Why was the JUPITER trial stopped early?

The independent Data Safety Monitoring Board recommended early termination after a median 1.9 years of follow-up because the prespecified Haybittle-Peto stopping boundary had been crossed, indicating a statistically overwhelming benefit for rosuvastatin. Early stopping can overestimate treatment effects relative to what a completed trial might show.

What were the inclusion criteria for JUPITER?

Participants needed LDL-C <130 mg/dL, hsCRP ≥2 mg/L, and no prior cardiovascular disease or diabetes. Men had to be 50 or older, women 60 or older. Those already on lipid-lowering therapy were excluded.

What was the absolute risk reduction in JUPITER?

The primary endpoint rate dropped from 0.77 to 0.37 per 100 person-years, yielding an NNT of approximately 95 over the 1.9-year median follow-up to prevent one composite cardiovascular event.

Did JUPITER show a mortality benefit for rosuvastatin?

All-cause mortality showed a hazard ratio of 0.80 (95% CI 0.67, 0.97, p = 0.02), suggesting a mortality benefit. Cardiovascular death alone showed HR 0.53, but with wider confidence intervals due to fewer events.

Did JUPITER find any diabetes risk with rosuvastatin?

Physician-reported new-onset diabetes was higher in the rosuvastatin arm (3.0% vs 2.4%). This signal, combined with data from other statin trials, contributed to FDA label updates regarding diabetes risk with statin therapy.

How did JUPITER change statin prescribing guidelines?

JUPITER expanded the evidence base for primary prevention beyond LDL-based criteria. The 2013 ACC/AHA guidelines incorporated CRP as a "risk enhancer" within broader ASCVD risk calculation, rather than adopting CRP-triggered prescribing as standalone policy.

Who funded the JUPITER trial?

AstraZeneca, the manufacturer of rosuvastatin (Crestor), sponsored the trial. Endpoint adjudication was performed by an independent committee, and the DSMB operated independently of the sponsor.

Can JUPITER results be applied to younger adults?

No. Men under 50 and women under 60 were excluded. The trial provides no direct evidence for primary prevention with rosuvastatin in younger age groups with elevated hsCRP.

How does JUPITER relate to the CANTOS trial?

CANTOS (2017) tested the inflammatory hypothesis from a pharmacological angle, using canakinumab to lower IL-1β-driven inflammation without affecting lipids. CANTOS validated the biological premise behind JUPITER's CRP-based enrollment while demonstrating that anti-inflammatory therapy alone reduces cardiovascular events.

Inside the JUPITER Methodology: What Most Summaries Skip

Q: What was the primary endpoint of the JUPITER trial?

The primary endpoint was the first occurrence of a five-part composite: nonfatal myocardial infarction, nonfatal stroke, arterial revascularization, hospitalization for unstable angina, or confirmed cardiovascular death. All events were adjudicated by an independent blinded committee.

Q: What was the absolute risk reduction in JUPITER?

The primary endpoint rate dropped from 0.77 to 0.37 per 100 person-years, yielding an NNT of approximately 95 over the 1.9-year median follow-up to prevent one composite cardiovascular event.

Q: Did JUPITER show a mortality benefit for rosuvastatin?

All-cause mortality showed a hazard ratio of 0.80 (95% CI 0.67, 0.97, p = 0.02), suggesting a mortality benefit. Cardiovascular death alone showed HR 0.53, but with wider confidence intervals due to fewer events.

Q: Did JUPITER find any diabetes risk with rosuvastatin?

Physician-reported new-onset diabetes was higher in the rosuvastatin arm (3.0% vs 2.4%). This signal, combined with data from other statin trials, contributed to FDA label updates regarding diabetes risk with statin therapy.

Q: How did JUPITER change statin prescribing guidelines?

JUPITER expanded the evidence base for primary prevention beyond LDL-based criteria. The 2013 ACC/AHA guidelines incorporated CRP as a "risk enhancer" within broader ASCVD risk calculation, rather than adopting CRP-triggered prescribing as standalone policy.

Q: Who funded the JUPITER trial?

AstraZeneca, the manufacturer of rosuvastatin (Crestor), sponsored the trial. Endpoint adjudication was performed by an independent committee, and the DSMB operated independently of the sponsor.

Q: Can JUPITER results be applied to younger adults?

No. Men under 50 and women under 60 were excluded. The trial provides no direct evidence for primary prevention with rosuvastatin in younger age groups with elevated hsCRP.

Q: How does JUPITER relate to the CANTOS trial?

CANTOS (2017) tested the inflammatory hypothesis from a pharmacological angle, using canakinumab to lower IL-1β-driven inflammation without affecting lipids. CANTOS validated the biological premise behind JUPITER's CRP-based enrollment while demonstrating that anti-inflammatory therapy alone reduces cardiovascular events.

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | N | 17,802 | | Intervention | Rosuvastatin 20 mg daily | | Comparator | Matching placebo | | Duration | Median 1.9 years (planned maximum 5 years) | | Primary Endpoint | First major cardiovascular event (MI, stroke, arterial revascularization, hospitalization for unstable angina, or CV death) | | Key Result | HR 0.56 (95% CI 0.46, 0.69), p <0.00001 |

The Enrollment Filter That Defined Everything

Most trial summaries describe JUPITER as a statin trial. It is more accurately a biomarker-selection trial that happened to use a statin. The inclusion criteria required two simultaneous conditions: LDL-C <130 mg/dL (a level at which statins were not indicated under 2001 ATP III guidelines) and hsCRP ≥2 mg/L (a marker of systemic inflammation with no guideline-based treatment threshold at the time). Men had to be ≥50 years old, women ≥60. This dual filter created a population that no prior statin trial had specifically targeted.

The exclusion list was extensive. Prior cardiovascular disease, diabetes, uncontrolled hypertension (systolic >190 mmHg), triglycerides >500 mg/dL, inflammatory conditions requiring immunosuppression, current use of lipid-lowering therapy, and hepatic dysfunction all disqualified participants. The diabetes exclusion is particularly consequential. By removing patients with fasting glucose ≥126 mg/dL or HbA1c diagnostic of diabetes, JUPITER enriched for a metabolically "intermediate" population, not yet diabetic but carrying inflammatory burden. Post-hoc analyses later showed a higher rate of physician-reported diabetes in the rosuvastatin arm, a finding that generated substantial debate about statin-associated dysglycemia.

The hsCRP threshold of 2 mg/L was not arbitrary. It was drawn from prior epidemiological work by the same investigators linking CRP levels above this cutoff to increased cardiovascular risk. Critics noted this creates a self-reinforcing evidence loop: the threshold was proposed by investigators who then designed the trial around it.

Randomization and Blinding Architecture

JUPITER used a 1:1 randomization scheme, centrally administered through an interactive voice-response system. Stratification was by clinical site. This is standard large-trial methodology, and the baseline characteristics table in the primary publication confirms balance across treatment arms for age, sex, race, BMI, blood pressure, lipid levels, and metabolic markers.

Blinding was double-blind with matching placebo. Neither participants nor site investigators knew treatment assignment. The trial maintained blinding through the use of identical capsules and a centralized randomization system. An independent Data Safety Monitoring Board (DSMB) reviewed unblinded data at prespecified intervals. This DSMB ultimately recommended early termination, a decision that warrants its own analysis (see below).

One methodological strength worth noting: lipid panels were not routinely returned to investigators or participants during the trial. This is important. If physicians had seen dramatic LDL-C drops in some patients, functional unblinding would have occurred, potentially triggering differential use of co-interventions. The protocol specified that lipid results be withheld, preserving the integrity of the comparison.

The Composite Primary Endpoint: Five Components, Unequal Clinical Weight

JUPITER's primary endpoint was the first occurrence of any component in a five-part composite: nonfatal myocardial infarction, nonfatal stroke, arterial revascularization, hospitalization for unstable angina, or confirmed cardiovascular death. All events were adjudicated by an independent endpoint committee blinded to treatment assignment.

Composite endpoints increase statistical power by accumulating events faster, enabling smaller sample sizes or shorter follow-up. The tradeoff is interpretive ambiguity. A revascularization procedure and a cardiovascular death are not equivalent clinical events, yet in a composite analysis they contribute equally. In JUPITER, revascularization comprised a substantial share of total events.

The original NEJM report did present individual component analyses. MI alone showed a hazard ratio of 0.46 (95% CI 0.30, 0.70). Stroke alone showed HR 0.52 (0.34, 0.79). Cardiovascular death showed HR 0.53 (0.29, 0.97). Revascularization showed HR 0.54 (0.41, 0.72). All components moved in the same direction, which is reassuring. Some large cardiovascular trials have shown composite reductions driven almost entirely by softer endpoints like revascularization while harder endpoints like death remained neutral. JUPITER does not have that problem, though the relatively small number of deaths (12 vs 22 in the primary analysis) limits the precision of the mortality estimate.

The unstable angina component is the most subjective of the five. Hospitalization thresholds vary by institution and geography. The independent adjudication committee mitigates this concern, but does not eliminate it.

Statistical Design and the Early-Stop Question

The trial was designed with 90% power to detect a 25% reduction in the primary endpoint at a two-sided alpha of 0.05, assuming event rates calibrated to the Framingham risk profiles of the target population. The planned maximum follow-up was 5 years with interim analyses at approximately 12-month intervals.

The DSMB recommended early termination after a median follow-up of only 1.9 years. By that point, 393 primary endpoint events had occurred (142 in the rosuvastatin arm, 251 in the placebo arm). The prespecified stopping boundary used the Haybittle-Peto rule, which requires an extremely low p-value (typically p <0.001) at interim looks to justify early stopping. JUPITER crossed this boundary convincingly.

Early termination in cardiovascular trials is contentious for specific reasons. Treatment effects observed at interim analyses tend to overestimate the true long-term benefit, a phenomenon well-documented in statistical literature. The CORONA trial (rosuvastatin in heart failure) and other statin trials that ran to completion showed more modest effects. Had JUPITER continued for 4 or 5 years, the hazard ratio might have remained significant but attenuated. The 44% relative reduction is the interim-stop estimate, not a confirmed long-term treatment effect.

Early termination also truncates the safety observation window. The diabetes signal, small elevations in HbA1c, and musculoskeletal complaints all had limited follow-up. The FDA label for rosuvastatin now includes diabetes risk language, informed partly by JUPITER and partly by subsequent meta-analyses.

Absolute vs Relative Risk: The Numbers That Change the Conversation

The 44% relative risk reduction is the headline number. The absolute numbers tell a different story. The primary endpoint rate was 0.77 per 100 person-years in the placebo group and 0.37 per 100 person-years in the rosuvastatin group. Over the 1.9-year median follow-up, this translates to a number needed to treat (NNT) of approximately 95 to prevent one primary endpoint event over that timeframe. Extrapolated to 5 years (assuming constant hazard rates, which is itself an assumption), the NNT drops to roughly 25.

For cardiovascular death specifically, the NNT over 1.9 years was approximately 500. For all-cause mortality, the trial showed HR 0.80 (95% CI 0.67, 0.97, p = 0.02), yielding an NNT of about 167 over the trial duration.

These absolute risk numbers are central to the clinical translation debate. A 44% relative reduction sounds dramatic. An NNT of 95 over two years means treating 95 patients to prevent one event. Whether that equation justifies population-level CRP screening and statin therapy depends on cost, side-effect tolerance, and how decision-makers weight a composite endpoint vs hard mortality.

The Estimand Problem: What Was JUPITER Actually Estimating?

Modern trial methodology, guided by the ICH E9(R1) addendum, asks trialists to define their estimand explicitly: the precise treatment effect being estimated, including how intercurrent events (treatment discontinuation, use of open-label statins, death from non-cardiovascular causes) are handled.

JUPITER predates this framework, but the estimand question still applies retrospectively. The primary analysis was intention-to-treat (ITT), meaning all randomized participants were analyzed regardless of adherence. Adherence in the trial was high (approximately 75% at end of study in both arms), but some placebo-arm participants started open-label statins during follow-up, and some rosuvastatin-arm participants discontinued therapy. Both of these intercurrent events dilute the ITT estimate. The true on-treatment effect may be larger than 44%, but the ITT analysis gives the more conservative, policy-relevant number.

The trial did not report a formal per-protocol or treatment-policy estimand comparison. Contemporary re-analyses using methods like inverse-probability-of-censoring weighting could recover a cleaner estimate, but the raw ITT result remains the basis for guidelines.

Generalizability Constraints

Several design features limit how broadly JUPITER's results should be applied.

The CRP filter. Roughly 60% of screened individuals failed the hsCRP ≥2 mg/L criterion. The trial tells us nothing about rosuvastatin's benefit in people with low hsCRP and low LDL-C. The 2013 ACC/AHA cholesterol guideline ultimately de-emphasized CRP as a gating criterion, instead using 10-year ASCVD risk calculation as the primary decision tool. CRP became a "risk enhancer" rather than a standalone trigger.

Age floors. Men under 50 and women under 60 were excluded. JUPITER cannot inform primary prevention decisions in younger adults.

Ethnic representation. Approximately 71% of participants were White, 12% Hispanic, 10% Black. The trial included participants from 26 countries, which adds geographic diversity but may dilute country-specific practice patterns.

Excluded populations. Patients with diabetes, prior CVD, or lipid-lowering therapy were excluded. JUPITER applies only to a specific "gap" population: elevated inflammatory risk, normal lipids, no diabetes, no prior events.

What the Critics Got Right

The most substantive criticisms of JUPITER fall into three categories.

First, the early-termination effect inflation. As noted, stopping a trial at the peak of the Kaplan-Meier separation tends to overestimate the long-term treatment benefit. This is not fraud. It is a known statistical phenomenon that the DSMB stopping rule is designed to account for, but it still biases the point estimate upward relative to what a completed trial would have shown.

Second, the investigator-sponsor overlap. The trial was sponsored by AstraZeneca (manufacturer of rosuvastatin under the brand name Crestor) and led by investigators who had published extensively on CRP as a cardiovascular biomarker. This does not invalidate the results. Endpoint adjudication was independent, randomization was centralized, and blinding was maintained. But the alignment of scientific hypothesis, trial design, and commercial interest warranted disclosure and scrutiny, both of which occurred publicly.

Third, the clinical actionability question. Even accepting the results at face value, universal hsCRP screening followed by statin therapy for all who qualify represents a substantial expansion of the treatment population. The USPSTF statin recommendations incorporate JUPITER's findings indirectly through pooled risk calculators but do not endorse CRP-triggered prescribing as standalone policy. Cost-effectiveness analyses have reached mixed conclusions depending on CRP assay cost, statin generic pricing, and the willingness-to-pay threshold used.

Where JUPITER Fits in the Evidence Hierarchy

JUPITER is one trial. It is a large, well-executed trial, but it is not the only evidence informing CRP-guided prevention. The CANTOS trial (canakinumab, 2017) later proved the inflammatory hypothesis from a different angle, showing that reducing IL-1β-driven inflammation lowered cardiovascular events independent of lipid lowering. CANTOS validated the biological premise behind JUPITER's enrollment criterion while using a completely different therapeutic mechanism.

The practical legacy of JUPITER is not CRP screening. It is the broader recognition that inflammatory risk markers add predictive information beyond LDL-C, and that some patients who would not qualify for statins under traditional lipid criteria still benefit from them. That insight was incorporated into guideline revisions. The trial did exactly what a well-designed study should: it changed the conversation without becoming the final word.

Frequently asked questions

›

References

Ridker PM, Danielson E, Fonseca FA, et al. Rosuvastatin to Prevent Vascular Events in Men and Women with Elevated C-Reactive Protein. N Engl J Med. 2008;359(21):2195-2207. https://pubmed.ncbi.nlm.nih.gov/18997196/
Ridker PM, Pradhan A, MacFadyen JG, et al. Cardiovascular benefits and diabetes risks of statin therapy in primary prevention: an analysis from the JUPITER trial. Lancet. 2012;380(9841):565-571. https://pubmed.ncbi.nlm.nih.gov/20167683/
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. https://pubmed.ncbi.nlm.nih.gov/24239923/
Ridker PM, Everett BM, Thuren T, et al. Antiinflammatory Therapy with Canakinumab for Atherosclerotic Disease (CANTOS). N Engl J Med. 2017;377(12):1119-1131. https://pubmed.ncbi.nlm.nih.gov/28845751/
US Food and Drug Administration. Rosuvastatin (Crestor) Prescribing Information. https://www.accessdata.fda.gov/drugsatfda_docs/label/2023/021366s045lbl.pdf
US Preventive Services Task Force. Statin Use for the Primary Prevention of Cardiovascular Disease in Adults: Recommendation Statement. JAMA. 2022;328(8):746-753. https://pubmed.ncbi.nlm.nih.gov/35040390/