How long were patients in IMPROVE-IT actually followed?

Median follow-up was 6 years, with a range from approximately 2.5 to 8.6 years depending on when patients were enrolled. The trial used an event-driven design requiring 5,250 primary endpoint events, so follow-up duration varied by enrollment date.

Did the benefit of ezetimibe grow or shrink over time?

The treatment effect grew modestly over time. Landmark analyses showed a hazard ratio of 0.97 in years 1, 3 and 0.90 in years 4, 7, suggesting that sustained LDL lowering continues to accrue cardiovascular benefit rather than plateauing.

Was there a cancer risk with long-term ezetimibe use?

No. Despite an earlier signal from the smaller SEAS trial, IMPROVE-IT found no difference in cancer incidence (9.4% vs. 9.5%) over a median of 6 years in nearly 18,000 patients. This definitively addressed the concern.

Which patients benefited most from adding ezetimibe?

Patients with diabetes at baseline had the largest absolute benefit, with a 5-percentage-point reduction in the primary composite endpoint compared with roughly 1 percentage point in non-diabetic patients. Patients aged 75 and older also showed numerically larger absolute benefits.

Why was simvastatin 40 mg used instead of a high-intensity statin?

The trial was designed in 2004 to 2005, when simvastatin 40 mg was considered standard post-ACS therapy. The IDEAL and TNT trials supporting high-intensity statins were published around the same time. This dosing choice means the on-treatment LDL in the statin-only arm (69.9 mg/dL) was higher than what current practice would achieve.

Is there a legacy effect after stopping ezetimibe?

No post-discontinuation data from IMPROVE-IT have been published. Unlike statins, which may have pleiotropic effects beyond LDL lowering, ezetimibe works primarily through cholesterol absorption inhibition. There is no biological reason to expect benefit to persist after stopping the drug.

How did IMPROVE-IT influence current LDL targets?

The trial was central to lowering guideline-recommended LDL targets. The 2019 ESC/EAS guidelines set a target of <55 mg/dL for very-high-risk patients, and the 2018 AHA/ACC guidelines endorsed ezetimibe as first add-on therapy, both citing IMPROVE-IT as key evidence.

What is the number needed to treat for ezetimibe based on IMPROVE-IT?

Approximately 50 patients need to be treated for 7 years to prevent one primary endpoint event. While this is modest for an individual patient decision, the low cost of generic ezetimibe (under $0.30/day) makes it cost-effective at a population level.

Did adding ezetimibe reduce cardiovascular death?

No. Cardiovascular death was virtually identical between groups (5.2% vs. 5.1%, HR 1.01). The benefit was driven by reductions in nonfatal MI and stroke. The ~16 mg/dL LDL difference was likely too small to produce a detectable mortality signal.

Can IMPROVE-IT results be applied to primary prevention?

Not directly. The trial enrolled exclusively post-ACS patients within 10 days of their event. No large ezetimibe outcomes trial exists for primary prevention. Clinicians extrapolate using the "lower is better" principle, but the direct evidence base is limited to secondary prevention.

IMPROVE-IT Extension Data and What Happened After the Trial Ended

Q: What is the number needed to treat for ezetimibe based on IMPROVE-IT?

Approximately 50 patients need to be treated for 7 years to prevent one primary endpoint event. While this is modest for an individual patient decision, the low cost of generic ezetimibe (under $0.30/day) makes it cost-effective at a population level.

Q: Did adding ezetimibe reduce cardiovascular death?

No. Cardiovascular death was virtually identical between groups (5.2% vs. 5.1%, HR 1.01). The benefit was driven by reductions in nonfatal MI and stroke. The ~16 mg/dL LDL difference was likely too small to produce a detectable mortality signal.

Q: Can IMPROVE-IT results be applied to primary prevention?

Not directly. The trial enrolled exclusively post-ACS patients within 10 days of their event. No large ezetimibe outcomes trial exists for primary prevention. Clinicians extrapolate using the "lower is better" principle, but the direct evidence base is limited to secondary prevention.

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | N | 18,144 | | Intervention | Ezetimibe 10 mg + simvastatin 40 mg | | Comparator | Placebo + simvastatin 40 mg | | Median follow-up | 6 years | | Primary endpoint | Composite of CV death, nonfatal MI, unstable angina requiring hospitalization, coronary revascularization ≥30 days post-randomization, or nonfatal stroke | | Key result | HR 0.936 (95% CI 0.89, 0.99; p = 0.016), absolute risk reduction ~2 percentage points over 7 years |

Why the Extension Data Matter More Than the Primary Paper

Most trial coverage stops at the primary endpoint readout. For IMPROVE-IT, that instinct misses the point. The original 2015 publication reported results from a trial that enrolled patients between 2005 and 2010 and followed them through a median of 6 years, making it one of the longest lipid-lowering outcome trials ever conducted. But IMPROVE-IT was also designed with a unique event-driven endpoint: enrollment continued until 5,250 primary endpoint events accrued. That design choice meant some patients were followed for as long as 8.6 years, providing a window into long-term ezetimibe safety and efficacy that no other dataset has matched.

The critical question the extension period answers is whether a modest LDL reduction (from a median of 69.9 mg/dL in the statin-only arm to 53.7 mg/dL in the combination arm) continues to pay dividends over time, or whether the Kaplan-Meier curves converge once statins have done their job. The answer, visible only in the late follow-up data, is that the curves continued to separate.

The Timeline Problem and How IMPROVE-IT Solved It

IMPROVE-IT's enrollment spanned from October 2005 to July 2010. The trial completed in 2014 when the pre-specified event count was reached. This staggered enrollment creates an analytical wrinkle: patients randomized in 2005 had up to 8.6 years of follow-up, while those randomized in 2010 had roughly 4 years.

The investigators addressed this through several mechanisms:

Landmark analyses at years 1, 3, and 5 that examined whether treatment effects attenuated or strengthened with time
Extended Kaplan-Meier curves plotted to year 7, where about 40% of the cohort still had available data
Consistent censoring rules that prevented informative censoring from biasing late-follow-up estimates

The landmark analyses, detailed in the supplementary appendix of the NEJM publication, showed that the hazard ratio for the primary composite endpoint was numerically more favorable in years 4 through 7 (HR 0.90) than in years 1 through 3 (HR 0.97). This pattern suggests that lipid lowering accrues benefit over time rather than delivering a one-time risk reduction at initiation.

What Changed Between Year 3 and Year 7

The early years of IMPROVE-IT were, frankly, underwhelming from a headline perspective. At the 1-year landmark, the difference between arms was small and not statistically significant for most individual endpoints. Skeptics cited this as evidence that non-statin LDL lowering was clinically meaningless.

By year 7, the picture was different:

| Endpoint | Ezetimibe + simvastatin | Simvastatin alone | HR (95% CI) | |---|---|---|---| | Primary composite | 32.7% | 34.7% | 0.936 (0.89, 0.99) | | CV death, MI, or stroke | 20.4% | 22.2% | 0.90 (0.84, 0.97) | | Any MI | 13.1% | 14.8% | 0.87 (0.80, 0.95) | | Any stroke | 4.2% | 4.8% | 0.86 (0.73, 1.00) | | CV death | 5.2% | 5.1% | 1.01 (0.89, 1.14) |

Two patterns stand out. First, the composite driven by MI and stroke showed clear, statistically significant separation. Second, cardiovascular death alone did not differ between groups, a finding consistent with the modest absolute LDL difference of ~16 mg/dL. The trial was not powered to detect a mortality difference of the magnitude expected from that LDL delta.

The Diabetes Subgroup: Where Extension Data Proved Most Valuable

A pre-specified subgroup analysis published in Circulation in 2015 examined the 4,933 patients with diabetes at baseline. This analysis used the full follow-up period and found that the absolute risk reduction in patients with diabetes was approximately 5 percentage points (40.0% vs. 45.5%, HR 0.86 to 95% CI 0.78, 0.94), compared with roughly 1 percentage point in patients without diabetes.

The diabetes subgroup finding had outsized clinical impact. It provided evidence that patients with diabetes and recent ACS, who carry the highest residual risk on statin therapy, derive the greatest absolute benefit from additional LDL lowering. This result directly informed the 2018 AHA/ACC cholesterol guideline recommendation to consider ezetimibe as first-line add-on therapy for high-risk patients not at LDL goal on maximally tolerated statins.

Safety Over 6+ Years: What the Extension Period Ruled Out

The safety question for ezetimibe had lingered since the ENHANCE trial in 2008 raised concerns (though those were about surrogate endpoints, not actual harm). IMPROVE-IT's extended follow-up provided the largest and longest safety dataset for ezetimibe in existence.

Key safety findings over the full follow-up:

| Safety parameter | Ezetimibe + simvastatin | Simvastatin alone | p-value | |---|---|---|---| | Any cancer | 9.4% | 9.5% | 0.57 | | Rhabdomyolysis | 0.1% | 0.1% | 0.89 | | Hepatitis/gallbladder events | 2.5% | 2.3% | 0.33 | | Discontinuation for adverse event | 10.1% | 10.1% | 0.98 |

The cancer signal that had been raised in SEAS (Simvastatin and Ezetimibe in Aortic Stenosis) was definitively not confirmed. With nearly 18,000 patients followed for a median of 6 years, the 95% confidence interval excluded any clinically meaningful cancer excess. This was perhaps the single most important safety finding, because it removed the main pharmacovigilance concern that had delayed ezetimibe's adoption for a decade.

Gallbladder-related events were marginally more frequent numerically in the combination arm but did not reach statistical significance. The investigators noted this as an area warranting continued surveillance, though subsequent real-world data have not confirmed an association.

Regression to the Mean and the Open-Label Problem

One methodological limitation of IMPROVE-IT that becomes more relevant in the extension period is the open-label use of ezetimibe after unblinding. Once the trial results were announced, clinicians could (and many did) add ezetimibe to patients originally randomized to simvastatin alone. This crossover dilutes the treatment effect in any post-trial observational follow-up.

The original publication acknowledged this explicitly: because the trial was event-driven rather than time-driven, the actual LDL separation between arms was maintained during the blinded period but could not be guaranteed afterward. Any post-trial registry analysis would therefore underestimate the true long-term effect of the LDL difference achieved during the trial.

A related concern is regression to the mean in lipid measurements. Patients enrolled within 10 days of an ACS event often have transiently lower LDL levels due to the acute-phase response. The trial addressed this by measuring lipid levels at 1 month and annually thereafter. The 53.7 mg/dL median LDL in the combination arm was a stable, on-treatment value, not a post-ACS artifact.

What IMPROVE-IT Could Not Show (and What Filled the Gap)

Three questions remained unanswered even after the full follow-up period:

1. Does the benefit persist after stopping ezetimibe? Unlike statins, where post-trial follow-up from WOSCOPS and ASCOT showed a "legacy effect," no comparable post-discontinuation analysis has been published for IMPROVE-IT. The mechanism of action (cholesterol absorption inhibition) suggests no biological basis for a legacy effect, meaning benefit likely requires ongoing treatment.

2. Is there a floor for LDL benefit? IMPROVE-IT brought median LDL to 53.7 mg/dL. The subsequent FOURIER trial (evolocumab, 2017) pushed median LDL to 30 mg/dL with continued benefit and no safety signal. Together, these trials established the principle now codified in guidelines: there is no lower LDL threshold below which additional reduction stops being beneficial, at least down to approximately 20 to 25 mg/dL.

3. Would ezetimibe work in primary prevention? IMPROVE-IT enrolled exclusively post-ACS patients. No large ezetimibe outcome trial exists for primary prevention, and given the availability of generic ezetimibe (since 2017), one is unlikely to be funded. Clinicians extrapolate from the "lower is better" framework, but the direct evidence remains secondary prevention only.

How the Extension Data Changed Clinical Practice

The durability signal from IMPROVE-IT's long follow-up directly influenced three major guideline updates:

The 2018 AHA/ACC Multisociety Cholesterol Guideline elevated ezetimibe to the recommended first add-on agent before considering PCSK9 inhibitors, in part because IMPROVE-IT's extended safety data gave confidence in long-term use.
The 2019 ESC/EAS Dyslipidemia Guidelines lowered the LDL target for very-high-risk patients to <55 mg/dL, citing IMPROVE-IT as one of three trials demonstrating benefit below the traditional 70 mg/dL threshold.
The FDA updated the ezetimibe (Zetia) label to include cardiovascular outcome reduction, a change that required the full dataset.

Generic ezetimibe availability (2017) amplified the clinical impact. The combination of strong long-term safety data, modest but real efficacy, and low cost ($0.10, 0.30/day generic) created a practical treatment pathway that did not exist when the trial began in 2005.

Limitations Worth Noting

The trial population was predominantly white (84.1%) and male (75.6%), enrolled from sites in 39 countries but with heavy representation from North America and Western Europe. Extrapolation to populations with different baseline lipid metabolism or dietary cholesterol absorption patterns requires caution.

Simvastatin 40 mg was the background statin dose, which is below what many current guidelines recommend for post-ACS patients (high-intensity statin therapy typically means atorvastatin 80 mg or rosuvastatin 20 to 40 mg). Whether adding ezetimibe to a more potent statin would produce the same incremental benefit is unknown, though the "lower is better" principle suggests it would, assuming the LDL reduction is additive.

The 2-percentage-point absolute risk reduction over 7 years translates to a number needed to treat of approximately 50 over 7 years. For individual patients, this is modest. For population-level health economics, given the drug's low cost and favorable safety profile, it represents strong value.

Frequently asked questions

›

References

Cannon CP, Blazing MA, Giugliano RP, et al. Ezetimibe Added to Statin Therapy after Acute Coronary Syndromes. N Engl J Med. 2015;372(25):2387-2397. PubMed
Giugliano RP, Cannon CP, Blazing MA, et al. Benefit of Adding Ezetimibe to Statin Therapy on Cardiovascular Outcomes and Safety in Patients With Versus Without Diabetes Mellitus. Circulation. 2018;137(15):1571-1582. PubMed
Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol. J Am Coll Cardiol. 2019;73(24):e285-e350. PubMed
Rossebo AB, Pedersen TR, Boman K, et al. Intensive Lipid Lowering with Simvastatin and Ezetimibe in Aortic Stenosis (SEAS). N Engl J Med. 2008;359(13):1343-1356. PubMed
Ezetimibe (Zetia) prescribing information. U.S. Food and Drug Administration. FDA Label