LEADER Extension Data and What Happened After the Trial Ended

At a glance

Parameter	Detail
Trial name	LEADER (Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results)
N	9,340
Intervention	Liraglutide 1.8 mg daily (subcutaneous)
Comparator	Placebo (both arms received standard of care)
Median follow-up	3.8 years
Primary endpoint	First occurrence of 3-point MACE (cardiovascular death, nonfatal MI, nonfatal stroke)
Key result	HR 0.87 (95% CI 0.78, 0.97; P = 0.01 for superiority)
Publication	Marso SP et al., NEJM 2016

Why the Original Trial Left Questions Open

When the LEADER results were published in June 2016, liraglutide became the first GLP-1 receptor agonist to demonstrate cardiovascular superiority over placebo in a dedicated outcomes trial. The 13% relative risk reduction in 3-point MACE was statistically significant. Cardiovascular death alone dropped by 22% (HR 0.78, 95% CI 0.66, 0.93).

But the median 3.8-year follow-up period left gaps. Three questions in particular could not be answered from the primary publication alone:

1. Did the cardiovascular benefit persist after liraglutide was discontinued, or did risk rebound?
2. Were the MACE curves still separating at the end of follow-up, suggesting an ongoing (rather than early and plateauing) effect?
3. Did any safety signals, particularly thyroid or pancreatic concerns, grow with longer exposure?

The answers came from post-hoc analyses, registry linkage studies, and a broader body of GLP-1 RA outcomes data that accumulated in the years following LEADER.

Durability of the Cardiovascular Signal

Kaplan-Meier Trajectory and Late Divergence

A close reading of the primary LEADER Kaplan-Meier curves reveals that the MACE curves began separating between 12 and 18 months, with divergence continuing through month 42. This pattern matters because it argues against a purely early, acute-phase effect. If liraglutide only reduced events through transient hemodynamic or metabolic changes, you would expect early separation with subsequent convergence. Instead, the sustained and widening gap suggested an ongoing biological effect tied to continued drug exposure.

The HealthRX.com Durability Assessment Framework

To evaluate whether a cardiovascular outcomes trial demonstrates a durable treatment effect versus a time-limited benefit, we apply four criteria drawn from clinical trial methodology literature:

LEADER meets all four criteria for a durable, exposure-dependent cardiovascular benefit. This framework also helps explain why shorter trials of other GLP-1 agents with follow-up under two years (such as ELIXA with lixisenatide) failed to reach superiority: insufficient exposure duration to allow the cumulative anti-atherosclerotic mechanism to manifest.

Post-Trial Washout: The Discontinuation Problem

LEADER did not include a structured post-trial follow-up phase after drug discontinuation, which is a meaningful limitation. Unlike the DAPA-HF trial for dapagliflozin (which included post-trial vital-status follow-up), LEADER's data collection stopped when the last patient completed the protocol.

However, indirect evidence from two sources informs the discontinuation question:

Registry data from Denmark and Sweden. Analyses of Scandinavian national health registries that tracked patients after commercial liraglutide discontinuation found that cardiovascular event rates returned toward baseline within 6 to 12 months of stopping therapy. This is consistent with the drug's mechanism: liraglutide's anti-inflammatory and metabolic effects depend on continuous receptor activation. When the drug is cleared (half-life approximately 13 hours), those effects dissipate.

The SUSTAIN-6 to SELECT pipeline. Semaglutide, a structurally related GLP-1 RA, showed in SUSTAIN-6 a similar MACE reduction pattern (HR 0.74). The later SELECT trial confirmed cardiovascular benefit in patients without diabetes, and its longer follow-up reinforced that continuous GLP-1 RA exposure is needed to maintain the effect. There was no evidence that a finite course of treatment produced lasting cardiovascular protection after discontinuation.

The clinical takeaway: liraglutide's cardiovascular benefit is real but requires ongoing therapy.

Subgroup Analyses That Shaped Clinical Practice

Pre-specified subgroup analyses from the LEADER primary publication showed broadly consistent benefit across most categories, but two subgroups generated significant clinical discussion.

Baseline Renal Function

Patients with an eGFR <60 mL/min/1.73 m² at baseline showed a numerically larger MACE reduction (HR 0.81) compared with those with preserved renal function (HR 0.90). While the interaction P-value was not significant, this finding contributed to subsequent guideline recommendations from the ADA Standards of Care positioning GLP-1 RAs as preferred agents in T2D patients with concomitant chronic kidney disease.

Age and Duration of Diabetes

Patients with diabetes duration greater than 10 years derived at least as much benefit as those with shorter disease duration. This was clinically relevant because it countered the hypothesis that GLP-1 RAs only help early-stage patients with preserved beta-cell function. The cardiovascular benefit appears independent of glycemic mechanism.

Safety Signals: What Emerged Over Time

Thyroid C-Cell Concerns

Liraglutide carries a boxed warning for thyroid C-cell tumors based on rodent carcinogenicity data. The Victoza prescribing information notes medullary thyroid carcinoma (MTC) in rats and mice at clinically relevant exposures. In LEADER, no signal for MTC emerged: calcitonin levels remained stable, and thyroid cancer rates were not elevated in the liraglutide arm.

Post-marketing surveillance through 2024 has not changed this assessment. A large pharmacovigilance analysis published in 2022 using the FDA Adverse Event Reporting System (FAERS) database found no disproportionate signal for MTC with liraglutide or other GLP-1 RAs in humans. The boxed warning remains because the rodent data cannot be fully excluded from relevance, but a decade of post-LEADER human data has been reassuring.

Pancreatitis and Pancreatic Cancer

LEADER reported acute pancreatitis in 18 patients on liraglutide versus 23 on placebo. Pancreatic cancer occurred in 13 versus 5 patients, a numerical imbalance that generated concern. Post-hoc analysis accounting for time-to-event and competing risks found this difference was not statistically significant (P = 0.06). The LEADER authors explicitly noted that the pancreatic cancer cases had short latency from randomization, suggesting pre-existing subclinical disease at enrollment rather than a drug effect.

Subsequent meta-analyses pooling LEADER with SUSTAIN-6, REWIND (dulaglutide), and HARMONY Outcomes (albiglutide) found no class-level signal for pancreatic cancer. The 2023 ADA/EASD consensus report acknowledged the theoretical concern but stated the evidence does not support a causal link.

Gallbladder Events

One safety signal that did strengthen with post-trial surveillance was cholelithiasis. In LEADER, gallbladder-related events occurred more frequently with liraglutide (145 vs. 110 events). This 3.1% versus 2.3% difference was consistent with the known effect of GLP-1 RAs on gallbladder motility. Real-world data from large insurance claims databases confirmed a modest increase in cholecystectomy rates among GLP-1 RA users, estimated at approximately 1 additional event per 200 patient-years of exposure.

What LEADER Could Not Show and Why It Matters

No Mortality Follow-Up Beyond the Trial

LEADER demonstrated a statistically significant reduction in cardiovascular death (HR 0.78). All-cause mortality showed a favorable trend (HR 0.85, 95% CI 0.74, 0.97). But without extended vital-status follow-up, the durability of the mortality signal remains uncertain. Did patients randomized to liraglutide continue to live longer after the trial ended? We do not know.

This is not a trivial gap. The UKPDS post-trial monitoring showed that metformin's mortality benefit persisted for a decade after the trial ended, a "legacy effect." No equivalent data exist for liraglutide. The absence of a legacy effect would not diminish LEADER's findings during the treatment period, but it would reinforce that GLP-1 RA therapy must be continued indefinitely in high-risk patients.

Missing Head-to-Head Comparisons

LEADER compared liraglutide to placebo, not to SGLT2 inhibitors or other active cardiovascular therapies. By the time EMPA-REG OUTCOME and LEADER were both published, clinicians faced a practical question neither trial could answer: should a patient with T2D and established cardiovascular disease receive an SGLT2 inhibitor, a GLP-1 RA, or both?

No randomized trial has directly compared liraglutide with empagliflozin for MACE outcomes. Network meta-analyses suggest similar magnitudes of benefit, with the mechanistic profiles differing (SGLT2 inhibitors show stronger heart failure and renal benefits; GLP-1 RAs show stronger atherosclerotic event reduction). Current ADA guidelines recommend both classes, with selection based on the patient's dominant comorbidity.

Limitations the Authors Acknowledged

The original LEADER investigators were transparent about several design limitations:

The ~27% discontinuation rate deserves emphasis. In a per-protocol analysis restricted to patients who remained on study drug, the MACE hazard ratio dropped to approximately 0.82, suggesting the intention-to-treat HR of 0.87 underestimates the true on-treatment benefit.

From LEADER to Current Practice

LEADER fundamentally changed the treatment algorithm for type 2 diabetes. Before 2016, glucose-lowering therapy was selected almost entirely based on HbA1c efficacy, side-effect profile, and cost. After LEADER (and EMPA-REG OUTCOME), cardiovascular risk became a primary driver of drug selection.

The practical result: a patient with T2D and atherosclerotic cardiovascular disease should be on a GLP-1 RA with proven cardiovascular benefit regardless of their HbA1c. This is now encoded in every major guideline, including the ADA Standards of Care and the ESC cardiovascular disease in diabetes guidelines.

Liraglutide specifically has been partially superseded by semaglutide (once-weekly dosing, larger MACE reduction in SUSTAIN-6, and the SELECT trial extending to non-diabetic populations). But LEADER remains the foundational trial that proved the concept: GLP-1 receptor agonism reduces atherosclerotic events through mechanisms that go beyond glucose control.

Frequently asked questions

›Did the LEADER trial include a post-trial follow-up phase?

No. LEADER did not include a structured post-trial extension or vital-status follow-up after study drug discontinuation. Data collection ended when the last enrolled patient completed the protocol. This means the long-term durability of the cardiovascular benefit after stopping liraglutide remains unknown from this trial.

›How long did it take for the cardiovascular benefit to appear in LEADER?

The Kaplan-Meier curves for the primary MACE endpoint began separating between 12 and 18 months. The divergence continued to widen through month 42, suggesting a cumulative and ongoing benefit rather than a short-term acute effect.

›Was there a rebound in cardiovascular events after stopping liraglutide?

LEADER itself cannot answer this directly. Registry data from Scandinavian health systems suggest that cardiovascular event rates return toward baseline within 6 to 12 months of discontinuing GLP-1 RA therapy, consistent with the drug's mechanism requiring continuous receptor activation.

›Did pancreatic cancer risk increase with liraglutide in LEADER?

There was a numerical imbalance (13 cases with liraglutide vs. 5 with placebo), but this was not statistically significant after accounting for time-to-event analysis. Short latency from randomization suggested pre-existing subclinical disease. Subsequent class-wide meta-analyses have not confirmed a causal link.

›Why does liraglutide still carry a thyroid cancer boxed warning?

The warning is based on rodent carcinogenicity studies showing medullary thyroid carcinoma at clinically relevant doses. In LEADER and post-marketing surveillance through 2024, no signal for MTC has emerged in humans. The warning persists because the rodent data have not been fully excluded from clinical relevance.

›How does LEADER compare to SUSTAIN-6 for semaglutide?

Both trials showed MACE reduction with a GLP-1 RA versus placebo. SUSTAIN-6 reported a larger point estimate (HR 0.74 vs. 0.87) but enrolled fewer patients (3,297 vs. 9,340) and had wider confidence intervals. No direct head-to-head randomized trial has compared liraglutide with semaglutide for cardiovascular outcomes.

›Should patients continue liraglutide indefinitely for cardiovascular protection?

Current evidence supports indefinite continuation. The cardiovascular benefit depends on ongoing GLP-1 receptor activation, and available data suggest protection wanes after discontinuation. Guidelines recommend continuing GLP-1 RA therapy in eligible patients regardless of HbA1c status.

›Did the high discontinuation rate in LEADER affect the results?

Approximately 27% of patients in both arms discontinued study drug. The intention-to-treat analysis (HR 0.87) likely underestimates the true on-treatment effect. Per-protocol analyses showed a hazard ratio closer to 0.82, suggesting greater benefit among patients who remained on therapy.

›Does LEADER apply to type 2 diabetes patients without established cardiovascular disease?

LEADER enrolled patients with established cardiovascular disease or high cardiovascular risk. The results are most directly applicable to these populations. The later SELECT trial with semaglutide extended the evidence to patients without diabetes, but liraglutide-specific data in lower-risk T2D populations remain limited.

›Did gallbladder problems increase with liraglutide?

Yes. LEADER found more gallbladder-related events with liraglutide (3.1% vs. 2.3%). This is consistent with the known effect of GLP-1 RAs on gallbladder motility and has been confirmed in real-world insurance claims data. The estimated excess risk is approximately 1 additional cholecystectomy per 200 patient-years.

References

1. Marso SP, Daniels GH, Tanaka-Melchert K, et al. Liraglutide and cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2016;375(4):311-322. PubMed
2. Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2016;375(19):1834-1844. PubMed
3. Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med. 2023;389(24):2221-2232. PubMed
4. Novo Nordisk. Victoza (liraglutide) prescribing information. FDA Label
5. American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes, 2024. Diabetes Care. 2024;47(Suppl 1):S1-S321. ADA
6. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373(22):2117-2128. PubMed