid and non-lipid risk factors. In multivariate analyses, the risk of future myocardial infarction was 80% higher for participants with baseline sICAM-1 concentrations in the highest quartile (relative risk 1·8 [1·1–2·8], p=0·02). Similar risk estimates were seen among non-smokers. We found slight but significant correlations between sICAM-1 and fibrinogen, high-density-lipoprotein cholesterol, homocysteine, triglycerides, tissue-type plasminogenactivator antigen, and C-reactive protein, but adjustment for these altered the risk little. The risk of myocardial infarction associated with raised concentrations of sICAM-1 seemed to increase with length of follow-up. Divisions of Cardiovascular Disease (P M Ridker MD), Preventive Medicine (P M Ridker, C H Hennekens MD), and Channing Laboratory (M J Stampfer MD), Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, USA; Department of Ambulatory Care and Prevention, Harvard Medical School, Cambridge, Massachusetts (C H Hennekens); Departments of Epidemiology (C H Hennekens, M J Stampfer) and Nutrition (M J Stampfer), Harvard School of Public Health, Boston, Massachusetts; and R & D Systems Inc, Minneapolis, Minnesota (B Roitman-Johnson BS, J Allen MD) Correspondence to: Dr Paul M Ridker, Brigham and Women’s Hospital, 900 Commonwealth Ave East, Boston, MA 02115, USA (e-mail: [email protected])
88
Interpretation Our data support the hypothesis that cellular mediators of inflammation have a role in atherogenesis and provide a clinical basis to consider antiadhesion therapies as a novel means of cardiovascular disease prevention.
Lancet 1998; 351: 88–92
Introduction Adhesion of circulating leucocytes to the endothelial cell and subsequent transendothelial migration is suggested to be an important step in the initiation of atherosclerosis.1 In part, this process is mediated by cellular adhesion molecules (CAMs),2,5 expressed on the endothelial membrane, in response to several inflammatory cytokines, including interleukin-1, tumour necrosis factor, and interferon.6 Pathological studies have shown increased CAM expression in several components of the atherosclerotic plaque7–11 and data suggest a role for adhesion molecules in acute atherothrombotic syndromes.12 Although their pathological role is uncertain, soluble forms of some CAMs can be found in plasma. Clinical data about soluble CAMs are sparse, but raise the possibility that plasma concentrations may be higher among patients with atherosclerosis13,14 and 15 dyslipidaemia. However, since no prospective analytical studies of even small sample sizes have assessed baseline concentrations of soluble CAMs and the risks of future vascular occlusion, it is uncertain whether relations between inflammatory adhesion molecules and atherosclerotic disease are causal or due simply to confounding by other factors, such as cigarette smoking.16 To address these issues directly, we assessed in a large prospective cohort of apparently healthy men whether raised baseline concentrations of soluble intercellular adhesion molecule-1 (sICAM-1) are associated with future risks of myocardial infarction. We further aimed to find out whether any observed association was dependent on other lipid and non-lipid cardiovascular risk factors, particularly the association between sICAM-1 and C-reactive protein, a liver-derived marker of systemic inflammation previously shown to be an independent risk factor for myocardial infarction.17
Methods We used prospective data from the Physicians’ Health Study,18 a randomised, double-blind, placebo-controlled trial of aspirin and -carotene in the primary prevention of cardiovascular disease and cancer among 22 071 healthy US male physicians. All study participants were free from previous myocardial infarction,
THE LANCET • Vol 351 • January 10, 1998
ARTICLES
Mean (SD) age (years)*
Cases (n=474)
Controls (n=474)
58·1 (8·5)
57·9 (8·5)
p ··
Smoking status (%)* Never Past Current
·· 44·4 40·8 14·8
44·4 40·8 14·8
Mean (SD) body-mass index (kg/m2)
25·4 (3·4)
25·0 (2·9)
0·02
History (%) Diabetes History of high cholesterol History of hypertension Parental history of coronary artery disease
5·1 14·2 28·6 18·0
2·3 10·0 17·6 11·7
0·03 0·06 0·001 0·01
*Matching factor.
Table 1: Baseline characteristics of study participants stroke, transient ischaemic attack, and cancer at study entry. Before randomisation, between August, 1982, and December, 1984, we asked eligible participants to provide baseline blood samples, which were collected in edetic acid and stored at ⫺80°C.17 14 916 (67·6%) provided baseline plasma samples for analysis and were randomised. For all cases of myocardial infarction reported during study follow-up, hospital records (and death certificates and necropsy reports for fatal events) were reviewed by a committee of physicians unaware of treatment assignment. Reported myocardial infarction was confirmed if symptoms met WHO criteria and the event was associated with raised plasma concentration of cardiac enzymes or characteristic electrocardiogram changes. Silent myocardial infarctions were not included since they could not be dated. Deaths due to myocardial infarction were confirmed by necropsy, symptoms, circumstances of death, and a history of coronary-artery disease. Each participant who provided an adequate baseline plasma sample for analysis and had a confirmed myocardial infarction during follow-up for 9 years was matched with one control. Controls were participants who also provided adequate baseline plasma samples and remained free of vascular disease at the time of myocardial infarction in the matched case. Controls were randomly selected from participants who met the matching criteria of age (within 1 year), smoking status (current, past, never), and length of follow-up. Overall, 474 case-control pairs were included in this analysis. Stored plasma for each patient and each control was thawed and assayed for sICAM-1 with a commercially available ELISA (R&D systems, Minneapolis, MN, USA). Laboratory personnel were unaware of case or control status. In pilot data for this study, the mean coefficient of variation for sICAM-1 in these stored frozen plasma samples was 6·3%. As well as providing blood samples, participants also reported at the baseline cardiovascular risk factors of age, smoking status, height, weight, systolic and diastolic blood pressure, history of hypercholesterolaemia, parental history of myocardial infarction,
diabetes, and exercise frequency. Plasma concentration of total and high-density-lipoprotein cholesterol, lipoprotein(a), tissuetype plasminogen-activator antigen, D-dimer, fibrinogen, total plasma homocysteine, and C-reactive protein were assayed.17,19–22 We used the Student’s t test to investigate the significance of any differences in mean baseline cardiovascular risk factors, and the 2 statistic for differences in proportions. We used logistic regression analyses, conditioned on the matching variables, to estimate relative risks and 95% CI and to control for randomised treatment assignment, as well as potential confounding variables. We tested for trend to assess any relation of increasing concentration of sICAM-1 with risk of future myocardial infarction after separation of the concentrations into quartiles, defined by the distribution of the control values. Because preliminary analyses suggested a non-linear association between sICAM-1 and risk, we also calculated estimates for patients with concentrations above and below the 75th percentile of the control distribution of sICAM-1 (260 ng/mL). Analyses were planned a priori for all participants and for the subgroup of nonsmokers at baseline, since smoking increases inflammation and sICAM-1 concentrations. We calculated age-adjusted correlation coefficients between baseline concentrations of sICAM-1 and concentrations of total and high-density-lipoprotein cholesterol, triglycerides, C-reactive protein, fibrinogen, homocysteine, tissue-type plasminogen-activator antigen; lipoprotein(a), and Ddimer. Additional logistic regression analyses were used to find out whether any observed association between sICAM-1 and risk of myocardial infarction was due to confounding by any of these factors. All p values were two-tailed.
Results Baseline clinical characteristics of cases and controls are shown in table 1. As expected, cases had higher rates of cardiovascular risk factors at baseline. sICAM-1 concentrations were normally distributed in the two groups (range 59–675 ng/mL). Smokers had significantly higher mean concentrations of sICAM-1 than did non-smokers (283·9 vs 229·0 ng/mL; p<0·001). In crude matched-pair analyses, the risk of future myocardial infarction was 60% higher for participants with baseline sICAM-1 concentrations in the highest quartile than for those with concentrations in the lowest quartile (relative risk 1·6 [95% CI 1·1–2·4], p=0·02) and the test for trend for increasing sICAM-1 concentrations was significant (p=0·006). These relations were similar in crude analysis limited to non-smokers and analyses for all participants and non-smokers only after adjustment for other cardiovascular risk factors (table 2). In crude and adjusted analyses, the increase in risk was almost entirely due to an excess of cases with plasma concentrations of sICAM-1 exceeding the 75th percentile (>260 ng/mL).
Quartile of sICAM-1 (range, ng/mL)
p trend
1 (<193)
2 (193–225)
3 (226–260)
4 (>260)
Crude analysis all participants* Relative risk (95% CI) p
1·0 ··
0·8 (0·5–1·2) 0·3
1·1 (0·8–1·7) 0·6
1·6 (1·1–2·4) 0·02
0·006
Crude analysis non-smokers* Relative risk (95% CI) p
1·0 ··
0·8 (0·5–1·3) 0·4
1·1 (0·8–1·7) 0·5
1·6 (1·0–2·5) 0·04
0·003
Adjusted analysis all participants† Relative risk (95% CI) p
1·0 ··
0·9 (0·6–1·4) 0·6
1·3 (0·8–2·1) 0·2
1·8 (1·1–2·8) 0·02
0·007
Adjusted analysis non-smokers† Relative risk (95% CI) p
1·0 ··
0·9 (0·5–1·5) 0·6
1·4 (0·8–2·2) 0·2
1·8 (1·1–3·0) 0··03
0·01
*Matched on age and smoking status, controlled for random treatment assignment. †Matched on age and smoking status, controlled for random treatment assignment, body-mass index, history of hyperlipidaemia, history of hypertension, and family history of coronary-artery disease at ages younger than 60 years.
Table 2: Crude and adjusted relative risks of first myocardial infarction according to baseline concentration of sICAM-1
THE LANCET • Vol 351 • January 10, 1998
89
ARTICLES
3
2·0 p<0·001
Relative risk
C-reactive protein (mg/L)
1·5 1·0 0·5
2
1
0 –0·5 0 0–1
–1·0 –1·5 0
100
200 300 sICAM-1 (ng/mL)
400
500
Figure 1: Relation of baseline concentration of sICAM-1 and log-normalised C-reactive protein Dotted lines represent 95% CI.
We found slight but significant age-adjusted correlations between baseline concentrations of sICAM-1 and fibrinogen (p=0·004), homocysteine (p=0·003), high-density-lipoprotein cholesterol (p=0·001), triglyceride concentration (p=0·02), and concentration of tissue-type plasminogen-activator antigen (p=0·03); the strongest correlation was for log-normalised concentrations of C-reactive protein (p<0·001, figure 1). We found no correlation between sICAM-1 and total cholesterol (p=0·5), D-dimer (p=0·9), or log-normalised concentration of lipoprotein(a) (p=0·9). The relative risks of future myocardial infarction associated with raised concentrations of sICAM-1 changed little after adjustment for any of these variables (table 3); therefore, our results were unlikely to be due simply to confounding. In analysis of lipid concentrations we found that among men with total cholesterol concentrations of less than the median value of 220 mg/dL, the relative risk of future myocardial infarction associated with raised concentrations of sICAM-1 was 1·6 (1·0–2·5; p=0·04), whereas the risk among men with total cholesterol concentrations of 220 mg/dL or more was 1·5 (1·0–2·2; p=0·03). Effects were similar for high or low values of the ratio of total cholesterol to high-density-lipoprotein cholesterol. Finally, to assess whether the effect of sICAM-1 on risks of myocardial infarction varied over time, we stratified the analysis by years of follow-up (figure 2). The Adjusted variable
Relative risk (95% CI)
p
Total cholesterol HDL cholesterol Triglycerides C-reactive protein Fibrinogen Homocysteine tPA antigen Lipoprotein(a) D-dimer
1·7 (1·2–2·3) 1·5 (1·1–2·1) 1·6 (1·1–2·4) 1·6 (1·1–2·4) 1·4 (0·9–2·3) 1·6 (1·1–2·2) 1·5 (1·0–2·3) 1·7 (1·2–2·3) 1·6 (1·1–2·2)
0·002 0·01 0·01 0·02 0·06 0·005 0·05 0·003 0·01
HDL=high-density lipoprotein; tPA=tissue-type plasminogen-activator.
Table 3: Relative risks of first myocardial infarction associated with sICAM-1 concentrations of more than 260 ng/mL after control for other lipid and non-lipid risk factors
90
1–2 2–4 Years of study follow-up
4–8
Figure 2: Relative risks (95% CI) for first myocardial infarction associated with baseline concentrations of sICAM-1 of more than 260 ng/dL according to year of study follow-up
relative risk of future myocardial infarction associated with sICAM-1 concentrations of more than 260 ng/mL seemed to increase with increasing length of follow-up. Specifically, among events that occurred in the first year of follow-up, no increase in risk was seen (1·1 [0·5–2·3], p=0·8), whereas a non-significant increase was seen for events occurring in the second year of follow-up (1·6 [0·9–3·0], p=0·1). By contrast, among events occurring after 2 years, we saw a significant two-fold increase in risk (2·0 [1·4–2·7], p<0·001). There was no evidence of interaction between aspirin, sICAM-1 concentration, and risk of future myocardial infarction.
Discussion These data show that baseline sICAM-1 concentrations are raised among apparently healthy men at risk of future myocardial infarction. Moreover, in these data, the risks of first myocardial infarction associated with raised sICAM-1 concentrations were not modified by smoking and were altered little by control for lipid and non-lipid cardiovascular risk factors. Since cellular adhesion molecules like ICAM-1 mediate a series of cell-cell interactions, including the adhesion and transmigration of leucocytes on the vascular endothelial wall,2–5 our data support the hypothesis that endothelial activation and inflammation are important steps in the initiation and progression of atherosclerosis.1 The strongest plasma correlate of sICAM-1 in these data was C-reactive protein, a sensitive marker of systemic inflammation previously shown in this cohort to be an independent risk factor for myocardial infarction.17 An association between sICAM-1 and C-reactive protein is plausible because several immune cytokines are involved in the production of both sICAM-1 and C-reactive protein.6,23 However, unlike C-reactive protein, which is hepatically derived and has an uncertain biological function, CAMs such as ICAM-1 have a direct and potentially critical role in the early immune response. Our finding that sICAM-1 concentrations are raised among apparently healthy men at risk of future myocardial infarction extends previous observations about inflammation and vascular risk because our data implicate a direct mediator of the inflammatory process,
THE LANCET • Vol 351 • January 10, 1998
ARTICLES
not simply a systemic marker. Our finding of an association between sICAM-1 concentration and C-reactive protein is also of interest in light of a report that C-reactive protein may itself have a role in the shedding process of another adhesion molecule, L-selectin.24 We also found a significant correlation between sICAM-1 concentration and plasma fibrinogen concentrations; studies suggest that fibrinogen mediates leucocyte adhesion to the vascular endothelium through an ICAM-1-dependent pathway.25,26 Furthermore, interactions between fibrinogen and sICAM-1 that may promote cellular proliferation have been suggested.27 By contrast, we found no correlation between sICAM-1 and total cholesterol, and there was no evidence that lipid concentrations modified the effect of sICAM-1 on risk of future myocardial infarction. Therefore, the risk associated with raised concentrations of sICAM-1 was not limited to patients with hyperlipidaemia.15 The risks of future myocardial infarction associated with raised concentrations of sICAM-1 seemed to increase with length of follow-up. Although this difference may have been due to chance, such a timedependent effect may implicate a role for sICAM-1 early in the atherosclerotic process, at least compared with Creactive protein, for which risks remain stable over time. Also, by contrast with our observations on C-reactive protein,17 there was no evidence of interaction in this analysis between random aspirin assignment and sICAM1. However, because of the apparent time-dependent effects of sICAM-1 on vascular risk, the power to detect such an interaction may have been limited, since aspirin use in the Physicians’ Health Study occurred only in the first 5 years of follow-up, during which time the predictive value of sICAM-1 was slight. Potential limitations of these data merit consideration. First, since they are based on assessment of sICAM-1 concentration in plasma samples stored at ⫺80°C for up to 12 years, we cannot exclude the possibility of protein degradation. However, this possibility seems unlikely because concentrations of sICAM-1 in our participants are similar to those reported in studies of fresh plasma. Second, because baseline plasma samples were obtained on only one occasion, we could not take into account variation in concentrations that may have occurred over time. The impact of this potential limitation would be to increase random misclassification, an effect that, if present, could lead to an underestimation of the magnitude of any relation. Finally, smokers had significantly higher concentrations of sICAM-1 than nonsmokers, an observation that raises the possibility of confounding.16 However, the association we found between sICAM-1 and future risk of myocardial infarction was present among non-smokers and, therefore, our results could not be due to confounding by smoking. The source of sICAM-1 found in our population is uncertain, although circulating forms of adhesion molecules may be derived from vascular-wall components, including endothelial and smooth-muscle cells.28,29 Previous studies have found sICAM-1 in plasma of patients with chronic inflammatory disorders,4 and increased expression of adhesion molecules has been reported in patients with unstable coronary syndromes.12 However, participants in our study were apparently healthy when blood samples were obtained, which would
THE LANCET • Vol 351 • January 10, 1998
keep to a minimum the likelihood that increased concentrations of sICAM-1 in this population were due to acute ischaemia or a clinically detectable chronic inflammatory condition. Our observation of increased concentrations of sICAM-1 several years before the occurrence of coronary occlusion supports the hypothesis that endothelial activation and inflammation occur early in the atherothrombotic process. In addition, because patients with the higher concentrations of sICAM-1 were at increased risk of future myocardial infarction, these data support the possibility that therapies aimed at antiadhesion may provide a novel approach to prevention and treatment of cardiovascular diseases.2,3,33 Experimental models of anti-ICAM antibodies involving vascular balloon injury, myocardial ischaemia, and reperfusion have had promising results.30–32 Finally, these data raise the possibility that cellular adhesion molecules can serve as a molecular marker for early atherosclerosis,2,3,33 a hypothesis that requires further testing. Contributors Paul Ridker was the lead investigator, conceived the hypothesis, managed the study, did the primary data analyses, and wrote the first draft of the paper. Charles Hennekens and Meir Stampfer made critical contributions to the design and analysis of these data and were involved in the writing and epidemiological assessment. Beatrice Roitman-Johnson and Jean Allen did all the biochemical analyses, including critical preliminary studies to address tissues of variance, stability, and reproducibility. All investigators contributed to the final writing of the manuscript.
Acknowledgments This study was funded by grants from the National Heart Lung and Blood Institute, Bethesda, Maryland, USA and from the American Heart Association, Dallas, Texas.
References 1 2 3
4
5 6
7
8
9
10
11
12 13
Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 1993; 363: 801–09. Springer TA. Adhesion molecules of the immune system. Nature 10090; 346: 425–34. Cybulsky MI, Gimbrone MA Jr. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science 1991; 25: 788–91. Bevilacqua MP, Nelson RM, Mannori G, Cecconi O. Endothelialleukocyte adhesion molecules in human disease. Annu Rev Med 1994; 45: 361–78. Adams DH, Shaw S. Leucocyte-endothelial interactions and regulation of leucocyte migration. Lancet 1994; 343: 831–36. Pober JS, Gimbrone MA Jr, Lapiette LA, et al. Overlapping patterns of activation of human endothelial cells by interleukin 1, tumour necrosis factor, and immune interferon. J Immunol 1986; 137: 1893–96. Poston RN, Haskard DO, Coucher JR, Gall NP, Johnson-Tidey RR. Expression of intercellular adhesion molecule-1 in atherosclerotic plaques. Am J Pathol 1992; 140: 665–73. Printseva OY, Peclo MM, Gown AM. Various cell types in human atherosclerotic lesions express ICAM-1: further immunocytochemical and immunochemical studies employing monoclonal antibody 10F3. Am J Pathol 1992; 140: 889–96. van der Wal AC, Das PK, Tigges AJ, Becker AE. Adhesion molecules on the endothelium and mononuclear cells in human atherosclerotic lesions. Am J Pathol 1992; 141: 1427–33. Davies MJ, Gordon JL, Gearing AJH, et al. The expression of the adhesion molecules ICAM-1, VACM-1, PECAM, and E-selectin in human atherosclerosis. J Pathol 1993; 171: 223–29. O’Brien KD, McDonald TO, Chait A, Allen MD, Alpers CE. Neovascular expression of E-selectin, intercellular adhesion molecule1, and vascular cell adhesion molecule-1 in human atherosclerosis and their relation to intimal leukocyte content. Circulation 1996; 93: 672–82. Jang Y, Lincoff AM, Plow EF, Topol EJ. Cell adhesion molecules in coronary artery disease. J Am Coll Cardiol 1994; 24: 1591–601. Peter K, Naworth P, Conradt C, et al. Circulating vascular cell adhesion molecule-1 correlates with the extent of human atherosclerosis in contrast to circulating intercellular adhesion molecule-1, E-selectin, P-selectin, and thrombomodulin. Arterioscler Thromb Vasc Biol 1997; 17: 505–12.
91
ARTICLES 14 Blann AD, McCollum CN. Circulating endothelial cell/leukocyte adhesion molecules in atherosclerosis. Thromb Haemost 1994; 74: 151–54. 15 Hackman A, Abe Y, Insull W, et al. Levels of soluble cell adhesion molecules in patients with dyslipidemia. Circulation 1996; 93: 1334–38. 16 Blann AD, Steele C, McCollum CN. The influence of smoking on soluble adhesion molecules and endothelial cell markers. Thromb 1997; 85: 433–38. 17 Ridker PM, Cushman M, Stampfer MJ, Tracey RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997; 336: 973–79. 18 Steering Committee of the Physicians’ Health Study Research Group. Final report of the aspirin component of the ongoing Physicians; Health Study. N Engl J Med 1989; 321: 129–35. 19 Ridker PM, Hennekens CH, Stampfer MJ. A prospective study of lipoprotein(a) and the risk of myocardial infarction. JAMA 1993; 270: 2195–99. 20 Ridker PM, Vaughan DE, Stampfer MJ, Manson JE, Hennekens CH. Endogenous tissue-type plasminogen activator and risk of myocardial infarction. Lancet 1993; 341: 1165–68. 21 Ridker PM, Hennekens CH, Cerskus A, Stampfer MJ. Plasma concentration of cross linked fibrin degradation product (D-dimer) and the risk of future myocardial infarction among apparently health men. Circulation 1994; 90: 2236–40. 22 Ridker PM, Hennekens CH, Selhub J, Miletich JP, Malinow MR, Stampfer MJ. Interrelation of hyperhomocyst(e)inemia, factor V Leiden, and risks of future venous thromboembolism. Circulation 1997; 95: 1777–82. 23 Dustin ML, Rothlein R, Bhan AR, Dinarello CA, Springer TA. Induction by IL-1 and interferon, tissue distribution, biochemistry and function of a natural adherence molecule (ICAM-1). J Immunol 1986; 137: 245–54.
92
24 Zouki C, Beauchamp M, Baron C, Filep JG. Prevention of in vitro neutrophil adhesion to endothelial cells through shedding of L-selectin by C-reactive protein and peptides derived from C-reactive protein. J Clin Invest 1997; 100: 522–29. 25 Languino LR, Plescia J, Duperray A, et al. Fibrinogen mediates leukocyte adhesion to vascular endothelium through an ICAM-1 dependent pathway. Cell 1993; 73: 1423–34. 26 Duperray A, Languino LR, Plescia, et al. Molecular identification of a novel fibrinogen binding site on the first domain of ICAM-1 regulating leukocyte-endothelium bridging. J Biol Chem 1997; 272: 435–41. 27 Gardiner EE, D’Souza SE. A mitogenic action for fibrinogen mediated through intercellular adhesion molecule-1. J Biol Chem 1997; 272: 15474–80. 28 Leewenberg JFM, Smeets EF, Neefjes JJ, et al. E-selectin and intercellular adhesion molecule-1 are released by activated human endothelial cells in vitro. Immunology 1992; 77: 543–49. 29 Pigott R, Dillon LP, Hemingway IH, Gearing AJH. Soluble forms of E-selectin, ICAM-1 and VCAM-1 are present in the supernatants of cytokine activated cultured endothelial cells. Biochem Biophys Res Commun 1992; 187: 584–89. 30 Ma X, Lefer DJ, Lefer AM, Rothlin R. Coronary endothelial and cardiac protective effects of a monoclonal antibody to intercellular adhesion molecule-1 in myocardial ischemia and reperfusion. Circulation 1992; 86: 937–46. 31 Yasukawa H, Imaizumi T, Matsuoka H, Nakashima A, Moritmatsu M. Inhibition of intimal hyperplasia after balloon injury by antibodies to intercellular adhesion molecule-1 and lymphocyte functionassociated antigen-1. Circulation 1997; 95: 1515–22. 32 Plow EF, D’Souza SE. A role for intercellular adhesion molecule-1 in restenosis. Circulation 1997; 95: 1355–56. 33 Libby P. Molecular bases of the acute coronary syndromes. Circulation 1995; 91: 2844–50.
THE LANCET • Vol 351 • January 10, 1998
Plasma concentration of soluble intercellular adhesion molecule 1 and risks of future myocardial infarction in apparently healthy men
Paul M Ridker, Charles H Hennekens, Beatrice Roitman-Johnson, Meir J Stampfer, Jean Allen
Summary Background The intercellular adhesion molecule ICAM-1 mediates adhesion and transmigration of leucocytes to the vascular endothelial wall, a step proposed to be critical in the initiation and progression of atherosclerosis. Whether concentrations of soluble ICAM-1 (sICAM-1) are raised in apparently healthy individuals who later suffer acute myocardial infarction is unknown. Methods We obtained baseline plasma samples from a prospective cohort of 14 916 healthy men enrolled in the Physicians’ Health Study. With a nested case-control design, we measured sICAM-1 concentrations for 474 participants who developed a first myocardial infarction, and 474 controls (participants who remained healthy throughout the 9-year follow-up). Cases were matched to controls according to age and smoking status at the time of myocardial infarction. Findings We found a significant association between increasing concentration of sICAM-1 and risk of future myocardial infarction (p=0·003), especially among participants with baseline sICAM-1 concentrations in the highest quartile (>260 ng/mL; relative risk 1·6 [95% CI 1·1–2·4], p=0·02). This association was present overall as well as among non-smokers, and persisted after control for lipid and non-lipid risk factors. In multivariate analyses, the risk of future myocardial infarction was 80% higher for participants with baseline sICAM-1 concentrations in the highest quartile (relative risk 1·8 [1·1–2·8], p=0·02). Similar risk estimates were seen among non-smokers. We found slight but significant correlations between sICAM-1 and fibrinogen, high-density-lipoprotein cholesterol, homocysteine, triglycerides, tissue-type plasminogenactivator antigen, and C-reactive protein, but adjustment for these altered the risk little. The risk of myocardial infarction associated with raised concentrations of sICAM-1 seemed to increase with length of follow-up. Divisions of Cardiovascular Disease (P M Ridker MD), Preventive Medicine (P M Ridker, C H Hennekens MD), and Channing Laboratory (M J Stampfer MD), Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, USA; Department of Ambulatory Care and Prevention, Harvard Medical School, Cambridge, Massachusetts (C H Hennekens); Departments of Epidemiology (C H Hennekens, M J Stampfer) and Nutrition (M J Stampfer), Harvard School of Public Health, Boston, Massachusetts; and R & D Systems Inc, Minneapolis, Minnesota (B Roitman-Johnson BS, J Allen MD) Correspondence to: Dr Paul M Ridker, Brigham and Women’s Hospital, 900 Commonwealth Ave East, Boston, MA 02115, USA (e-mail: [email protected])
88
Interpretation Our data support the hypothesis that cellular mediators of inflammation have a role in atherogenesis and provide a clinical basis to consider antiadhesion therapies as a novel means of cardiovascular disease prevention.
Lancet 1998; 351: 88–92
Introduction Adhesion of circulating leucocytes to the endothelial cell and subsequent transendothelial migration is suggested to be an important step in the initiation of atherosclerosis.1 In part, this process is mediated by cellular adhesion molecules (CAMs),2,5 expressed on the endothelial membrane, in response to several inflammatory cytokines, including interleukin-1, tumour necrosis factor, and interferon.6 Pathological studies have shown increased CAM expression in several components of the atherosclerotic plaque7–11 and data suggest a role for adhesion molecules in acute atherothrombotic syndromes.12 Although their pathological role is uncertain, soluble forms of some CAMs can be found in plasma. Clinical data about soluble CAMs are sparse, but raise the possibility that plasma concentrations may be higher among patients with atherosclerosis13,14 and 15 dyslipidaemia. However, since no prospective analytical studies of even small sample sizes have assessed baseline concentrations of soluble CAMs and the risks of future vascular occlusion, it is uncertain whether relations between inflammatory adhesion molecules and atherosclerotic disease are causal or due simply to confounding by other factors, such as cigarette smoking.16 To address these issues directly, we assessed in a large prospective cohort of apparently healthy men whether raised baseline concentrations of soluble intercellular adhesion molecule-1 (sICAM-1) are associated with future risks of myocardial infarction. We further aimed to find out whether any observed association was dependent on other lipid and non-lipid cardiovascular risk factors, particularly the association between sICAM-1 and C-reactive protein, a liver-derived marker of systemic inflammation previously shown to be an independent risk factor for myocardial infarction.17
Methods We used prospective data from the Physicians’ Health Study,18 a randomised, double-blind, placebo-controlled trial of aspirin and -carotene in the primary prevention of cardiovascular disease and cancer among 22 071 healthy US male physicians. All study participants were free from previous myocardial infarction,
THE LANCET • Vol 351 • January 10, 1998
ARTICLES
Mean (SD) age (years)*
Cases (n=474)
Controls (n=474)
58·1 (8·5)
57·9 (8·5)
p ··
Smoking status (%)* Never Past Current
·· 44·4 40·8 14·8
44·4 40·8 14·8
Mean (SD) body-mass index (kg/m2)
25·4 (3·4)
25·0 (2·9)
0·02
History (%) Diabetes History of high cholesterol History of hypertension Parental history of coronary artery disease
5·1 14·2 28·6 18·0
2·3 10·0 17·6 11·7
0·03 0·06 0·001 0·01
*Matching factor.
Table 1: Baseline characteristics of study participants stroke, transient ischaemic attack, and cancer at study entry. Before randomisation, between August, 1982, and December, 1984, we asked eligible participants to provide baseline blood samples, which were collected in edetic acid and stored at ⫺80°C.17 14 916 (67·6%) provided baseline plasma samples for analysis and were randomised. For all cases of myocardial infarction reported during study follow-up, hospital records (and death certificates and necropsy reports for fatal events) were reviewed by a committee of physicians unaware of treatment assignment. Reported myocardial infarction was confirmed if symptoms met WHO criteria and the event was associated with raised plasma concentration of cardiac enzymes or characteristic electrocardiogram changes. Silent myocardial infarctions were not included since they could not be dated. Deaths due to myocardial infarction were confirmed by necropsy, symptoms, circumstances of death, and a history of coronary-artery disease. Each participant who provided an adequate baseline plasma sample for analysis and had a confirmed myocardial infarction during follow-up for 9 years was matched with one control. Controls were participants who also provided adequate baseline plasma samples and remained free of vascular disease at the time of myocardial infarction in the matched case. Controls were randomly selected from participants who met the matching criteria of age (within 1 year), smoking status (current, past, never), and length of follow-up. Overall, 474 case-control pairs were included in this analysis. Stored plasma for each patient and each control was thawed and assayed for sICAM-1 with a commercially available ELISA (R&D systems, Minneapolis, MN, USA). Laboratory personnel were unaware of case or control status. In pilot data for this study, the mean coefficient of variation for sICAM-1 in these stored frozen plasma samples was 6·3%. As well as providing blood samples, participants also reported at the baseline cardiovascular risk factors of age, smoking status, height, weight, systolic and diastolic blood pressure, history of hypercholesterolaemia, parental history of myocardial infarction,
diabetes, and exercise frequency. Plasma concentration of total and high-density-lipoprotein cholesterol, lipoprotein(a), tissuetype plasminogen-activator antigen, D-dimer, fibrinogen, total plasma homocysteine, and C-reactive protein were assayed.17,19–22 We used the Student’s t test to investigate the significance of any differences in mean baseline cardiovascular risk factors, and the 2 statistic for differences in proportions. We used logistic regression analyses, conditioned on the matching variables, to estimate relative risks and 95% CI and to control for randomised treatment assignment, as well as potential confounding variables. We tested for trend to assess any relation of increasing concentration of sICAM-1 with risk of future myocardial infarction after separation of the concentrations into quartiles, defined by the distribution of the control values. Because preliminary analyses suggested a non-linear association between sICAM-1 and risk, we also calculated estimates for patients with concentrations above and below the 75th percentile of the control distribution of sICAM-1 (260 ng/mL). Analyses were planned a priori for all participants and for the subgroup of nonsmokers at baseline, since smoking increases inflammation and sICAM-1 concentrations. We calculated age-adjusted correlation coefficients between baseline concentrations of sICAM-1 and concentrations of total and high-density-lipoprotein cholesterol, triglycerides, C-reactive protein, fibrinogen, homocysteine, tissue-type plasminogen-activator antigen; lipoprotein(a), and Ddimer. Additional logistic regression analyses were used to find out whether any observed association between sICAM-1 and risk of myocardial infarction was due to confounding by any of these factors. All p values were two-tailed.
Results Baseline clinical characteristics of cases and controls are shown in table 1. As expected, cases had higher rates of cardiovascular risk factors at baseline. sICAM-1 concentrations were normally distributed in the two groups (range 59–675 ng/mL). Smokers had significantly higher mean concentrations of sICAM-1 than did non-smokers (283·9 vs 229·0 ng/mL; p<0·001). In crude matched-pair analyses, the risk of future myocardial infarction was 60% higher for participants with baseline sICAM-1 concentrations in the highest quartile than for those with concentrations in the lowest quartile (relative risk 1·6 [95% CI 1·1–2·4], p=0·02) and the test for trend for increasing sICAM-1 concentrations was significant (p=0·006). These relations were similar in crude analysis limited to non-smokers and analyses for all participants and non-smokers only after adjustment for other cardiovascular risk factors (table 2). In crude and adjusted analyses, the increase in risk was almost entirely due to an excess of cases with plasma concentrations of sICAM-1 exceeding the 75th percentile (>260 ng/mL).
Quartile of sICAM-1 (range, ng/mL)
p trend
1 (<193)
2 (193–225)
3 (226–260)
4 (>260)
Crude analysis all participants* Relative risk (95% CI) p
1·0 ··
0·8 (0·5–1·2) 0·3
1·1 (0·8–1·7) 0·6
1·6 (1·1–2·4) 0·02
0·006
Crude analysis non-smokers* Relative risk (95% CI) p
1·0 ··
0·8 (0·5–1·3) 0·4
1·1 (0·8–1·7) 0·5
1·6 (1·0–2·5) 0·04
0·003
Adjusted analysis all participants† Relative risk (95% CI) p
1·0 ··
0·9 (0·6–1·4) 0·6
1·3 (0·8–2·1) 0·2
1·8 (1·1–2·8) 0·02
0·007
Adjusted analysis non-smokers† Relative risk (95% CI) p
1·0 ··
0·9 (0·5–1·5) 0·6
1·4 (0·8–2·2) 0·2
1·8 (1·1–3·0) 0··03
0·01
*Matched on age and smoking status, controlled for random treatment assignment. †Matched on age and smoking status, controlled for random treatment assignment, body-mass index, history of hyperlipidaemia, history of hypertension, and family history of coronary-artery disease at ages younger than 60 years.
Table 2: Crude and adjusted relative risks of first myocardial infarction according to baseline concentration of sICAM-1
THE LANCET • Vol 351 • January 10, 1998
89
ARTICLES
3
2·0 p<0·001
Relative risk
C-reactive protein (mg/L)
1·5 1·0 0·5
2
1
0 –0·5 0 0–1
–1·0 –1·5 0
100
200 300 sICAM-1 (ng/mL)
400
500
Figure 1: Relation of baseline concentration of sICAM-1 and log-normalised C-reactive protein Dotted lines represent 95% CI.
We found slight but significant age-adjusted correlations between baseline concentrations of sICAM-1 and fibrinogen (p=0·004), homocysteine (p=0·003), high-density-lipoprotein cholesterol (p=0·001), triglyceride concentration (p=0·02), and concentration of tissue-type plasminogen-activator antigen (p=0·03); the strongest correlation was for log-normalised concentrations of C-reactive protein (p<0·001, figure 1). We found no correlation between sICAM-1 and total cholesterol (p=0·5), D-dimer (p=0·9), or log-normalised concentration of lipoprotein(a) (p=0·9). The relative risks of future myocardial infarction associated with raised concentrations of sICAM-1 changed little after adjustment for any of these variables (table 3); therefore, our results were unlikely to be due simply to confounding. In analysis of lipid concentrations we found that among men with total cholesterol concentrations of less than the median value of 220 mg/dL, the relative risk of future myocardial infarction associated with raised concentrations of sICAM-1 was 1·6 (1·0–2·5; p=0·04), whereas the risk among men with total cholesterol concentrations of 220 mg/dL or more was 1·5 (1·0–2·2; p=0·03). Effects were similar for high or low values of the ratio of total cholesterol to high-density-lipoprotein cholesterol. Finally, to assess whether the effect of sICAM-1 on risks of myocardial infarction varied over time, we stratified the analysis by years of follow-up (figure 2). The Adjusted variable
Relative risk (95% CI)
p
Total cholesterol HDL cholesterol Triglycerides C-reactive protein Fibrinogen Homocysteine tPA antigen Lipoprotein(a) D-dimer
1·7 (1·2–2·3) 1·5 (1·1–2·1) 1·6 (1·1–2·4) 1·6 (1·1–2·4) 1·4 (0·9–2·3) 1·6 (1·1–2·2) 1·5 (1·0–2·3) 1·7 (1·2–2·3) 1·6 (1·1–2·2)
0·002 0·01 0·01 0·02 0·06 0·005 0·05 0·003 0·01
HDL=high-density lipoprotein; tPA=tissue-type plasminogen-activator.
Table 3: Relative risks of first myocardial infarction associated with sICAM-1 concentrations of more than 260 ng/mL after control for other lipid and non-lipid risk factors
90
1–2 2–4 Years of study follow-up
4–8
Figure 2: Relative risks (95% CI) for first myocardial infarction associated with baseline concentrations of sICAM-1 of more than 260 ng/dL according to year of study follow-up
relative risk of future myocardial infarction associated with sICAM-1 concentrations of more than 260 ng/mL seemed to increase with increasing length of follow-up. Specifically, among events that occurred in the first year of follow-up, no increase in risk was seen (1·1 [0·5–2·3], p=0·8), whereas a non-significant increase was seen for events occurring in the second year of follow-up (1·6 [0·9–3·0], p=0·1). By contrast, among events occurring after 2 years, we saw a significant two-fold increase in risk (2·0 [1·4–2·7], p<0·001). There was no evidence of interaction between aspirin, sICAM-1 concentration, and risk of future myocardial infarction.
Discussion These data show that baseline sICAM-1 concentrations are raised among apparently healthy men at risk of future myocardial infarction. Moreover, in these data, the risks of first myocardial infarction associated with raised sICAM-1 concentrations were not modified by smoking and were altered little by control for lipid and non-lipid cardiovascular risk factors. Since cellular adhesion molecules like ICAM-1 mediate a series of cell-cell interactions, including the adhesion and transmigration of leucocytes on the vascular endothelial wall,2–5 our data support the hypothesis that endothelial activation and inflammation are important steps in the initiation and progression of atherosclerosis.1 The strongest plasma correlate of sICAM-1 in these data was C-reactive protein, a sensitive marker of systemic inflammation previously shown in this cohort to be an independent risk factor for myocardial infarction.17 An association between sICAM-1 and C-reactive protein is plausible because several immune cytokines are involved in the production of both sICAM-1 and C-reactive protein.6,23 However, unlike C-reactive protein, which is hepatically derived and has an uncertain biological function, CAMs such as ICAM-1 have a direct and potentially critical role in the early immune response. Our finding that sICAM-1 concentrations are raised among apparently healthy men at risk of future myocardial infarction extends previous observations about inflammation and vascular risk because our data implicate a direct mediator of the inflammatory process,
THE LANCET • Vol 351 • January 10, 1998
ARTICLES
not simply a systemic marker. Our finding of an association between sICAM-1 concentration and C-reactive protein is also of interest in light of a report that C-reactive protein may itself have a role in the shedding process of another adhesion molecule, L-selectin.24 We also found a significant correlation between sICAM-1 concentration and plasma fibrinogen concentrations; studies suggest that fibrinogen mediates leucocyte adhesion to the vascular endothelium through an ICAM-1-dependent pathway.25,26 Furthermore, interactions between fibrinogen and sICAM-1 that may promote cellular proliferation have been suggested.27 By contrast, we found no correlation between sICAM-1 and total cholesterol, and there was no evidence that lipid concentrations modified the effect of sICAM-1 on risk of future myocardial infarction. Therefore, the risk associated with raised concentrations of sICAM-1 was not limited to patients with hyperlipidaemia.15 The risks of future myocardial infarction associated with raised concentrations of sICAM-1 seemed to increase with length of follow-up. Although this difference may have been due to chance, such a timedependent effect may implicate a role for sICAM-1 early in the atherosclerotic process, at least compared with Creactive protein, for which risks remain stable over time. Also, by contrast with our observations on C-reactive protein,17 there was no evidence of interaction in this analysis between random aspirin assignment and sICAM1. However, because of the apparent time-dependent effects of sICAM-1 on vascular risk, the power to detect such an interaction may have been limited, since aspirin use in the Physicians’ Health Study occurred only in the first 5 years of follow-up, during which time the predictive value of sICAM-1 was slight. Potential limitations of these data merit consideration. First, since they are based on assessment of sICAM-1 concentration in plasma samples stored at ⫺80°C for up to 12 years, we cannot exclude the possibility of protein degradation. However, this possibility seems unlikely because concentrations of sICAM-1 in our participants are similar to those reported in studies of fresh plasma. Second, because baseline plasma samples were obtained on only one occasion, we could not take into account variation in concentrations that may have occurred over time. The impact of this potential limitation would be to increase random misclassification, an effect that, if present, could lead to an underestimation of the magnitude of any relation. Finally, smokers had significantly higher concentrations of sICAM-1 than nonsmokers, an observation that raises the possibility of confounding.16 However, the association we found between sICAM-1 and future risk of myocardial infarction was present among non-smokers and, therefore, our results could not be due to confounding by smoking. The source of sICAM-1 found in our population is uncertain, although circulating forms of adhesion molecules may be derived from vascular-wall components, including endothelial and smooth-muscle cells.28,29 Previous studies have found sICAM-1 in plasma of patients with chronic inflammatory disorders,4 and increased expression of adhesion molecules has been reported in patients with unstable coronary syndromes.12 However, participants in our study were apparently healthy when blood samples were obtained, which would
THE LANCET • Vol 351 • January 10, 1998
keep to a minimum the likelihood that increased concentrations of sICAM-1 in this population were due to acute ischaemia or a clinically detectable chronic inflammatory condition. Our observation of increased concentrations of sICAM-1 several years before the occurrence of coronary occlusion supports the hypothesis that endothelial activation and inflammation occur early in the atherothrombotic process. In addition, because patients with the higher concentrations of sICAM-1 were at increased risk of future myocardial infarction, these data support the possibility that therapies aimed at antiadhesion may provide a novel approach to prevention and treatment of cardiovascular diseases.2,3,33 Experimental models of anti-ICAM antibodies involving vascular balloon injury, myocardial ischaemia, and reperfusion have had promising results.30–32 Finally, these data raise the possibility that cellular adhesion molecules can serve as a molecular marker for early atherosclerosis,2,3,33 a hypothesis that requires further testing. Contributors Paul Ridker was the lead investigator, conceived the hypothesis, managed the study, did the primary data analyses, and wrote the first draft of the paper. Charles Hennekens and Meir Stampfer made critical contributions to the design and analysis of these data and were involved in the writing and epidemiological assessment. Beatrice Roitman-Johnson and Jean Allen did all the biochemical analyses, including critical preliminary studies to address tissues of variance, stability, and reproducibility. All investigators contributed to the final writing of the manuscript.
Acknowledgments This study was funded by grants from the National Heart Lung and Blood Institute, Bethesda, Maryland, USA and from the American Heart Association, Dallas, Texas.
References 1 2 3
4
5 6
7
8
9
10
11
12 13
Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 1993; 363: 801–09. Springer TA. Adhesion molecules of the immune system. Nature 10090; 346: 425–34. Cybulsky MI, Gimbrone MA Jr. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science 1991; 25: 788–91. Bevilacqua MP, Nelson RM, Mannori G, Cecconi O. Endothelialleukocyte adhesion molecules in human disease. Annu Rev Med 1994; 45: 361–78. Adams DH, Shaw S. Leucocyte-endothelial interactions and regulation of leucocyte migration. Lancet 1994; 343: 831–36. Pober JS, Gimbrone MA Jr, Lapiette LA, et al. Overlapping patterns of activation of human endothelial cells by interleukin 1, tumour necrosis factor, and immune interferon. J Immunol 1986; 137: 1893–96. Poston RN, Haskard DO, Coucher JR, Gall NP, Johnson-Tidey RR. Expression of intercellular adhesion molecule-1 in atherosclerotic plaques. Am J Pathol 1992; 140: 665–73. Printseva OY, Peclo MM, Gown AM. Various cell types in human atherosclerotic lesions express ICAM-1: further immunocytochemical and immunochemical studies employing monoclonal antibody 10F3. Am J Pathol 1992; 140: 889–96. van der Wal AC, Das PK, Tigges AJ, Becker AE. Adhesion molecules on the endothelium and mononuclear cells in human atherosclerotic lesions. Am J Pathol 1992; 141: 1427–33. Davies MJ, Gordon JL, Gearing AJH, et al. The expression of the adhesion molecules ICAM-1, VACM-1, PECAM, and E-selectin in human atherosclerosis. J Pathol 1993; 171: 223–29. O’Brien KD, McDonald TO, Chait A, Allen MD, Alpers CE. Neovascular expression of E-selectin, intercellular adhesion molecule1, and vascular cell adhesion molecule-1 in human atherosclerosis and their relation to intimal leukocyte content. Circulation 1996; 93: 672–82. Jang Y, Lincoff AM, Plow EF, Topol EJ. Cell adhesion molecules in coronary artery disease. J Am Coll Cardiol 1994; 24: 1591–601. Peter K, Naworth P, Conradt C, et al. Circulating vascular cell adhesion molecule-1 correlates with the extent of human atherosclerosis in contrast to circulating intercellular adhesion molecule-1, E-selectin, P-selectin, and thrombomodulin. Arterioscler Thromb Vasc Biol 1997; 17: 505–12.
91
ARTICLES 14 Blann AD, McCollum CN. Circulating endothelial cell/leukocyte adhesion molecules in atherosclerosis. Thromb Haemost 1994; 74: 151–54. 15 Hackman A, Abe Y, Insull W, et al. Levels of soluble cell adhesion molecules in patients with dyslipidemia. Circulation 1996; 93: 1334–38. 16 Blann AD, Steele C, McCollum CN. The influence of smoking on soluble adhesion molecules and endothelial cell markers. Thromb 1997; 85: 433–38. 17 Ridker PM, Cushman M, Stampfer MJ, Tracey RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997; 336: 973–79. 18 Steering Committee of the Physicians’ Health Study Research Group. Final report of the aspirin component of the ongoing Physicians; Health Study. N Engl J Med 1989; 321: 129–35. 19 Ridker PM, Hennekens CH, Stampfer MJ. A prospective study of lipoprotein(a) and the risk of myocardial infarction. JAMA 1993; 270: 2195–99. 20 Ridker PM, Vaughan DE, Stampfer MJ, Manson JE, Hennekens CH. Endogenous tissue-type plasminogen activator and risk of myocardial infarction. Lancet 1993; 341: 1165–68. 21 Ridker PM, Hennekens CH, Cerskus A, Stampfer MJ. Plasma concentration of cross linked fibrin degradation product (D-dimer) and the risk of future myocardial infarction among apparently health men. Circulation 1994; 90: 2236–40. 22 Ridker PM, Hennekens CH, Selhub J, Miletich JP, Malinow MR, Stampfer MJ. Interrelation of hyperhomocyst(e)inemia, factor V Leiden, and risks of future venous thromboembolism. Circulation 1997; 95: 1777–82. 23 Dustin ML, Rothlein R, Bhan AR, Dinarello CA, Springer TA. Induction by IL-1 and interferon, tissue distribution, biochemistry and function of a natural adherence molecule (ICAM-1). J Immunol 1986; 137: 245–54.
92
24 Zouki C, Beauchamp M, Baron C, Filep JG. Prevention of in vitro neutrophil adhesion to endothelial cells through shedding of L-selectin by C-reactive protein and peptides derived from C-reactive protein. J Clin Invest 1997; 100: 522–29. 25 Languino LR, Plescia J, Duperray A, et al. Fibrinogen mediates leukocyte adhesion to vascular endothelium through an ICAM-1 dependent pathway. Cell 1993; 73: 1423–34. 26 Duperray A, Languino LR, Plescia, et al. Molecular identification of a novel fibrinogen binding site on the first domain of ICAM-1 regulating leukocyte-endothelium bridging. J Biol Chem 1997; 272: 435–41. 27 Gardiner EE, D’Souza SE. A mitogenic action for fibrinogen mediated through intercellular adhesion molecule-1. J Biol Chem 1997; 272: 15474–80. 28 Leewenberg JFM, Smeets EF, Neefjes JJ, et al. E-selectin and intercellular adhesion molecule-1 are released by activated human endothelial cells in vitro. Immunology 1992; 77: 543–49. 29 Pigott R, Dillon LP, Hemingway IH, Gearing AJH. Soluble forms of E-selectin, ICAM-1 and VCAM-1 are present in the supernatants of cytokine activated cultured endothelial cells. Biochem Biophys Res Commun 1992; 187: 584–89. 30 Ma X, Lefer DJ, Lefer AM, Rothlin R. Coronary endothelial and cardiac protective effects of a monoclonal antibody to intercellular adhesion molecule-1 in myocardial ischemia and reperfusion. Circulation 1992; 86: 937–46. 31 Yasukawa H, Imaizumi T, Matsuoka H, Nakashima A, Moritmatsu M. Inhibition of intimal hyperplasia after balloon injury by antibodies to intercellular adhesion molecule-1 and lymphocyte functionassociated antigen-1. Circulation 1997; 95: 1515–22. 32 Plow EF, D’Souza SE. A role for intercellular adhesion molecule-1 in restenosis. Circulation 1997; 95: 1355–56. 33 Libby P. Molecular bases of the acute coronary syndromes. Circulation 1995; 91: 2844–50.
THE LANCET • Vol 351 • January 10, 1998