The CORONARY ARlZRY DtsEAsE American Journal of cardiology MAY 15, 1992, VOL. 69, NO. 16 Predictive Value of Apolipoproteins in a Prospective Survey of Coronary Artery Disease in Men Gunnar Sigurdsson, MD, Asdis Baldursdottir, BSC, Helgi Sigvaldason, PhD, Uggi Agnarsson, MD, Gudmundur Thorgeirsson, MD, and Nikulas Sigfusson, MD Some studies have suggested that measurements of apolipoproteins may be valuable in the clinical assessment of susceptibility to coronary artery disease, over and above the lipoprotein lipids. Only a few of these studies have been prospective in nature and further knowledge is therefore need- ed to clarify the issue. The independent prognostic value of apolipoproteins (apo-B, apo-Al and apo[a]) with regard to coronary artery disease was estimated from a prospective survey among 1,332 randomly setected Icelandic men, aged 45 to 72 years, participating in a health survey from 1979 to 1981. The group was followed for 8.6 years, and during that period 104 men had fatal or non- fatal myocardial infarction. The Cox’s proportion- al hazards model was used to estimate the signifi- cance of independent variables. The resuits of multivariate analysis showed that ape(a) was a stgnificant independent risk factor (odds ratio 1.22 for 1 SD), but ape-Al was a stronger nega- tive risk factor (odds ratio 0.70 for 1 SD). Apo-B was a highly signtficant risk factor in a univariate analysis, but not in a multivariate analysis when serum cholesterol was included. Previous popula- tion surveys in Iceland have confirmed the impor- tance of cigarette smoking, cholesterol, triglycer- ides and bland pressure as risk factors for coro- nary artery disease. The present results illustrate additional importance of apo-Al and ape(a) con- centrations in predicting coronary artery disease among Icelandic men, whereas apo-B did not con- tribute anything further to the prediction than se- rum total cholesterol. (Am J Cardid 1992;69:1251-1254) From the Icelandic Heart Association, Lagmula; and Department of Medicine, Reykjavik City Hospital, Reykjavik, Iceland. Manuscript received September 16.1991; revised manuscript received and accepted January 13,1992. Address for reprints: Gunnar Sigurdsson, MD, PhD, Department of Medicine, Reykjavik City Hospital, 108 Reykjavik, Iceland. S ome studies have suggested that measurements of apolipoproteins may be valuable in the clinical as- sessment of susceptibility to coronary artery dis- ease, over and above the lipoprotein lipids.lw5 Only a few of these studies have been prospective in nature and further knowledge is therefore needed to clarify the is- sue. There is a substantial evidence for an association between serum lipoprotein (Lp[a]) concentration and myocardial infarction.6y7 This lipoprotein is assembled from low density lipoprotein and a large glycoprotein, apolipoprotein(a) (apo[a]). However, most of this evi- dence is from case control rather than prospective stud- ies. Similarly some studies have suggested that hyper- apobeta(B) lipoproteinemia (the main apoprotein of low-density lipoprotein) is an independent risk factor and more predictive for future coronary artery disease than low-density lipoprotein cholesterol.*,4*5 However, most of these studies have also been retrospective case control studies with the inherent limitations of that ex- perimental approach. High-density lipoprotein choles- terol has been substantiated as an independent negative risk factor in several, although not in all, prospective and retrospective studies.8*9However, a limited knowl- edge is available on the use of apo-AI (about 75% of the protein component of high-density lipoprotein) concen- tration as a risk factor for coronary artery disease. We have therefore used the opportunity to measure ape(a) (the apoprotein of Lp[a]), apo-B and apo-AI concentration in the serum of 1,332 Icelandic men at baseline, who have subsequently been followed for 8.6 years on average with regard to myocardial infarction and coronary death. Concomitantly, we have measured the better recognized risk factors, such as serum choles- terol, serum triglycerides, systolic blood pressure, smok- ing history and height and weight. The Cox’s propor- tional hazards model was used to estimate the indepen- dent prognostic significance of variables for myocardial infarction. METHODS §ubjeetsr Participants in this survey (stage IV of the Reykjavik Study) were male residents in the Reykjavik area and selected from the national roster, born on the 1 APOLIPOPROTEINS AND CORONARY ARTERY DISEASE 1251