JAMA SEPTEMBER 24, 2019

Effect of Linagliptin vs Glimepiride on Major Adverse Cardiovascular Outcomes in Patients With Type 2 Diabetes: The CAROLINA Randomized Clinical Trial

Julio Rosenstock, Steven E. Kahn, Odd Erik Johansen, et al. for the CAROLINA Investigators

Bottom Line

In patients with type 2 diabetes and elevated cardiovascular risk, linagliptin was noninferior to glimepiride regarding major adverse cardiovascular events (MACE) over a median follow-up of 6.3 years, while demonstrating a significantly lower risk of hypoglycemia.

Key Findings

1. The primary composite outcome (cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke) occurred in 11.8% of the linagliptin group compared with 12.0% of the glimepiride group, confirming noninferiority (hazard ratio [HR] 0.98; 95.47% confidence interval [CI], 0.84-1.14; P < 0.001 for noninferiority).
2. Linagliptin was not superior to glimepiride for the primary MACE endpoint (P = 0.76 for superiority).
3. Linagliptin was associated with a significantly lower risk of any hypoglycemic adverse event (10.6% vs. 37.7%; HR 0.23; 95% CI, 0.21-0.26; P < 0.001).
4. Patients in the linagliptin group experienced modest weight reduction compared with those in the glimepiride group (weighted average mean difference, -1.54 kg; P < 0.05).

Study Design

Design
RCT
Double-Blind
Sample
6,042
Patients
Duration
6.3 yr
Median
Setting
Multicenter, international
Population Adults with relatively early type 2 diabetes and elevated cardiovascular risk or established cardiovascular disease.
Intervention Linagliptin (5 mg once daily).
Comparator Glimepiride (1–4 mg once daily).
Outcome Time to first occurrence of a composite of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke.

Study Limitations

The trial was an active-comparator study comparing two glucose-lowering agents rather than comparing an agent against placebo, which complicates direct comparison with other landmark CVOTs.
A significant proportion of the study population did not have established cardiovascular disease at baseline (approx. 63% were at high risk but lacked prior events), which may dilute the observation of cardiovascular benefits.
The primary endpoint was modified during the trial (hospitalization for unstable angina was removed), though this was a planned, blinded adjustment.

Clinical Significance

The CAROLINA trial provides critical evidence that for patients with type 2 diabetes who require additional glucose-lowering therapy beyond metformin, linagliptin is a safe alternative to sulfonylureas like glimepiride regarding cardiovascular risk, with a superior safety profile specifically concerning hypoglycemia and weight gain.

Historical Context

Following the FDA's 2008 mandate requiring cardiovascular safety trials for all new type 2 diabetes medications, most trials utilized a placebo-controlled design. CAROLINA stands out as one of the few large-scale cardiovascular outcome trials (CVOTs) that directly compared a DPP-4 inhibitor against an active sulfonylurea comparator, addressing long-standing clinical concerns regarding the safety and tolerability of older agents like glimepiride.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

Contrast the physiological mechanisms of insulin secretion stimulated by linagliptin (a DPP-4 inhibitor) versus glimepiride (a sulfonylurea) to explain why their hypoglycemia profiles differ.

Key Response

Sulfonylureas like glimepiride bind to the SUR1 subunit of ATP-sensitive potassium channels in pancreatic beta cells, causing them to close and trigger insulin release regardless of ambient glucose levels. In contrast, DPP-4 inhibitors like linagliptin prevent the degradation of GLP-1; GLP-1 only stimulates insulin secretion in a glucose-dependent manner (when blood sugar is elevated). Consequently, linagliptin has a 'ceiling' effect on insulin secretion that significantly reduces the risk of hypoglycemia compared to the constant stimulation provided by sulfonylureas.

Resident
Resident

In a patient with Type 2 Diabetes and high cardiovascular risk who is already on metformin, how does the CAROLINA trial influence your selection between adding a sulfonylurea versus a DPP-4 inhibitor?

Key Response

The CAROLINA trial demonstrated that linagliptin is noninferior to glimepiride regarding Major Adverse Cardiovascular Events (MACE). However, it also showed a significantly lower risk of hypoglycemia and weight neutrality with linagliptin. Therefore, while both are cardiovascularly safe, linagliptin is generally the preferred second-line agent over glimepiride for patients where avoiding hypoglycemia is a priority (e.g., the elderly or those with vocational risks), provided cost is not the primary constraint.

Fellow
Fellow

CAROLINA utilized an active-comparator design (linagliptin vs. glimepiride) rather than a placebo-controlled design. How does this study's outcome reconcile with the findings of the LEADER or EMPA-REG trials when managing a patient with established atherosclerotic cardiovascular disease (ASCVD)?

Key Response

CAROLINA establishes that linagliptin is 'CV neutral' relative to glimepiride but does not provide the 'CV benefit' (superiority) seen with GLP-1 receptor agonists (LEADER) or SGLT2 inhibitors (EMPA-REG). In the hierarchy of evidence for ASCVD, SGLT2i and GLP-1RA remain the preferred add-on therapies to metformin. CAROLINA's value lies in proving that if those primary agents cannot be used, a DPP-4 inhibitor is a safe alternative that avoids the hypoglycemia associated with the older, active comparator glimepiride.

Attending
Attending

For decades, the UGDP study suggested that sulfonylureas might increase cardiovascular mortality. How does the 6.3-year follow-up data from CAROLINA definitively shift the teaching paradigm regarding the safety of modern sulfonylureas?

Key Response

CAROLINA provides the most robust, long-term evidence to date (median 6.3 years) that a modern sulfonylurea (glimepiride) does not increase the risk of MACE compared to a known CV-neutral class (DPP-4 inhibitors). This effectively exonerates glimepiride from the historical concerns of cardiotoxicity, allowing clinicians to teach that the primary clinical drawback of sulfonylureas is hypoglycemia and weight gain, rather than intrinsic cardiovascular harm.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

Assess the implications of the 1.3 noninferiority margin used in the CAROLINA trial. Does this statistical threshold sufficiently guard against Type II error when comparing two active agents in a high-risk population?

Key Response

The 1.3 margin was the standard regulatory threshold established by the 2008 FDA guidance for diabetes drugs. While it provides a clear framework for ruling out unacceptable excess risk, using such a margin in an active-comparator trial might lack the sensitivity to detect subtle differences in CV outcomes that a narrower margin or a placebo-controlled trial might reveal. However, the long duration (6.3 years) and high number of events in CAROLINA lend significant weight to the noninferiority conclusion despite the 1.3 upper bound.

Journal Editor
Journal Editor

The CAROLINA trial reported a high rate of study drug discontinuation (nearly 25%). As a reviewer, how would you evaluate the impact of this 'off-treatment' time on the validity of the intention-to-treat (ITT) analysis for MACE?

Key Response

High discontinuation rates in long-term trials can dilute the observed treatment effect, potentially biasing results toward noninferiority (making the drugs look more similar than they are). A tough reviewer would demand a sensitivity analysis (such as a 'per-protocol' or 'on-treatment' analysis) to ensure that the cardiovascular neutrality of glimepiride wasn't simply a result of patients stopping the drug before events occurred, though the long median exposure in CAROLINA partially mitigates this concern.

Guideline Committee
Guideline Committee

Current ADA Standards of Care prioritize SGLT2i and GLP-1RA for patients with high CV risk. Should CAROLINA’s evidence of glimepiride’s CV safety move sulfonylureas higher in the treatment algorithm for this specific population?

Key Response

While CAROLINA confirms glimepiride's CV safety (Level A evidence for noninferiority), it does not demonstrate the mortality or morbidity benefits required to displace SGLT2i or GLP-1RA. However, the findings should influence guidelines to specify that if cost is a barrier, glimepiride is a safe option. It also reinforces the recommendation to prefer DPP-4 inhibitors over sulfonylureas when the goal is to minimize hypoglycemia (Grade A), as the trial quantified a 5-fold higher hypoglycemia risk with the sulfonylurea.

Clinical Landscape

Noteworthy Related Trials

2013

SAVOR-TIMI 53 Trial

n = 16,492 · NEJM

Tested

Saxagliptin

Population

T2DM patients with history of or risk factors for cardiovascular disease

Comparator

Placebo

Endpoint

Composite of cardiovascular death, myocardial infarction, or ischemic stroke

Key result: Saxagliptin did not increase or decrease the rate of ischemic events, though it was associated with an increased risk of hospitalization for heart failure.
2013

EXAMINE Trial

n = 5,380 · NEJM

Tested

Alogliptin

Population

T2DM patients who had a recent acute coronary syndrome event

Comparator

Placebo

Endpoint

Composite of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke

Key result: Alogliptin was non-inferior to placebo with respect to the primary cardiovascular outcome in this high-risk population.
2015

TECOS Trial

n = 14,671 · NEJM

Tested

Sitagliptin

Population

T2DM patients with established cardiovascular disease

Comparator

Placebo

Endpoint

Composite of cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, or unstable angina hospitalization

Key result: Sitagliptin added to usual care did not increase the risk of major adverse cardiovascular events or hospitalization for heart failure.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis