Follow-Up Study of Patients Randomized

in the Scandinavian Simvastatin Survival

Study (4S) of Cholesterol Lowering
Terje R. Pedersen, MD, Lars Wilhelmsen, MD, Ole Færgeman, MD,
Timo E. Strandberg, MD, Gudmundur Thorgeirsson, MD, Linda Troedsson, MScPHARM,
Johan Kristianson, MScPHARM, Kåre Berg, MD, Thomas J. Cook, MS,
Torben Haghfelt, MD, John Kjekshus, MD, Tatu Miettinen, MD, Anders G. Olsson, MD,
Kalevi Pyörälä, MD, and Hans Wedel, PhD, on behalf of the
Scandinavian Simvastatin Survival Study Group*

The Scandinavian Simvastatin Survival Study (4S) and mmol/L (200 mg/dl), or placebo. After the double-blind
other randomized clinical trials have demonstrated that period, most patients in both treatment groups received
cholesterol-lowering treatment with statins improves simvastatin as open-label prescription. Of the 1,967
prognosis in patients with coronary atherosclerosis com- patients originally treated with placebo and surviving
pared with placebo. The effect of therapy with statins the double-blind period, 97 (4.9%) died during the fol-
beyond the typical 5 to 6 years’ duration of the trials, in lowing 2 years. In the group randomized to simvastatin
particular regarding the risk of cancer, has not been the corresponding number was 74 of the 2,039 survi-
investigated. This study examines the long-term effects vors (3.6%). Adding these deaths to those occurring
of simvastatin for up to 8 years on cause-specific mor- during the original trial, the total was 353 (15.9%) and
tality in patients with coronary heart disease (CHD). We 256 (11.5%) deaths in the groups originally randomized
performed an observational, government registry-based
to placebo and simvastatin, respectively. The relative
study of mortality in the groups originally randomized to
risk was 0.70 (95% confidence interval 0.60 to 0.82,
simvastatin or placebo in the 4S over an additional
p ⴝ 0.00002). The total number of cancer deaths was
2-year follow-up period, so that the median total fol-
68 (3.1%) in the placebo group and 52 (2.3%) in the
low-up period was 7.4 years (range 6.9 to 8.3 in sur-
viving patients). Randomization took place at outpatient simvastatin group (relative risk 0.73, 95% confidence
clinics at 94 clinical centers in Denmark, Finland, Ice- interval 0.51 to 0.05, p ⴝ 0.087), and the numbers of
land, Norway, and Sweden from 1988 to 1989. Of noncardiovascular and other deaths were similar in both
4,444 patients with CHD, 2,223 and 2,221 were ran- groups. We therefore conclude that treatment with sim-
domized to treatment with placebo or simvastatin ther- vastatin for up to 8 years in patients with CHD is safe
apy, respectively. Patients received treatment with sim- and yields continued survival benefit. 䊚2000 by Ex-
vastatin, starting at 20 mg/day, with titration to 40 cerpta Medica, Inc.
mg/day at 12 or 24 weeks if total cholesterol was >5.2 (Am J Cardiol 2000;86:257–262)

S everal randomized controlled trials have now

demonstrated that long-term treatment with cer-
tain hydroxy-methyl-glutaryl-coenzyme A reductase
inhibitors (statins) to lower serum cholesterol can
reduce the incidence of cardiovascular events.1– 6 The
trials have had a duration of 5 to 6 years, and the
Kaplan-Meier curves for coronary heart disease
From the Aker Hospital, University of Oslo, Oslo, Norway; Sahlgrenska (CHD) events have started to separate after 1 to 2
Hospital, University of Gothenburg, Gothenburg, Sweden; Århus Univer- years. No trial has found increased risk of cancer or
sity Hospital, Århus, Denmark; Helsinki University Hospital, Helsinki, any other life-threatening adverse effects of statin
Finland; Landspitalinn University Hospital, Reykjavik, Iceland; MSD
Scandinavia, Stockholm, Sweden; Ullevål Hospital, University of
treatment for up to 6 years. Nevertheless, there is still
Oslo, Norway; Merck Research Laboratories, Rahway, New Jersey; some concern that cholesterol-lowering treatment for
Odense University Hospital, Odense, Denmark; The National Hospi- longer periods may increase the risk of noncardiovas-
tal, University of Oslo, Oslo, Norway; Linköping University Hospital, cular mortality including cancer and thereby offset the
Linköping, Sweden; Kuopio University Hospital, Kuopio, Finland, and beneficial effects on CHD.7–10 After completion of the
Nordic School of Public Health, Gothenburg, Sweden. This study was
Scandinavian Simvastatin Survival Study (4S),1 in
supported by Merck & Co Inc., Rahway, New Jersey. Manuscript
received November 1, 1999; revised manuscript received and ac- which the median duration of follow-up was 5.4 years,
cepted February 21, 2000. patients were followed for an additional 2 years, with
Address for reprints: Terje R. Pedersen, MD, Cardiology Depart- treatment at the discretion of the patient’s physician.
ment, University of Oslo, Aker Hospital, N 0514 Oslo, Norway. Because appreciable separation of the Kaplan-Meier
E-mail: t.r.pedersen@ioks.uio.no. survival curves representing the simvastatin and pla-
*Members of the Scandinavian Simvastatin Survival Group have been cebo groups was not seen until about 2 years after
listed in a previous publication.1 randomization, we hypothesized that continued sur-

©2000 by Excerpta Medica, Inc. All rights reserved. 0002-9149/00/$–see front matter 257
The American Journal of Cardiology Vol. 86 August 1, 2000 PII S0002-9149(00)00910-3
vival benefit, expressed as maintenance or further the original double-blind period, based on the death
widening of the gap between the 2 mortality curves, certificate information in the registries. The study pro-
would be observed regardless of the proportion of tocol with its extension was approved by regional, or
patients in the 2 treatment groups receiving simvasta- if applicable, national ethics committees and by the
tin or other lipid-lowering therapy. In this report, we regulatory agencies in each of the participating Scan-
document the mortality risk reduction attributable to dinavian countries.
simvastatin therapy after up to 8.3 years of follow-up Statistical analyses: A pilot study comparing cause
(median 7.4 years). of death in the original study as determined by the End
point Classification Committee with that recorded on
METHODS death certificates had revealed inconsistencies, mainly
Double-blind period: The design and results of the in the cause of death subcategories. For this reason
original study have been published previously.1 In deaths were categorized into only 4 classes, which in
brief, this was a randomized, double-blind, placebo- the pilot study reduced the error rate to 2.7%. The
controlled trial performed in 94 clinical centers groups originally randomized to placebo and simva-
throughout the 5 Nordic countries: Denmark, Finland, statin were compared according to the intention-to-
Iceland, Norway, and Sweden. Men and women aged treat principle. Treatment group differences, relative
35 to 70 years with a history of myocardial infarction risk, and 95% confidence intervals (CIs) were calcu-
or angina pectoris and serum total cholesterol in the lated using the Cox proportional-hazards regression
range 5.5 to 8.0 mmol/L (212 to 309 mg/dl) and model. Covariates used in these analyses were sex,
triglycerides ⱕ2.5 mmol/L (221 mg/dl) and no exclu- age, hypertension, and history of myocardial infarc-
sion criteria were randomized to treatment with sim- tion, smoking, and diabetes. Two-sided p values
vastatin 20 mg/day or placebo. Patients in the simva- ⱕ0.05 were regarded as significant.
statin group with serum total cholesterol ⬎5.2
mmol/L after 6 or 18 weeks had their dosage increased RESULTS
to 40 mg (37% of patients required titration, 72% Four thousand forty-four patients with CHD were
reached the goal of 5.2 mmol/L at 12 months follow- randomized in 1988 and 1989 (Figure 1). The placebo
up). and simvastatin treatment groups were well matched
The primary end point was total mortality, but at baseline (Table I). Median follow-up at the com-
secondary end points included cause-specific mortal- pletion of the double-blind period was 5.4 years
ity, which was classified by an independent end point (range of those surviving 4.9 to 6.3). In this period,
classification committee using hospital records, death 13% of the patients in the placebo group and 10% in
certificates, and, if necessary, through contacts with the simvastatin group stopped taking their study med-
physicians and relatives. ication. Adverse events were the reported reason in
Extended follow-up: After the cutoff date of the 6% of the randomized patients in both groups. Over
double-blind study, August 1, 1994, surviving patients the entire double-blind duration of the study, mean
returned to the clinic for an interview and physical changes from baseline in the simvastatin group in
examination. The steering committee advised that all total, low-density lipoprotein (LDL) and high-density
patients should be treated with simvastatin 20 mg/day lipoprotein cholesterol, and triglycerides were ⫺25%,
until the study results had been published. After pub- ⫺35%, ⫹8%, and ⫺10%, respectively. The corre-
lication on November 19, 1994, patients were invited sponding changes in the placebo group were ⫹1%,
to attend a briefing on the results of the study. Further ⫹1%, ⫹1%, and ⫹7%, respectively. These changes
treatment was left to the discretion of each patient’s are by intention-to-treat analysis, including patients
doctor. who stopped taking the study drug but continued to
The cutoff date of the extended follow-up was set contribute with blood samples.
to August 1, 1996, 2 years after the cutoff date of the Open-label lipid-lowering treatment: Questionnaires
base study. One year later, when governmental were sent to 3,731 patients reported to be alive in the
records of vital status were complete as of the cutoff autumn of 1997, 1,824 in the group originally ran-
date, a comparison of these records and the 4S patient domized to placebo and 1,907 in the original simva-
register was made. Patients reported to be alive at this statin group. The response rate was 88.4% in the
comparison were sent letters 3.4 years after comple- original placebo group and 89.4% in the original sim-
tion of the double-blind period, informing them about vastatin group. Of those responding, 81.5% in the
the follow-up study and asking them to complete an original placebo group reported that they were taking
enclosed questionnaire about the use of cholesterol- cholesterol-lowering drugs, versus 85.5% of the
lowering drugs and their latest cholesterol level. The former simvastatin group (p ⫽ 0.002) (Table II). Of
4S patient registers were compared with the national those taking cholesterol-lowering drugs, simvastatin
death by cause registers, which are derived from death was used by 96.4% in the former placebo group and
certificate information. The comparison was made 2 96.3% in the former simvastatin group. The mean
years after the completion of the extended follow-up extension values for serum cholesterol in all patients
to allow for updating and completion of the registers. reporting such data were 5.16 and 5.11 mmol/L in the
Cause of deaths occurring in the 2-year extension groups originally randomized to placebo and simva-
period was ascertained by 2 of the investigators (TRP statin, respectively. The drug reimbursement rules of
and LW) without knowledge of treatment allocation in governments differ among the participating countries.

258 THE AMERICAN JOURNAL OF CARDIOLOGY姞 VOL. 86 AUGUST 1, 2000

 TABLE I Selected Baseline Characteristics of Randomized
Patients
Placebo Simvastatin
Baseline Characteristics (n ⫽ 2,223) (n ⫽ 2,221)

No (%) of patients
Men 1,803 (81) 1,814 (82)
Women 420 (19) 407 (18)
Age ⱖ65 years 503 (23) 518 (23)
LDL cholesterol 3.0–4.5 628 (28) 651 (29)
mmol/L
Qualifying diagnosis (%)
Angina only 456 (21) 462 (21)
Myocardial infarction 1,767 (79) 1,759 (79)
Secondary diagnoses (%)
Systemic hypertension 584 (26) 570 (26)
Claudication 123 (6) 130 (6)
Diabetes mellitus 97 (4) 105 (5)
Cholesterol (mmol/L) Mean (SD)
Total 6.75 (0.66) 6.74 (0.67)
High-density lipoprotein 1.19 (0.29) 1.18 (0.30)
Low-density lipoprotein 4.87 (0.65) 4.87 (0.66)
Triglycerides (mmol/L) 1.51 (0.52) 1.49 (0.49)

TABLE II Use of Cholesterol-Lowering Drugs 3.4 Years After

Completion of Double-Blind Study Among Patients’
Returning Questionnaires
Previous Previous
Placebo Group Simvastatin Group
Drugs Used (n ⫽ 1,824) (n ⫽ 1,907) p Value

Simvastatin 1,268 (78.6)* 1,404 (82.2)* 0.005

Other statin 89 (5.5) 68 (4.0) 0.046
Fibrate 6 (0.4) 11 (0.6) 0.261
Resin 6 (0.4) 12 (0.7) 0.186
Other 5 (0.3) 7 (0.4) 0.616
No drug 298 (18.5) 248 (14.5) 0.004
Unknown 211 (11.6)† 201 (10.5)† 0.317

*Percentage of patients responding to the questionnaire.

†
Percentage of those that were sent questionnaires.
Patients could be counted under ⬎1 drug.
Values are expressed as number (%) of patients.

For the extended follow-up, the vital status of all

patients randomized to placebo or simvastatin treat-
FIGURE 1. Profile of 4S. ment was confirmed as of August 1, 1996. At this
date, 97 patients (4.9%) of the 1967 originally treated
with placebo and surviving the double-blind study
Mainly for this reason, open-label prescription of sim- period had died. In the group randomized to simva-
vastatin occurred less frequently in Denmark and Fin- statin the corresponding number was 74 of the 2,039
land than in the other countries. survivors (3.6%). Adding these deaths to those occur-
Median follow-up time for the double-blind plus ring during the double-blind study period, there were
the extension period was 7.4 years (range of those 353 deaths (15.9%) in the group originally random-
surviving 6.9 to 8.3 years). Vital status was confirmed ized to placebo and 256 (11.5%) in the original sim-
in all randomized patients and cause of death was vastatin group (Table III). The relative risk was 0.70
classified in all cases. (95% CI 0.60 to 0.82, p ⫽ 0.00002). The Kaplan-
Mortality: During the double-blind period, 256 of Meier 8-year (96 months) survival rate (Figure 2) was
the patients (11.5%) in the placebo group and 182 0.84 in the original placebo group and 0.88 in the
(8.2%) in the simvastatin group died (Table III). The original simvastatin group. The increase in the abso-
relative risk in the simvastatin group was 0.70 (95% lute difference was due not only in fewer coronary
CI 0.58 to 0.85, p ⫽ 0.0003) due to a 42% risk deaths in the simvastatin group, but also in less cancer
reduction in coronary deaths. The numbers of deaths deaths (Table III), although the cancer death differ-
from cerebrovascular and other cardiovascular causes ence did not reach conventional levels of statistical
were similar, as were numbers for deaths caused by significance (p ⫽ 0.087) (Figure 3.). Given the ob-
cancer, suicide, and violence. served difference in favor of simvastatin, the proba-

CORONARY ARTERY DISEASE/CHOLESTEROL, CORONARY HEART DISEASE, PROGNOSIS 259

 TABLE III Mortality and Causes of Death
Placebo (n ⫽ 2,223) Simvastatin (n ⫽ 2,221)
Relative Risk*
Causes of Death In-Trial Extension Total In-Trial Extension Total (95% CI)

All coronary 189 50 239 (10.8) 111 42 153 (6.9) 0.62 (0.51–0.76)
Other cardiovascular 18 10 28 25 7 32
All cardiovascular 207 59 267 (12.0) 136 49 185 (8.3) 0.67 (0.56–0.81)
Malignancy 35 33 68 (3.1) 33 19 52 (2.3) 0.73 (0.51–1.05)
Other noncardiovascular 14 4 18 (0.8) 13 6 19 (0.9)
All noncardiovascular 49 37 86 (3.9) 46 25 71 (3.2) 0.80 (0.58–1.09)
All deaths 256 97 353 (15.9) 182 74 256 (11.5) 0.70 (0.60–0.82)

*Relative risk calculated by Cox regression analysis.

Values are expressed as number (%) of patients.

FIGURE 2. Kaplan-Meier curves for all-cause mortality. FIGURE 3. Kaplan-Meier curves for cancer mortality.

bility that simvastatin increases the risk of cancer

within the time frame of the study by ⱖ10% is very quently treated with nonstudy cholesterol-lowering
small (⬍5%). For other cardiovascular and noncardio- drugs after discontinuation (35 of 291 patients in the
vascular deaths the numbers were similar for the 2 placebo group, and 2 of 230 patients in the simvastatin
groups. group). The mortality in the simvastatin group during
Mortality was analyzed separately for patients aged the trial was about twice as high for withdrawn pa-
ⱖ65 years at the time of randomization and for pa- tients than for those not withdrawn from study therapy
tients with baseline serum LDL cholesterol ⬍4.5 (Table IV). In the extension period, patients not with-
mmol/L (174 mg/dl), because these are categories of drawn from the simvastatin group continued to have a
patients for which there is still some debate about the much lower mortality than those withdrawn; this was
survival benefit of cholesterol-lowering therapy. In also the case in the placebo group.
patients ⱖ65 years at randomization, the number of
deaths during the entire 8-year follow-up was 120 in DISCUSSION
the placebo group (23.9%) and 92 in the simvastatin The 4S with the 2-year extension represents the
group (17.8%); the relative risk was 0.72 (95% CI longest clinical trial experience to date in a large
0.55 to 0.94, p ⫽ 0.02) (Figure 4). In patients with cohort of patients taking statins. Although nonfatal
LDL cholesterol less than the median value of 4.5 events could not be studied after completion of the
mmol/L, the number of deaths during 8 years was 96 double-blind trial, mortality is the ultimate test of
(15.3%) and 72 (11.1%) in the placebo and simvasta- long-term safety. With no suggestion of harmful ef-
tin groups, respectively (relative risk 0.70, 95% CI fects, the study provides reassuring evidence of the
0.52 to 0.96, p ⫽ 0.02) (Figure 5). long-term safety of the drug. In particular, there were
During the double-blind part of the trial, patients fewer cancer deaths with long-term simvastatin ther-
who stopped taking study medication were infre- apy, although the difference was not significant. This

260 THE AMERICAN JOURNAL OF CARDIOLOGY姞 VOL. 86 AUGUST 1, 2000

FIGURE 4. Kaplan-Meier curves for mortality in patients aged FIGURE 5. Kaplan-Meier curves for mortality in patients with LDL
>65 years at baseline. cholesterol <4.5 mmol/L at baseline.

result is in contrast to that of the extended follow-up TABLE IV Mortality Among Patients Withdrawn from Study
of the Helsinki Heart Study9,10; after a total observa- Medication and Those Not Withdrawn
tion period of 8.5 years, there was a trend to a higher Placebo Simvastatin
mortality rate in the gemfibrozil group (101 deaths)
than in the placebo group (83 deaths). This was attrib- Patients withdrawn n ⫽ 291 n ⫽ 230
Deaths in trial* 50 (17.2) 39 (17.0)
utable mainly to more cancer deaths in the gemfibrozil Deaths in extension 24 (10.0)† 16 (8.4)†
group (30 vs 18 in the placebo group, p ⫽ 0.08). Total deaths 74 (25.4) 55 (23.9)
Reduction in all-cause mortality was demonstrated Patients not withdrawn n ⫽ 1,932 n ⫽ 1,991
in patients with relatively low LDL cholesterol levels, Deaths in trial‡ 206 (10.7) 143 (7.2)
⬍4.5 mmol/L (174 mg/dl) (n ⫽ 1,279) at baseline. Deaths in extension 73 (4.2)† 58 (3.1)†
Patients in the simvastatin group experienced a 33% Total deaths 279 (14.4) 201 (10.1)
mean reduction in LDL cholesterol from a mean base- *Deaths occurring after ⬎28 days after withdrawal.
line level of 4.10 mmol/L at baseline to a mean level †
Denominator is number of patients living at the end of the original study.
‡
of 2.73 mmol/L at year 1. As in the entire study Deaths occurring while on study medication or within 28 days of with-
population, there was a 30% risk reduction (p ⫽ drawal.
Values are expressed as number (%) of patients.
0.024) by the end of the extension period. Reducing
cholesterol to low levels therefore does not seem to
introduce any long-term risk. Patients aged 65 to 70
years at baseline (n ⫽ 1,021) also had increased sur- or other statins was probably more widespread among
vival with simvastatin therapy (28% relative risk re- patients not withdrawn from the original simvastatin
duction, p ⫽ 0.016). These patients were aged 73 to group. The low mortality in these patients during the
78 years at the completion of the extended follow-up, extension adds further evidence to the safety and
and were thus at an age when cancer is becoming long-term protective effect of simvastatin. (A statisti-
more prevalent. There were no signs of excess risk of cal comparison of mortality in patients withdrawn
cancer or other causes of mortality in this subgroup from the placebo and simvastatin groups would not be
either. a randomized comparison, and would therefore be
Patients withdrawn from double-blind therapy had problematic.)
a much higher mortality rate than those adhering to the The absolute mortality difference increased from
study regimen. Possibly, development of disease 3.3% to 4.4%, maintaining the 30% reduction in the
tended to cause withdrawal. The deaths among these mortality risk in the simvastatin group recorded at the
patients accounted for 22% of the deaths in the sim- end of the base study. Post hoc analysis of the relation
vastatin group and 25% in the placebo group during between simvastatin-induced changes in serum li-
the extension. These patients were unlikely to be poprotein concentrations and the subsequent risk of
given statins after the double-blind trial, although in- CHD during the double-blind trial indicated that re-
formation about the use of statins by patients who had duction in LDL cholesterol by simvastatin could ex-
died was unavailable. However, the use of simvastatin plain most of the effect,11 with some contribution from

CORONARY ARTERY DISEASE/CHOLESTEROL, CORONARY HEART DISEASE, PROGNOSIS 261

raising HDL cholesterol. A meta-analysis by Law et 5. Downs JR, Clearfield M, Weis S, Whitney E, Shapiro DR, Beere PA,
Langendorfer A, Stein EA, Kruyer W, Gotto AM. Primary prevention of acute
al12 of several large cohort studies showed that a coronary events with lovastatin in men and women with average cholesterol
decrease in serum cholesterol concentration of 0.6 levels. Results of AFCAPS/TexCAPS. JAMA 1998;279:1615–1622.
mmol/L (23 mg/dl) (approximately 10%) was associ- 6. Gould AL, Rossouw JE, Santanello NC, Heyse JF, Furberg CD. Cholesterol
reduction yields clinical benefit. Impact of statin trials. Circulation 1998;97:946 –
ated with a decrease in the incidence of ischemic heart 952.
disease, ranging from 54% at the age of 40 years to 7. Committee of Principal Investigators. A co-operative trial in the primary
19% at 80 years. This suggests that even greater prevention of ischemic heart disease using clofibrate. Br Heart J 1978;40:1069 –
preventive effects could be obtained if treatment were 1118.
8. Newman TB, Hulley SB. Carcinogenicity of lipid-lowering drugs. JAMA
started earlier in life in patients at risk of developing 1996;275:55– 60.
atheroscleotic disease compared with the benefits 9. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, Huttunen JK,
achieved in randomized trials. Because of the evi- Kaitaniemi P, Koskinen P, Manninen V, Maenpaa H, Malkonen M, Manttani M,
Norola S, Pasternack A, Pikkarainen J, Romo M, Sjöblom T. Helsinki Heart
dence for the long-term safety of simvastatin derived Study. Primary-prevention trial with gemfibrozil in middle-aged men with dys-
from 4S13 and 10 years of postmarketing surveillance, lipidemia (safety of treatment, changes in risk factors, and incidence of coronary
this approach is not unreasonable. heart disease). N Engl J Med 1987;317:1237–1245.
10. Huttunen JK, Heinonen OP, Manninen V, Koskinen O, Hakulines T, Teppo
L, Mänttäri M, Frick MH. The Helsinki Heart Study: an 8.5-year safety and
mortality follow-up. J Intern Med 1994;235:31–39.
1. Scandinavian Simvastatin Survival Study Group. Randomized trial of choles- 11. Pedersen TR, Olsson AG, Faergeman O, Kjekshus J, Wedel H, Berg K,
terol lowering in 4444 patients with coronary heart disease: the Scandinavian Wilhelmsen L, Haghfelt T, Thorgeirsson G, Pyörälä K, Miettinen TA, Christo-
Simvastatin Survival Study (4S). Lancet 1994;344:1383–1389.
phersen B, Tobert JA, Musliner TA, Cook TJ. Lipoprotein changes and reduction
2. The Long-Term Intervention With Pravastatin In Ischaemic Disease (LIPID)
in the incidence of major coronary heart disease events in the Scandinavian
Study Group. Prevention of cardiovascular events and death with pravastatin in
patients with coronary heart disease and a broad range of initial cholesterol levels. Simvastatin Survival Study (4S). Circulation 1998;97:1453–1460.
N Engl J Med 1998;339:1349 –1357. 12. Law MR, Wald NJ, Thompson SG. By how much and how quickly does
3. Shepherd J, Cobbe SM, Ford I, Isles, CG, Lorimer AR, McFarlane PW, reduction in serum cholesterol concentration lower risk of ischaemic heart dis-
McKillop JH, Packard CJ. Prevention of coronary heart disease with pravastatin ease? Br Med J 1994;308:367–372.
in men with hypercholesterolemia. N Engl J Med 1995;333:1301–1307. 13. Pedersen TR, Berg K, Cook TJ, Færgeman O, Haghfelt T, Kjekshus J,
4. Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG, Miettinen TA, Musliner TA, Olsson AG, Pyörälä K, Thorgeirsson G, Tobert JA,
Brown L, Warnica JW, Arnold JMO, Wun C, Davis BR, Braunwald E. The effect Wedel H, Wilhelmsen L. Safety and tolerability of cholesterol lowering with
of pravastatin on coronary events after myocardial infarction in patients with simvastatin during 5 years in the Scandinavian Simvastatin Survival Study. Arch
average cholesterol levels. N Engl J Med 1996;335:1001–1009. Intern Med 1996;156:2085–2092.

262 THE AMERICAN JOURNAL OF CARDIOLOGY姞 VOL. 86 AUGUST 1, 2000