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, Bernard Zinman, Mark A. Espeland, Hans J. Woerle, et al.

Bottom Line

In adults with early type 2 diabetes and elevated cardiovascular risk, the DPP-4 inhibitor linagliptin was noninferior to the sulfonylurea glimepiride regarding major adverse cardiovascular events, but offered significant benefits in reducing hypoglycemia and preventing weight gain.

Key Findings

1. The primary composite outcome (3P-MACE) occurred in 11.8% of patients (356 of 3,023) in the linagliptin group compared with 12.0% (362 of 3,010) in the glimepiride group (Hazard Ratio 0.98; 95% CI, 0.84 to 1.14; P < 0.001 for noninferiority; P = 0.76 for superiority).
2. There was no significant difference in the incidence of cardiovascular death, which occurred in 4.3% of the linagliptin arm and 4.2% of the glimepiride arm.
3. Incident hypoglycemia of any severity was significantly lower in patients treated with linagliptin compared to glimepiride (10.6% vs 37.7%, P < 0.001), representing an absolute risk reduction of 27.1%.
4. Moderate or severe hypoglycemic events occurred at a rate of 1.4 per 100 person-years with linagliptin versus 8.4 per 100 person-years with glimepiride.
5. Linagliptin treatment was associated with a modest but statistically significant relative weight reduction compared to glimepiride, with a weighted average mean difference in body weight of -1.54 kg.

Study Design

Design
Randomized Clinical Trial
Double-Blind
Sample
6,033
Patients
Duration
6.3 yr
Median
Setting
Multicenter, 43 countries
Population Adults with type 2 diabetes (HbA1c 6.5% to 8.5%) and elevated cardiovascular risk (defined by established atherosclerotic cardiovascular disease, multiple CV risk factors, age ≥70 years, or microvascular complications)
Intervention Linagliptin 5 mg once daily
Comparator Glimepiride 1 to 4 mg once daily
Outcome Time to first occurrence of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke (3P-MACE)

Study Limitations

The trial participants generally had relatively early type 2 diabetes (median duration ~6.2 years) and modest baseline glycemic elevation (HbA1c 6.5% to 8.5%), which may limit generalizability to populations with more advanced disease and poorer glycemic control.
The results specifically reflect the safety profiles of linagliptin and glimepiride; it remains unclear if these findings represent a class effect for all DPP-4 inhibitors or all sulfonylureas.
While the 6.3-year median follow-up is robust for modern diabetes outcome trials, atherosclerotic disease progression is a slow process, and longer observation might be required to detect diverging cardiovascular impacts.

Clinical Significance

CAROLINA provides critical head-to-head evidence that the sulfonylurea glimepiride does not increase cardiovascular risk compared to a modern DPP-4 inhibitor. This reassures clinicians who utilize sulfonylureas as a cost-effective second-line therapy. However, the dramatically lower incidence of hypoglycemia and the absence of weight gain strongly favor linagliptin as a safer alternative in patients where cost is not the primary barrier, particularly for frail patients, the elderly, or those operating heavy machinery where hypoglycemia poses severe risks.

Historical Context

Following the rosiglitazone controversy, the FDA mandated in 2008 that all new type 2 diabetes medications demonstrate cardiovascular safety. Most subsequent cardiovascular outcome trials (CVOTs) compared new agents against placebo (e.g., SAVOR-TIMI 53, EMPA-REG OUTCOME, LEADER). CAROLINA stands out as a unique active-comparator trial. Furthermore, it directly addressed a 50-year-old debate stemming from the 1970 University Group Diabetes Program (UGDP) study, which first raised alarms regarding the cardiovascular safety of sulfonylureas. Together with CARMELINA (which established linagliptin's safety versus placebo), CAROLINA definitively mapped the comparative safety landscape of DPP-4 inhibitors and modern sulfonylureas.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How do the mechanisms of action of linagliptin and glimepiride explain the significant difference in hypoglycemia and weight changes observed in the CAROLINA trial?

Key Response

Glimepiride is a sulfonylurea that binds to the SUR1 receptor on pancreatic beta cells, causing continuous, glucose-independent insulin release, which inherently risks hypoglycemia and promotes weight gain. Linagliptin is a DPP-4 inhibitor that prevents the breakdown of incretins like GLP-1 and GIP; these hormones stimulate insulin secretion only in a glucose-dependent manner, thereby minimizing hypoglycemia risk and remaining weight-neutral.

Resident
Resident

Given the noninferiority in cardiovascular outcomes but superiority in side effect profile for linagliptin demonstrated in CAROLINA, in which specific clinical scenarios would you still consider prescribing a sulfonylurea like glimepiride for a patient with type 2 diabetes?

Key Response

While DPP-4 inhibitors have a safer profile regarding hypoglycemia and weight, sulfonylureas remain highly efficacious for rapid glycemic lowering and are significantly cheaper. A resident must weigh the side effect profile against socioeconomic factors, utilizing glimepiride in patients where medication cost or insurance coverage is a major barrier to adherence, provided the patient is counseled on hypoglycemia management.

Fellow
Fellow

How does the CAROLINA trial's selection of an active comparator (glimepiride) rather than placebo contextualize the cardiovascular safety of both DPP-4 inhibitors and sulfonylureas, particularly when contrasted with the cardiovascular benefits seen with SGLT2 inhibitors and GLP-1 receptor agonists?

Key Response

CAROLINA proved that a modern sulfonylurea does not increase CV risk compared to a DPP-4 inhibitor, helping to debunk older fears (stemming from the UGDP study) about SU-induced CV mortality. However, it also highlights that unlike SGLT2is or GLP-1 RAs which demonstrate active cardiovascular risk reduction, DPP-4is and SUs merely possess neutral cardiovascular safety, reinforcing the preferential use of the former in high-risk patients.

Attending
Attending

The CAROLINA trial challenges historical dogma regarding the cardiovascular danger of sulfonylureas. How should this trial change your approach to deprescribing or maintaining glimepiride in older, stable patients who are not experiencing hypoglycemia?

Key Response

Attendings often face pressure to deprescribe sulfonylureas due to historical fears of cardiovascular harm and hypoglycemia. CAROLINA provides reassurance that in established, stable patients without a history of hypoglycemia, glimepiride is cardiovascularly safe, allowing for a more nuanced, patient-centered approach that avoids unnecessarily disrupting a well-tolerated, cost-effective regimen.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The CAROLINA trial utilized an active-comparator, non-inferiority design with a pre-specified non-inferiority margin of 1.3 for the hazard ratio. What are the methodological vulnerabilities of this design regarding assay sensitivity, and how does the constancy assumption affect the interpretation of the results?

Key Response

Using an active comparator assumes the comparator has a known, stable effect (constancy assumption) compared to placebo. Establishing assay sensitivity is challenging without a placebo arm, as the trial must rely on historical data to ensure that neither drug is worse than a putative placebo. If both drugs were equally harmful or equally ineffective, non-inferiority could still be met, which requires rigorous statistical justification of the chosen margin.

Journal Editor
Journal Editor

Given the relatively long median follow-up of over 6 years in the CAROLINA trial, how does the differential rate of medication discontinuation and the introduction of unblinded, off-protocol rescue medications (such as SGLT2is or GLP-1 RAs) threaten the internal validity of the intention-to-treat analysis for the primary MACE endpoint?

Key Response

Over a prolonged follow-up, patients often require therapy escalation. If one arm preferentially receives modern cardioprotective agents as rescue therapy due to worse glycemic control or side effects, the intention-to-treat analysis could be biased toward the null, masking a true difference in MACE. Peer reviewers must heavily scrutinize the balance and timing of rescue medications to ensure the outcome reflects the randomized drugs.

Guideline Committee
Guideline Committee

In light of the CAROLINA trial demonstrating cardiovascular safety but no cardiovascular benefit for both linagliptin and glimepiride, how should current ADA/EASD guidelines position these two drug classes in the treatment algorithm for patients with high ASCVD risk versus those primarily requiring cost-effective interventions?

Key Response

Current ADA guidelines prioritize SGLT2is and GLP-1 RAs for patients with ASCVD due to proven benefit. CAROLINA reinforces that DPP-4is and SUs should not be used for CV risk reduction (providing Level A evidence for neutrality). However, it directly informs guidelines that when cost is a major barrier, SUs are a safe alternative for glycemic control without conferring excess cardiovascular risk, validating their position in cost-conscious guideline algorithms.

Clinical Landscape

Noteworthy Related Trials

2013

SAVOR-TIMI 53 Trial

n = 16,492 · NEJM

Tested

Saxagliptin 5mg or 2.5mg daily

Population

T2DM patients with a history of or risk for CV events

Comparator

Placebo

Endpoint

3-point MACE

Key result: Saxagliptin did not increase or decrease the rate of ischemic events but significantly increased the risk of hospitalization for heart failure.
2015

TECOS Trial

n = 14,671 · NEJM

Tested

Sitagliptin 100mg or 50mg daily

Population

T2DM patients and established cardiovascular disease

Comparator

Placebo

Endpoint

4-point MACE

Key result: Sitagliptin was non-inferior to placebo for major adverse cardiovascular events and did not increase heart failure hospitalizations.
2019

CARMELINA Trial

n = 6,979 · JAMA

Tested

Linagliptin 5mg daily

Population

T2DM patients with high CV and renal risk

Comparator

Placebo

Endpoint

3-point MACE

Key result: Linagliptin was non-inferior to placebo for 3-point MACE, demonstrating cardiovascular safety but no superiority.

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