Effects of pioglitazone on risks for cardiovascular events and diabetes in patients with prediabetes and a history of stroke or transient ischemic attack

Effects of pioglitazone on risks for cardiovascular events and diabetes in patients with prediabetes and a history of stroke or transient ischemic attack

By Aly Becraft, MS and Kevin C Maki, PhD


Insulin resistance is an established risk factor for stroke and other adverse cardiovascular events.1,2 As many as 50% of patients with stroke or transient ischemic attack have insulin resistance without being classified as having diabetes.3 Furthermore, insulin resistance is associated with cardiovascular risk factors such as increased blood pressure, elevated levels of triglycerides and inflammatory markers and reduced high-density lipoprotein concentration.4 Pioglitazone is an insulin-sensitizing medication that works to lower insulin resistance by activating peroxisome proliferator-activated receptors (PPAR)-γ, and slightly activating PPAR-α, which have potential cardioprotective effects by promoting fatty acid uptake and oxidation.5-9 In the Insulin Resistance Intervention After Stroke (IRIS) trial, pioglitazone was shown to reduce the risk of stroke or myocardial infarction (MI) by 24% compared to placebo in patients with insulin resistance and a history of stroke or transient ischemic attack.10 Treatment with pioglitazone also reduced new-onset diabetes by half.10

Spence and colleagues published a post-hoc analysis of the IRIS trial to investigate the effect of pioglitazone in those patients with good adherence (taking ≥80% of the protocol dose over the duration of the study) and with prediabetes defined using the American Diabetes Association (ADA) definition.11 In the IRIS trial, patients were enrolled based on their homeostasis model assessment of insulin resistance (HOMA-IR) score,10 which is not routinely measured in clinical practice, whereas the ADA definition considers patients to have prediabetes if their glycated hemoglobin (HbA1c) level is 5.7-6.4% or fasting plasma glucose level is 100-125 mg/dL.  The primary outcome was recurrent fatal or nonfatal stroke or myocardial infarction. Secondary outcomes included stroke; acute coronary syndrome; the composite of stroke, MI, hospitalization for heart failure; and the progression to diabetes.

In the IRIS trial, patients were randomized to receive either 15 mg/d pioglitazone titrated up to a maximum dose of 45 mg/d, or a matched placebo. In this analysis, 2885 of the 3876 participants enrolled in the IRIS trial were classified as have prediabetes; 1456 were in the pioglitazone group and 1429 in the placebo group. Among these, 1454 were also classified as having good adherence; 644 were in the pioglitazone group and 810 were in the placebo group. Median follow-up time was 4.8 years.

In those patients with ADA-defined prediabetes and good adherence, the relative risk reductions (RRR) with pioglitazone vs. placebo were 40% for stroke + MI, 33% for stroke, 52% for acute coronary syndrome, and 38% for stroke + MI + hospitalization for heart failure. The relative risk for new-onset diabetes was also reduced by 80% for pioglitazone vs. placebo. Adverse events in the pioglitazone group included weight gain of ≥10% of body weight (29.8% vs. 12% in placebo group; p < 0.001), edema (29.2% vs. 21.6% in placebo group; p < 0.001), and serious bone fractures (3.6% vs. 2.8% in placebo group; p = 0.08). These adverse effects were also observed in the full IRIS trial analysis.12


Comment: An initial requirement of enrollment in the IRIS trial was HOMA-IR score ≥3; therefore, the findings from this trial can only be extended to patients with prediabetes that meet this criterion. That said, this post-hoc analysis provides evidence that patients with prediabetes and established stroke or transient ischemic attack have improved clinical outcomes when treated early, particularly when adherence to treatment is high. Edema was a large contributor to weight gain observed with pioglitazone treatment, which may be less with lower dosages than were used in this trial. For instance, a dose of 7.5 mg/d has been associated with low incidence of weight gain and edema.12 The IRIS investigators conclude that the benefit of pioglitazone treatment demonstrated in this and in the original analysis10 appear to outweigh the observed risks. Additional research is warranted to assess the effects of lower dosage pioglitazone therapy for cardiovascular risk reduction in a wider range of patients than were studied in IRIS.



  1. Kernan WN, Inzucchi SE, Viscoli CM, et al. Insulin resistance and risk for stroke. Neurology. 2002;59:809-815.
  2. Burchfiel CM, Curb JD, Rodriguez BL, Abbott RD, Chiu D, Yano K. Glucose intolerance and 22-year stroke incidence. The Honolulu Heart Program. Stroke. 1994;25:951-957
  3. Kernan WN, Inzucchi SE, Viscoli CM, et al. Impaired insulin sensitivity among nondiabetic patients with a recent TIA or ischemic stroke. Neurology. 2003;60:1447-1451.
  4. Semenkovich CF. Insulin resistance and atherosclerosis. J Clin Invest. 2006;116:1813-1822.
  5. Lee M, Saver JL, Liao HW, Lin CH, Ovbiagele B. Pioglitazone for secondary stroke prevention: a systematic review and meta-analysis. Stroke. 2017;48:388-393.
  6. Yki-Järvinen H. Thiazolidinediones. N Engl J Med. 2004;351:1106-1118.
  7. Spencer M, Yang L, Adu A, et al. Pioglitazone treatment reduces adipose tissue inflammation through reduction of mast cell and macrophage number and by improving vascularity. PLoS One. 2014;9:e102190.
  8. Zhang MD, Zhao XC, Zhang YH, et al. Plaque thrombosis is reduced by attenuating plaque inflammation with pioglitazone and is evaluated by fluorodeoxyglucose positron emission tomography. Cardiovasc Ther. 2015;33:118-126.
  9. Berger J, Moller DE. The mechanisms of action of PPARs. Annu Rev Med. 2002;53:409-435.
  10. Kernan WN, Viscoli CM, Furie KL, et al; IRIS Trial Investigators. Pioglitazone after ischemic stroke or transient ischemic attack. N Engl J Med. 2016;374:1321-1331.
  11. Spence JD, Viscoli CM, Inzucchi SE, Dearborn-Tomazos J, Ford GA, Gorman M, Furie KL, Lovejoy AM, Young LH, Kernan WN. Pioglitazone therapy in patients with stroke and prediabetes: a post hoc analysis of the IRIS randomized clinical trial. JAMA Neurol. 2019; Epub ahead of print.
  12. Adachi H, Katsuyama H, Yanai H. The low dose (7.5 mg/day) pioglitazone is beneficial to the improvement in metabolic parameters without weight gain and an increase of risk for heart failure. Int J Cardiol. 2017;227:247-248.
Photo by Louis Reed

ODYSSEY Outcomes Trial: Topline Results and Clinical Implications

ODYSSEY Outcomes Trial: Topline Results and Clinical Implications

By Kevin C Maki, PhD, CLS, FNLA; Kristen N Smith, PhD, RD, LD; Mary R Dicklin, PhD


Cardiovascular disease (CVD) event risk is high in those with recent acute coronary syndromes (ACS), despite treatment with evidence-based preventive therapies. Prior research has shown that CVD event risk is lowered when low-density lipoprotein cholesterol (LDL-C) is lowered through various means, such as:

  • Statin therapy (compared with placebo)2
  • High-intensity statin therapy (compared with moderate-intensity statin therapy)3
  • Ezetimibe added to statin therapy (compared with placebo)4
  • Anacetrapib added to statin therapy (compared with placebo)5
  • Evolocumab added to statin therapy (compared with placebo)6

Alirocumab is a fully human monoclonal antibody against proprotein convertase subtilisin kexin type 9 (PCSK9), a validated target for risk reduction in patients with stable atherosclerotic CVD.7-8. Research outcomes on alirocumab show that it reduces LDL-C (sustained reductions) and other atherogenic lipoproteins7 with documented safety and tolerability.8

The hypothesis of the Evaluation of Cardiovascular Outcomes after an Acute Coronary Syndrome During Treatment with Alirocumab (ODYSSEY Outcomes) trial was that alirocumab, versus placebo, reduces cardiovascular (CV) morbidity and mortality after recent ACS in patients with elevated levels of atherogenic lipoproteins despite intensive or maximum-tolerated statin therapy.9


This study was a randomized, double-blind, placebo-controlled, parallel group study of 18,924 patients randomized at 1315 sites in 57 countries between November 2, 2012 and November 11, 2017.1,9

Key inclusion criteria:

  • Age ≥40 years
  • ACS
    • 1 to 12 months prior to randomization [acute myocardial infarction (MI) or unstable angina]
  • High-intensity statin therapy
    • Atorvastatin 40 to 80 mg daily, or
    • Rosuvastatin 20 to 40 mg daily, or
    • Maximum tolerated dose of one of these agents for ≥2 weeks
    • Patients not taking statins were authorized to participate if tolerability issues were present and documented
  • Inadequate control of lipids
    • LDL-C ≥70 mg/dL (1.8 mmol/L), or
    • Non-high-density lipoprotein cholesterol (non-HDL-C) ≥100 mg/dL (2.6 mmol/L), or
    • Apolipoprotein B ≥80 mg/dL


Key exclusion criteria:

  • Uncontrolled hypertension
  • NYHA class III or IV heart failure; left ventricular ejection fraction <25% if measured
  • History of hemorrhagic stroke
  • Fasting triglycerides >400 mg/dL (4.52 mmol/L)
  • Use of fibrates, other than fenofibrate or fenofibric acid
  • Recurrent ACS within 2 weeks prior to randomization
  • Coronary revascularization performed within 2 weeks prior to or after randomization
  • Liver transaminases >3 x upper limit of normal; hepatitis B or C infection
  • Creatine kinase >3 x upper limit of normal
  • Estimated glomerular filtration rate <30 mL/min/1.73 m2
  • Positive pregnancy test


Primary Efficacy Outcome

Major Secondary Efficacy Endpoints

Other Secondary and Safety Endpoints

Time of first occurrence:

§  Coronary heart disease (CHD) death, or

§  Non-fatal MI, or

§  Fatal or non-fatal ischemic stroke, or

§  Unstable angina requiring hospitalization

Tested in the following hierarchical sequence:

§  CHD event: CHD death, non-fatal MI, unstable angina requiring hospitalization, or ischemia-driven coronary revascularization

§  Major CHD event: CHD death or non-fatal MI

§  CV event: CV death, non-fatal CHD event, or non-fatal ischemic stroke

§  All-cause death, non-fatal MI, non-fatal ischemic stroke

§  CHD death

§  CV death

§  All-cause death

Secondary endpoints:

§  Components of the primary endpoint considered individually:

·       CHD death

·       Non-fatal MI

·       Fatal and non-fatal ischemic stroke

·       Unstable angina requiring hospitalization

§  Ischemia-driven coronary revascularization

§  Congestive heart failure requiring hospitalization


Safety endpoints:

§  Adverse events

§  Laboratory assessments

Patients screened for this study completed a run-in period of 2 to 16 weeks on high-intensity or maximum-tolerated dose of atorvastatin or rosuvastatin. If at least one lipid entry criterion was met, then the subject was randomized to receive either subcutaneous alirocumab (75 or 150 mg) or placebo every 2 weeks. In order to maximize the number of patients in the target LDL-C range (25-50 mg/dL), alirocumab was blindly titrated or subjects were blindly switched to placebo if they were either substantially above or below (<15 mg/dL for LDL-C) the target range.


Of the 18,924 patients randomized for this study:

  • 9462 were assigned to alirocumab and 9462 received placebo
  • Median follow-up was 2.8 years (interquartile range limits 2.3-3.4 years)
  • 8242 (44%) patients with potential follow-up ≥3 years
  • 1955 patients experienced a primary endpoint; 726 patients died

Topline results showed that treatment with alirocumab was associated with significant reductions in LDL-C, and these reductions remained consistent over time.  Mean baseline and on-treatment LDL-C values are shown in the table below.




(n = 9462)


(n = 9462)


  87.0 mg/dL

87.0 mg/dL

4 months

  93.3 mg/dL

39.8 mg/dL

12 months

  96.4 mg/dL

48.0 mg/dL

48 months

101.4 mg/dL

66.4 mg/dL

The “on-treatment” analysis showed that mean LDL-C was lowered by 55.7 mg/dL (-62.7%) in the alirocumab group vs. placebo at 4 months, 54.1 mg/dL (-61.0%) at 12 months and 48.1 mg/dL (-54.7%) at 48 months.

Several endpoints were significantly less frequent in the alirocumab group vs. placebo. Major adverse cardiac events (MACE; includes CHD death, non-fatal MI, ischemic stroke, or unstable angina requiring hospitalization) are shown in the table below, along with other endpoints.




(n = 9462)

n (%)


(n = 9462)

n (%)

Hazard Ratio (95% Confidence Interval)




903 (9.5)

1052 (11.1)

0.85 (0.78, 0.93)


     CHD death

205 (2.2)

222 (2.3)

0.92 (0.76, 1.11)


     Non-fatal MI

626 (6.6)

722 (7.6)

0.86 (0.77, 0.96)


     Ischemic stroke

111 (1.2)

152 (1.6)

0.73 (0.57, 0.93)


     Unstable angina

37 (0.4)

60 (0.6)

0.61 (0.41, 0.92)




     CHD event

1199 (12.7)

1349 (14.3)

0.88 (0.81, 0.95)


     Major CHD


793 (8.4)

899 (9.5)

0.88 (0.80, 0.96)


     CV event

1301 (13.7)

1474 (15.6)

0.87 (0.81, 0.94)


     Death, MI,

     ischemic stroke

973 (10.3)

1126 (11.9)

0.86 (0.79, 0.93)


     CHD death

205 (2.2)

222 (2.3)

0.92 (0.76, 1.11)


     CV death

240 (2.5)

271 (2.9)

0.88 (0.74, 1.05)


     All-cause death

334 (3.5)

392 (4.1)

0.85 (0.73, 0.98)



Several pre-specified subgroup analyses for the primary outcome variable were presented, including, notably, an analysis by baseline LDL-C categories of <80, 80-99, and ≥100 mg/dL.  Although the test for heterogeneity of response across subgroups was not statistically significant (p = 0.09), the hazard ratio (HR) for the comparison of alirocumab to placebo was numerically lower for the subgroup with baseline LDL-C ≥100 mg/dL [HR 0.76, 95% confidence interval (CI) 0.65 to 0.87] than for those with baseline LDL-C <80 mg/dL (HR 0.86, 95% CI 0.74 to 1.01) or 80-99 mg/dL (HR 0.96, 95% CI 0.92 to 1.14).

Comment by Kevin C Maki, PhD, CLS, FNLA:

When compared with placebo, the use of alirocumab 75 or 150 mg every two weeks, aiming for LDL-C levels of 25-50 mg/dL (and allowing levels as low as 15 mg/dL), led to reduced MACE, MI and ischemic stroke, and was associated with reduced all-cause death. Treatment was safe and well tolerated. CV and CHD death were not significantly reduced. Therefore, the results for total mortality should be viewed with caution, since roughly 70% of total mortality was attributable to CV causes. 

Subgroup analyses identified numerically larger benefits for the primary outcome in subjects with baseline levels of LDL-C ≥100 mg/dL (median LDL-C 118 mg/dL). However, the test for heterogeneity of effect across subgroups was not statistically significant (p = 0.09).  The proportional risk reductions for subjects with baseline LDL-C <80, 80-99 and ≥100 mg/dL were 14%, 4% and 24%, respectively. Only the subgroup with LDL-C ≥100 mg/dL showed a statistically significant reduction in the alirocumab group compared with placebo. That was also true for all-cause mortality, which was reduced by 29% in the alirocumab group vs. placebo in subjects with baseline LDL-C ≥100 mg/dL, but was not significantly reduced in the other subgroups. This finding should also be interpreted with caution because the test for heterogeneity was, again, not significant (p = 0.12). 

The US Institute for Clinical and Economic Review (ICER) provided new estimates of the price range that would be acceptable for the drug, based on the results of the ODYSSEY Outcomes trial.10 ICER calculated two updated value-based price benchmarks, net of rebates and discounts, for alirocumab in patients with a recent acute coronary event:  $2300-$3400 per year if used to treat all patients who meet trial eligibility criteria, and $4500-$8000 per year if used to treat higher-risk patients with LDL-C ≥100 mg/dL despite intensive statin therapy. The manufacturers of alirocumab (Sanofi and Regeneron Pharmaceuticals, Inc.) have announced plans for a reduced price for the drug that will be in the range of the $4500-$8000 identified by ICER, which is substantially below the “list price” of approximately $14,000 per year that had been charged initially.

A number of commentaries from experts in the field have suggested that this class of medications is appropriate for use mainly in patients in whom LDL-C is ≥100 mg/dL, based on the results from the analyses presented at the American College of Cardiology meeting (and not yet published in a peer reviewed journal), including the cost-effectiveness evaluation by ICER. For example, the highly respected cardiologist Milton Packer, MD wrote a piece in which he stated:11

“…the benefit in the entire trial was driven entirely by the effect seen in 5,629 patients who started with LDL cholesterol >100 mg/dL. There was no benefit in patients with lower values for baseline LDL cholesterol.”

With all due respect to Dr. Packer and others who hold this opinion, I view this conclusion as premature for several reasons. The test for heterogeneity (treatment by subgroup interaction) across baseline LDL-C categories was not significant at an alpha of 5%, showing a p-value of 0.09.  Moreover, the study was not designed with sufficient statistical power to reliably differentiate effects across baseline LDL-C categories. This lack of statistical power for tests of heterogeneity of treatment effects argues for caution in the clinical application of such findings, even when the test for heterogeneity is pre-specified and/or when it does reach statistical significance.

Sir Richard Peto, the eminent biostatistician and epidemiologist from the University of Oxford has quipped that only one thing is worse than doing subgroup analyses for a clinical trial, and that is believing the results! To demonstrate the potential unreliability, Peto reported on a set of subgroup analyses from the Second International Study of Infarct Survival (ISIS-2).12  In the trial overall, the survival advantage produced by aspirin for patients with suspected myocardial infarction was 23%, which was highly statistically significant (p < 0.000001).13 ISIS-2 patients were divided into 12 subgroups according to their astrological sign, and the treatment effect of aspirin compared with placebo was calculated in each subgroup. The results ranged from no apparent effect of aspirin in two subgroups (Libra and Gemini) to aspirin being associated with a halving of the mortality in another (Capricorn).

Results from subgroup analyses are useful for generating hypotheses to test prospectively, but should not, in most cases, be applied as the sole basis for clinical practice decisions without replication in other trials, and, ideally, prospective testing in one or more trials designed for the purpose of evaluating possible differences across subgroups in clinical response.  Regarding the results for the subgroup with baseline LDL-C <100 mg/dL in ODYSSEY Outcomes, it should be noted that the authors of the paper from the Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER) trial with the other approved PCSK9 inhibitor agent, evolocumab, report the following:6

“The benefits were also consistent across quartiles of baseline LDL cholesterol levels, from patients in the top quartile, who had a median LDL cholesterol level of 126 mg per deciliter (interquartile range, 116 to 143) (3.3 mmol per liter [interquartile range, 3.0 to 3.7]) at baseline, down to those in the lowest quartile, who had a median LDL cholesterol level of 74 mg per deciliter (interquartile range, 69 to 77) (1.9 mmol per liter [interquartile range, 1.8 to 2.0]) at baseline.”

Results from other trials such as the Improved Reduction of Outcomes: Vytorin Efficacy International Trial4 (IMPROVE-IT; statin plus ezetimibe) and the HPS3/TIMI55 Randomized Evaluation of the Effects of Anacetrapib through Lipid-modification5 (REVEAL; statin plus anacetrapib) have shown benefits with atherogenic cholesterol lowering that are generally consistent with the results observed in statin trials based on the Cholesterol Treatment Trialists’ (CTT) analyses,14 i.e., a HR of 0.78 (22% reduction) per mmol/L (38.7 mg/dL) reduction in LDL-C, despite average baseline LDL-C levels below 100 mg/dL (~94 mg/dL in IMPROVE-IT and 61 mg/dL in HPS2/TIMI55-REVEAL).  Notably, in both IMPROVE-IT and HPS3/TIMI55-REVEAL, the placebo and active treatment group Kaplan-Meier curves did not clearly separate for the first 2.0 to 2.5 years. In a prior study with evacetrapib [Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes (ACCELERATE)], no CVD event benefit (or harm) was observed over a median follow-up period of 2.2 years, despite modest lowering of LDL-C.15  The median follow-up period for both of the PCSK9 inhibitor trials (FOURIER, 2.2 years and ODYSSEY Outcomes, 2.8 years) was short compared with those from most trials of statins, which had median follow-up periods that averaged roughly 5 years.14 In fact, the FOURIER investigators reported that:

“… in FOURIER, the magnitude of the risk reduction with regard to the key secondary end point appeared to grow over time, from 16% during the first year to 25% beyond 12 months, which suggests that the translation of reductions in LDL cholesterol levels into cardiovascular clinical benefit requires time.”

Findings from studies of genetic variants that alter atherogenic cholesterol levels suggest that the benefits of maintaining lower levels may not be fully apparent after only a few years of intervention. The prototypical example of this is one of the findings that led to the development of the PCSK9 inhibitor class of lipid-altering agents. Cohen et al.16 reported that a nonsense loss-of-function mutation in the PCSK9 gene was associated with a 38 mg/dL (0.98 mmol/L) lower average level of LDL-C, and a missense mutation was associated with a 21 mg/dL (0.54 mmol/L) reduction in LDL-C. Based on the CTT relationship, the predicted reductions in CVD event risk would have been roughly 22% and 13%, respectively. However, the observed reductions in CVD (CHD and stroke) incidence were approximately 50% and 37%, respectively (estimated from data presented in the paper). The reductions in risk were most evident for CHD, where HRs were 0.11 (89% reduction) and 0.50 (50% reduction) for those with the nonsense and missense mutations, respectively. These results, and those from many other studies of lipid-altering genetic variants, suggest a greater CVD event risk reduction than would be predicted from the effects of statin and other lipid-altering therapies on risk.17 A likely explanation is that genetic variants produce differences that are maintained over decades, rather just a few years duration, as is the case in randomized, controlled intervention trials. 

Thus, there are reasons to believe that “lower is probably better” for atherogenic cholesterol levels with regard to CVD event reduction in high-risk patients, even if the baseline level of LDL-C is less than 100 mg/dL. Atherosclerosis is a disease that develops and progresses over decades. Thus, it seems possible, and indeed, likely, that benefits will be observed with therapy to further reduce atherogenic cholesterol among those with LDL-C less than 100 mg/dL over follow-up periods longer than the 2- to 3-year median durations in the ODYSSEY Outcomes and FOURIER trials. At present, this is a hypothesis that remains to be verified with additional clinical research. Given that the subgroup with baseline LDL-C <100 mg/dL who received placebo in the ODYSSEY Outcomes trial experienced an event rate above 3% per year, substantial residual risk is present in such patients. Dr. Packer ended his commentary by saying “The ODYSSEY trial shows that we may have reached the limits of what we can achieve by lowering lipids.” My view is that the potential for aggressive atherogenic cholesterol reduction to lower CVD event risk in those with recent ACS (and other high-risk patients) has not been fully evaluated. Accordingly, efforts to understand the effects of atherogenic cholesterol lowering in high-risk patients with LDL-C levels <100 mg/dL should remain an important priority.


  1. Schwartz GG, Szarek M, Bhatt DL, et al. The ODYSSEY OUTCOMES trial: topline results. Alirocumab in patients after acute coronary syndrome. Presented at ACC.18 67th Annual Scientific Session & Expo. Acc.18 Joint ACC/JACC Late-breaking clinical trials. Accessed at https://accscientificsession.acc.org/features/2018/03/video-sanofi-regeneron on March 15, 2018.
  2. Schwartz GG, Olsson AG, Ezekowitz MD, et al. Effects of atorvastatin on early recurrent ischemic events in acute coronary syndromes: the MIRACL study: a randomized controlled trial. JAMA. 2001;285(13):1711-1718.
  3. Cannon CP, Braunwald E, McCabe CH, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med. 2004;350(15):1495-1504.
  4. Cannon CP, Blazing MA, Giugliano RP, et al. Ezetimibe added to statin therapy after acute coronary syndromes. N Engl J Med. 2015;372(25):2387-2397.
  5. HPS3/TIMI55-REVEAL Collaborative Group, Bowman L, Hopewell JC, et al. Effects of anacetrapib in patients with atherosclerotic vascular disease. N Engl J Med. 2017;377(13):1217-1227.
  6. Sabatine MS, Giugliano RP, Keech AC, et al. Evolocumab and clinical outcomes in patients with cardiovascular disease. N Engl J Med. 2017;376(18):1713-1722.
  7. Robinson JG, Farnier M, Krempf M, et al. Efficacy and safety of alirocumab in reducing lipids and cardiovascular events. N Engl J Med. 2015;372(16):1489-1499.
  8. Robinson JG, Rosenson RS, Farnier M, et al. Safety of very low low-density lipoprotein cholesterol levels with alirocumab: pooled data from randomized trials. J Am Coll Cardiol. 2017;69(5):471-482.
  9. Schwartz GG, Bessac L, Berdan LG, et al. Effect of alirocumab, a monoclonal antibody to PCSK9, on long-term cardiovascular outcomes following acute coronary syndromes: rationale and design of the ODYSSEY outcomes trial. Am Heart J. 2014;168(5):682-689.
  10. Institute for Clinical and Economic Review, 2018. Alirocumab for treatment of high cholesterol: effectiveness and value. Preliminary New Evidence Update. March 10, 2018. Accessed at https://icer-review.org/wp-content/uploads/2018/03/Alirocumab-Preliminary-New-Evidence-Update_03102018.pdf on March 23, 2018.
  11. Packer, M. Confessions and omens from the ODYSSEY Trial - Milton Packer assesses his predictions in the future of lipid research. MEDPAGE TODAY, March 14, 2018. Accessed at https://www.medpagetoday.com/blogs/revolutionandrevelation/71755 on March 23, 2018.
  12. Peto R. Current misconception 3: that subgroup-specific trial mortality results often provide a good basis for individualising patient care. Br J Cancer. 2011;104:1057-1058.
  13. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet. 1988;2(8607):349-360.
  14. Baigent C, Keech A, Kearney PM, et al. Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet. 2005;366(9493):1267-1278.
  15. Lincoff AM, Nicholls SJ, Riesmeyer JS, et al. Evacetrapib and cardiovascular outcomes in high-risk vascular disease. N Engl J Med. 2017;376(20):1933-1942.
  16. Cohen JC, Boerwinkle E, Mosley TH, Jr., et al. Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med. 2006;354(12):1264-1272.

17.       Ference BA. Mendelian randomization studies: using naturally randomized genetic data to fill evidence gaps. Curr Opin

Redefining a Healthful Diet: New Results from the Largest Observational Study Ever Conducted on Nutrition and Heart Health Challenge Current Advice

Redefining a Healthful Diet: New Results from the Largest Observational Study Ever Conducted on Nutrition and Heart Health Challenge Current Advice

By Orsolya M. Palacios, RD, PhD and Kevin C. Maki, PhD

Cardiovascular disease is a major cause of morbidity and mortality worldwide.  Its relationship to healthful diet and lifestyle practices has been an area of active research for decades since these represent modifiable behaviors that have the potential to affect cardiovascular disease risk and overall health.  Within dietary recommendations, health authorities advise reducing total and saturated dietary fats while increasing carbohydrates from whole grains, as well as intakes of fruits, vegetables, nuts, seeds and legumes (DGA, 2015). These recommendations are based on studies that have been mostly observational in nature and conducted in high-income countries such as the U.S. and those in Western Europe.  The advice to lower saturated fat intake and replace it with unsaturated fat sources stems largely from the linear relationship between saturated fat intake and a low-density lipoprotein cholesterol (LDL-C) level.  LDL-C is a risk factor for cardiovascular disease, and thus, the working concept is that reducing saturated fat intake reduces LDL-C levels which, in turn, reduce cardiovascular disease risk.  In conjunction with lowering saturated fat intake, health authorities recommend reducing intakes of animal products such as meat and dairy products to accommodate higher intake of plant foods.  However, the advice to increase fruit, vegetable and legume intake also stems from observational studies conducted mainly in high-income nations.  Research on the associations of fruit, vegetable and legume intakes with health outcomes in other nations is sparse and inconclusive.

The dietary habits of populations within wealthy countries are generally one of excess, which significantly differs from the dietary habits of populations in low- and middle-income countries, where intake of certain nutrients, including adequate intake of complete proteins, may be sub-optimal.  Furthermore, dietary habits are strongly rooted in cultural practices, which can also vary greatly among countries, regardless of income status.  Since cardiovascular disease is a leading cause of morbidity and mortality in low- and middle-income countries as well, understanding the link between currently recommended dietary patterns, cardiovascular disease events and/or mortality in more globally-represented populations is crucial in providing accurate and meaningful guidelines for healthful food consumption.

The Prospective Urban Rural Epidemiology (PURE) study was conducted to address this topic.  Researchers recruited individuals aged 35 to 70 years of age in 18 low-income, middle-income and high-income countries between January 1, 2003 and March 31, 2013 to participate in the PURE prospective cohort study to assess the association between total mortality and major cardiovascular events and diet choices.  Habitual dietary intake data was analyzed from 135,355 individuals using validated food frequency questionnaires, which documented energy intake from fat (including total, saturated, monounsaturated and polyunsaturated fat), carbohydrate and protein as well as daily intake of fruit, vegetable and legume servings.  Demographic information, socioeconomic status, lifestyle, physical activity, health history and medication use questionnaires were also distributed and assessed.  Trained physicians using standard definitions completed standardized case-report forms to report mortality and major cardiovascular events.  The primary outcomes in this study were total mortality and major cardiovascular events (fatal cardiovascular disease, non-fatal myocardial infarction, stroke and heart failure) and secondary outcomes were all myocardial infractions, stroke, cardiovascular disease mortality, and non-cardiovascular disease mortality.

To assess associations between macronutrient energy contribution and cardiovascular disease events and/or mortality, participants were categorized into quintiles based on the dietary percentage of energy from total fat, individual fats, carbohydrates and protein; hazard ratios (HRs) were calculated using a multivariable Cox frailty model.  To assess the associations between daily fruit, vegetable and legume servings and cardiovascular disease events and/or mortality, Cox frailty models with random effects were also employed and HRs calculated.

Median follow up of participants was 7.4 years during which time 5796 deaths and 4784 major cardiovascular disease events were recorded.  Regarding macronutrient intake, higher carbohydrate intake was associated with a significantly higher total mortality risk for the highest quintile versus the lowest quintile, but there was no significant association between carbohydrate intake and cardiovascular disease, myocardial infarction, stroke or cardiovascular disease mortality.

Results from this study indicate that total fat, as well as saturated, monounsaturated and polyunsaturated fats all were significantly associated with a lower risk of mortality for the highest quintile versus the lowest quintile of total and individual fat intake.  Specifically, the HR for the highest versus the lowest quintile of fat intake was 0.77 (95% confidence interval [CI] 0.67-0.87) for total fat, 0.86 (95% CI 0.76-0.99) for saturated fat, 0.81 (95% CI 0.71-0.92) for monounsaturated fat, and 0.80 (95% CI 0.71-0.89) for polyunsaturated fat.  For cardiovascular disease events, the highest quintile of saturated fat intake was associated with a significantly lower risk of stroke (HR 0.79, 95% CI 0.64-0.98) compared to the lowest quintile of saturated fat intake.  Neither total fat nor any of the individual fats were associated with myocardial infarction risk or cardiovascular disease mortality.

Like total fat, the highest quintile versus the lowest quintile of total protein intake was significantly and inversely associated with total mortality risk (HR 0.88, 95% CI0.77-1.00) and non-cardiovascular disease mortality (HR 0.85, 95% CI 0.73-0.99).  Animal protein intake was associated with a significantly lower risk of total mortality whereas plant protein intake had no significant association with total mortality.

Total Fat HR (5th Quintile vs. 1st Quintile) 95% CI P-trend
Total Mortality 0.77 0·67–0·87 <0.0001
CVD Mortality 0.92 0·72–1·16         0.50
Non-CVD Mortality 0.70 0·60–0·82       <0.0001
Major CVD Events 0.95 0·83–1·08         0.33
Saturated Fat HR (1st Quintile vs. 5th Quintile) 95% CI P-trend
Total Mortality 0.86 0·76–0·99 0.0088
CVD Mortality 0.83 0·65–1·07         0.20
Non-CVD Mortality 0.86 0·73–1·01 0.0108
Major CVD Events 0.95 0·83–1·10         0.49
Fruits, Vegetables & Legumes HR (< 1 serving/day vs. 3-4 servings/day) 95% CI P-trend
Total Mortality 0.78 0.69–0.88 0.0001
CVD Mortality 0.81 0.65–1.02 0.0568
Non-CVD Mortality 0.77 0.66–0.89 0.0038
Major CVD Events 1.06 0.92–1.22 0.1301

Abbreviations: CVD, cardiovascular disease; HR, hazard ratio

Adapted from: Ramsden et al. Lancet (2017) S0140 (17)32241-9; Toledo et al.  Lancet (2017) S0140-6736(17)32251-1.

The mean fruit, vegetable and legume intake was 3.91 (standard deviation 2.77) daily servings.  When the researchers assessed the links between fruit, vegetable and legume intakes and outcomes, they found that higher fruit, vegetable and legume intake was significantly inversely associated with major cardiovascular disease, myocardial infarction, cardiovascular mortality, non-cardiovascular mortality and total mortality after adjustments for age, sex and random effects.  However, these effects were diminished after multivariable adjustments.  The HR for total mortality was lowest for those consuming three to four daily servings of fruit, vegetables and legumes (HR 0.78, 95% CI 0.69-0.88) compared to the reference group, who consumed less than one serving of these foods per day.  Higher intakes of fruits, vegetables and legumes were not associated with further lowering of risk.  When assessed independently, fruit intake was associated with lower mortality, including total mortality, cardiovascular mortality and non-cardiovascular mortality.  Raw vegetables were strongly linked to lower mortality risk whereas cooked vegetables had a modest association with lower risk.  Legume intake was inversely associated with non-cardiovascular death and total mortality.


To date, the PURE study is the largest observational study to assess the link between nutrient intakes, food group intakes, cardiovascular disease events (including death) and overall mortality.  It encompassed data from over 135,000 participants in 18 countries across five continents from low-, mid- and high-income nations.  The results of this study align with some general recommendations (e.g,. emphasize consumption of fruits, vegetables and legumes) but are in conflict with some others.  For example, health authorities recommend increasing intakes of fruits, vegetables and legumes at the expense of animal foods (DGA, 2015).  However, the results of this study suggest that the association between increased fruit, vegetable and legume intake plateaus after three to four daily servings, and the median fruit, vegetable and legume intake among participants was already 3.9 daily servings.  Thus, as a whole, participants were theoretically obtaining the maximal benefit from intake of these foods and the incremental benefit beyond the median level of intake in the populations studied is uncertain.  The study’s finding that higher energy intake from animal protein is linked to reduced total mortality, while plant proteins, such as those found in legumes, showed no significant association, does not align with the some aspects of the current Dietary Guidelines for Americans (DGA, 2015).  Although it recommends an increase in seafood, The Dietary Guidelines for Americans also recommends strategies such as using legumes, nuts and seeds in place of meat and poultry in mixed dishes to attain protein needs and to increase vegetable intake while cutting back on foods such as some meats, poultry and cheeses to help lower saturated fat intake (DGA, 2015).

However, the PURE study results challenge the emphasis on reducing intake of saturated fat.  Higher energy intake from fat and each individual type of fat, including saturated fat, was associated with lower total mortality, as well as lower risk for some cardiovascular disease events.  Carbohydrate energy intake either showed no association on assessed outcomes or was associated with an increased risk for mortality.  However, it should be emphasized that intakes of saturated fats were generally low, with mean values ranging from 5.7% in China to 10.9% in Europe and North America.  Across countries, total and saturated fat intakes are positively associated with socioeconomic status.  Thus, in countries with higher intakes of total and saturated fat, and thus lower intakes of carbohydrate, higher socioeconomic status, with resulting access to higher quality healthcare, is a potential confounder.

Taken together, the results from PURE raise questions about current dietary guidance, which is largely based on results from observational studies completed in the U.S. and Europe.  Unfortunately, very few randomized, controlled trials have been completed to assess the influence of dietary guidance on long-term health and disease incidence.  While difficult and expensive, these are essential for fully evaluating the potential benefits and risks of dietary recommendations (Maki, 2014).  The strongest recommendations should be limited to those instances where results from randomized, controlled trials align with findings from observational studies.  While the results from PURE are at odds with some current dietary recommendations, they are consistent with the age-old adage “everything in moderation.”

 PURE Study References

Dehghan M, Mente A, Zhang X, et al. Associations of fats and carbohydrate intake with cardiovascular disease and mortality in 18 countries from five continents (PURE): a prospective cohort study.  Lancet. 2017; S0140-6736(17)32252-3.

Miller V, Mente A, Dehghan M, et al. Fruit, vegetable, and legume intake, and cardiovascular disease and deaths in 18 countries (PURE): a prospective cohort study.  Lancet. 2017; S0140-6736(17)32253-5.

Ramsden CE, Domenichiello AF. PURE study challenges the definition of a healthy diet: but key questions remain.  Lancet. 2017; S0140-6736(17)32241-9.

Toledo E, Martinez-Gonzalez MA. Fruits, vegetables, and legumes: sound prevention tools.  Lancet. 2017; S0140-6736(17)32251-1.

Additional References

Maki KC, Slavin JL, Rains TM, Kris-Etherton PM.  Limitations of observational evidence: implications for evidence-based dietary recommendations.  Adv Nutr. 2014;5(1)7-14.

U.S. Department of Health and Human Services and U.S. Department of Agriculture. 2015 – 2020 Dietary Guidelines for Americans. 8th Edition. December 2015. Available at https://health.gov/dietaryguidelines/2015/guidelines/.

Photo by Eaters Collective