What was the primary endpoint of the 4S trial?

All-cause mortality. The investigators chose this over cardiovascular-specific mortality because it requires no adjudication and cannot be biased by endpoint classification decisions. The result was a 30% relative risk reduction (p = 0.0003) favoring simvastatin over placebo.

How many patients were enrolled in the 4S trial?

A total of 4,444 patients with established coronary heart disease were randomized: 2,223 to simvastatin and 2,221 to placebo. All had baseline total cholesterol between 5.5 and 8.0 mmol/L after an eight-week dietary run-in.

How was blinding maintained despite dose titration?

Dose adjustments (from 20 mg to 40 mg simvastatin or matching placebo) were managed centrally. Local investigators received titration instructions without seeing the cholesterol values that triggered the change, reducing the risk that lipid responses would unblind them.

Did the 4S trial include women?

Yes, but women made up only 19% of the study population (827 of 4,444). Subgroup analysis showed a consistent trend toward benefit, but the trial was not powered to demonstrate a statistically significant mortality reduction in women alone. Later trials like HPS confirmed statin benefits in women.

Was 4S a primary or secondary prevention trial?

Strictly secondary prevention. All patients had a prior myocardial infarction or stable angina pectoris at enrollment. Primary prevention evidence came later from WOSCOPS (1995) and ASCOT-LLA (2003).

What was the absolute risk reduction in mortality?

The absolute risk reduction was 3.3 percentage points over the median 5.4-year follow-up (11.5% placebo mortality minus 8.2% simvastatin mortality). This translates to a number needed to treat (NNT) of approximately 30 patients treated for 5.4 years to prevent one death.

Did simvastatin increase non-cardiovascular deaths?

No. Non-cardiovascular deaths were nearly identical between groups (67 placebo vs. 71 simvastatin). This finding was critical because earlier cholesterol-lowering trials had raised concerns about excess cancer or violent deaths with lipid reduction.

How was the sample size determined?

The trial used event-driven design, requiring approximately 440 total deaths to achieve 90% power to detect a 30% relative mortality reduction at a two-sided alpha of 0.05. The final trial accumulated 438 deaths, closely matching the target.

What simvastatin dose was used in 4S?

Patients started at 20 mg daily. If total cholesterol remained above 5.2 mmol/L at the 6-week or 18-week visit, the dose was increased to 40 mg. About 37% of simvastatin-group patients were uptitrated. This tested a treat-to-target strategy rather than a fixed-dose approach.

How does the 4S methodology compare to modern statin trials?

Modern trials like PROVE IT-TIMI 22 use fixed high-intensity dosing, composite cardiovascular endpoints, and shorter follow-up. 4S used dose titration, an all-cause mortality primary endpoint, and over five years of follow-up. The 4S approach was more conservative but produced a result that was harder to question.

Inside the 4S Methodology: What Most Summaries Skip

By HealthRX.com Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

Clinical image for Inside the 4S Methodology: What Most Summaries Skip Image: HealthRX.com clinical image

At a glance

| Parameter | Detail | |---|---| | Trial name | Scandinavian Simvastatin Survival Study (4S) | | N | 4,444 | | Intervention | Simvastatin 20 mg daily (titrated to 40 mg if total cholesterol remained above 5.2 mmol/L at 6 or 18 weeks) | | Comparator | Matching placebo | | Duration | Median 5.4 years | | Primary endpoint | All-cause mortality | | Key result | 30% relative risk reduction in all-cause mortality (RR 0.70, 95% CI 0.58 to 0.85, p = 0.0003) |

Why the Design Mattered More Than the Number

Before 4S, clinicians had cholesterol-lowering drugs but no proof that using them kept patients alive. Earlier trials with fibrates and resins showed lipid improvements without mortality benefits. Some even raised safety concerns about non-cardiovascular deaths. The 4S investigators understood that only an all-cause mortality endpoint, not cardiovascular death or lipid changes, would settle the debate.

That decision carried a cost. Powering a trial for all-cause mortality requires more patients, longer follow-up, and more money than powering for a surrogate marker. The 4S team accepted that trade-off. It is the reason the result held up under decades of scrutiny while earlier lipid trials did not.

Enrollment: Who Got In and Who Didn't

The inclusion criteria were deliberately narrow. Patients had to be 35 to 70 years old with a history of angina pectoris or prior myocardial infarction (secondary prevention only). Baseline total cholesterol had to fall between 5.5 and 8.0 mmol/L (roughly 213 to 310 mg/dL) after an eight-week dietary run-in period.

The exclusion list was extensive:

Premenopausal women
Patients with secondary hyperlipidemia (hypothyroidism, nephrotic syndrome)
Unstable angina or MI within the preceding six months
Heart failure requiring treatment (NYHA class III or IV)
Patients already on lipid-lowering therapy
Significant hepatic or renal disease

This created a population of stable, moderate-to-high-risk coronary patients with elevated but not extreme cholesterol. The investigators were not testing whether statins work in the general population. They selected a group where the biological hypothesis (lower LDL reduces atherosclerotic events) had the highest prior probability of producing a detectable mortality signal.

The HRX Enrollment Stringency Framework

When reading any RCT, the enrollment criteria tell you exactly who the result applies to and, just as importantly, who it does not. We score enrollment on three axes:

| Axis | 4S Rating | Implication | |---|---|---| | Disease-stage specificity | High (established CHD only) | Result cannot be extrapolated to primary prevention without separate trials | | Biomarker window | Moderate (TC 5.5-8.0 mmol/L) | Excludes patients with very high or near-normal cholesterol | | Comorbidity exclusion | High (heart failure, liver disease, renal disease excluded) | Cleaner signal but limits applicability to sicker real-world patients |

This framework explains why 4S alone did not change primary prevention guidelines. That required WOSCOPS (1995) and subsequent trials with broader enrollment.

Randomization and Blinding

Randomization was 1:1, stratified by clinical center. The original publication describes a computer-generated randomization sequence with sealed envelopes, standard practice for the era. A total of 2,223 patients received simvastatin and 2,221 received placebo.

Blinding was maintained through matched tablets. The dose-titration protocol (escalation from 20 mg to 40 mg based on cholesterol response at 6 and 18 weeks) posed a potential unblinding risk because clinicians measuring cholesterol could infer treatment assignment from the lipid response. The investigators addressed this by having dose adjustments handled centrally, with local investigators receiving titration instructions without seeing the underlying lipid values that drove the decision.

This is a detail most summaries skip. The dose-titration design was necessary because the investigators wanted to achieve a target cholesterol reduction (below 5.2 mmol/L). But any titration based on a biomarker that the drug obviously changes creates a crack in the blind. The central titration model was a reasonable patch, though not a perfect seal.

The Eight-Week Dietary Run-In

Before randomization, all patients underwent an eight-week dietary stabilization period following a lipid-lowering diet (NCEP Step I equivalent). This served two purposes. First, it established stable baseline lipid values unaffected by recent dietary changes. Second, it screened out non-adherent patients. Those who could not follow a simple dietary protocol for eight weeks were unlikely to take a pill for five years.

Run-in periods improve internal validity by enriching for compliant patients. They also reduce generalizability. The 4S population was, by design, more adherent than average clinical practice. This is worth remembering when comparing the trial's event rates to real-world registry data.

Primary Endpoint: All-Cause Mortality

The primary endpoint was total mortality from any cause. This was not a composite. It was not cardiovascular mortality. It was the hardest, most conservative endpoint available in clinical research.

Why this matters: a cardiovascular mortality endpoint requires adjudication. Someone has to decide whether a death was "cardiac" or not. That process introduces judgment calls and potential bias. All-cause mortality requires only a death certificate. The patient is alive or dead. There is no adjudication ambiguity.

The 4S investigators reported 256 deaths in the placebo group versus 182 in the simvastatin group over the median 5.4-year follow-up. The 30% relative risk reduction (p = 0.0003) was driven primarily by reduced coronary deaths (189 placebo vs. 111 simvastatin), but the all-cause framing meant that any excess in non-cardiovascular deaths (cancer, suicide, trauma) would have diluted the signal. It did not. Non-cardiovascular mortality was nearly identical between groups (67 placebo vs. 71 simvastatin), directly addressing the concern from earlier cholesterol trials that lipid lowering might increase non-cardiac deaths.

Secondary Endpoints and the Full Results Picture

| Endpoint | Placebo (n = 2,221) | Simvastatin (n = 2,223) | RR (95% CI) | p-value | |---|---|---|---|---| | All-cause mortality | 256 (11.5%) | 182 (8.2%) | 0.70 (0.58-0.85) | 0.0003 | | Coronary death | 189 (8.5%) | 111 (5.0%) | 0.58 (0.46-0.73) | <0.00001 | | Major coronary events | 622 (28.0%) | 431 (19.4%) | 0.66 (0.59-0.75) | <0.00001 | | Any coronary event | 822 (37.0%) | 559 (25.1%) | 0.63 (0.57-0.70) | <0.00001 | | Revascularization (CABG or PTCA) | 383 (17.2%) | 252 (11.3%) | 0.63 (0.54-0.74) | <0.00001 |

The consistency across endpoints is notable. Simvastatin reduced coronary deaths, non-fatal MIs, and the need for revascularization procedures by roughly similar proportions. This internal coherence strengthened the causal interpretation.

Statistical Design and Power Calculations

The trial was designed as event-driven, not time-driven. The protocol specified that enrollment would continue until enough deaths accumulated to provide 90% power to detect a 30% relative reduction in total mortality at a two-sided alpha of 0.05. The required number of events was approximately 440 deaths.

The actual trial accumulated 438 deaths, almost exactly hitting the target. This is a sign of good trial management and reasonable initial assumptions about event rates. The placebo group's mortality rate (11.5% over 5.4 years) was close to what the investigators predicted based on Scandinavian registry data for stable CHD patients in the late 1980s.

One interim analysis was planned and conducted by an independent Data Safety Monitoring Board. The stopping boundary used an O'Brien-Fleming approach, which is conservative at early looks (requiring very extreme results to stop early) and spends little alpha, preserving nearly full statistical power at the final analysis. The DSMB did not recommend early stopping, and the trial ran to its planned conclusion.

The Estimand Question (Applied Retrospectively)

The 4S trial predates the ICH E9(R1) addendum on estimands by 25 years. But applying the estimand framework retrospectively clarifies what the trial actually measured.

The treatment policy estimand (intention-to-treat) asks: what happens when you assign patients to simvastatin versus placebo, regardless of whether they actually take the drug? The 4S ITT analysis answered this question. Compliance was high (simvastatin group adherence around 89% at year 5), so the ITT estimate was close to the on-treatment effect.

The trial did not pre-specify per-protocol or principal stratum analyses. In a trial with ~90% adherence, the ITT and per-protocol estimates converge, making this less of a concern than in trials with high dropout rates. The intercurrent event most relevant to interpretation is crossover: patients in the placebo group who began open-label statins. The published data show minimal crossover during the blinded phase, partly because statins were not yet standard of care when 4S was running (1988 to 1994).

Limitations the Authors Acknowledged

The original investigators were transparent about several constraints.

Generalizability. The trial enrolled patients in Scandinavia only (94 centers in Denmark, Finland, Iceland, Norway, and Sweden). The population was overwhelmingly white and male (81% men). Whether the mortality benefit extended to women, non-white populations, and patients with lower baseline cholesterol required additional trials. The Heart Protection Study (2002) later demonstrated statin benefits in a broader population, including women and patients with lower LDL.

Dose titration. The two-dose design (20 or 40 mg) means the trial tested a titration strategy, not a fixed dose. About 37% of simvastatin patients were uptitrated to 40 mg. Modern statin prescribing tends toward fixed high-intensity dosing based on the 2013 ACC/AHA cholesterol guidelines, a shift partly informed by later trials like PROVE IT-TIMI 22.

Era-specific care. Background medical therapy in 4S did not include routine use of ACE inhibitors, beta-blockers at current doses, or dual antiplatelet therapy. The absolute risk reduction observed in 4S (3.3 percentage points for mortality) might be smaller today because background cardiovascular mortality is lower. The relative risk reduction, though, has been remarkably consistent across subsequent statin trials.

No lipid-lowering run-in. Unlike some later trials, 4S did not use a statin run-in to exclude patients with side effects before randomization. This means the trial's adverse-event profile reflects real-world tolerability more accurately than trials with active run-in periods, but it also means some randomized patients may never have achieved therapeutic drug levels.

What Changed Because of This Trial

The FDA approved simvastatin in 1991 for hypercholesterolemia, but the 4S results in 1994 transformed statins from lipid-lowering agents into life-saving drugs. Within two years of publication, statin prescribing rates in Scandinavia tripled. Guidelines from the AHA and ESC shifted from treating cholesterol numbers to treating cardiovascular risk, with statins as first-line therapy.

The trial's influence extended beyond simvastatin. By proving the mortality hypothesis, 4S justified the subsequent wave of large-scale statin trials (CARE, LIPID, WOSCOPS, HPS, ASCOT-LLA) that expanded the evidence base to primary prevention, lower-risk patients, and different statins. Every modern guideline recommendation for statin therapy traces a direct line back to 4S.

Frequently asked questions

›

References

Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet. 1994;344(8934):1383-1389. PubMed
Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals. Lancet. 2002;360(9326):7-22. PubMed
Stone NJ, Robinson JG, Lichtenstein AH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults. Circulation. 2014;129(25 Suppl 2):S1-S45. PubMed
Cannon CP, Braunwald E, Murphy SA, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes (PROVE IT-TIMI 22). N Engl J Med. 2004;350(15):1495-1504. PubMed
ICH E9(R1) Expert Working Group. ICH E9(R1) addendum on estimands and sensitivity analysis in clinical trials. Stat Med. 2020;39(27):3981-3992. PubMed