Large reductions in hsCRP appeared to foretell improved outcomes.
An important CANTOS trial substudy bolsters evidence that inflammation plays an independent role in cardiovascular disease (CVD) and may point toward a new targeted approach to treatment of CVD and perhaps even more profoundly impact cancer treatment.
Canakinumab is a monoclonal antibody that targets IL-1ß and has little or no effect on LDL or other cholesterol parameters. The main results of the large CANTOS trial, released earlier this year showing a significant benefit to the drug in ACS, were widely interpreted as providing the first proof of a role for inflammation in CVD.
But that benefit was modest in the total patient population, resulting in a 15% reduction in major adverse CV events compared to placebo.
The Subgroup
Ridker said that the substudy set out to ask whether there was an easily-identified clinical subgroup that had a large enough benefit to outweigh the hazards and would be worth the cost (although the cost of the drug for a CV indication is unknown at this time).
The new analysis looked at the group of CANTOS patients who achieved hsCRP levels below 2 mg/L and compared them with subjects in the study who had smaller reductions. (All patients in the trial had hsCRP levels above 2 at the start of the trial.)
Although baseline risk factors did not appear to determine efficacy, the reduction in hsCRP after a single dose of canakinumab was a strong predictor of canakinumab efficacy.
Study patients with hsCRP levels that stayed above 2 did not have a significant reduction in CV events (HR 0.95, CI 0.84-1.09, P=0.48). By contrast, those with hsCRP levels that dropped below 2 had a significant 25% reduction in major events (HR 0.75, CI 0.66-0.85, P<0.0001).
The authors also reported a highly significant 31% reduction in both CV death (P=0.0004) and overall mortality (P=0.0001) in the responder group, while there was no significant effect in the nonresponder group.
In the overall trial, the number needed to treat (NNT) to prevent one myocardial infarction, stroke, coronary revascularization, or death was 24.
In the substudy, the NNT was 16 for the responder group and 57 for the non-responders. The results were “virtually identical” for analyses substituting on-treatment IL-6 levels for hsCRP levels.
The major side effect of canakinumab in CANTOS was a small increase in fatal infection. Fatal infection rates were 0.27 per 100 person-years in the responder group, 0.35 in the non responder group, and 0.18 in the placebo group.
“These data support the use of an on-treatment measure of hsCRP as a simple clinical mechanism to differentiate between candidates for sustained canakinumab treatment and individuals much less likely to benefit from continued treatment,” Ridker’s group concluded.
Canakinumab’s Future in CVD
There has been widespread speculation that Novartis, the manufacturer of canakinumab, might pursue a strategy in which the initial dose of the drug would be given for free.
Only patients with a significant reduction in hsCRP would then continue to take — and pay for — the drug.
“Patients with residual inflammatory risk represent a separate and distinct group from patients with residual cholesterol risk who probably require different personalized approaches to treatment,” the authors wrote in their paper.
“We believe the clinical approach of targeting treatment to those who truly benefit on the basis of biological response represents a major step toward personalized medicine and rational resource utilization.”
The authors cautioned that this was not based on a randomized comparison and “could reflect, in part, differences in baseline characteristics.”
Sanjay Kaul (Cedars-Sinai) echoed this point in commenting to CardioBrief: “This is a responder analysis based on an observational dataset and therefore it is vulnerable to confounding, especially by unmeasured confounders, that no multivariable adjustment can overcome.”
He said that these “types of subgroup analyses that are based on a post-randomization variable are considered ‘improper’ and not reliable for regulatory or clinical decision-making.”
In an accompanying editorial, Roger Blumenthal and Erin Michos (Johns Hopkins) wrote that, in the overall study, the modest “reduction in the primary endpoint likely would not justify its routine use in all patients post myocardial infarction with elevated hsCRP.”
In his talk, Ridker described a targeted approach to preventing CV events in patients with known atherosclerosis already receiving statins.
He said there were two broad phenotypes: one with residual cholesterol risk who have elevated LDL levels, and another with residual inflammatory risk with elevated hsCRP.
The first would receive additional cholesterol treatment, including ezetimibe (Zetia) or a PCSK9 inhibitor, while the second would be eligible for additional inflammation reduction, such as with canakinumab or, down the road, perhaps other anti-inflammatory agents.
Lung Cancer Prevention
In his presentation, Ridker also reported that reduction in hsCRP was associated with a large reduction in lung cancer. There was a 71% reduction in patients who achieved below-median hsCRP levels, but there was no significant reduction in those above the median.
At a briefing for news media, a Novartis spokesperson said that in discussions with regulatory agencies it appeared likely that the cancer findings could not be included in the efficacy section but that they might be mentioned in the safety section.
It is widely expected that the FDA will schedule an advisory committee meeting to review the main CANTOS study and the substudy. Novartis has stated that it wants to pursue a targeted strategy, which would clearly require using the hsCRP substudy.
But the main trial by itself did not reach FDA pre-specified requirement for use as a single study, and the substudy would likely generate a lot of resistance for an indication based on the hsCRP strategy.
Novartis has indicated that it might be reluctant to pursue a broad, untargeted indication.