Potential for cholesterol lowering in secondary ... - Atherosclerosis

Atherosclerosis 153 (2000) 505 – 517 www.elsevier.com/locate/atherosclerosis

Potential for cholesterol lowering in secondary prevention of coronary heart disease in Europe: findings from EUROASPIRE study Diego Vanuzzo a, Lorenza Pilotto a, Giovanni B. Ambrosio b, Kalevi Pyo¨ra¨la¨ c,*, Seppo Lehto c, Dirk De Bacquer d, Guy De Backer d, David Wood e, on behalf of the EUROASPIRE Study Group 1 a

Centre for Cardio6ascular Diseases, A.S.S. 4 ‘Medio Friuli’, Udine, Italy b First Department of Medicine, Venice City Hospital, Venice, Italy c Department of Medicine, Kuopio Uni6ersity Hospital, PO Box 1627, 70211 Kuopio, Finland d Department of Public Health, Uni6ersity of Ghent, Ghent, Belgium e Cardiac Medicine, Clinical Epidemiology Study Group, National Heart and Lung Institute, Imperial College of Science, Technology and Medicine, London, UK Received 2 November 1999; received in revised form 31 May 2000; accepted 5 July 2000

Abstract We have examined the potential for cholesterol lowering in secondary prevention of coronary heart disease based on data from the European Action on Secondary Prevention through Intervention to Reduce Events (EUROASPIRE) study carried out in 1995–1996 in nine European centres (Czech Republic, Finland, France, Germany, Hungary, Italy, The Netherlands, Slovenia and Spain). Consecutive patients aged 5 70 years in four diagnostic categories — coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, acute myocardial infarction, and acute myocardial ischaemia without infarction — were identified from hospital records and invited for an interview and risk factor assessment at least 6 months after hospital admission. Plasma lipid measurements were carried out in a central laboratory. Combining patients from all centres and diagnostic categories (n= 2749) the medians (interquartile ranges) for plasma lipids were: total cholesterol 5.36 (4.76 – 6.03) mmol/l, high density lipoprotein (HDL) cholesterol 1.19 (1.01–1.42) mmol/l, triglycerides 1.55 (1.15 – 2.24) mmol/l, and low density lipoprotein (LDL) cholesterol 3.32 (2.76–3.91) mmol/l. Only 33% of the patients received lipid-lowering drugs. If the therapeutic goal given in the 1998 European recommendations, total cholesterol B5.0 mmol/l, were applied, 67% of these patients would have needed an intensified cholesterol-lowering action, and with an even stricter goal, total cholesterol B 4.5 mmol/l, this proportion would have been as high as 84%. © 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Cholesterol; LDL cholesterol; Coronary heart disease; Lipid-lowering drugs; Secondary prevention

1. Introduction Research evidence indicating that cholesterol lowering reduces the risk of recurrent coronary events and retards the progression of coronary atherosclerosis in patients with clinically established coronary heart dis* Corresponding author. Tel.: +358-17-172159; fax: + 358-17173993. E-mail address: [email protected] (K. Pyo¨ra¨la¨). 1 Investigators and participating centres of the EUROASPIRE study are given in Ref. [8]

ease (CHD) has become importantly strengthened during the last decade. The benefit from cholesterol lowering in secondary prevention of CHD was already shown at the beginning of the 1990s by a meta-analysis of clinical endpoint trials [1] and by the results of coronary angiographic studies using different treatment modalities, including diet, lipid-lowering drugs and ileal bypass surgery [2]. Introduction of a new class of powerful lipid-lowering drugs, HMG Coenzyme A reductase inhibitors, statins, however, gave better possibilities than ever before to test the benefit of cholesterol lowering in both primary and secondary prevention of

0021-9150/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0021-9150(00)00596-7

506

D. Vanuzzo et al. / Atherosclerosis 153 (2000) 505–517

CHD. The first major statin trial in secondary prevention, the Scandinavian Simvastatin Survival Study (4S) published its results in 1994 and showed convincingly that in CHD patients with serum total cholesterol levels of 5.5–8.0 mmol/l simvastatin treatment reduced allcause mortality, coronary mortality, the risk of recurrent non-fatal myocardial infarction and other acute CHD events, and the need for myocardial revascularizations [3]. Two other large secondary prevention trials using another statin drug, pravastatin — Cholesterol and Recurrent Events (CARE) published in 1996 [4] and Long-Term Intervention with Pravastatin in Ischemic Disease (LIPID) published in 1998 [5] — confirmed these findings in CHD patients with average serum total cholesterol levels. In 1994, three European scientific societies in the field of cardiovascular medicine, namely the European Society of Cardiology (ESC), European Atherosclerosis Society (EAS) and European Society of Hypertension (ESH), published their joint recommendations on the prevention of CHD in clinical practice [6] and in the United States the National Cholesterol Education Program (NCEP) Expert Panel updated its recommendations [7], but these recommendations were formulated before the results of 4S had become available. On the basis of already existing evidence, however, both the European and American recommendations emphasized that the top priority of medical practitioners should be to concentrate on aggressive risk factor management in patients with clinically established CHD. The European recommendations did not give any specific treatment goals for total or low density lipoprotein (LDL) cholesterol in secondary prevention of CHD, whereas the NCEP Expert Panel recommended that LDL cholesterol should be lowered below 2.6 mmol/l in patients with clinically established CHD. In 1995–1996, upon the initiative of the ESC Working Group on Epidemiology and Prevention, a survey on secondary prevention of CHD — EUROASPIRE (European Action on Secondary Prevention through Intervention to Reduce Events) — was carried out in nine European countries (Czech Republic, Finland, France, Germany, Hungary, Italy, The Netherlands, Slovenia and Spain) [8]. Its design was largely based on a similar study carried out in 1994 – 1995 in the UK [9]. With regard to cholesterol management, EUROASPIRE study, based on a study population of more than 3500 CHD patients who were interviewed and whose risk factors were measured at least 6 months after hospitalization, showed that almost one half of the patients had plasma total cholesterol E 5.5 mmol/l (lower limit of cholesterol window in 4S). Only one third of all patients received lipid-lowering drugs and among those receiving these drugs one third had plasma cholesterol ] 5.5 mmol/l [8].

Because important new trial evidence had emerged in the field of lipid lowering, not only in secondary prevention but also in primary prevention of CHD [10,11], as well as in some other areas of CHD prevention, the three European Societies — ESC, EAS and ESH — took the view that the 1994 recommendations have to be revised. The revised recommendations were published in October 1998 [12]. These revised European recommendations give a definite therapeutic goal for cholesterol lowering; in both primary and secondary prevention of CHD the goal for total cholesterol is B 5.0 mmol/l and for LDL cholesterol B 3.0 mmol/l. The European recommendations, however, emphasize that in primary prevention the need for aggressive LDL cholesterol lowering has to be considered on the basis of multifactorial assessment of absolute risk of CHD. In the US, the therapeutic LDL cholesterol goal in secondary prevention given already in the 1994 NCEP Expert Panel recommendations, B 2.6 mmol/l, has been reinforced also in more recent statements [13,14]. In this article, we describe the lipid distributions in the EUROASPIRE study population in more detail and our findings concerning differences between countries and diagnostic categories in the use of lipid-lowering drugs and factors that correlated with their use. We have also analysed EUROASPIRE data with regard to the potential for cholesterol lowering in secondary prevention of CHD considering the therapeutic goals given in the 1998 European recommendations.

2. Methods

2.1. Study population The formation of the study population of EUROASPIRE has been previously described in detail [8] and will be only briefly described here. Within each of the nine countries one geographical area was selected and all hospitals serving that population identified. The area included at least one hospital offering interventional cardiology and cardiac surgery, and one or more acute hospitals receiving patients with acute myocardial infarction and ischaemia. Within each hospital, consecutive male or female patients (570 years) hospitalized with the diagnoses or treatments of CHD described below were identified retrospectively over a period of not less than 6 months and up to 48 months prior to the start of the survey. The dates of the hospital index events ranged from 1 July 1991 to 30 June 1995. The following hierachical order for patient identification was followed: 1. Coronary artery bypass grafting (CABG): patients having their first elective or emergency CABG. 2. Percutaneous transluminal coronary angioplasty (PTCA): patients having their first elective or emer-


gency PTCA, excluding patients who had had CABG at any time. 3. Acute myocardial infarction (AMI): patients with hospital diagnosis of AMI (first or recurrent), excluding patients with a previous CABG or PTCA. 4. Acute myocardial ischaemia without infarction (Ischaemia): patients with hospital diagnosis of acute myocardial ischaemia (first or recurrent) but no evidence of infarction, excluding patients with a previous CABG, PTCA or AMI. In each country, the aim was to obtain 400 patients for interview and risk factor assessment. Thus, to allow for deaths and non-response, at least 125 consecutive patients were identified for CABG, PTCA and Ischaemia categories and 150 patients for AMI category to allow for higher mortality in that category. Altogether 4863 patient records were reviewed in the EUROASPIRE study and of these patients 3569 participated in the interview and risk factor assessment between 1 June 1995 and 1 April 1996. After allowing for deaths and losses from follow-up, the adjusted overall response rate to the interview was 84.7% (3569/4215) [8]. The study population of the present study comprised 2749 patients (77.0% of the interviewed patients) for whom complete plasma lipid measurements carried out at the EUROASPIRE central laboratory were available. We examined possible selection bias in the

507

formation of this study population by comparing it with respect to age, gender and diagnostic category distribution with those 2114 patients on whom we had no interview data or no lipid measurements and with those 820 patients who were interviewed but on whom we did not have lipid measurements. These comparisons showed no statistically significant differences with respect to age and gender distribution but a slight and statistically significant bias in the diagnostic category distribution (PB0.0001, P= 0.002, respectively), characterized mainly by a somewhat lower proportion of the Ischaemia category in the present study population (21.8% versus 26.5 and 27.5%). The size of the study population by centre and diagnostic category is shown in Table 1. Given the small number of patients in different diagnostic categories in individual centres, the results by diagnostic categories will be reported for the whole study population. The gender and age distribution of the patients and the median time since hospital admission by diagnostic category and in the whole study population is illustrated in Table 2.

2.2. Study protocol The study protocol at patient interview and examination has been previously described [8]. In brief, it

Table 1 Size of the study population by centre and diagnostic categorya CZE/PP

FIN/KUO

FRA/LLR

GER/MUN HUN/BUD ITA/UTV

NET/ROT

SLO/LJU

SPA/BAR

Total

81 77 74 60

108 106 101 99

104 110 102 71

96 90 108 89

64 56 66 5

39 44 42 36

35 25 48 62

100 99 104 98

88 85 97 80

715 692 742 600

292

414

387

383

191

161

170

401

350

2749

Diagnostic category CABG PTCA AMI Ischaemia All

CZE/PP, Czech Republic/Pilsen, Prague; FIN/KUO, Finland/Kuopio; FRA/LLR, France/Lille, Lomme, Roubaix; GER/MU8 N, Germany/ Mu¨nster; HUN/BUD, Hungary/Budapest; IT/UTV, Italy/Udine, Treviso, Verona; NET/ROT, The Netherlands/Rotterdam; SLO/LJU, Slovenia/ Ljubljana; SPA/BAR, Spain/Barcelona. a

Table 2 Gender and age distribution and median time since hospital admission by diagnostic category and in the whole study population

Women (%) Median age at interview (years) Median age at hospital admission (years) Median time since hospital admission (years) Age]60 years (%) Men Women

CABG

PTCA

AMI

Ischaemia

All

17.2 61.6 60.2 1.2

20.7 58.3 56.4 1.6

23.2 61.3 59.3 1.7

36.0 62.0 60.0 1.7

13.8 60.9 59.2 1.5

25.3 32.5

14.9 18.9

21.4 39.5

25.3 35.2

21.5 32.3

508


included: personal and demographic details, personal and family history of CHD, information on lifestyle advice given by health professionals and medication, assessment of smoking status (checked by measurement of breath carbon monoxide), measurement of height and weight, calculation of body mass index (BMI, weight (kilograms)/height (meters) squared), measurement of blood pressure and plasma lipids.

skewed distribution. To evaluate factors associated with non-use of lipid-lowering drugs, we performed a multiple logistic regression analysis, adjusting for centre. All the statistical analyses were performed at the Department of Public Health, University of Ghent, Belgium.

2.3. Lipid measurements

Table 3 shows the lipid distributions by centre, diagnostic category and gender and in all thses subjects and in the whole study population. Although differences in the median values for different centres were rather small, statistical testing for equality in lipid distributions between centres, adjusting for gender, age, BMI, smoking status and diagnostic category, showed highly significant variation (P B 0.0001 for all lipid variables). Highest median values for total cholesterol were observed in Italy, France and Slovenia, and the lowest in Spain, Finland and Germany. There was also highly significant variation in total and HDL cholesterol levels between diagnostic categories (P= 0.03 for total cholesterol, P= 0.04 for HDL cholesterol, adjusting for gender, age, BMI, smoking status and centre). The highest median value for total cholesterol was observed in Ischaemia category and the lowest ones in CABG and PTCA categories. Female patients had significantly higher total and HDL cholesterol than male patients (PB 0.0001 for both, adjusting for age, BMI, smoking status, centre and diagnostic category). Triglycerides, however, were significantly lower in women (PB 0.0001), while LDL cholesterol levels were comparable. As a substantial proportion of patients used lipid-lowering drugs, as will be explained below, it has to be taken into account that the lipid distributions of patients in different centres and diagnostic categories may in part reflect treatment patterns. At the interview the patients were asked a question ‘Have you ever been told that your blood cholesterol or other blood lipid levels are raised?’ and, furthermore, they were asked by whom their blood lipid levels were found to be raised (general practitioner, hospital physician, other health professional, e.g. nurse, dietician). Table 4 shows the proportion of patients who were aware of hyperlipidaemia among those 1803 subjects whose total cholesterol was ] 5.5/mmol/l or who were using lipid-lowering drugs at interview by centre, diagnostic category, gender and and in all these subjects in the whole subpopulation and the distribution of the first diagnosis of hyperlipidaemia between health professionals. In the whole sub-population 79.0% of such patients knew that they had hyperlipidaemia; the awareness varied between centres from 66.7% in the Netherlands to 91.1% in Finland. The awareness varied between diagnostic categories from 69.7% in Ischaemia category to 82.1% in PTCA category. There was, how-

For plasma lipid measurements a venous blood sample (fasting or non-fasting) was drawn into a tube containing potassium EDTA as anticoagulant (Vacutainer, Becton Dickinson) from which plasma was obtained and stored at − 20°C or below and then transported to the EUROASPIRE central laboratory (Clinical Research Laboratories Europe, Zaventem, Belgium) for measurement of total cholesterol, high density lipoprotein (HDL) cholesterol, and triglycerides using a Hitachi 747 analyser. Cholesterol was analysed using Raichem Enzymatic Cholesterol Reagent and triglycerides (with blanking when appropriate) using Technicon Triglyceride Reagent. HDL cholesterol was measured after precipitation of other lipoproteins with manganese/heparin. All analyses were standardised by the Centre for Disease Control/Control NHLBI Lipid Standardization Program. Pooled plasma lipid controls were analysed with the samples, and during the course of the study, coefficients of variation for cholesterol were B 1.8%, triglycerides B2.4% and HDL cholesterol B 3.7%. LDL cholesterol was calculated using the Friedewald formula [15], excluding 89 patients whose triglyceride level was \4.6 mmol/l. Since blood samples were taken after overnight fast in only five centres (in Czech Republic, Finland, Germany, in one subcentre in Hungary, and in Slovenia), comprising 1580 patients (57.5% of the whole study population), medians and other estimates describing triglyceride distributions in the whole study population will be slightly overestimated, whereas LDL cholesterol levels will be slightly underestimated.

2.4. Statistical methods Regarding the positive skewness of the plasma lipid distributions, medians and interquartile ranges are given to characterize their distributions. Variations in lipid distributions between centres, diagnostic categories and genders have been tested by analysis of covariance adjusting for gender, age, body mass index, smoking status, and centre or diagnostic category, as appropriate, respectively. Variation in frequency estimates between centres, diagnostic categories and genders was tested by chi-square test. Triglycerides were log-transformed for statistical calculations due to their

3. Results


509

Table 3 Distribution of plasma lipids by centre, diagnostic category and gender, and in the whole study populationa Median (Interquartile range) TC (mmol/l)

HDL-C (mmol/l)

TG (mmol/l)

LDL-C (mmol/l)

Centre CZE/PP FIN/KUO FRA/LLR GER/MU8 N HUN/BUD ITA/UTV NET/ROT SLO/LJU SPA/BAR

5.31 5.25 5.49 5.25 5.41 5.62 5.41 5.46 5.18

1.16 1.26 1.26 1.19 1.13 1.19 1.13 1.19 1.16

1.68 1.46 1.69 1.35 1.75 2.13 1.84 1.54 1.50

3.29 3.22 3.30 3.31 3.38 3.44 3.38 3.43 3.26

Significanceb

PB0.0001

PB0.0001

PB0.0001

PB0.0001

Diagnostic category CABG PTCA AMI Ischaemia

5.31 5.31 5.33 5.49

1.19 1.21 1.16 1.24

1.55 1.58 1.54 1.54

3.33 3.25 3.30 3.35

Significanceb

P= 0.03

P =0.04

P=0.54

P= 0.09

Gender Men Women

5.31 (4.74–5.95) 5.59 (4.92–6.21)

1.16 (0.98–1.34) 1.37 (1.13–1.63)

1.58 (1.16–2.28) 1.50 (1.10–2.11)

3.31 (2.77–3.88) 3.33 (2.71–4.00)

Significanceb

PB0.0001

PB0.0001

PB0.0001

P =0.71

Whole study population

5.36 (4.76–6.03)

1.19 (1.01–1.42)

1.55 (1.15–2.24)

3.32 (2.76–3.91)

(4.79–5.93) (4.61–5.98) (4.89–6.13) (4.74–5.95) (4.84–6.13) (5.12–6.26) (5.02–6.03) (4.87–6.19) (4.53–5.72)

(4.74–6.01) (4.71–5.95) (4.76–6.01) (4.92–6.11)

(0.95–1.39) (1.08–1.52) (1.03–1.55) (0.98–1.42) (0.98–1.37) (0.98–1.42) (0.98–1.29) (1.01–1.39) (1.01–1.34)

(1.01–1.37) (1.01–1.42) (0.98–1.42) (1.03–1.50)

(1.18–2.38) (1.14–1.96) (1.18–2.35) (0.98–1.86) (1.18–2.84) (1.54–2.97) (1.27–2.53) (1.16–2.19) (1.10–1.98)

(1.17–2.21) (1.15–2.27) (1.11–2.27) (1.16–2.20)

(2.78–3.91) (2.62–3.78) (2.74–3.88) (2.79–3.91) (2.83–4.02) (2.86–3.92) (2.91–3.81) (2.85–4.11) (2.64–3.83)

(2.76–3.91) (2.68–3.84) (2.78–3.89) (2.81–3.99)

a TC, total cholesterol; HDL-C, HDL cholesterol; TG, triglycerides; LDL-C, LDL cholesterol (calculated by the Friedewald formula, excluding 89 patients with triglyceride level\4.6 mmol/l). b Significance of differences between centres, diagnostic categories, and gender, adjusted for each other, and for age, BMI and smoking.

Table 4 Awareness about hyperlipidaemia and its first diagnosis by health professionals by centre, diagnostic category and gender among 1803 subjects with TC]5.5 mmol/l or using lipid-lowering drugs at interview Proportion of patients aware of hyperlipidaemia% (n)

Who first found that blood lipids are raised? (%) General practitioner

Hospital physician

Other health professional

(133/174) (257/282) (203/286) (199/245) (84/112) (95/110) (78/117) (211/272) (164/205)

38.4 27.2 59.6 63.8 29.8 64.2 28.2 43.1 47.0

54.1 48.6 30.0 35.7 65.5 33.7 66.7 55.4 27.4

7.5 24.1 10.3 0.5 4.8 2.1 5.1 1.4 25.6

Diagnostic category CABG 82.0 (402/490) PTCA 82.1 (408/497) AMI 79.6 (366/460) Ischaemia 69.7 (248/356)

49.2 43.4 38.0 52.8

38.1 47.8 51.6 37.5

12.7 8.8 10.4 9.7

Gender Men Women All

46.0 43.1 45.3

42.8 48.4 44.2

11.1 8.5 10.5


76.4 91.1 71.0 81.2 75.0 86.4 66.7 77.9 80.0

79.8 (1060/1328) 76.6 (364/475) 79.0 (1424/1803)

767 (36.6) 179 (27.4) 946 (34.4)

(31.4) (28.2) (38.0) (40.7)

(40.4) (31.9) (26.1) (36.0) (41.3) (31.6) (31.2) (32.2) (41.4) (29.3) (28.6) (30.5) (41.7)

(29.8) (29.2) (32.0) (28.8) (36.7) (45.4) (31.8) (37.1) (27.4)

625 (29.8) 259 (39.6) 884 (32.2)

210 198 226 250

87 121 124 110 70 73 54 149 96

(c) TC]5.5 mmol/l

(25.9) (28.2) (20.4) (10.0)

(15.8) (27.8) (25.1) (25.6) (12.7) (6.2) (24.1) (18.5) (22.4)

458 (21.9) 133 (20.3) 591 (21.5)

185 195 151 60

46 115 97 98 24 10 41 74 79

(d) TCB5.5 mmol/l

LLD treatment

(13.3) (15.0) (11.2) (7.7)

(14.0) (11.1) (16.8) (9.7) (9.4) (12.4) (12.9) (12.2) (8.5)

245 (11.7) 83 (12.7) 328 (11.9)

95 104 83 46

41 46 65 37 18 20 22 49 30

(e) TC]5.5 mmol/l

33.6 24.5 33.4

39.2 43.2 31.5 17.7

29.8 38.9 41.9 35.2 22.0 23.0 37.1 30.6 31.0

(d+e)/a

% LLD treatedb

34.8 38.4 35.7

33.9 34.8 35.5 43.4

47.1 28.6 40.1 27.4 42.9 54.1 34.9 39.8 27.5

e/(d+e)

63.4 72.6 64.5

68.5 71.8 62.0 59.3

59.6 68.1 73.9 64.0 58.6 63.9 68.8 67.8 58.2

(c+d+e)/a

% LLD treated TC % needing LLD ] 5.5 mmol/l treatment

52.9 45.5 51.0

57.1 60.2 50.9 29.8

50.0 57.1 56.6 55.1 37.5 29.1 53.8 45.2 53.1

(d+e)/c+d+e)

% of those needing who received LLD treatment

b

LLD, lipid-lowering drugs; TC, total cholesterol. Testing differences in proportion of LLD treated in a multiple logistic model with additional adjustment for age, BMI and smoking: centre: PB0.0001, diagnostic category: PB0.0001, and gender: P =0.07.

a

2095 654 2749

225 195 282 244

Diagnostic category CABG 715 PTCA 692 AMI 742 Ischaemia 600

Gender Men Women Whole study population

118 132 101 138 79 51 53 129 145

(b) TCB5.5 mmol/l

292 414 387 383 191 161 170 401 352


(a) Total

No LLD treatment

Table 5 Lipid treatment status by centre, diagnostic category and gender, and in the whole study populationa

510 D. Vanuzzo et al. / Atherosclerosis 153 (2000) 505–517


Fig. 1. The distribution of the use of lipid-lowering drugs (LLD) according to total cholesterol level in the whole EUROASPIRE study population.

ever, virtually no gender difference in the awareness. As to the responsibility for the first diagnosis of hyperlipidaemia, there was a marked variation between centres in the role of general practitioners and hospital physicians in the detection of hyperlipidaemia. In France, Germany, Italy and Spain general practitioners had the most important role, whereas in other countries hospital physicians had made a greater contribution to detection of hyperlipidaemia than general practitioners. In only two countries, Finland and Spain, other health professionals contributed substantially to the detection

511

of hyperlipidaemia. There was rather little difference between diagnostic categories with regard to the role of different health professionals in the detection of hyperlipidaemia. In the CABG and Ischaemia categories, however, the role of general practitioners was somewhat more prominent than in the two other diagnostic categories. There was no gender difference in this respect. Table 5 shows the proportion of patients using lipidlowering drugs according to total cholesterol level, using 5.5 mmol/l as a cut-off point, by centre, diagnostic category and gender and in the whole study population. This cut-off point, corresponding to the lower limit for inclusion of CHD patients in 4S trial [3] was chosen, because evidence based medicine concerning the use of lipid-lowering drugs did not extend to lower total cholesterol levels at the time of EUROASPIRE survey in 1995–1996. In the whole study population, 919 of the 2749 patients (33.4%) used lipid-lowering drugs. Of these 919 patients 591 (64.3%) had total cholesterol B 5.5 mmol/l and 328 (35.7%) ] 5.5 mmol/l. There was a marked variation in the use of lipid-lowering drugs between centres (PB 0.0001), ranging from 22.0% in Hungary to 41.9% in France. There were also marked differences in this respect between diagnostic categories (PB 0.0001); the proportion of patients using lipid-lowering drugs was highest in CABG and PTCA categories, 39.2% and 43.2%, respectively, and lowest,

Table 6 Distribution of lipid-lowering drug therapy according to drug class by centre, diagnostic category and gender among 919 patients using these drugsa Number of patients using lipid-lowering drugs

Proportion of patients by drug class (%)b Statin

Fibrate

Other


87 161 162 135 42 37 63 123 109

24.1 90.0 49.4 90.4 66.6 78.4 85.7 78.1 78.9

71.3 9.4 42.0 7.4 31.0 16.2 14.3 17.9 20.2

4.6 0.6 8.6 2.2 2.4 5.4 0.0 4.1 0.9

Diagnostic category CABG PTCA AMI Ischaemia

280 299 234 106

71.4 73.9 69.2 73.6

24.6 24.1 26.9 21.7

4.0 2.0 3.9 4.7

Gender Men Women All

703 216 919

70.4 76.8 71.9

26.2 19.9 24.7

3.4 3.2 3.4

a Testing equality of distributions of lipid-lowering therapy across centres: PB0.001; across diagnostic categories: P =0.68, across gender: P =0.17. b Patients receiving combination treatment including a statin drug have been placed into statin class.

512


Table 7 Multivariate analysis of factors associated with non-use versus use of lipid-lowering drugs in patients with total cholesterol ]5.5 mmol/l Odds ratio (95% CI) Age]60 years Female Time since index event]2.5 years Body mass index]30 kg/m2 History of hypertension History of hyperlipidaemia History of CHD prior to index event CHD event following index event Current smoking Education56 years Currently employed Positive family history of CHD Diabetes Currently under care of GPa Index event = CABG or PTCA

1.18 1.09 1.36 1.10 1.09 0.43 1.43 0.71 1.62 0.95 0.67 0.91 1.04 1.43 0.43

(0.94–1.49) (0.85–1.38) (1.01-1.83)b (0.87-1.38) (0.87-1.37) (0.32-0.57)d (1.11-1.84)c (0.56-0.90)c (1.23-2.13)d (0.65-1.40) (0.51-0.88)c (0.75-1.14) (0.81-1.43) (1.09-1.79)c (0.34-0.56)d

a

GP, general practitioner. PB0.05. c PB0.01. d PB0.001. b

17.7% in Ischaemia category. Of men 33.6% but only 24.5% of women used lipid-lowering drugs (P B 0.07) Fig. 1 illustrates the distribution of the use of lipidlowering drugs according to total cholesterol level in the whole study population in more detail. As expected, and due to the cholesterol-lowering effect of drugs, the proportion of patients using lipid-lowering drugs was largest in lowest total cholesterol categories. A striking finding was that in the highest total cholesterol level categories, with total cholesterol levels ]6.5 mmol/l, two thirds of the patients did not receive lipid-lowering drugs. Table 6 shows the distribution of lipid-lowering drug therapy according to drug class by centre, diagnostic category and gender among 919 patients using these drugs. Of all the patients using lipid-lowering drugs, 71.9% used a statin drug either alone or as part of combination treatment, 24.7% used a fibrate drug either alone or as part of combination treatment, and 3.4% used some other lipid-lowering drug. In the majority of centres statins formed the most widely used drug class. In the Czech Republic, however, fibrates formed the most commonly used drug class, and also in France and Hungary fibrates were used rather extensively. The variation between the centres was statistically significant (PB 0.001). There was no statistically significant difference between diagnostic categories (P =0.68) or gender difference in the distribution of drug classes (P=0.17). Table 7 shows the results of a multivariate (multiple logistic regression) analysis examining factors associ-

ated with the non-use of lipid-lowering drugs. This analysis comprised 884 patients not using lipid-lowering drugs and with total cholesterol ]5.5 mmol/l and 920 patients using lipid-lowering drugs. A long time since index event, history of CHD before the index event, being a current smoker, and being under the care of a general practitioner were factors significantly associated with the non-use of lipid-lowering drugs, whereas factors inversely and significantly associated with the nonuse of lipid-lowering drugs, i.e. favouring their use, were: history of hyperlipidaemia (noted in hospital records), history of CHD event following index event, being currently employed, and having undergone CABG or PTCA. In Table 8, we have applied the therapeutic goals defined by the 1998 European recommendations [12] for total cholesterol (B 5.0 mmol/l) and LDL cholesterol (B 3.0 mmol/l) to see how large a proportion of patients among those EUROASPIRE patients who did not receive lipid-lowering drugs, and on the other hand, among those receiving lipid-lowering drugs, would have needed more intensive action for cholesterol lowering. We have also made these calculations by lowering the therapeutic goals by 0.5 mmol/l, which brings them near to the goals defined for CHD patients by the U.S. recommendations [13,14]. The 1998 European recommendations proposed that a triglyceride level \2.0 mmol/l and an HDL cholesterol level B 1.0 mmol/l be used as indicators of increased risk to be taken into account in the assessment of CHD risk. A total cholesterol/HDL cholesterol \5.0 is another lipid indicator of high risk We have applied these cut-off points for lipid indicators of high risk to the EUROASPIRE study population and the outcome is given in Table 8. Because in EUROASPIRE study blood samples had been drawn in fasting state in only 57% of the patients, we made all these calculations also separately for this subpopulation of patients. With the exception of data concerning triglycerides, the results were, however, essentially similar in the whole study population and in the subpopulation with fasting blood samples. Therefore, the results given in Table 8 are based on the whole study population, dividing patients into those not using lipid-lowering drugs and those using them. The results with regard to triglycerides in the fasting subpopulation are also given. Applying the therapeutic goal for total cholesterol B 5.0 mmol/l, 66.8% of the whole EUROASPIRE study population would have needed a more aggressive action for cholesterol lowering; this was the case in 71.6% of the patients not using lipid-lowering drugs and in 57.2% of the patients using such drugs. Definition of therapeutic goal as LDL cholesterol B3.0 mmol/l gave almost similar results. If instead the lower therapeutic goal for total cholesterol, B 4.5 mmol/l, was applied, 83.8% of the whole EUROASPIRE study population, and 87.5% of those patients not using and


76.4% of those patients using lipid-lowering drugs, would have needed an intensified action for cholesterol lowering. Again, the outcome was rather similar, when the lower therapeutic goal for LDL cholesterol, B 2.5 mmol/ l, was applied. As to elevated plasma triglycerides, \ 2 mmol/l, the results from the subpopulation of patients with fasting blood samples probably give the best approximation of the prevalence of this lipid indicator of high risk. Of all the fasting patients, 27.4% had plasma triglycerides \ 2.0 mmol/l; the prevalence of this lipid abnormality was only slightly higher among patients not using lipidlowering drugs than among patients using them. Among patients not using lipid lowering drugs there was no difference between men and women in the prevalence of elevated triglycerides, but among patients using lipidlowering drugs the prevalence of elevated triglycerides was somewhat lower in women than in men. The prevalence of low HDL cholesterol, B 1.0 mmol/l, was 23.6% in the whole EUROASPIRE study population and somewhat higher among patients not using lipidlowering drugs than among those using them. There was, however, a marked gender difference in the prevalence of low HDL cholesterol; using the same cut-off point for men and women, the prevalence was in men more than two times higher than in women, reflecting the gender difference in the HDL cholesterol distribution shown in Table 3.

513

The prevalence of total cholesterol/HDL cholesterol ratio \ 5 was 34.0% in the whole EUROASPIRE study population; clearly higher among men than in women, as could be expected on the basis of gender difference in HDL cholesterol distribution. The prevalence of this lipid indicator of high risk was also clearly higher among patients not using lipid-lowering drugs than among patients using them.

4. Discussion This nine-country European survey carried out in 1995–1996 demonstrated that the lipid pattern and the management of hypercholesterolaemia in patients with clinically established CHD were far from optimal. For all the EUROASPIRE patients median values for plasma lipids were: total cholesterol 5.36 mmol/l, HDL cholesterol 1.19 mmol/l, triglycerides 1.55 mmol/l, and LDL cholesterol 3.32 mmol/l. Because in 43% of the patients plasma lipids were measured from non-fasting blood samples, the median value for triglycerides may have become slightly overestimated and the median value for LDL cholesterol calculated by the Friedewald formula slightly underestimated. Although there was statistically significant variation in the distributions of plasma lipids between centres, the differences were surprisingly small; the difference between the lowest and

Table 8 Proportion of patients with too high total or LDL cholesterol levels defined by two different levels of therapeutic goals or with other lipid indicators of high risk %(n)a Whole study population Not using LLDb

Using LLD

All

Total cholesterol ]5.0 mmol/l ]4.5 mmol/l

71.6 (1311/1830) 87.5 (1601/1830)

57.2 (526/919) 76.4 (702/919)

66.8 (1837/2749) 83.8 (2303/2749)

LDL cholesterol ]3.0 mmol/l ]2.5 mmol/l

70.3 (1245/1770) 89.1 (1577/1770)

52.7 (469/890) 77.9 (693/890)

64.4 (1714/2660) 85.3 (2270/2660)

Triglycerides\2 mmol/l Men Women

33.2 (608/1830) 34.0 (473/1392) 30.8 (135/438)

30.1 (277/919) 32.4 (228/703) 22.7 (49/216)

32.2 (885/2749) 33.5 (701/2095) 28.1 (184/654)

Triglycerides\2 mmol/l ( fasting) Men (fasting) Women (fasting)

28.8 (306/1062) 28.7 (225/785) 29.2 (81/277)

24.5 (127/518) 26.5 (101/381) 19.0 (26/137)

27.4 (433/1580) 28.0 (326/1166) 25.8 (107/414)

HDL cholesterolB1 mmol/l Men Women

25.3 (463/1830) 28.9 (402/1392) 13.9 (61/438)

20.2 (186/919) 23.9 (168/703) 8.3 (18/216)

23.6 (649/2749) 27.2 (570/2095) 12.1 (79/654)

Total/HDL cholesterol\5 Men Women

38.4 (703/1830) 41.3 (575/1392) 29.2 (128/438)

25.4 (233/919) 29.2 (205/703) 13.0 (28/216)

34.0 (936/2749) 37.2 (780/2095) 23.8 (156/654)

a b

Calculated by the Friedewald formula, excluding 89 patients whose triglyceride level was \4.6 mmol/l. LLD, lipid lowering drugs.

514


highest total cholesterol median values was only 0.44 mmol/l. The relatively low median total cholesterol levels for CHD patients observed in EUROASPIRE and small differences in median total cholesterol between centres may seem surprising at first sight and therefore need some comments considering possible underlying factors. Population-based studies carried out before the introduction of statins in countries with relatively high cholesterol levels showed that mean total cholesterol levels of patients with myocardial infarction or angina pectoris were only slightly higher than or rather similar to the population mean levels for total cholesterol [16,17]. The explanation for this finding is that in such populations the majority of new patients with CHD come from the large majority of people with cholesterol levels around the population mean value. The WHO MONICA Project has recently reported population mean levels for total cholesterol in the 1990s and their 10-year trends in a number of European countries based on repeated surveys of representative samples of the general population [18]. Total cholesterol levels have shown declining trends in the majority of European countries, most markedly in countries with high initial levels, and consequently differences between populations have become smaller, the majority of the European populations having mean levels for total cholesterol between 5.5 and 6.2 mmol/l. The WHO MONICA Project report included population samples from six countries participating in the EUROASPIRE study (Czech Republic, Finland, France, Germany, Italy, and Spain). In all these six countries, the total cholesterol mean values of the EUROASPIRE patients were somewhat lower than those observed in the general population. One probable explanation for this finding and for the generally low position of the total cholesterol distributions of the EUROASPIRE patients, and possibly also for the relatively small differences between centres, is the effect of lipid-lowering drugs, because 33% of all EUROASPIRE patients used such drugs, although there was variation from 22 to 42% in their use between centres. Furthermore, it is likely that improvement in the diet of the patients has had some lowering effect on the cholesterol distributions and also contributed to the reduction of differences between centres. Patients who had been admitted because of an attack of acute ischaemia without infarction (Ischaemia category) appeared to be a particularly neglected group with regard to cholesterol-lowering action. In this diagnostic category only 18% of the patients used lipid-lowering drugs and their median total and LDL cholesterol levels were higher than those observed in CABG, PTCA and AMI categories. The majority of the EUROASPIRE patients with elevated cholesterol levels were aware of their hyperlip-

idaemia, with variation from 67 to 91% between centres and from 70 to 82% between diagnostic categories. In some countries, i.e. France, Germany, Italy and Spain, the first diagnosis of hyperlipidaemia had in most instances been made by a general practitioner, whereas in other countries a hospital physician had most often made the diagnosis. In the interpretation of this finding it has to be taken into account that CHD patients of the EUROASPIRE study were not representative of all CHD patients in the population, because they were identified among patients hospitalized for acute CHD events or for revascularization procedures. This may have overemphasized the role of hospital physicians in the detection of hyperlipidaemia as part of care of patients with CHD. Unfortunately, it was not possible to perform a detailed dietary survey as part of the core protocol of the EUROASPIRE study, but some information was collected at interview on general lifestyle advice, including dietary advice, related to raised blood lipid levels given to patients who were not using lipid-lowering drugs. Presuming that those EUROASPIRE patients who were using lipid-lowering drugs had received at least some dietary advice before being put on drug treatment and adding them to those patients who were not on lipid-lowering drugs but said that they had received lifestyle advice, it could be estimated that 58% of all EUROASPIRE patients had received some dietary advice, with a variation from 46 to 71% between centres. A detailed dietary survey was carried out only in the Finnish EUROASPIRE centre [19]. This survey showed that 35% of the Finnish patients had achieved the recommended goal for total fat intake (B 30 E%), but only 25% the goal for saturated fat intake (B10 E%), whereas 83% had achieved the recommended cholesterol intake (B 300 mg/day). Interestingly, patients using lipid-lowering drugs were more compliant with cholesterol-lowering diet than those not using such drugs. In 4S, documenting the benefit from cholesterol lowering with statin treatment in the reduction of all cause mortality, coronary mortality and the risk of non-fatal major CHD events, the baseline mean total cholesterol of CHD patients was 6.7 mmol/l and the lower total cholesterol limit for inclusion 5.5 mmol/l [3]. In the whole EUROASPIRE study population, only one third of the patients used lipid-lowering drugs and of those using these drugs one third still had total cholesterol ]5.5 mmol/l. Of the patients not using lipid-lowering drugs, one half had total cholesterol ] 5.5 mmol/l. Presuming that all EUROASPIRE patients using lipidlowering drugs would, have without drug treatment, have had total cholesterol ] 5.5 mmol/l and adding to them those patients who were not using lipid-lowering drugs and had total cholesterol ] 5.5 mmol/l, we obtained an estimate of the proportion of patients in the


EUROASPIRE study population who would have fulfilled the rather conservative 4S inclusion criterion, and applying the results of 4S, would have needed lipid-lowering drug treatment. In the whole EUROASPIRE study population this proportion was 64% and varied between centres from 58 to 74%. Of those patients in the whole EUROASPIRE study population who fulfilled the 4S inclusion criterion, however, only 51% received lipid-lowering drugs, with variation from 29 to 57% between centres; in other words, the penetration of scientific evidence derived from 4S into clinical practice had been slow and far from satisfactory. Statins formed the most frequently used class of lipid-lowering drugs; 72% of the EUROASPIRE patients using lipid-lowering drugs received a statin drug either alone or as part of a combination treatment. Relatively frequent use of fibrate drugs in some centres (Czech Republic, France, and Hungary) and their infrequent use in some other centres (Finland and Germany) evidently reflect differences in medical traditions which may have been influenced by marketing strategies of local and international pharmaceutical industries. Multivariate analysis of factors associated with the non-use of lipid-lowering drugs among EUROASPIRE patients for whom these drugs would, in the light of the results of 4S, have been indicated, showed some interesting and intriguing associations. A long time (] 2.5 years) since hospitalization for index CHD event and a history of CHD prior to index event were associated with the non-use of lipid-lowering drugs, possibly reflecting the treatment tradition prevailing before the publication of 4S results. Association of current smoking with the non-use of lipid-lowering drugs may reflect poor compliance of smokers with regard to treatment recommendations. Being under the care of a general practitioner increased the likelihood of the non-use of lipid-lowering drugs. A slower spreading of trial evidence among general practitioners than among hospital physicians and a greater uncertainty of general practitioners in the prescription of lipid-lowering drug treatment may explain this finding. As expected, recognition of hyperlipidaemia in hospital medical records favoured the use of lipid-lowering drugs. It is also understandable that a recurrent CHD event after index hospitalisation increased the likelihood of the use of lipid-lowering drugs. Patients who had undergone CABG or PTCA were more likely to receive lipid-lowering drugs than patients in AMI and Ischaemia categories, although the use of lipid-lowering drugs would have been equally well indicated in the two latter patient categories. The positive effect of being employed on the use of lipidlowering drugs may in part be due to better ability to pay the drug expenses, because in several countries health insurance systems reimburse only a part of the relatively high retail price of lipid-lowering drugs, particularly statins.

515

CARE and LIPID trials examined the benefit from statin treatment in CHD patients with lower total cholesterol levels than 4S. In CARE baseline total cholesterol levels were below 6.2 mmol/l, with mean total cholesterol of 5.4 mmol/l, and LDL cholesterol from 3.0 to 4.5 mmol/l [4]. LIPID recruited patients with total cholesterol levels from 4.0 to 7.0 mmol/l, with mean total cholesterol of 5.6 mmol/l [5]. Both CARE and LIPID demonstrated a significant reduction in the incidence of composite outcome of CHD death and non-fatal myocardial infarction, and LIPID also a significant reduction in all cause mortality. Based on the results of 4S, CARE and LIPID trials and other available information, the 1998 European recommendations [12] took the view that in secondary prevention the therapeutic goal for total cholesterol is B 5.0 mmol/ l and for LDL cholesterol B 3.0 mmol/l. When we applied this total cholesterol goal to the EUROASPIRE study population, 67% of the whole study population would have needed an intensified action for cholesterol lowering; of the patients not using lipid-lowering drugs 72% and of those already using them 57% would have needed such action. The results were almost similar, when LDL cholesterol goal B 3.0 mmol/l was used. The US recommendations [13,14] give a lower therapeutic goal for LDL cholesterol, B 2.6 mmol/l, in secondary prevention of CHD. This view has gained some support from the results of angiographic PostCABG trial [20] and post hoc epidemiological observations from 4S [21]. To get an approximate estimate of the increase in the proportion of patients needing intensified treatment using the US definition of therapeutic goal for LDL cholesterol, we calculated the proportions of patients in the EUROASPIRE study population with total cholesterol level B 4.5 mmol/l and LDL cholesterol level B 2.5 mmol/l. Using this definition for total cholesterol goal, 84% of the whole EUROASPIRE study population would have needed intensification of cholesterol-lowering therapy; of the patients not using lipid-lowering drugs 88% and of those already using them 76% would have needed an intensified action. The use of LDL cholesterol goal B 2.5 mmol/l in these calculations gave essentially similar results. The 1998 European recommendations [12] did not define any treatment goals for triglycerides and HDL cholesterol, because such goals could not be derived from trial evidence. Triglyceride level \ 2.0 mmol/l and HDL cholesterol level B 1.0 mmol/l were, however recommended as indicators of high risk to be considered, in addition to other risk factors, in the assessment of the risk of future CHD events. The US recommendations [13,14] have defined triglycerides B 2.3 mmol/l and HDL cholesterol \ 1.0 mmol/l as ‘secondary lipid goals’. In the EUROASPIRE study population, about

516


one third of CHD patients had triglyceride level \ 2 mmol/l, with very little difference in the prevalence between men and women, and almost one quarter of all EUROASPIRE patients had HDL cholesterol B 1.0 mmol/l, but with markedly higher prevalence in men than in women. Both the European and U.S. recommendations have for practical reasons simplified the definition of low HDL cholesterol, not taking into account the gender difference in these plasma lipids which exists also in CHD patients, as the EUROASPIRE study data on HDL cholesterol distributions in male and female CHD patients demonstrate; the median HDL cholesterol was 0.2 mmol/l higher in female patients than in male patients. The implication of this gender difference in HDL cholesterol in CHD patients for the prediction of the risk of future CHD events in female CHD patients needs to be assessed in prospective studies. Recently results have been reported from Veterans Affairs High-Density Lipoprotein Intervention Trial (VA-HIT), a trial of secondary prevention of CHD using a fibrate drug, gemfibrozil, in men with HDL cholesterol levels 5 1.0 mmol/l, LDL cholesterol levels 5 3.6 mmol/l and triglyceride levels 5 3.4 mmol/l [22]. Gemfibrozil treatment decreased the mean HDL cholesterol level by 6%, the mean triglyceride level by 31%, but did not change the mean LDL cholesterol level significantly. During the 5.1-year trial period gemfibrozil treatment reduced the incidence of the primary trial endpoint, non-fatal myocardial infarction or CHD death, by 22% (P=0.006); the event rate was 21.7% in the placebo group and 17.3% in the gemfibrozil group. Of those male EUROASPIRE patients who were not using lipid-lowering drugs, 16% would have fulfilled the VA-HIT trial inclusion criteria. In conclusion, EUROASPIRE, a nine-country European survey of the practice of secondary prevention of CHD in 1995–1996 showed that the majority of CHD patients had too high plasma total and LDLcholesterol levels, and that only one third of them were receiving cholesterol-lowering drugs. Application of the therapeutic goal given in the 1998 revision of the European recommendations, total cholesterol B5.0 mmol/l (LDL cholesterol B3.0 mmol/l), to the EUROASPIRE study data reveals the potential for cholesterol lowering in secondary prevention in Europe: with this therapeutic goal two-thirds of the European CHD patients would need an intensified cholesterol-lowering action. This intensified action has to include, in addition to a more effective dietary management, a more extensive and effective use of lipid-lowering drugs. Evidence from trials of cholesterol-lowering in secondary prevention of CHD suggests that a full use of this potential in the management of European patients with clinically established CHD would lead to substantial reductions in mortality and in the occurrence of recur-

rent non-fatal coronary events, as well as in the need for revascularization procedures. Application of an even stricter therapeutic goal, total cholesterol B4.5 mmol/l (LDL cholesterol B 2.5 mmol/l), still pending confirmation from ongoing trials, would mean that more than 80% of the patients should be subjected to intensified cholesterol-lowering action. Acknowledgements EUROASPIRE Study Group is grateful to the administrative staff, physicians, nurses and other personnel at all the hospitals in which the study was carried out. We are also grateful to the patients who participated in the survey. The EUROASPIRE study was supported by an educational grant made by Merck, Sharp & Dohme to the European Society of Cardiology. References [1] Roussow JE, Lewis B, Rifkind BM. The value of lowering cholesterol after myocardial infarction. N Engl J Med 1990;323:1112– 9. [2] Vos J, de Feyter J, Simoons ML, Tiijsen JGP, Deckers JW. Retardation and arrest of progression or regression of coronary artery disease: a review. Progr Cardiovasc Dis 1993;35:435–54. [3] The Scandinavian Simvastatin Survival Study Group. Randomized trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S), Lancet 1994;344;1383– 9. [4] Sacks FM, Pfeffer MA, Moye LA et al., for the Cholesterol and Recurrent Events Trial Investigators. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels, N Engl J Med 1996;335:1001–9. [5] The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels, N Engl J Med 1998; 339:1349 – 57. [6] Pyo¨ra¨la¨ K, De Backer G, Graham I, Poole-Wilson P, Wood D. On behalf of the Task Force. Prevention of coronary heart disease in clinical practice. Recommendations of the Task Force of the European Society of Cardiology, European Atherosclerosis Society and European Society of Hypertension, Eur Heart J 1994;15:1300 – 31. [7] National Cholesterol Education Program. Second Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II), Circulation 1994;89:1329 – 445. [8] EUROASPIRE Study Group. EUROASPIRE. A European Society of Cardiology survey of secondary prevention of coronary heart disease: principal results, Eur Heart J 1997;18:1569–82. [9] ASPIRE Steering Group. A British Cardiac Society survey of the potential for the secondary prevention of coronary heart disease: ASPIRE (Action on Secondary Prevention Through Intervention to Reduce Events). Principal results, Heart 1996;75:334–42. [10] Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. West of Scotland Coronary Prevention Study Group. N Engl J Med 1995;333:1301– 7.

D. Vanuzzo et al. / Atherosclerosis 153 (2000) 505–517 [11] Downs JR, Clearfield M, Wis S, et al. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: result of AFCAPS/TexCAPS. Air Force/Texas Coronary Atherosclerosis Prevention Study. JAMA 1998;279:1615– 22. [12] Wood D, De Backer G, Faergeman O, Graham I, Mancia K, Pyo¨ra¨la¨ K, together with members of the Task Force. Prevention of coronary heart disease in clinical practice. Recommendations of the Second Task Force of European and other Societies on Coronary Prevention, Eur Hear J 1998;19:1434–1503. [13] Smith SC, Blair SN, Criqui MH, et al. AHA Consensus Panel Statement. Preventing heart attack and death in patients with coronary disease. The Secondary Prevention Panel. J Am Coll Cardiol 1995;26:292–4. [14] Grundy SM, Baladt GJ, Criqui MH, et al. When to start cholesterol-lowering therapy in patients with coronary heart disease. A statement for healthcare professionals from the American Heart Association Task Force on Risk Reduction. Circulation 1997;95:1683 – 5. [15] Friedewald WT, Levy RI, Fredrickson DS. Estimation of low density lipoprotein cholesterol concentration without the use of the preparative ultracentrifuge. Clin Chem 1972;18:499–502. [16] Lehto S, Paloma¨ki P, Miettinen H, et al. Serum cholesterol and high density lipoprotein cholesterol distributions in patients with acute myocardial infarction and in the general population of Kuopio province, Eastern Finland. J Intern Med 1993;233:179 – 85.

.

517

[17] Rosengren A, Hagman M, Wedel H, et al. Serum cholesterol and long-term prognosis in middle-aged men with myocardial infarction and angina pectoris. A 16-year follow-up of the Primary Prevention Study in Go¨teborg, Sweden. Eur Heart J 1997;18:754 – 61. [18] Kuulasmaa K, Tunstall-Pedoe H, Dobson A et al., for the WHO MONICA Project. Estimation of contribution of changes in classic risk factors to trends in coronary-event rates across the WHO MONICA Project populations, Lancet 2000; 355: 675–87. [19] Erkkila¨ AT, Sarkkinen ES, Koukkunen H, et al. Concordance of diet with the recommended cholesterol lowering diet in patients with coronary heart disease. Eur J Clin Nutr 1998;52:279–85. [20] Post Coronary Artery Bypass Graft Trial Investigators. The effect of aggressive lowering of low-density lipoprotein cholesterol levels and low-dose anticoagulation on obstructive changes in saphenous-vein bypass grafts, N Engl J Med 1997;336:153– 62. [21] Pedersen TR, Olsson AG, Faergeman O et al, for the Scandinavian Simvastatin Survival Study Group. Lipoprotein changes and reduction in the incidence of major coronary heart disease events in the Scandinavian Simvastatin Survival Study (4S), Circulation 1998;97:1453 – 60. [22] Rubins HB, Robins SJ, Collins D et al. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Intervention Trial Study Group, N Engl J Med 1999;341:410 – 8.