Accepted Manuscript

Anti-Inflammatory Therapy with Canakinumab for the Prevention and Management ofDiabetes

Brendan M. Everett, MD, MPH, Marc Y. Donath, MD, Aruna D. Pradhan, MD, MPH,Tom Thuren, MD, Prem Pais, MD, Jose C. Nicolau, MD, Robert J. Glynn, ScD, PeterLibby, MD, Paul M. Ridker, MD, MPH

PII: S0735-1097(18)33483-1

DOI: 10.1016/j.jacc.2018.03.002

Reference: JAC 24732

To appear in: Journal of the American College of Cardiology

Received Date: 14 February 2018

Revised Date: 28 February 2018

Accepted Date: 1 March 2018

Please cite this article as: Everett BM, Donath MY, Pradhan AD, Thuren T, Pais P, Nicolau JC,Glynn RJ, Libby P, Ridker PM, Anti-Inflammatory Therapy with Canakinumab for the Preventionand Management of Diabetes, Journal of the American College of Cardiology (2018), doi: 10.1016/j.jacc.2018.03.002.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Anti-Inflammatory Therapy with Canakinumab for the Prevention and Management of Diabetes

Brendan M. Everett, MD, MPH,a,b Marc Y. Donath, MD,c Aruna D. Pradhan, MD, MPH,a,d Tom Thuren, MD,e Prem Pais, MD,f Jose C. Nicolau,g MD, Robert J. Glynn, ScD,a Peter Libby, MD,b

Paul M Ridker, MD, MPHa,b From the aCardiovascular Disease Prevention and the bDivision of Cardiovascular Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston; cEndocrinology, Diabetes & Metabolism, University Hospital Basel, Basel, Switzerland; dDivision of Cardiovascular Medicine, Department of Medicine, VA Boston Medical Center, West Roxbury, Massachusetts; eNovartis, East Hanover, NJ, and Basel, Switzerland; fSt. John’s Research Institute, Bangalore, India; gInstituto do Coracao (InCor), Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, Brazil. Brief Title: Canakinumab for Diabetes Prevention and Management Funding: Funded by Novartis, Basel, Switzerland Disclosures: PMR, RJG, PP, and JCN received research grant support from Novartis Pharmaceuticals to conduct the CANTOS trial. PMR, BME, and PL have served as a consultants to Novartis and BME has grant support from Novartis for work unrelated to CANTOS. PMR is listed as a co-inventor on patents held by the Brigham and Women’s Hospital that relate to the use of inflammatory biomarkers in cardiovascular disease and diabetes that have been licensed to AstraZeneca and Siemens. TT is an employee of, and holds stock in, Novartis Pharmaceuticals. All other authors declare no competing interests. Address for Correspondence: Brendan M. Everett, MD, MPH Brigham and Women’s Hospital Divisions of Preventive and Cardiovascular Medicine 900 Commonwealth Ave Boston, MA 02215 Telephone: 857-307-1990 Fax: 857-307-1955 E-mail: beverett@bwh.harvard.edu Twitter: @_brendan_

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Abstract Background: Subclinical inflammation mediated in part by interleukin-1 beta (IL-1β) participates in peripheral insulin resistance and impaired pancreatic insulin secretion. Objectives: We tested the hypothesis that canakinumab, an IL-1β inhibitor, reduces incident diabetes. Methods: The Canakinumab Antiinflammatory Thrombosis Outcomes Study (CANTOS) randomized 10,061 patients with prior MI and high-sensitivity C-reactive protein (hsCRP) ≥2 mg/L to placebo or canakinumab at doses of 50mg, 150mg, or 300mg subcutaneously once every three months. We tested the effects of canakinumab on major cardiovascular events in patients with and without diabetes at baseline, and evaluated as a pre-specified analysis whether canakinumab would reduce the risk of adjudicated cases of new onset type 2 diabetes among those with protocol-defined pre-diabetes at trial entry. We also evaluated the effect of canakinumab on fasting plasma glucose and HbA1c in patients with and without established diabetes. Results: Of CANTOS participants, 4,057 (40.3%) had baseline diabetes, 4,960 (49.3%) had pre-diabetes, and 1,044 (10.4%) had normal glucose. Among those without diabetes, increasing tertiles of hsCRP at baseline associated with an increased risk of developing diabetes during the median follow-up period of 3.7 years (incidence rates 3.2, 4.1, and 4.4 per 100 person years, P = 0.003). Canakinumab 150mg as compared to placebo had similar magnitude effects on major cardiovascular event rates among those with diabetes (HR 0.85, 95% CI 0.70-1.03), pre-diabetes (0.86, 95% CI 0.70-1.06), and normoglycemia (HR 0.81, 95% CI 0.49-1.35). Despite large reductions in hsCRP and interleukin-6, canakinumab did not reduce the incidence of new onset diabetes, with rates per 100 person years in the placebo, 50mg, 150mg, and 300mg canakinumab groups of 4.2, 4.2, 4.4, and 4.1, respectively (Log-rank P=0.84). The hazard ratio comparing all canakinumab doses to placebo was HR 1.02 (95% CI 0.87-1.19, P=0.82). Canakinumab reduced HbA1c during the first 6-9 months of treatment but no consistent long-term benefits on HbA1c or fasting plasma glucose were observed. Conclusions: While IL-1β inhibition with canakinumab had similar effects on major cardiovascular events among those with and without diabetes, treatment over a median period of 3.7 years did not reduce incident diabetes. CANTOS ClinicalTrials.gov number, NCT01327846. Condensed Abstract: We tested the hypothesis that canakinumab, an IL-1β inhibitor, prevents new onset diabetes among patients with pre-diabetes in the CANTOS trial. In CANTOS, 4,057 (40.3%) had baseline diabetes, 4,960 (49.3%) had pre-diabetes, and 1,044 (10.4%) had normoglycemia. Baseline hsCRP and interleukin-6 predicted new onset diabetes among those without diabetes at baseline, and canakinumab 150mg vs. placebo had similar magnitude effects on major cardiovascular event rates among those with diabetes [HR 0.85 (0.70-1.03)], pre-diabetes [0.86 (0.70-1.06)], and normoglycemia [0.81 (0.49-1.35)]. Canakinumab reduced hsCRP and interleukin-6, as well as HbA1c over the short-term, but did not prevent new onset diabetes [HR 1.02 (0.87-1.19, P=0.82]. Key Words: Inflammation, diabetes, cardiovascular disease, randomized trial Abbreviations: BMI, body mass index; CANTOS, Canakinumab Antiinflammatory Thrombosis Outcomes Study; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; IL-1, interleukin-1 beta: IL-6, interleukin-6, NF-κB, nuclear factor kappa-B; NLRP3, NOD-like receptor pyrin-3

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Introduction

Substantial data support a pathologic role for subclinical inflammation in both insulin

sensitivity and impaired insulin production from pancreatic beta cells that promotes initiation and

progression of type 2 diabetes.(1-3) Cellular and animal experiments suggest prolonged

hyperglycemia and the deposition of amyloid polyprotein in pancreatic islet cells lead to

induction of the NOD-like receptor pyrin-3 (NLRP3) inflammasome, followed by activation of

interleukin (IL)-1β and local auto-induction of IL-1 in pancreatic islets.(4-6) Macrophage

recruitment and local inflammation could then lead to impaired insulin secretion and beta cell

death, a critical step in progression from pre-diabetes to type 2 diabetes.(2,4,7) Clinically, a

randomized trial of anakinra, an IL-1 receptor antagonist, showed improvements in beta cell

function and peripheral glucose sensitivity, and reductions in HbA1c, in patients with established

type 2 diabetes.(8) Longer term follow up of these patients revealed persistence of these

beneficial effects.(9) Other small studies with IL-1 inhibitors have shown modest reductions in

HbA1c.(10,11) Yet, while these data support the inflammation hypothesis of diabetes, no data

inform whether IL-1 inhibition can prevent incident type 2 diabetes in a pre-diabetic population.

A pre-specified secondary endpoint of the recently reported Canakinumab Anti-inflammatory

Thrombosis Outcome Study (CANTOS) specifically addressed this hypothesis.(12) In addition,

the CANTOS trial structure included many patients with preexisting diabetes, and thus afforded

the additional opportunity to evaluate whether IL-1β inhibition slows the progression of type 2

diabetes as assessed by HbA1c and fasting plasma glucose.

Methods

Study Design and Participants

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

CANTOS was a randomized, double-blind, placebo-controlled trial of three doses of

canakinumab (50 mg, 150mg, or 300 mg) compared to placebo that enrolled 10,061 patients with

prior myocardial infarction and high-sensitivity C-reactive protein (hsCRP) ≥2 mg/L.(12) The

study protocol pre-specified that the occurrence of type 2 diabetes among those with pre-diabetes

at baseline was a key secondary endpoint of the trial.

Before randomization, participants were assessed for the presence of either diabetes (type

1 or 2) as well as protocol-defined pre-diabetes. The assessment of the effect of canakinumab on

the occurrence of type 2 diabetes included only those without diabetes at baseline, and the

primary analysis was the protocol pre-specified effect of canakinumab among those with pre-

diabetes at baseline. The definition of diabetes at baseline included any patient with a medical

history of type 2 diabetes, or any patient currently on an anti-diabetic medication, or any patient

with an HbA1c ≥ 6.5% at screening and randomization; or a fasting plasma glucose (FPG) of

≥126 mg/dL at screening and randomization; or a combination of FPG ≥126 mg/dL and a HbA1c

≥6.5% at screening or randomization; or any patient subsequently determined by the Clinical

Endpoints Committee to have developed type 2 diabetes with a date of onset equal to or prior to

the date of randomization. Pre-diabetes was defined by an HbA1c of 5.7 to <6.5% at screening

or randomization; a fasting plasma glucose of 100 to 125 mg/dL (5.6-6.9 mmol/L) at screening

or randomization; a fasting plasma glucose ≥126 mg/dL at either screening or randomization, but

not both; or a HbA1c ≥6.5% at either screening or randomization, but not both.(13) Patients with

an HbA1c <5.7 and a fasting plasma glucose < 100 mg/dL at screening and randomization were

defined as normoglycemic.

Procedures

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Patients enrolled in CANTOS had blood collected for assessment of fasting plasma

glucose and HbA1c at screening and randomization, and then again at months 1.5, 3, 6, 9, 12,

and every 12 months thereafter. Patients were asked about a new diagnosis of type 2 diabetes and

medications were assessed for the addition of any new anti-diabetic medications at in-person

visits every 3 months. Baseline and 3 month concentrations of hsCRP and interleukin (IL)-6

assessed inflammation inhibition with canakinumab for individual participants.

Washout Period

To determine whether any effect of canakinumab on the development of type 2 diabetes

was maintained after study drug discontinuation, patients with pre-diabetes at baseline and at the

time of trial completion had blood collected for HbA1c and fasting plasma glucose 6 months

after study drug discontinuation. Other trial participants entered an open-label extension period.

Endpoints

The CANTOS primary cardiovascular endpoint was the composite of non-fatal

myocardial infarction, non-fatal stroke, or cardiovascular death. The CANTOS secondary

cardiovascular endpoint was the primary endpoint plus hospitalization for unstable angina with

unplanned coronary revascularization. Both endpoints were assessed by a clinical endpoint

committee composed of cardiologists and neurologists blinded to study drug allocation. New

onset type 2 diabetes was adjudicated by an endpoints committee of endocrinologists who were

blinded to study drug allocation. The definition of type 2 diabetes utilized by the adjudication

committee required an HbA1c ≥6.5% or a fasting plasma glucose ≥126 mg/dL, or the

combination, on two blood draws within 6 weeks of one another, or a new prescription of an

anti-diabetic medication. The date of onset was the date of the first laboratory abnormality if the

diagnosis was made using laboratory parameters, or the date of anti-diabetic medication

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

prescription. In a sensitivity analysis, we expanded the definition of type 2 diabetes to include

physician-diagnosed type 2 diabetes, even if the diagnosis did not meet the definition used by the

Clinical Endpoints Committee. Results for the primary and key secondary cardiovascular

endpoints were previously published,(12) whereas results for the other key secondary endpoint

of incident diabetes in participants with pre-diabetes are presented here so as to include results

from the washout period.

Statistical Analysis

We first identified the effect of canakinumab on the occurrence of the trial primary and

secondary cardiovascular endpoints in patients with and without baseline diabetes. Kaplan-

Meier graphs and unadjusted log-rank P-values were used to evaluate differences between

groups. Evidence for heterogeneity of treatment across baseline diabetes strata was tested by

means of a likelihood ratio test by adding multiplicative interaction terms for randomized

treatment (all canakinumab doses combined vs. placebo) x baseline pre-diabetes, diabetes, or

normoglycemia status to a proportional hazards model.

We then attempted to replicate previously reported associations between baseline tertiles

of hsCRP and IL-6 and new onset type 2 diabetes among the CANTOS patients without diabetes

at study entry.(1) We utilized Kaplan-Meier graphs, unadjusted log-rank P-values and adjusted

Cox proportional hazards models to compare the rates of new onset type 2 diabetes by tertile of

baseline hsCRP or IL-6. Of the 6,004 patients without diabetes at entry, baseline hsCRP and IL-

6 measures were available in 6,001 and 2,928, respectively. Consistent with prior work, models

were adjusted for age, sex, race, body mass index (BMI), family history of diabetes, smoking,

physical activity, alcohol use, and randomized treatment allocation.(1)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

On a protocol pre-specified basis, we compared unadjusted incidence rates of

adjudicator-determined type 2 diabetes in the placebo and canakinumab 50mg, 150mg, and 300

mg treatment groups among patients with pre-diabetes at baseline. Kaplan-Meier graphs and

unadjusted log-rank P-values evaluated differences between randomly allocated treatment

groups, and Cox proportional hazards were used to calculate the risk of new onset type 2

diabetes among those randomly allocated to each of the canakinumab doses vs. placebo, as well

as for all canakinumab doses combined vs. placebo. We evaluated the effect of canakinumab vs.

placebo on HbA1c and fasting plasma glucose over the course of the trial in patients with

baseline pre-diabetes, established diabetes, and normoglycemia using linear regression mixed

models with percent change from baseline as the outcome and adjusted for baseline measure.

Among patients with diabetes at study entry, a protocol pre-specified analysis compared

HbA1c by treatment group over the course of the study. In a sensitivity analysis, we analyzed

changes in HbA1c among participants with a baseline diabetes and a HbA1c ≥8.0%. A second

protocol pre-specified analysis compared the time to failure of glycemic control, defined as the

time to a HbA1c ≥ 7.5% for those with baseline diabetes and a HbA1c <7%. We compared the

mean number of classes of diabetes medications by treatment group, and the proportion of

patients who initiated insulin among those with baseline diabetes in the placebo vs. canakinumab

groups. The diabetes medication classes were insulin, thiazolidinediones, metformin/guanides,

sulfonlyureas, alpha-glucosidase inhibitors, meglitinides, dipeptidyl peptidase 4 inhibitors,

glucagon-like peptide-1 receptor agonists, and sodium-glucose co-transporter 2 inhibitors.

Role of the Funding Source

CANTOS was funded by Novartis Pharmaceuticals. The sponsor was involved in the

design of the trial protocol and the collection of trial data. Drs. Everett and Ridker had full

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

access to the trial databases, generated trial analyses, and made the decision to submit the

manuscript for publication. Dr. Everett prepared the first draft of the manuscript.

Results

Of the 10,061 patients randomized in CANTOS, 4,057 had diabetes at baseline, 4,960

had protocol defined pre-diabetes, and 1,044 had normal glucose. Compared to placebo,

canakinumab had similar efficacy in preventing major cardiovascular events among those with

and without diabetes at trial entry. The estimated risk of the CANTOS primary cardiovascular

endpoint for patients randomly allocated to all doses of canakinumab vs. placebo was 0.90 (95%

CI 0.77-1.05) among those with baseline diabetes, and 0.86 (0.73-1.01) with baseline pre-

diabetes, and 0.81 (0.56-1.19) among those with normal glucose at baseline (P-heterogeneity =

0.86; Online Table 1). Comparable estimates for the secondary cardiovascular endpoint were

0.90 (0.78-1.04) for patients with diabetes, 0.81 (0.70-0.95) for patients with pre-diabetes, and

0.77 (0.54-1.11) for those with normal glucose at baseline (P-heterogeneity = 0.56; Online Table

2). Compared to placebo, randomly allocated canakinumab was associated with a similar

reduction in the absolute risk of the primary cardiovascular endpoint in patients with baseline

diabetes (0.55), pre-diabetes (0.55), and normoglycemia (0.64).

Baseline concentrations of hsCRP and IL-6 significantly predicted new onset type 2

diabetes among those without diabetes at study entry (Figure 1). In models adjusted for age, sex,

race, and study drug allocation, relative risks in the lowest to highest tertile of hsCRP were 1.0,

(referent), 1.29 (1.08-1.56) and 1.36 (1.13-1.63), P-trend <0.001. Comparable tertile data for IL-

6 were 1.0 (referent), 1.42 (1.09-1.84), and 1.93 (1.50-2.48), P-trend<0.001. These results were

modestly attenuated but remained statistically significant after adjustment for BMI, family

history of diabetes, smoking, exercise, alcohol use, and randomly allocated therapy. They were

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

attenuated and no longer statistically significant for hsCRP, but remained significant for IL-6

after the addition of baseline HbA1c to the model. (Online Table 3).

The baseline characteristics of the 4,960 patients with protocol defined pre-diabetes at

study entry were well balanced across randomized treatment assignment (Online Table 4).

Similar to observations made in the CANTOS trial as a whole, the median percent

reduction (IQR) in hsCRP after the first dose of canakinumab, compared to placebo, among

those with protocol-specified prediabetes in the 50, 150, and 300 mg groups were -49.2 (-20.0 to

-67.2), -61.5 (-33.3 to -75.8), and -67.1 (-43.2 to -80.6), respectively. Comparable data for IL-6

were -25.7 (0.7 to -46.6), -37.4 (-9.1 to -54.9), and -43.4 (-21.0 to -60.0). Among those with

diabetes at baseline, a similar patterns was seen (Online Table 5).

In spite of these significant, dose dependent reductions in inflammation with

canakinumab, IL-1β inhibition did not reduce rates of new onset diabetes, a major pre-specified

secondary endpoint of CANTOS. The incidence rates of adjudicated cases of new-onset diabetes

were 4.20, 4.24, 4.35, and 4.12 in the placebo, 50mg, 150mg and 300 mg canakinumab groups,

respectively (Table 1, all P-values ≥ 0.70). When all canakinumab doses were combined and

compared to placebo, the hazard ratio was 1.02 (95% CI 0.87-1.18; P-value = 0.85)(Figure 2A

and 2B). The use of physician reported diabetes as an endpoint yielded similar neutral results

(Table 1 and Online Figure 1). We repeated these analyses in the subgroup of patients with

normal glucose at study entry (Online Table 6) and saw no benefit of canakinumab on rates of

new onset diabetes. We also found no evidence that compared to placebo, canakinumab reduced

the risk of new onset diabetes among those who achieved hsCRP <2mg/L or an IL-6 less than the

median (1.54 ng/L).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Participants with protocol-defined prediabetes randomized to active canakinumab

demonstrated modest but statistically significant reductions in HbA1c over the first 6-9 months,

an effect that attenuated over time such that there was no difference in HbA1c at 48 months

(Figure 3A and 3B, Online Table 7). A similar pattern was observed among those with diabetes

(Figure 3C and 3D, Online Table 7) or normoglycemia at trial entry (Figure 3E and 3F; Online

Table 7). No consistent effects were observed for fasting plasma glucose (Online Figure 2 and

Online Table 8). In a sensitivity analysis conducted among 1224 patients with baseline diabetes

and HbA1c ≥8.0%, we observed an initial fall in HbA1c and fasting plasma glucose in all 4

treatment groups for the first 6 months of the study, reflecting regression to the mean or more

aggressive treatment, and similar HbA1c and fasting glucose levels for the remainder of the

study (Online Figure 3 and Online Table 9). There were no significant differences in the time to

HbA1c ≥7.5% among patients with baseline diabetes and a baseline HbA1c <7% (Online Figure

4). Among patients with baseline diabetes, the mean number of diabetes medication classes did

not differ between active and placebo treatment groups, (Online Table 10) and a similar

proportion of patients with type 2 diabetes randomly allocated to placebo (11.4 percent) and

canakinumab (11.8 percent) initiated insulin during the trial (P=0.40).

Washout Period

The baseline characteristics of the 1770 patients enrolled in the washout period were well

balanced across treatment groups (Online Table 11). At the beginning of the washout period, the

median (IQR) HbA1c for placebo and 50 mg, 150 mg, and 300 mg treatment groups were 5.9

(5.7-6.2), 5.9 (5.7-6.2), 5.9 (5.7-6.1), and 5.9 (5.7-6.1), respectively. No clinically substantive

changes in HbA1c or fasting plasma glucose were observed 6 months after discontinuation of

canakinumab (Table 2).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Adverse events

Irrespective of drug allocation, rates of serious infections were higher among patients

with diabetes than without diabetes (incidence rate per 100 person years, 3.94 vs 2.49; HR 1.58 ,

95% CI 1.41-1.78, P<0.0001), as were rates of fatal infections (incidence rate per 100 person

years, 0.41 vs. 0.17, HR 2.42, 95% CI 1.62-3.62, P<0.0001). Among patients with diabetes,

canakinumab as compared to placebo was not associated with an increased risk of infection (HR

1.05, 95% CI 0.88-1.25, P=0.63) but was associated with a significant increase in fatal infection

(HR 1.86, 95% CI 1.01-3.43) as was seen in the trial as a whole.

Discussion

This randomized, double blind, placebo controlled trial of IL-1β inhibition affirmed that

baseline concentrations of hsCRP and IL-6 predict incident diabetes, and found that

canakinumab has similar efficacy for major cardiovascular events in patients with and without

diabetes. However, contrary to our protocol pre-specified hypothesis, IL-1β inhibition over a 5-

year period did not limit the development of new onset diabetes. Canakinumab reduced HbA1c

over the first 6-9 months, an effect that attenuated over time such that there was no difference in

HbA1c at 48 months. In contrast to our prior publications for atherosclerosis and cancer, where

we saw greater risk reductions with higher doses of canakinumab, we saw no dose response

effect of IL-1β inhibition for incident diabetes.(12,14)

These data have several important implications for both clinical practice and

pathophysiology. First, our neutral data for incident diabetes indicate that previously reported

benefits on major cardiovascular events in CANTOS as a whole do not relate to the prevention of

diabetes or to changes in glucose metabolism. These data add to prior evidence from CANTOS

which showed no effect of canakinumab on low-density lipoprotein cholesterol levels, and thus

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

reinforce that the benefits of IL-1β antagonism on atherosclerosis appear due to inflammation

inhibition alone (Central Illustration).(12,15)

Second, while canakinumab did not reduce the risk of new onset type 2 diabetes in

CANTOS, nor did IL-1β neutralization increase the risk of diabetes. These analyses reassure that

current strategies to treat atherosclerosis through inflammation inhibition do not appear to hasten

the onset of diabetes in those already at risk.(16)

Third, careful interpretation of these results is warranted given the a priori strength of the

hypothesis relating inflammation and diabetes. In our study, over the first 6 to 9 months,

canakinumab led to small decreases in HbA1c in the first 6-9 months that were similar in

magnitude to prior work in patients with type 2 diabetes.(8,10,11,17) However, this effect was

attenuated over time. The reason for this attenuation is not clear, but may be due to the design of

the study, which allowed lifestyle interventions and alterations in other anti-diabetic therapies.

Indeed, in other cardiovascular outcome studies using “classical” antidiabetic drugs (dipeptidyl

peptidase 4 and sodium–glucose cotransporter 2 inhibitors), similar patterns and magnitude of

effects were observed.(18,19)

Our data do not address the possibility that cytokines beyond IL-1 may drive the

progression from prediabetes to type 2 diabetes. Adipose tissue can elaborate tumor necrosis

factor, a cytokine that causes insulin resistance in mouse models of type 2 diabetes.(20)

However, the role of tumor necrosis factor in humans with type 2 diabetes or metabolic

syndrome remains to be tested in well-designed studies.(21-23) Salicylate, a less-targeted anti-

inflammatory agent that is thought to inhibit NF-κB activity by activating adenosine

monophosphate-activated protein kinase, modestly lowered HbA1c and inflammatory mediators

in patients with type 2 diabetes.(24) NF-κB integrates and amplifies the inflammatory signals

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

transmitted by key chemokines and cytokines known to associate with type 2 diabetes and

insulin resistance, including IL-1β and IL-6, two key cytokines directly affected by canakinumab

administration.(25) Agents with less targeted anti-inflammatory actions that nonetheless alter

NF-κB activity, such as salicylate, may be more likely to alter the key inflammatory pathways

contributing to insulin resistance and progression from pre-diabetes to type 2 diabetes.(3) The

ongoing Cardiovascular Inflammation Reduction Trial of low-dose methotrexate in patients with

established coronary artery disease and diabetes or the metabolic syndrome is tracking incident

type 2 diabetes.(26) Just as anti-inflammatory agents targeting alternative, non-IL-1β pathways

failed to show a benefit on major cardiovascular events prior to CANTOS,(27,28) drugs

targeting alternative inflammatory pathways may yet show benefit in diabetes prevention.

More than 20 years ago, the “common soil” hypothesis postulated that a shared

antecedent for both type 2 diabetes and cardiovascular disease.(29) While inflammation remains

important for both disorders, data from CANTOS suggest divergence of inflammatory pathways

in cardiovascular disease and diabetes.

CANTOS included more than 9,000 patients with pre-diabetes and diabetes within a

randomized, double-blind, placebo controlled trial. Nonetheless, the CANTOS eligibility criteria

limit the generalizability of our findings to those with prior myocardial infarction and an hsCRP

≥ 2 mg/L. As shown here, these criteria selected for a population with a remarkably high

prevalence of abnormal glucose tolerance (90% of all participants), emphasizing the utility of

hsCRP, a measure of subclinical inflammation, in identifying patients with abnormalities in

glucose metabolism.(30) CANTOS was not designed specifically to test the effect of

canakinumab among patients with established diabetes, so the results of glucose control in that

population should be considered hypothesis-generating.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Conclusions

In conclusion, random allocation to canakinumab had similar efficacy for cardiovascular

events in patients with and without diabetes at study entry. Yet, IL-1β inhibition did not reduce

the risk of new onset diabetes in spite of significant reductions in hsCRP and IL-6, a transient

improvement of HbA1c, nor did it have long-lasting effects on glycemia among those with

diabetes.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Perspectives: Clinical competencies and translational outlook

Competency in Medical Knowledge 1: Canakinumab provides similar cardiovascular event

reduction in patients with diabetes, pre-diabetes, and normoglycemia.

Competency in Medical Knowledge 2: Markers of subclinical inflammation, including high-

sensitivity C-reactive protein and interleukin-6, predict the development of type 2 diabetes

Competency in Medical Knowledge 3: In spite of large reducctions in high-sensitivyt C-

reactive protein and interleukin-6, and modest short-term reductions in hemoglobin A1c,

canakinumab does not prevent the development of new onset diabetes among those with pre-

diabetes at study entry.

Translational Outlook 1: Although there is evidence from human and animal experiments that

subclinical inflammation plays a pathologic role in insulin sensitivity and impaired insulin

production, canakinumab, and inhibitor of IL-1β, did not prevent the development of type 2

diabetes. Whether other anti-inflammatory therapies might be more effective will need to be

addressed in future studies.

Translational Outlook 2: While canakinumab lowered HbA1c over the first 6-9 months, an

effect that attenuated over time, the biological basis for why it stopped lowering HbA1c, and the

mechanism by which it lowered HbA1c, such as increased insulin production or enhanced insulin

sensitivity, remains unknown.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

References

1. Pradhan AD, Manson JE, Rifai N, Buring JE, Ridker PM. C-reactive protein, interleukin

6, and risk of developing type 2 diabetes mellitus. JAMA 2001;286:327-34.

2. Donath MY. Targeting inflammation in the treatment of type 2 diabetes: time to start. Nat

Rev Drug Discov 2014;13:465-76.

3. Goldfine AB, Shoelson SE. Therapeutic approaches targeting inflammation for diabetes

and associated cardiovascular risk. J Clin Invest 2017;127:83-93.

4. Maedler K, Sergeev P, Ris F et al. Glucose-induced beta cell production of IL-1beta

contributes to glucotoxicity in human pancreatic islets. J Clin Invest 2002;110:851-60.

5. Boni-Schnetzler M, Thorne J, Parnaud G et al. Increased interleukin (IL)-1beta

messenger ribonucleic acid expression in beta -cells of individuals with type 2 diabetes

and regulation of IL-1beta in human islets by glucose and autostimulation. J Clin

Endocrinol Metab 2008;93:4065-74.

6. Masters SL, Dunne A, Subramanian SL et al. Activation of the NLRP3 inflammasome by

islet amyloid polypeptide provides a mechanism for enhanced IL-1beta in type 2

diabetes. Nat Immunol 2010;11:897-904.

7. Maedler K, Spinas GA, Lehmann R et al. Glucose induces beta-cell apoptosis via

upregulation of the Fas receptor in human islets. Diabetes 2001;50:1683-90.

8. Larsen CM, Faulenbach M, Vaag A et al. Interleukin-1-receptor antagonist in type 2

diabetes mellitus. N Engl J Med 2007;356:1517-26.

9. Larsen CM, Faulenbach M, Vaag A, Ehses JA, Donath MY, Mandrup-Poulsen T.

Sustained effects of interleukin-1 receptor antagonist treatment in type 2 diabetes.

Diabetes Care 2009;32:1663-8.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10. Cavelti-Weder C, Babians-Brunner A, Keller C et al. Effects of gevokizumab on

glycemia and inflammatory markers in type 2 diabetes. Diabetes Care 2012;35:1654-62.

11. Sloan-Lancaster J, Abu-Raddad E, Polzer J et al. Double-blind, randomized study

evaluating the glycemic and anti-inflammatory effects of subcutaneous LY2189102, a

neutralizing IL-1beta antibody, in patients with type 2 diabetes. Diabetes Care

2013;36:2239-46.

12. Ridker PM, Everett BM, Thuren T et al. Antiinflammatory Therapy with Canakinumab

for Atherosclerotic Disease. N Engl J Med 2017;377:1119-1131.

13. American Diabetes Association. Diagnosis and classification of diabetes mellitus.

Diabetes Care 2014;37 Suppl 1:S81-90.

14. Ridker PM, MacFadyen JG, Thuren T et al. Effect of interleukin-1beta inhibition with

canakinumab on incident lung cancer in patients with atherosclerosis: exploratory results

from a randomised, double-blind, placebo-controlled trial. Lancet 2017;390:1833-1842.

15. Ridker PM, MacFadyen JG, Everett BM et al. Relationship of C-reactive protein

reduction to cardiovascular event reduction following treatment with canakinumab: a

secondary analysis from the CANTOS randomised controlled trial. Lancet 2017.

16. Ridker PM, Pradhan A, MacFadyen JG, Libby P, Glynn RJ. Cardiovascular benefits and

diabetes risks of statin therapy in primary prevention: an analysis from the JUPITER trial.

Lancet 2012;380:565-71.

17. Ridker PM, Howard CP, Walter V et al. Effects of interleukin-1beta inhibition with

canakinumab on hemoglobin A1c, lipids, C-reactive protein, interleukin-6, and

fibrinogen: a phase IIb randomized, placebo-controlled trial. Circulation 2012;126:2739-

48.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18. Green JB, Bethel MA, Armstrong PW et al. Effect of Sitagliptin on Cardiovascular

Outcomes in Type 2 Diabetes. N Engl J Med 2015;373:232-42.

19. Zinman B, Wanner C, Lachin JM et al. Empagliflozin, Cardiovascular Outcomes, and

Mortality in Type 2 Diabetes. N Engl J Med 2015;373:2117-28.

20. Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis

factor-alpha: direct role in obesity-linked insulin resistance. Science 1993;259:87-91.

21. Ofei F, Hurel S, Newkirk J, Sopwith M, Taylor R. Effects of an engineered human anti-

TNF-alpha antibody (CDP571) on insulin sensitivity and glycemic control in patients

with NIDDM. Diabetes 1996;45:881-5.

22. Dominguez H, Storgaard H, Rask-Madsen C et al. Metabolic and vascular effects of

tumor necrosis factor-alpha blockade with etanercept in obese patients with type 2

diabetes. J Vasc Res 2005;42:517-25.

23. Bernstein LE, Berry J, Kim S, Canavan B, Grinspoon SK. Effects of etanercept in

patients with the metabolic syndrome. Arch Intern Med 2006;166:902-8.

24. Goldfine AB, Fonseca V, Jablonski KA et al. Salicylate (salsalate) in patients with type 2

diabetes: a randomized trial. Ann Intern Med 2013;159:1-12.

25. Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. J Clin Invest

2006;116:1793-801.

26. Everett BM, Pradhan AD, Solomon DH et al. Rationale and design of the Cardiovascular

Inflammation Reduction Trial: A test of the inflammatory hypothesis of

atherothrombosis. Am Heart J 2013;166:199-207 e15.

27. Stability Investigators, White HD, Held C et al. Darapladib for preventing ischemic

events in stable coronary heart disease. N Engl J Med 2014;370:1702-11.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

28. O'Donoghue ML, Glaser R, Cavender MA et al. Effect of Losmapimod on

Cardiovascular Outcomes in Patients Hospitalized With Acute Myocardial Infarction: A

Randomized Clinical Trial. JAMA 2016;315:1591-9.

29. Stern MP. Diabetes and cardiovascular disease. The "common soil" hypothesis. Diabetes

1995;44:369-74.

30. Ridker PM, Wilson PW, Grundy SM. Should C-reactive protein be added to metabolic

syndrome and to assessment of global cardiovascular risk? Circulation 2004;109:2818-

25.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Figure Legends

Central Illustration: Canakinumab, inflammation reduction, and the prevention of type 2

diabetes and major cardiovascular events. (A) Elevated hsCRP associates with an increased

risk of new onset diabetes among those without diabetes at study entry in CANTOS. The

median percent reduction (IQR) in hsCRP after the first dose of canakinumab, compared to

placebo, among those with protocol-specified prediabetes in the 50, 150, and 300 mg groups

were -49.2 (-20.0 to -67.2), -61.5 (-33.3 to -75.8), and -67.1 (-43.2 to -80.6), respectively. (B)

Canakinumab reduces HbA1c in patients with pre-diabetes during the first 6-9 months of

therapy, but not consistent long-term benefits were observed. (C) Canakinumab did not prevent

the pregression of patients with pre-diabetes to type 2 diabetes. (D) Canakinumab was equally

effective in preventing major cardiovascular events in patients with diabetes, pre-diabetes, and

normoglycemia at study entry. Shown here are data for the 150 mg dose of canakinumab.

Figure 1: Risk of centrally adjudicated new onset type 2 diabetes according to baseline

tertiles of high-sensitivity C-reactive protein and interleukin (IL)-6 in CANTOS. The

cumulative incidence of centrally adjudicated incident type 2 diabetes among the CANTOS

participants who did not have diabetes at study entry is displayed. The groups are stratified by

baseline tertile of (A) high-sensitivity C-reactive protein or (B) interleukin-6. The Kaplan-Meier

curves are unadjusted, and the number at risk at each year is included in the table below each

figure. Risk estimates from Cox proportional hazards models that are adjusted for potential

confounders are available in Online Table 3.

Figure 2. Risk of centrally adjudicated new onset type 2 diabetes according to randomized

treatment group in CANTOS. The cumulative incidence of centrally adjudicated incident type

2 diabetes among patients with protocol-defined pre-diabetes at baseline is displayed. The

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

groups are stratified by randomly allocated treatment group of placebo, canakinumab 50mg,

canakinumab 150mg, or canakinumab 300mg (A) or random allocation to placebo or

canakinumab (all doses) (B). All analyses were conducted in patients with protocol-defined pre-

diabetes at trial entry, and utilized the endpoint of centrally adjudicated new onset type 2

diabetes.

Figure 3: HbA1c concentrations and percent change in HbA1c concentrations throughout

the CANTOS trial. CAPTION: The least square mean HbA1c concentrations, and least square

mean percent change from baseline in HbA1c concentrations, are presented according to

randomized treatment allocation in patients with pre-diabetes at trial entry (panels A and B),

patients with diabetes at trial entry (panels C and D), and patients with normal glucose at trial

entry (panels E and F).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table 1: Rates of the protocol pre-specified secondary endpoint of centrally adjudicated

new onset type 2 diabetes in patients with prediabetes at study entry. The rates of all

physician-reported new onset diabetes in patients with prediabetes at baseline are also presented.

Canakinumab dose

Placebo 50 mg 150 mg 300 mg All doses

combined

Adjudicated

new-onset

diabetes

N events/N at risk 246/1645 161/1089 171/1094 169/1132 501/3315

Incidence rate* 4.20 4.24 4.35 4.12 4.23

Hazard ratio (95%

CI)

- 1.04(0.85,1.27) 1.03(0.85-

1.26)

0.98(0.80-

1.19)

1.01 (0.87-

1.18)

P-value† - 0.70 0.75 0.80 0.86

All

physician-

reported

diabetes

N events/N at risk 279/1645 186/1089 191/1094 190/1132 567/3315

Incidence rate* 4.84 4.97 4.92 4.68 4.85

Hazard ratio (95%

CI)

- 1.06 (0.88-

1.27)

1.02(0.84-

1.22)

0.97(0.80-

1.16)

1.01(0.88-

1.17)

P-value† - 0.56 0.88 0.70 0.89

*Incidence rate per 100 person years of observation

†P-value for canakinumab vs. placebo comparison

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table 2. Median (IQR) HbA1c and fasting plasma glucose concentrations in the patients who entered the study drug washout from the

time of randomization, throughout the study, at the beginning of the study drug washout period, and at the end of the study drug

washout period.

Treatment group Baseline

6 months 12 months 24 months 36 months 48 months End of study† End of washout

HbA1c (%)

Placebo 5.7(5.6-5.9) 5.7(5.5-5.9) 5.7(5.5-5.9) 5.8(5.6-6.0) 5.9(5.7-6.1) 6.0(5.8-6.2) 6.0(5.8-6.2) 5.8(5.7-6.1)

Canakinumab 50 mg 5.8(5.6-6.0) 5.7(5.5-5.9) 5.7(5.5-5.9) 5.8(5.5-6.0) 5.9(5.7-6.1) 5.9(5.7-6.2) 6.0(5.7-6.2) 5.9(5.7-6.1)

Canakinumab 150 mg 5.7(5.6-5.9) 5.6(5.5-5.8) 5.7(5.5-5.9) 5.7(5.5-6.0) 5.8(5.6-6.0) 5.9(5.7-6.1) 5.9(5.7-6.2) 5.9(5.7-6.1)

Canakinumab 300 mg 5.7(5.5-5.9) 5.6(5.4-5.8) 5.6(5.4-5.9) 5.7(5.5-5.9) 5.8(5.6-6.1) 5.9(5.7-6.1) 5.9(5.8-6.1) 5.9(5.7-6.1)

Fasting plasma glucose

(mg/dL)

Placebo 103(97-109) 103(95-110) 103(95-108) 103(95-110) 103(95-110) 103(95-110) 105(97-110) 105(97-113)

Canakinumab 50 mg 103(97-110) 105(97-112) 103(95-110) 105(97-112) 105(99-112) 105(99-112) 105(99-112) 106(99-115)

Canakinumab 150 mg 101(95-108) 103(97-110) 103(95-108) 103(97-111) 105(97-112) 105(101-112) 106(99-113) 106(99-114)

Canakinumab 300 mg 103(97-112) 103(95-111) 103(95-110) 103(97-112) 106(99-112) 105(97-112) 106(99-114) 106(99-114)

† The end of study visit that marked the beginning of the washout period occurred before 48 months in many study participants.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

B C

D

1.00.75

Hazard Ratio for Canakinumab 150 mg vs. Placebo

Diabetes

Pre-Diabetes

Normal glucose

Relative Risk of MI, stroke, or Cardiovascular Death

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

ONLINE SUPPLMENT

Anti-Inflammatory Therapy with Canakinumab for the Prevention and Management of Diabetes

Brendan M. Everett, MD, MPH; Marc Y. Donath, MD; Aruna D. Pradhan, MD, MPH; Tom Thuren, MD;

Prem Pais, MD; J. Nicolau, MD; Robert J. Glynn, ScD Peter Libby, MD; Paul M Ridker, MD, MPH

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

TABLE OF CONTENTS

TABLE OF CONTENTS................................................................................................................................... 2

Online Figure 1. ........................................................................................................................................... 3

Online Figure 2. ........................................................................................................................................... 4

Online Figure 3. ........................................................................................................................................... 5

Online Figure 4. ........................................................................................................................................... 6

Online Table 1. ............................................................................................................................................ 7

Online Table 2. ............................................................................................................................................ 8

Online Table 3. ............................................................................................................................................ 9

Online Table 4. .......................................................................................................................................... 10

Online Table 5. .......................................................................................................................................... 12

Online Table 6. .......................................................................................................................................... 13

Online Table 7. .......................................................................................................................................... 14

Online Table 8. .......................................................................................................................................... 15

Online Table 9. .......................................................................................................................................... 16

Online Table 10. ........................................................................................................................................ 17

Online Table 11. ........................................................................................................................................ 18

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

Online Figure 1.

Cumulative incidence of all physician-reported new onset type 2 diabetes, stratified by the four

CANTOS randomized treatment groups (Panel A) and for all canakinumab doses combined vs. placebo

(Panel B). All analyses are conducted in patients with protocol-defined pre-diabetes at trial entry.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

Online Figure 2.

Fasting plasma glucose concentrations and percent change in fasting plasma glucose concentrations

throughout the CANTOS trial in patients with baseline pre-diabetes (panels A and B), diabetes (panels C

and D), and normal glucose (panels E and F) at trial entry.

98

100

102

104

106

108

110

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Fa

stin

g P

lasm

a G

luco

se (

mg

/dL)

Month of Study

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Pe

rce

nt

Ch

an

ge

in

Fa

stin

g P

lasm

a

Glu

cose

fro

m B

ase

lin

e

Month of Study

140

145

150

155

160

165

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Fa

stin

g P

lasm

a G

luco

se (

mg

/dL)

Month of Study

-4.0

1.0

6.0

11.0

16.0

21.0

26.0

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Pe

rce

nt

Ch

an

ge

in

Fa

stin

g P

lasm

a

Glu

cose

fro

m B

ase

lin

e

Month of Study

84

86

88

90

92

94

96

98

100

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Fa

stin

g P

lasm

a G

luco

se (

mg

/dL)

Month of Study

Placebo 50mg 150mg 300mg

-4.0

-2.0

0.0

2.0

4.0

6.0

8.0

10.0

0 0.5 1.3 3 6 9 12 18 24 30 36 42 48

Pe

rce

nt

Ch

an

ge

in

Fa

stin

g P

lasm

a

Glu

cose

fro

m B

ase

lin

e

Month of Study

Placebo 50mg 150mg 300mg

A B

C D

E F

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

Online Figure 3.

Mean HbA1c (Panel A) and fasting plasma glucose (panel B) values throughout the trial, stratified by

randomized treatment group, in patients with baseline diabetes and HbA1c ≥ 8.0%.

Panel A:

Panel B:

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

Online Figure 4.

Cumulative incidence of HbA1c ≥7.5% among patients with baseline diabetes and a baseline HbA1c

<7.0%, stratified by the four canakinumab treatment groups (Panel A) and for all canakinumab doses

vs. placebo (Panel B).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

Online Table 1.

The effect of random allocation to canakinumab on the CANTOS trial primary cardiovascular endpoint, stratified by diabetes, pre-diabetes,

and normoglycemia status at trial entry.

Placebo Canakinumab 50 mg Canakinumab 150 mg Canakinumab 300 mg All Canakinumab

Baseline Diabetes N events/N at

risk

258/1342 153/863 161/961 162/891 476/2715

Incidence Rate* 5.53 5.16 4.68 5.13 4.98

Hazard ratio

(95% CI) vs.

Placebo

- 0.93 (0.76-1.14) 0.85 (0.70-1.03) 0.92 (0.76-1.12) 0.90 (0.77-1.05)

P-value vs.

placebo

- 0.48 0.10 0.42 0.17

Baseline Pre-

Diabetes

N events/N at

risk

234/1645 137/1089 136/1094 137/1132 410/3315

Incidence Rate* 3.93 3.52 3.38 3.25 3.38

Hazard ratio

(95% CI) vs.

Placebo

- 0.90(0.73-1.11) 0.86(0.70-1.06) 0.83(0.67-1.02) 0.86(0.73-1.01)

P-value vs.

placebo

- 0.30 0.16 0.07 0.07

Baseline

Normoglycemia

N events/N at

risk

43/357 23/218 23/229 23/240 69/687

Incidence Rate* 3.43 3.02 2.78 2.60 2.79

Hazard ratio vs.

Placebo

- 0.89(0.53-1.47) 0.81(0.49-1.35) 0.75(0.45-1.24) 0.81(0.56-1.19)

P-value vs.

placebo

- 0.64 0.41 0.26 0.29

P-heterogeneity - 0.86

* Per 100 person years of observation

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

Online Table 2.

The effect of random allocation to canakinumab on the CANTOS trial secondary cardiovascular endpoint, stratified by diabetes, pre-

diabetes, and normoglycemia status at trial entry.

Placebo Canakinumab 50 mg Canakinumab 150 mg Canakinumab 300 mg All Canakinumab

Baseline Diabetes N events/N at

risk

281/1342 169/863 177/961 171/891 517/2715

Incidence Rate* 6.09 5.78 5.21 5.46 5.47

Hazard ratio

(95% CI) vs.

Placebo

- 0.95 (0.78-1.14) 0.86 (0.71-1.03) 0.90 (0.74-1.09)

0.90 (0.78-1.04)

P-value vs.

placebo

- 0.56 0.11 0.27 0.15

Baseline Pre-

Diabetes

N events/N at

risk

272/1645 151/1089 150/1094 153/1132 454/3315

Incidence Rate* 4.65 3.91 3.76 3.67 3.78

Hazard ratio

(95% CI) vs.

Placebo

0.84(0.68-1.02) 0.81(0.66-0.99) 0.79(0.65-0.96) 0.81(0.70-0.95)

P-value vs.

placebo

0.07 0.04 0.02 0.01

Baseline

Normoglycemia

N events/N at

risk

48/357 24/218 25/229 24/240 73/687

Incidence Rate* 3.86 3.17 3.04 2.73 2.97

Hazard ratio vs.

Placebo

0.82(0.51-1.35) 0.79(0.49-1.28) 0.70(0.43-1.14) 0.77(0.54-1.11)

P-value vs.

placebo

0.44 0.33 0.15 0.16

P-heterogeneity - 0.56

* Per 100 person years of observation

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

Online Table 3.

Risk of new onset type 2 diabetes in patients enrolled in CANTOS according to baseline tertile of hsCRP and IL-6. Among the 6,004 patients

without diabetes at baseline, 6,001 had hsCRP measures available at baseline, and 2,928 patients had IL-6 measures available at baseline.

Hazard Ratio (95% CI) for Adjudicated Type 2 Diabetes

Risk Marker and Model Tertile 1 Tertile 2 Tertile 3 P-value for

trend*

hsCRP

hsCRP range, mg/L <3.1 3.1 to <5.43 ≥5.43

N events/N at risk 204/1987 270/2014 288/2000

Incidence rate per 100

person years

2.84 3.72 4.06 0.0002

Model 1† 1.0 (referent) 1.30(1.08-1.56) 1.41(1.18-1.69) 0.0001

Model 2‡ 1.0 (referent) 1.22(1.01-1.47) 1.28(1.06-1.54) 0.0070

Model 3§ 1.0 (referent) 1.13(0.94-1.36) 1.08(0.89-1.30) 0.47

IL-6

IL-6 range, ng/L <1.93 1.93 to <3.18 ≥3.18

N events/N at risk 104/979 130/973 154/976

Incidence rate per 100

person years

2.75 3.66 4.62 <0.0001

Model 1† 1.0 (referent) 1.40(1.08-1.81) 1.84(1.43-2.37) <0.0001

Model 2‡ 1.0 (referent) 1.24(0.95-1.62) 1.62(1.25-2.11) 0.001

Model 3§ 1.0 (referent) 1.02(0.78-1.33) 1.35(1.04-1.75) 0.01

*P-value for trend is the log-rank P-value for the incidence rates, and the P-value for an ordinal term representing either hsCRP or IL-6

tertile that was added to the Cox proportional hazards model.

†Model 1 adjusted for age, sex, race, and random allocation to canakinumab or placebo.

‡Model 2 adjusted for the covariates in Model 1 plus body mass index, family history of diabetes, smoking, exercise, and alcohol use.

§Model 3 adjusted for the covariates in Model 2 plus baseline HbA1c.

Abbreviations: hsCRP, high-sensitivity C-reactive protein; IL-6, interleukin-6.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

Online Table 4.

Baseline characteristics of the 4,960 patients with pre-diabetes at study entry, stratified according to randomized treatment assignment.

Canakinumab

Placebo

(n=1645)

50 mg (n=1089) 150 mg

(n=1094)

300 mg

(n=1132)

All doses

(n=3315)

Age (yr) 61(54-68) 61(54-68) 61(54-68) 61(54-68) 61(54-68)

Female sex (%) 383(23%) 242(22%) 261(24%) 277(24%) 780(24%)

Smoking status (%)

Current smoking 428(26%) 295(27%) 290(27%) 295(26%) 880(27%)

Past smoking 792(48%) 515(47%) 524(48%) 535(47%) 1574(47%)

Never smoker 425(26%) 279(26%) 280(26%) 302(27%) 861(26%)

Body mass index (kg/m2) 29.1(26.1-32.7) 29.3(26.4-32.8) 29(25.9-32.5) 29(25.8-32.6) 29.1(26-32.7)

Waist circumference (cm) 102(95-112) 104(95-112) 102(94-112) 102(93-111) 102.5(94-112)

Alcohol use (%,>1/day) 83(5%) 46(4%) 43(4%) 50(4%) 139(4%)

Hypertension (%) 1256(76%) 839(77%) 809(74%) 840(74%) 2488(75%)

Daily Exercise (%) 281(17%) 199(18%) 179(16%) 190(17%) 568(17%)

hsCRP (mg/L) 4.1(2.8-6.6) 4.1(2.8-6.9) 4.2(2.8-6.9) 4.1(2.7-6.7) 4.1(2.8-6.8)

Interleukin-6 (ng/L) 2.5(1.8-3.9) 2.4(1.7-4) 2.4(1.7-3.7) 2.5(1.7-4) 2.5(1.7-3.9)

Total cholesterol (mg/dL) 163.2(140-

190.6)

160.9(138.1-

191)

162.4(139.2-

191.4)

162.8(138.1-

191.9)

162(138.4-

191.4)

LDL cholesterol (mg/dL) 85.1(67.5-110) 84.7(66-109.8) 86.2(67.7-110) 85.1(65.7-

109.4)

85.1(66.1-

109.8)

HDL cholesterol (mg/dL) 45.2(38-53.8) 44.9(38-53.4) 44.9(37.9-53.8) 45.2(37.9-54.5) 45(37.9-53.8)

Triglycerides (mg/dL) 132.9(98.2-186) 132(95-183.3) 132.9(100.1-

186)

133.7(98.3-186) 132.9(97.4-

184.2)

Fasting plasma glucose (mg/dL) 102.7(97-109.9) 104(97.3-111.7) 102.7(95.5-

109.9)

102.7(95.5-110) 102.7(95.5-

109.9)

Hemoglobin A1c (%) 5.7(5.5-6) 5.8(5.5-6) 5.8(5.5-6) 5.8(5.5-6) 5.8(5.5-6)

Impaired fasting glucose (≥100

mg/dL) (%)

1057(65%) 702(65%) 660(61%) 678(61%) 2040(62%)

eGFR (mL/min/1.73m2) 78(65-90) 78(64-91) 78(65-92) 77(65.5-90) 78(65-91)

Metabolic syndrome (%) 1011(62%) 690(64%) 656(60%) 682(61%) 2028(62%)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

Abbreviations: eGFR, estimated glomerular filtration rate; HDL, high density lipoprotein; LDL, low density lipoprotein

*Shown are median within group levels of characteristics for continuous variables, and percentages for dichotomous variables

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

Online Table 5.

Median (IQR) high-sensitivity C-reactive protein (mg/L) throughout the study in patients with baseline pre-diabetes and baseline diabetes.

Treatment group Baseline 12 months 24 months 36 months 48 months P-value*

Baseline pre-diabetes

Placebo 4.1(2.8-6.6) 3.3(2.0-5.6) 3.2(2.0-5.7) 3.5(2.1-5.8) 3.7(2.2-6.2) -

Canakinumab 50 mg 4.1(2.8-6.9) 2.2(1.2-4.0) 2.3(1.2-4.3) 2.4(1.3-4.3) 2.4(1.3-4.6) 0.08

Canakinumab 150 mg 4.2(2.8-6.9) 1.7(0.9-3.3) 1.7(1.0-3.4) 1.9(1.1-3.5) 1.8(1.0-3.5) <0.0001

Canakinumab 300 mg 4.1(2.7-6.7) 1.4(0.8-2.7) 1.6(0.9-3.2) 1.6(0.9-3.3) 1.8(1.0-3.7 <0.0001

Baseline diabetes

Placebo 4.3(2.9-7.6) 3.6(2.2-6.4) 3.6(2.1-6.5) 3.7(2.0-6.20 3.4(1.8-6.1) -

Canakinumab 50 mg 4.8(3.1-7.6) 2.4(1.4-4.8) 2.4(1.3-4.9) 2.6(1.4-5.1) 2.6(1.4-4.9) 0.11

Canakinumab 150 mg 4.5(3.0-7.5) 2.0(1.1-4.1) 2.0(1.1-4.1) 2.1(1.1-4.3) 1.9(1.1-4.00 <0.0001

Canakinumab 300 mg 4.5(3.1-7.8) 1.7(0.9-3.4) 1.8(1.0-3.6) 1.8(1.0-3.7) 2.1(1.0-4.0) <0.0001

* P-value for comparison of each active canakinumab treatment group vs. placebo

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

Online Table 6.

Rates of centrally adjudicated and all physician-reported new onset type 2 diabetes in patients with normal fasting plasma glucose

concentrations at study entry.

Canakinumab dose

Placebo 50 mg 150 mg 300 mg All doses

combined

Adjudicated

new-onset

diabetes

N events/N at risk 4/357 5/218 1/229 5/240 11/687

Incidence rate* 0.31 0.64 0.12 0.55 0.43

Hazard ratio (95% CI) - 2.09(0.56-7.77) 0.37(0.04-3.28) 1.73(0.47-6.46) 1.39(0.44-4.36)

P-value† - 0.27 0.33 0.41 0.56

All physician-

reported

diabetes

N events/N at risk 5/357 5/218 1/229 5/240 11/687

Incidence rate* 0.38 0.64 0.12 0.55 0.43

Hazard ratio (95% CI) - 1.67(0.48-5.75) 0.30(0.03-2.53) 1.41(0.41-4.88) 1.11(0.39-3.20)

P-value† - 0.42 0.21 0.59 0.84

*Incidence rate per 100 person years of observation

†P-value for canakinumab vs. placebo comparison

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

Online Table 7.

Median (IQR) HbA1c concentrations in patients randomly allocated to placebo or canakinumab 50, 150, or 300 mg across the duration of

the trial.

Treatment group Baseline 6 months 12 months 24 months 36 months 48 months P-value

Pre-Diabetes at Baseline

Placebo 5.7 (5.5-6.0) 5.7(5.5-6.0) 5.7(5.5-5.9) 5.8(5.5-6.0) 5.8(5.6-6.1) 5.9(5.7-6.2) -

Canakinumab 50 mg 5.8 (5.5-6.0) 5.7(5.4-5.9) 5.7(5.4-5.9) 5.7(5.5-6.0) 5.8(5.6-6.1) 5.9(5.6-6.2) <0.0001

Canakinumab 150 mg 5.8 (5.5-6.0) 5.6(5.4-5.9) 5.7(5.4-5.9) 5.7(5.5-6.0) 5.8(5.6-6.1) 5.8(5.6-6.1) <0.0001

Canakinumab 300 mg 5.8 (5.5-6.0) 5.6(5.4-5.9) 5.6(5.4-5.9) 5.7(5.5-6.0) 5.8(5.6-6.1) 5.8(5.6-6.1) <0.0001

Diabetes at Baseline

HbA1c (%)

Placebo 7.1(6.4-8.3) 6.9(6.2-8.0) 6.9(6.3-8.1) 7.1(6.4-8.2) 7.2(6.5-8.5) 7.3(6.5-8.4) -

Canakinumab 50 mg 7.0(6.3-8.3) 6.8(6.2-7.9) 6.9(6.3-8.1) 7.1(6.3-8.3) 7.2(6.4-8.4) 7.1(6.5-8.3) 0.95

Canakinumab 150 mg 7.1(6.4-8.2) 6.8(6.2-8.0) 6.9(6.2-8.1) 7.1(6.3-8.2) 7.3(6.4-8.4) 7.1(6.4-8.2) 0.92

Canakinumab 300 mg 7.2(6.5-8.3) 6.9(6.2-8.0) 6.9(6.2-8.0) 7.1(6.3-8.2) 7.1(6.3-8.4) 7.1(6.5-8.3) 0.30

Normoglycemia at Baseline

HbA1c (%)

Placebo 5.3(5.2-5.5) 5.4(5.2-5.6) 5.4(5.2-5.6) 5.5(5.2-5.7) 5.6(5.3-5.8) 5.6(5.3-5.8) -

Canakinumab 50 mg 5.3(5.2-5.5) 5.3(5.1-5.5) 5.3(5.2-5.5) 5.4(5.3-5.6) 5.5(5.3-5.7) 5.5(5.3-5.7) 0.23

Canakinumab 150 mg 5.3(5.1-5.5) 5.3(5.1-5.5) 5.2(5.1-5.4) 5.4(5.2-5.6) 5.5(5.3-5.7) 5.4(5.2-5.7) 0.0002

Canakinumab 300 mg 5.3(5.2-5.5) 5.3(5.1-5.5) 5.3(5.1-5.4) 5.4(5.1-5.5) 5.5(5.2-5.7) 5.5(5.2-5.7) 0.0001

* P-values derived from mixed repeated measures models testing for evidence of a significant effect of drug vs. placebo over the course of

the trial with percent change from baseline in HbA1c as the outcome, and adjusted for baseline measure of HbA1c.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

Online Table 8.

Median (IQR) fasting plasma glucose concentrations in patients randomly allocated to placebo or canakinumab 50, 150, or 300 mg across

the duration of the trial. Treatment group Baseline 6 months 12 months 24 months 36 months 48 months P-value*

Pre-Diabetes at Baseline

Placebo 103(97-110) 103(95-112) 103(95-109) 103(95-110) 103(95-110) 103(95-110) -

Canakinumab 50 mg 104(97-112) 105(97-112) 103(95-112) 104(95-112) 105(97-112) 103(97-112) 0.076

Canakinumab 150 mg 103(95-110) 103(95-110) 102(95-110) 103(95-112) 104(97-112) 103(97-110) 0.0009

Canakinumab 300 mg 103(95-110) 103(95-112) 101(94-110) 103(95-112) 104(97-112) 103(95-112) 0.036

Diabetes at Baseline

Placebo 137(116-179) 135(113-169) 134(112-168) 139(114-173) 141(115-177) 135(114-171) -

Canakinumab 50 mg 138(117-175) 137(115-173) 139(115-174) 139(117-173) 141(117-177) 141(118-171) 0.71

Canakinumab 150 mg 137(114-178) 135(115-171) 135(115-177) 141(114-180) 142(117-184) 139(119-175) 0.009

Canakinumab 300 mg 139(115-182) 135(113-170) 135(114-175) 139(114-176) 139(113-179) 137(134-175) 0.87

Normoglycemia at Baseline

Placebo 92(87-95) 94(88-99) 94(88-97) 95(90-101) 95(90-101) 94(88-99) -

Canakinumab 50 mg 90(86-95) 94(88-99) 92(88-98) 92(88-99) 95(90-99) 94(88-101) 0.38

Canakinumab 150 mg 92(87-95) 95(90-100) 92(88-99) 94(88-99) 95(90-101) 95(92-101) 0.82

Canakinumab 300 mg 91(86-95) 94(88-99) 92(86-97) 92(88-99) 95(88-101) 95(88-101) 0.25

*P-values derived from mixed repeated measures models testing for evidence of a significant effect of drug vs. placebo over the course of

the trial with percent change from baseline in fasting plasma glucose as the outcome, and adjusted for baseline measure of fasting plasma

glucose.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

Online Table 9.

Median (IQR) HbA1c and fasting plasma glucose concentrations among patients with baseline diabetes and a baseline HbA1c of 8.0% or

higher, stratified by randomized treatment group. A total of 1224 patients with baseline diabetes had a baseline HbA1c ≥ 8.0%.

Treatment group Baseline 6 months 12 months 24 months 36 months 48 months P-value*

HbA1c (%)

Placebo 9.2(8.4-10.5) 8.5(7.6-9.9) 8.6(7.5-9.7) 8.5(7.4-9.6) 8.8(7.5-10.0) 8.5(7.5-9.9) -

Canakinumab 50 mg 9.3(8.4-10.2) 8.6(7.4-9.8) 8.5(7.6-9.9) 8.8(7.6-9.9) 8.7(7.5-10.1) 8.4(7.3-9.5) 0.47

Canakinumab 150 mg 9.2(8.4-10.3) 8.5(7.6-9.5) 8.5(7.6-9.6) 8.5(7.5-10.1) 8.7(7.6-10.3) 8.3(7.5-9.5) 0.33

Canakinumab 300 mg 9.0(8.5-10.0) 8.4(7.4-9.4) 8.5(7.6-9.8) 8.4(7.4-10.0) 8.6(7.4-9.8) 8.2(7.2-9.4) 0.26

Fasting plasma glucose

(mg/dL)

Placebo 204(157-259) 176(132-231) 169(130-227) 168(126-205) 170(137-224) 163(123-223) -

Canakinumab 50 mg 198(155-245) 176(135-234) 169(132-224) 165(137-223) 169(133-214) 168(132-222) 0.64

Canakinumab 150 mg 196(159-254) 175(146-234) 180(137-229) 178(141-250) 182(140-236) 162(135-219) 0.76

Canakinumab 300 mg 195(149-245) 170(126-223) 178(128-238) 171(135-227) 165(125-229) 153(119-216) 0.24

* P-values derived from mixed repeated measures models testing for evidence of a significant effect of drug vs. placebo over the course of

the trial with percent change from baseline in HbA1c or fasting plasma glucose as the outcome, and adjusted for baseline measure of either

HbA1c or fasting plasma glucose, as appropriate.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

Online Table 10.

Mean (SD) number of diabetes medication classes taken by patients with diabetes at study entry throughout the trial.

Treatment group Baseline 12 months 24 months 36 months 48 months

Diabetes at Baseline

HbA1c (%)

Placebo 1.2(0.9) 1.3(1.0) 1.4(1.0) 1.4(1.0) 1.5(1.0)

Canakinumab 50 mg 1.2(0.9) 1.3(0.9) 1.4(1.0) 1.4(1.0) 1.4(1.0)

Canakinumab 150 mg 1.2(1.0) 1.4(1.0) 1.4(1.0) 1.4(1.0) 1.5(1.0)

Canakinumab 300 mg 1.2(0.9) 1.3(0.9) 1.4(1.0) 1.4(1.0) 1.5(1.0)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

Online Table 11.

Baseline characteristics of patients enrolled in the CANTOS washout period.

Characteristic Placebo 50 mg 150 mg 300 mg

N 601 412 371 386

Age 61 (54-67) 60 (52-66) 59 (54-65) 61 (55-67)

Female sex 151(25%) 88(21%) 82(22%) 93(24%)

Smoking status

Current 134(22%) 106(26%) 94(22%) 92(24%)

Past 297(49%) 198(48%) 187(50%) 185(48%)

Never 170(28%) 108(26%) 90(24%) 109(28%)

Body mass index (kg/m2) 29.3(26.3-32.7) 28.9(26.4-32.3) 28.7(25.8-32.1) 29.4(26.1-32.8)

Waist circumference (cm) 102(95-112) 104(94-111) 102(95-110) 103(95-112)

Alcohol use (%, >1 per day) 29(4%) 16(3%) 21(5%) 25(5%)

Hypertension (%) 539(73%) 374(76%) 323(71%) 358(75%)

Daily Exercise (%) 119(16%) 95(19%) 82(18%) 79(17%)

hsCRP (mg/L) 3.9(2.7-6.0) 3.8(2.6-6.1) 3.7(2.5-6.2) 3.8(2.7-6.2)

Interleukin-6 (ng/L) 2.4(1.7-3.8) 2.3(1.7-3.3) 2.1(1.5-3.1) 2.4(1.7-3.4)

Total cholesterol (mg/dL) 165(139-190) 158(137-186) 162(139-191) 163(138-189)

LDL cholesterol (mg/dL) 88(68-110) 82(65-106) 85(68-109) 84(67-106)

HDL cholesterol (mg/dL) 45(39-53) 45(39-53) 44(38-53) 46(39-55)

Triglycerides (mg/dL) 132(99-181) 129(95-180) 132(102-177) 131(92-182)

eGFR (mL/min/1.73m2) 78(67-90) 80(67-90) 81(70-94) 77(66-89)

Abbreviations: eGFR, estimated glomerular filtration rate; HDL, high density lipoprotein; LDL, low

density lipoprotein

*Shown are median within group levels of characteristics for continuous variables, and n followed by

percentages in parentheses for dichotomous variables