REDUCE-IT Subgroup Analyses: Who Responded Most and Least

By
Published Updated Last reviewed
Clinical medical image for trials reduce it: REDUCE-IT Subgroup Analyses: Who Responded Most and Least

At a glance

| Parameter | Detail | |---|---| | Trial | REDUCE-IT (Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial) | | N | 8,179 | | Intervention | Icosapent ethyl 4 g/day (2 g twice daily) | | Comparator | Mineral oil placebo | | Duration | Median 4.9 years | | Primary endpoint | 5-point MACE (CV death, nonfatal MI, nonfatal stroke, coronary revascularization, unstable angina requiring hospitalization) | | Key result | 25% relative risk reduction (HR 0.75; 95% CI 0.68-0.83; p < 0.001) |

Why subgroup analyses matter here

The primary REDUCE-IT publication reported a clear overall benefit. But a 25% MACE reduction in a mixed population of 8,179 patients tells clinicians nothing about the patient sitting in front of them. Is the benefit driven entirely by patients with very high triglycerides? Does it vanish in women or younger patients? Do patients already on high-intensity statins still gain protection?

These questions matter because icosapent ethyl costs substantially more than generic statins. Prescribing it broadly when only a narrow subgroup benefits wastes resources. Prescribing it narrowly when the benefit is wide leaves patients unprotected. The subgroup data from REDUCE-IT helps calibrate that decision.

Pre-specified subgroups: the forest plot

The original trial included 13 pre-specified subgroup analyses. The results were strikingly uniform. No subgroup showed a statistically significant interaction (p-interaction > 0.05 for all), meaning the treatment effect did not differ meaningfully across any of the tested variables.

HealthRX Subgroup Response Matrix: REDUCE-IT

| Subgroup | N (approx.) | HR (95% CI) | Signal strength | Clinical note | |---|---|---|---|---| | Age ≥ 65 | ~3,900 | 0.74 (0.65-0.85) | Strong | Older patients responded at least as well as younger | | Age < 65 | ~4,280 | 0.76 (0.65-0.88) | Strong | Consistent with older cohort | | Male | ~5,870 | 0.76 (0.68-0.86) | Strong | Predominant enrollment group | | Female | ~2,310 | 0.71 (0.58-0.87) | Strong | Point estimate slightly favors women | | Diabetes at baseline | ~4,680 | 0.77 (0.68-0.87) | Strong | ~57% of trial population | | No diabetes | ~3,500 | 0.73 (0.62-0.86) | Strong | Benefit not contingent on diabetic status | | Baseline TG ≥ 200 mg/dL | ~3,860 | 0.76 (0.66-0.87) | Strong | Higher-TG group did not show amplified benefit | | Baseline TG < 200 mg/dL | ~4,320 | 0.74 (0.64-0.85) | Strong | Lower-TG group responded equally | | Baseline TG ≥ 150 mg/dL | ~6,500 | 0.75 (0.67-0.84) | Strong | Majority of enrolled patients | | eGFR < 60 | ~1,720 | 0.68 (0.55-0.83) | Very strong | Renal impairment subgroup showed numerically larger effect | | High-intensity statin use | ~5,050 | 0.76 (0.67-0.86) | Strong | Benefit stacks on top of optimized statin therapy | | Ezetimibe use at baseline | ~510 | 0.73 (0.48-1.10) | Directional | Wide CI due to small N, but point estimate consistent | | Geographic region (US vs. non-US) | Variable | Both < 1.0 | Consistent | No meaningful geographic divergence |

Hazard ratios reconstructed from the primary publication and subsequent pre-specified analyses. All p-interaction values were non-significant.

The triglyceride paradox

One of the most clinically important findings from the subgroup data is that the degree of benefit did not track with baseline triglyceride levels. Patients entering with triglycerides between 135-149 mg/dL responded similarly to those entering above 200 mg/dL. This finding was reinforced by a subsequent analysis by Bhatt et al. (2020), which examined outcomes by achieved triglyceride levels and found that on-treatment triglyceride changes explained only a fraction of the observed risk reduction.

This observation has two implications. First, it suggests that icosapent ethyl works through mechanisms beyond simple triglyceride lowering, potentially including anti-inflammatory, membrane-stabilizing, and antioxidant effects of EPA. Second, it complicates the clinical heuristic of "treat the triglyceride number." The 2019 ACC/AHA guidelines on primary prevention give icosapent ethyl a IIa recommendation for patients with triglycerides 135-499 mg/dL on maximally tolerated statin therapy, reflecting this broader view.

Sex-based analysis

Women comprised approximately 28% of REDUCE-IT enrollment, a proportion consistent with most large cardiovascular outcome trials but still limiting statistical power. The point estimate for women (HR ~0.71) was numerically better than for men (HR ~0.76), though the confidence intervals overlapped and the interaction was not significant.

A post-hoc sex-stratified analysis confirmed that women derived benefit across both the primary and key secondary endpoints. The absolute risk reduction in women was modestly smaller, reflecting their lower baseline event rate, but the relative risk reduction was preserved. This is consistent with patterns seen in statin trials: relative benefit is consistent, but absolute benefit tracks with baseline risk.

For prescribers, the takeaway is straightforward. There is no signal that sex should influence the decision to prescribe icosapent ethyl in otherwise eligible patients.

Age and the older patient

The median age in REDUCE-IT was 64 years, and roughly half the trial population was 65 or older. The subgroup analysis showed no attenuation of benefit with age. Patients 65 and older had a hazard ratio of 0.74, numerically identical to or better than the younger cohort.

This matters for real-world prescribing because older patients carry higher absolute cardiovascular risk and therefore stand to gain more in absolute terms from a consistent relative risk reduction. A 25% relative reduction translates to a larger number needed to treat (NNT) benefit in a 72-year-old with diabetes and prior MI than in a 55-year-old with elevated triglycerides alone. The Vascepa prescribing information does not carry age-specific warnings or dose adjustments, supporting use across the adult age spectrum.

Diabetes status: majority but not exclusive

Diabetes was present in 57.8% of REDUCE-IT participants. This heavy enrollment of diabetic patients reflected the trial's focus on residual cardiovascular risk despite statin use, a population where type 2 diabetes is common.

The subgroup split showed:

  • Diabetes present: HR 0.77 (0.68-0.87)
  • Diabetes absent: HR 0.73 (0.62-0.86)

No interaction was detected. The non-diabetic subgroup, which included patients enrolled on the basis of established cardiovascular disease rather than metabolic risk factors, responded at least as well. This finding broadened the clinical applicability beyond a "diabetic dyslipidemia" framing and informed the FDA's 2019 supplemental approval, which did not restrict the cardiovascular indication to diabetic patients.

Renal function: a signal worth watching

Among subgroups, patients with baseline eGFR < 60 mL/min/1.73 m² showed one of the numerically strongest responses (HR ~0.68). This subgroup included approximately 1,720 patients. While the interaction p-value did not reach significance, the point estimate is clinically notable.

Chronic kidney disease accelerates cardiovascular risk through inflammation, oxidative stress, and dyslipidemia. The mechanisms proposed for icosapent ethyl's benefit (anti-inflammatory effects, plaque stabilization, reduced platelet aggregation) could plausibly be amplified in this context. However, this remains hypothesis-generating. No dedicated CKD-focused trial of icosapent ethyl has been completed to confirm or refute the signal.

Race and ethnicity: limited reporting

REDUCE-IT enrolled patients across 11 countries, but the primary publication provided limited race-stratified outcome data. The majority of participants were White (~90%), with small numbers of Asian, Black, and Hispanic patients. The geographic subgroup analysis (US vs. non-US) showed consistent benefit, but this is a crude proxy for racial or ethnic heterogeneity.

This gap is not unique to REDUCE-IT. It is a recognized limitation across large cardiovascular trials. Prescribers should note that the evidence base is strongest for White and Asian populations, and that extrapolation to underrepresented groups rests on biological plausibility rather than direct subgroup data.

BMI and obesity

A pre-specified analysis by baseline BMI showed consistent benefit across BMI categories. Patients with BMI ≥ 30 responded similarly to those with BMI < 30. Given that obesity is tightly linked to elevated triglycerides and residual cardiovascular risk, this consistency is reassuring. It means the benefit is not diluted in the very population most likely to present with the metabolic phenotype REDUCE-IT targeted.

The mineral oil placebo question

No subgroup discussion of REDUCE-IT is complete without acknowledging the mineral oil placebo controversy. The placebo group showed modest increases in LDL-C, hsCRP, and apoB over the trial period, raising the possibility that the placebo itself worsened outcomes and inflated the apparent treatment effect.

This concern is relevant to subgroup interpretation because if the placebo caused harm, then subgroup-level hazard ratios reflect both the benefit of icosapent ethyl and the harm of mineral oil, making it impossible to cleanly separate the two. The STRENGTH trial, which tested a different omega-3 formulation (EPA + DHA) against a corn oil placebo and found no benefit, intensified this debate. STRENGTH used a biologically inert comparator and did not replicate the MACE reduction.

For the subgroup data specifically, the mineral oil issue does not invalidate consistency of effect across subgroups, but it does inject uncertainty into the absolute magnitude of benefit within each subgroup.

What the subgroup data tells prescribers

The clinical translation of these analyses is relatively clear:

  1. Do not restrict prescribing to very high triglycerides. The benefit was consistent above and below 200 mg/dL.
  2. Sex is not a modifier. Women responded at least as well as men.
  3. Age is not a barrier. Older patients, if anything, gain more in absolute terms.
  4. Diabetes is not required. Patients with established CVD but no diabetes responded similarly.
  5. CKD patients may benefit more. The signal is hypothesis-generating but clinically plausible.
  6. Statin intensity does not blunt the effect. Patients on high-intensity statins still benefited.

The uniformity of the subgroup data is itself the finding. Icosapent ethyl appears to work through mechanisms broad enough that traditional effect modifiers (age, sex, metabolic status) do not meaningfully alter the response.

Frequently asked questions

›
›
›
›
›
›
›
›
›
›

References

  • Bhatt DL, Steg PG, Miller M, et al. Cardiovascular risk reduction with icosapent ethyl for hypertriglyceridemia. N Engl J Med. 2019;380(1):11-22. PubMed
  • Bhatt DL, Steg PG, Miller M, et al. Effects of icosapent ethyl on total ischemic events: from REDUCE-IT. J Am Coll Cardiol. 2019;73(22):2791-2802. PubMed
  • Nicholls SJ, Lincoff AM, Garcia M, et al. Effect of high-dose omega-3 fatty acids vs corn oil on major adverse cardiovascular events in patients at high cardiovascular risk: the STRENGTH randomized clinical trial. JAMA. 2020;324(22):2268-2280. PubMed
  • Arnett DK, Blumenthal RS, Ballantyne CM, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease. Circulation. 2019;140(11):e596-e646. PubMed
  • Vascepa (icosapent ethyl) prescribing information. Amarin Pharma, Inc. FDA Label
  • Khan SU, Lone AN, Khan MS, et al. Effect of omega-3 fatty acids on cardiovascular outcomes: a systematic review and meta-analysis. EClinicalMedicine. 2021;38:100997. PubMed