Study of the Effectiveness of Additional Reductions in Cholesterol and Homocysteine: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of the Effectiveness of Additional Reductions in Cholesterol and Homocysteine (NCT00124072), a randomized phase 3 study. The primary outcome is Major Vascular Events (MVE).

2. Clinical Question

3. Trial Design

SEARCH, formally the Study of the Effectiveness of Additional Reductions in Cholesterol and Homocysteine, was a phase 3 randomized trial with a factorial model. Its enrollment was 12,064, and its status is completed. The registry record lists the University of Oxford as sponsor. The trial began in 1998-07, and primary completion was in 2008-05.

The four described combinations paired simvastatin 20 mg or 80 mg daily with folic acid 2 mg plus vitamin B12 1 mg daily or matching placebo. Thus, the study addressed two treatment factors: simvastatin dose and the vitamin combination. The record describes the trial as double-blind and multi-centre in the United Kingdom, and lists quadruple masking. Details of who was masked, how assignment was generated, and how assignment was concealed are not reported in the registry record.

The primary endpoint was major vascular events during the scheduled study treatment period, with a median follow-up of 6.7 years. The posted outcome counts are 1553 for simvastatin 20 mg, 1477 for simvastatin 80 mg, 1537 for folic acid plus vitamin B12, and 1493 for placebo. The registry also lists secondary outcomes, including major coronary events and total strokes, as well as analyses by follow-up period, baseline LDL group, and presence or absence of the other factorial treatment. Numerical results for those secondary outcomes are not reported in the registry record.

The original target sample-size justification, the intended statistical analysis plan, and details of interim monitoring are not reported in the registry record. The planning illustrations associated with the posted comparisons are separate calculations and should not be mistaken for the trial's original design assumptions.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Patients entered a run-in period during which they received simvastatin 20 mg daily and placebo-vitamin tablets for 2 months. Eligible patients who completed run-in were then randomized in a 2x2 factorial blinded design between simvastatin 80 mg daily versus simvastatin 20 mg daily and folic acid 2 mg + vitamin B12 1 mg daily versus placebo.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Major Vascular Events (MVE), measured over 6.7 years median follow-up.

The primary endpoint, major vascular events (MVE), counted participants with an event during the scheduled study treatment period. The definition combined major coronary events—non-fatal myocardial infarction, coronary death, or coronary revascularisation—with fatal or non-fatal stroke and peripheral revascularization. Peripheral procedures included peripheral artery angioplasty or arterial surgery, including amputations.

This composite brings together several serious vascular outcomes in one endpoint. In general, a composite endpoint can capture more events than any one component alone and can provide a broad measure of vascular events relevant to the trial question. It also means that the overall count does not show how much each component contributed. The registry lists some component outcomes as secondary outcomes, but does not report their numerical results.

Because the primary measure is a count of participants with an event, the posted event totals do not show event timing or recurrent events. The median follow-up was 6.7 years, but the primary endpoint is defined over the scheduled treatment period rather than as a specific time-to-first-event estimate in the posted comparisons. Details of endpoint adjudication and any handling of events after treatment discontinuation are not reported in the registry record.

7. Results Posted in the Registry

• Simvastatin 20 mg Daily: 25.7% (1553/6033) versus 24.8% (1493/6031) on control; risk difference 1.0 percentage points (95% CI -0.6 to 2.5), risk ratio 1.04 (0.98–1.11); unadjusted two-proportion z-test p = 0.212.
• Simvastatin 80 mg Daily: 24.5% (1477/6031) versus 24.8% (1493/6031) on control; risk difference -0.3 percentage points (95% CI -1.8 to 1.3), risk ratio 0.99 (0.93–1.05); unadjusted two-proportion z-test p = 0.735.
• Folic Acid 2 mg + Vitamin B12 1 mg Daily: 25.5% (1537/6033) versus 24.8% (1493/6031) on control; risk difference 0.7 percentage points (95% CI -0.8 to 2.3), risk ratio 1.03 (0.97–1.09); unadjusted two-proportion z-test p = 0.361.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: a participant either had a major vascular event during the scheduled treatment period or did not. The registry's outcome summaries give counts of participants with an event and the numbers at risk for each listed group. A risk difference expresses the contrast between two event proportions on an absolute scale. A value above zero means the named intervention group has a higher observed event proportion than its comparator; a value below zero means it has a lower observed event proportion. Zero represents no difference in the proportions.

For the three posted comparisons, the reported risk differences are 0.010 for simvastatin 20 mg, -0.003 for simvastatin 80 mg, and 0.007 for folic acid plus vitamin B12, each relative to the listed placebo comparator. These estimates are unadjusted. The registry also posts relative risks and odds ratios, which express contrasts on different scales. In general, relative risks compare event proportions as a ratio, while odds ratios compare event odds. These measures can answer related but distinct questions, and their numerical values should not be treated as interchangeable.

A confidence interval describes a range of values compatible with the data and the method used to construct the interval. In repeated use of a valid interval procedure, intervals at a stated confidence level have the corresponding coverage property. It is not a probability statement that the particular fixed treatment effect lies inside the calculated interval. For the posted risk differences, each 95% confidence interval includes zero; the intervals also extend to values on both sides of zero. That pattern communicates uncertainty about both the direction and size of the differences.

A p-value evaluates how unusual data at least as discrepant as those observed would be under a specified null hypothesis and the analysis assumptions. It is not a measure of clinical importance, and a larger p-value does not demonstrate that two groups are equivalent. The posted unadjusted p-values are 0.212, 0.735, and 0.361 for the three comparisons. The registry record does not report a multiplicity adjustment for these or other analyses, so the p-values should not be read as adjusted for the number of questions examined.

Factorial designs study more than one treatment factor in a shared randomized trial. In general, this can permit efficient estimation of each factor's association with an outcome, provided the analysis respects the design and handles possible interaction appropriately. Here, the four treatment combinations pair two simvastatin doses with the vitamin combination or matching placebo. The posted comparison summaries use named intervention labels and placebo as comparator. The precise model, whether estimates are marginal or based on particular combinations, and how any interaction was assessed are not reported in the registry record. Those details matter when interpreting factorial results and when mapping a comparison to a specific treatment contrast.

There are 6.7 years median follow-up, but the posted estimates are risk differences for participants with an event, not hazard ratios. General time-to-event analyses account for when events occur and can handle differing follow-up through methods that use event timing and censoring assumptions. Those methods cannot be inferred from these counts, and a time-to-event analysis plan is not reported in the registry record.

The sample-size illustrations are planning calculations based on the observed effects, not the trial's own design calculations. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. In general, the normal approximation is a mathematical simplification for estimating how sample size relates to a target effect and statistical power. Actual design choices may also depend on expected follow-up, outcome frequency, attrition, and the planned analysis, but those original assumptions are not reported in the registry record.

Event counts and proportions also rely on complete and consistently classified outcome information. The registry record does not report the handling of missing outcome data, withdrawals, adherence, or changes in treatment. These omissions limit evaluation of the analysis populations and of whether the posted group summaries reflect an intention-to-treat analysis. None of these unreported choices should be inferred from the numerical results alone.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate a treatment contrast. The posted risk-difference intervals are 0.010 with limits -0.006 and 0.025 for simvastatin 20 mg; -0.003 with limits -0.018 and 0.013 for simvastatin 80 mg; and 0.007 with limits -0.008 and 0.023 for folic acid plus vitamin B12. Each interval covers zero and extends to values on either side. Consequently, each estimate leaves uncertainty about whether the difference favors the named intervention or the comparator.

In general, a confidence interval combines the estimated contrast with its sampling uncertainty. A narrow interval indicates greater precision for the scale and method used; a wide interval indicates that a wider range of effects remains compatible with the data. Precision is distinct from whether an estimate is clinically important. A confidence interval can be statistically informative while still leaving a decision about clinical importance unresolved, particularly when the relevant threshold has not been specified.

The intervals should be read alongside the event counts and the follow-up definition. The outcome counts describe participants with at least one recorded event under the primary endpoint definition, while the median follow-up was 6.7 years. The posted risk differences are not adjusted for other variables according to the displayed summaries. The registry record does not report further information about variation in follow-up, missing outcome data, or how the intervals were constructed. Such details could affect how fully the precision can be assessed.

The uncertainty is not captured by the point estimate alone. For example, a risk difference close to zero can coexist with an interval that allows meaningful differences in either direction. The posted intervals therefore support a cautious reading of the estimates, rather than a conclusion that the compared groups had exactly the same event frequency.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrations computed from the observed effect, not the original trial design calculations. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set at 100%, 75%, or 50% of the observed effect. Details of the trial's original sample-size justification are not reported in the registry record.

The 80% power column gives the estimated participants per arm under those assumptions for a design with 80% probability of detecting the specified effect at the stated significance level, if that effect and the assumptions are correct. The 90% power column gives the corresponding per-arm estimate for a design with 90% probability. Power is a planning property of a proposed study under specified assumptions, not a description of certainty in the result of a completed trial.

Each comparison has its own observed effect and therefore its own planning estimates. Planning for 75% or 50% of the observed effect asks what sample size would be needed to detect a smaller effect than the observed estimate. Smaller effects require more participants under the same assumptions. When the observed estimate is close to zero, a fractional target can also imply a very small effect to detect and consequently very large planning figures. These figures should be interpreted as illustrations, not as a recommendation for a future trial.

A sample-size curve generally visualizes how the required sample size changes as the target effect changes, holding the planning assumptions fixed. The curve does not show observed power for SEARCH. The registry record does not report the original design assumptions needed to reconstruct its own sample-size plan.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.248; treatment proportion p₁ = 0.257; average p̄ = 0.252.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6144 = 1.2042.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6143 = 0.5170.
4. n = (1.2042 + 0.5170)² / (0.0099)² = 30,453.6, rounded up to 30,454 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.247554302769027                          # control proportion
p1 <- 0.257417536880491                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(124072)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 30,454 (80% power), 40,769 (90% power)
 75% of observed effect: n per arm = 53,966 (80% power), 72,244 (90% power)
 50% of observed effect: n per arm = 121,027 (80% power), 162,020 (90% power)
power.prop.test: 30453.6 
Simulated power at n = 30,454 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and five secondary outcomes. Those secondary questions include major vascular events separately in year 1 and later years, MVE by baseline LDL group, MVE with and without the other factorial treatment, major coronary events, and total strokes. Numerical results for these outcomes are not reported in the registry record. The record lists six outcome measures and zero analyses, but that summary does not supply the analysis details needed to determine which comparisons were prespecified or how they were conducted.

Multiplicity arises when several hypotheses or outcome comparisons are considered. In general, testing many questions can increase the chance that at least one result appears notable by chance. Researchers may address this through a prespecified hierarchy or adjustment, but the statistical plan and any multiplicity procedure are not reported in the registry record. The displayed p-values are unadjusted, and no adjustment should be assumed.

The record does not report interim analyses, stopping rules, or whether the trial team had access to accumulating results during the study. It also does not report an analysis population definition, such as whether the primary analysis followed intention-to-treat principles. These details affect interpretation of the inferential results and cannot be reconstructed from the outcome counts.

Missing-data handling, loss to follow-up, treatment adherence, and the treatment of participants who discontinued assigned therapy are not reported in the registry record. No conclusion about the amount or impact of missingness can be drawn from the posted summaries. The registry record also does not describe the endpoint adjudication process or how disagreements about event classification were handled.

Because the design is factorial, interpretation also depends on how the two treatment factors and any interaction between them were analyzed. The treatment combinations are described, but the model specification, interaction assessment, and any adjustment for baseline variables are not reported in the registry record. These gaps counsel against treating the brief posted comparisons as a complete account of the planned analysis.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2.5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: other (58.8%), mve (25.7%).

The serious adverse event summaries list 5037 affected participants among 6033 at risk for simvastatin 20 mg daily, 5024 among 6031 for simvastatin 80 mg daily, 5068 among 6033 for folic acid plus vitamin B12, and 4993 among 6031 for placebo. These are counts, not rates calculated. The corresponding counts listed under other adverse events are 479, 493, 480, and 492, respectively.

The categories should be read as registry summaries, not as a complete comparative safety analysis. The record does not report how adverse events were defined, whether the counts represent participants or event episodes beyond the stated participant-with-event meaning, how events were collected, or whether severity and relatedness were assessed consistently. Death counts are not reported in the registry record.

The most frequent serious adverse event terms listed are Other at 58.8% and MVE at 25.7%. The registry record does not specify the denominator or whether participants could contribute to multiple terms. These figures therefore describe the posted terms but do not establish comparative risk or causation. They should be considered separately from the primary endpoint estimates and from any clinical decision.

15. What the Result Does, and Does Not, Mean

The posted primary outcome results quantify observed differences in the proportion of participants with a major vascular event under the listed comparisons. The estimates are modest in magnitude on the risk-difference scale, and all three confidence intervals include zero. The intervals also include values in both directions, so the data summarized in the registry do not precisely determine whether the event proportion was higher or lower in each named intervention group.

That statistical description is narrower than a conclusion about treatment benefit or harm. A confidence interval that includes zero is not proof of equivalence, and an unadjusted p-value does not state the probability that the null hypothesis is true. Conversely, an observed point estimate does not establish a reliable treatment effect. The registry does not provide a threshold for a clinically worthwhile effect, detailed adjusted analyses, or numerical results for the listed secondary endpoints.

The primary endpoint is a composite of coronary events, stroke, and peripheral revascularization. Its total count does not show the contribution of each component, nor does it describe timing or recurrence. The secondary outcome list indicates that some components were of interest, but their results are not reported in the registry record.

The trial's factorial design means that the two treatment factors were studied in a shared trial. However, the record does not explain the precise statistical model or how the posted comparisons relate to all four treatment combinations. Interpretation should therefore stay close to the labels and estimates that are actually posted, without assuming unreported adjustment or interaction analyses.

Overall, these results describe uncertain unadjusted comparisons over a median follow-up of 6.7 years. They do not by themselves establish that any treatment does or does not work, and they should not be treated as medical advice. The registry record lacks several details needed for a fuller evaluation, including its complete statistical plan, analysis populations, missing-data methods, and results for secondary outcomes.

16. Why This Trial Matters Statistically

SEARCH is statistically notable as a factorial trial addressing two treatment factors within a single randomized study. In general, factorial designs can examine more than one intervention question efficiently, while requiring careful interpretation of treatment contrasts and possible interaction. Here, simvastatin dose and folic acid plus vitamin B12 were paired in four described combinations. The record does not report the model used to analyze those factors, so the design's potential statistical efficiencies cannot be assessed from the brief summaries alone.

The primary endpoint is a composite of several vascular event types. Composite endpoints can provide a broader event measure, but their overall result may combine components with different frequencies and clinical meanings. The posted total counts do not give the component-specific results, so the overall estimates cannot reveal which events drove the observed pattern.

The trial also illustrates why effect estimates should be read with confidence intervals rather than p-values alone. Each posted interval spans zero and values in both directions. That means the results leave uncertainty about both direction and size; the p-values alone cannot convey this range. The planning illustrations further show how study size depends on the effect a future study seeks to detect, with smaller target effects requiring larger samples under fixed assumptions.

Finally, the distinction between a registry summary and a full statistical report matters. The record includes several outcome labels and group counts, but lacks details on multiplicity, missing data, analysis populations, interim monitoring, and the factorial model. These gaps limit how far readers can take the posted comparisons. The trial is useful for teaching the interpretation of absolute contrasts, interval uncertainty, factorial questions, and the limits of sparse reporting without treating a single p-value as a definitive verdict.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The posted summaries do not include the full statistical analysis plan, the original sample-size justification, interim monitoring details, or a description of allocation generation and concealment. These items are not reported in the registry record, so the planned inferential framework cannot be fully assessed. The registry lists quadruple masking but does not describe which groups of people were masked or how masking was maintained.

The outcome comparisons are unadjusted, and the record does not report whether the primary analysis followed intention-to-treat principles or another analysis population. Missing outcome data, loss to follow-up, adherence, discontinuation, and treatment switching are also not described. Without those details, it is not possible to assess how the posted counts relate to the randomized groups over follow-up.

Although the study is factorial, the registry does not explain the model used, whether the posted comparisons are marginal or specific to treatment combinations, or how interaction was assessed. The event counts do not include timing or recurrent events, and the registry does not report the methods used to calculate the intervals. Numerical results for the listed secondary outcomes are not reported in the registry record, and no multiplicity procedure is described.

The serious adverse event summaries lack detail on event definitions, collection methods, attribution, and the denominator for the listed event terms. Death counts are not reported in the registry record. These limitations do not invalidate the posted figures, but they constrain what can be inferred from them and make a complete assessment dependent on information outside the registry summary.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Study of the Effectiveness of Additional Reductions in Cholesterol and Homocysteine — NCT00124072.
• Registry record last updated 2012-02-01; status last verified 2012-01.
• Results first posted 2010-06-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

SEARCH was a completed phase 3 randomized factorial trial sponsored by the University of Oxford. It enrolled 12,064 participants with a prior myocardial infarction and compared simvastatin regimens and folic acid plus vitamin B12 with the listed placebo comparator. The study is described as double-blind, multi-centre, and conducted in the United Kingdom, with median follow-up of 6.7 years.

The primary outcome was major vascular events during the scheduled treatment period. The three posted unadjusted risk-difference estimates are 0.010 for simvastatin 20 mg, -0.003 for simvastatin 80 mg, and 0.007 for folic acid plus vitamin B12; each confidence interval includes zero. The registry record does not report the full statistical plan, analysis populations, missing-data handling, multiplicity methods, or numerical secondary-outcome results. The estimates therefore require cautious interpretation alongside those reporting limitations.