Review of the CANTOS Trial

The Canakinumab Anti-inflammatory Thrombosis Outcomes Study

N Engl J Med 2017;377:1119-1131

Background The hypothesis tested in the CANTOS Trial reflected an emerging view of atherosclerosis development and progression - that it was not simply a matter of plasma cholesterol but also systemic inflammation. In the manuscript’s introduction the authors cite data on the independent association of certain inflammatory markers with an increased risk of cardiovascular events independent of lipid levels.

For an anecdote that I suspect many cardiologists and general practitioners will relate to, I am involved with managing many younger women (<65 years of age) who lack “traditional” cardiovascular risk factors but have atherosclerotic CVD which typically presents as an ACS event. The thread that ties these women together is their history of a systemic inflammatory disorder like rheumatoid arthritis or inflammatory bowel disease. Since the publication of CANTOS and other trials, the relationship between systemic inflammatory disease and CVD has gained broader recognition but remains underrecognized. For a very general overview of this topic I would direct readers to the following link, which is free to read: Cardio-Rheumatology: Prevention of Cardiovascular Disease in Inflammatory Disorders - PubMed (nih.gov)

The Canakinumab Anti-inflammatory Thrombosis Outcomes Study (CANTOS) tested the hypothesis that Canakinumab, which is a fully human monoclonal antibody targeting interleukin-1 beta would reduce recurrent cardiovascular events in patients with a history of MI and persistent proinflammatory response.

Patients Eligible patients had a history of MI and a hs-CRP level >/= 2 mg/L despite use of aggressive secondary prevention strategies. The trial sought to enroll stable patients and thus, broad exclusion criteria were applied which included any patient with a history of recurrent infection or with any condition that could render a patient immunocompromised, such as ongoing use of other systemic inflammatory treatments. Thus, it should be appreciated that patients with severe systemic inflammation were excluded.

Baseline characteristics There were 17,482 patients with a previous MI who underwent screening in the central laboratory and 10,061 (57.6%) were randomized. Among these patients, the most common exclusions were due to hs-CRP <2 mg/L (46%), active tuberculosis or tuberculosis risk factors (25%) and exclusionary concomitant disorders (10%). The mean age of participants was 61 years, 26% were women and 41% had diabetes. Nearly all patients were taking antithrombotic agents and lipid lowering agents. The median hs-CRP was 4.2 mg/L and the median LDL was 82 mg/dL.

Procedures Patients were randomized 1.5:1:1:1 to receive placebo, canakinumab 50mg, canakinumab 150 mg or canakinumab 300 mg. All doses of canakinumab and placebo were administered subcutaneously once every 3 months; however, for the 300 mg canakinumab dose, the regimen was 300 mg every 2 weeks for the first 2 doses and then once every 3 months thereafter. Randomization was stratified by time from index MI.

Endpoints The primary efficacy endpoint was a composite endpoint that included time to first occurrence of cardiovascular death, nonfatal MI or nonfatal stroke in a time-to-event analysis. Secondary endpoints included individual components of the primary endpoint as well as others. Endpoint adjudication was done by a committee blinded to treatment assignment.

The trial was designed to have 90% power to detect a 20% lower event rate in the primary endpoint for at least 1 of the 3 doses of canakinumab compared to placebo. It was estimated that 17,200 patients would need to be enrolled in order to accrue 1,400 events over a 5 year period. While the trial was ongoing, at the request of the sponsor, the sample size was reduced to 10,000 and follow up was extended by 1 year in an effort to still achieve 1,400 events.

The formal evaluation of significance for individual doses was adjusted for multiple comparisons. The prespecified 2-sided P value thresholds for statistical significance for the primary endpoint were 0.01058 for the 300 mg canakinumab dose versus placebo and 0.02115 for tests of the other 2 canakinumab doses individually versus placebo.

Results Canakinumab was effective in lowering hs-CRP with 48 months reductions from baseline of 26%, 37% and 41% for 50 mg, 150 mg and 300 mg dose groups compared to placebo. Canakinumab did not change LDL or HDL cholesterol but did increase triglycerides by 4-5% compared to placebo.

Compared to placebo and according to the prespecified P value thresholds for the trial, the 150 mg canakinumab dose significantly reduced the primary endpoint of cardiovascular death, nonfatal MI and stroke (14% vs 16%; p=0.021); however, the 50 mg (14.4% vs 16%; p=0.30) and 300 mg (14.2% vs 16%; p=0.031) doses did not. The statistically significant difference for the 150 mg dose was driven by reduction in nonfatal MI (7.0% vs 8.7%; p=0.005) that was accentuated in the 150 mg dose group compared to the 50 and 300 mg doses.

No significant differences in nonfatal stroke (placebo rate [2.8%], 50 mg rate [2.7%], 150 mg rate [2.8%], and 300 mg rate [2.3%]), CV death (placebo rate [5.4%], 50 mg rate [4.3%], 150 mg rate [4.8%], and 300 mg rate [5.1%]) or all cause death (placebo rate [11.2%], 50 mg rate [10.5%], 150 mg rate [10.4%], and 300 mg rate [10.6%]) were seen in the trial.

When all canakinumab groups were combined, canakinumab significantly reduced the primary endpoint (14.2% vs 16%; p=0.02) and nonfatal MI (7.5% vs 8.7%; p=0.03) but not CV death or all cause death. These results should be de-emphasized as this was not the primary analysis framework for the trial.

The results of standard subgroup testing is not presented in the manuscript or supplement. Regarding adverse events, neutropenia was significantly more common in the canakinumab group as well as death from infection. Patients who died from infection were older and more likely to have diabetes. Thrombocytopenia was also more common in the canakinumab group but not major bleeding events. Canakinumab reduced arthritis and gout as well as cancer death.

Conclusions In patients with a previous MI and persistently elevated hs-CRP, the 150 mg dose of canakinumab reduced the primary composite endpoint. The estimated number of patients needed to treat with 150 mg of canakinumab to prevent 1 primary outcome event over 3.7 years is approximately 50. The difference in events was driven mainly by nonfatal MI. Significant differences in nonfatal stroke, CV death and all-cause death were not observed for the 150 mg group compared to placebo nor for any of the other canakinumab groups. When other canakinumab dose groups are compared to placebo or when all canakinumab groups are combined, the results generally mirror those of the 150 mg dose group.

Canakinumab significantly increased neutropenia, thrombocytopenia and death from infection compared to placebo. Information on subgroups is not provided.

Via its action on reducing systemic inflammation, canakinumab reduces cardiac events in well-selected patients with a history of MI and persistently elevated hs-CRP; however, the clinical effect is small (NNT = 50) and driven by nonfatal events. Serious adverse events, such as death from infection, are increased with canakinumab and it is possible that in unselected patient populations, particularly those with severe systemic inflammation and systemic inflammatory disorders that these adverse events could overwhelm any potential benefits.