Advertisement

Statins and Diabetes: Current Perspectives and Implications for Clinicians

Published:January 14, 2017DOI:https://doi.org/10.1016/j.amjmed.2016.12.022
      SEE RELATED EDITORIAL, p.499
      There is general consensus about the large and persuasive body of evidence that statins are effective and safe in reducing cardiovascular morbidity and mortality in secondary and primary prevention, as well as among patients with diabetes. In contrast, there is controversy about real and perceived risks of statins.
      • Collins R.
      • Reith C.
      • Emberson J.
      • et al.
      Interpretation of the evidence for the efficacy and safety of statin therapy.
      • Hennekens C.H.
      • Lieberman E.
      • Rubenstein M.
      • Hebert P.
      • DeMets D.
      • Pfeffer M.
      Lipid modification in the treatment and prevention of cardiovascular diseases: emerging clinical and public health challenges.
      In this Commentary, we address whether there is a valid statistical association between statin use and the risk of developing diabetes,
      • Hennekens C.H.
      • DeMets D.
      Statistical association and causation: contributions of different types of evidence.
      and if so, whether a causal judgment is warranted.
      • Hennekens C.H.
      • DeMets D.
      Statistical association and causation: contributions of different types of evidence.
      Finally, we address the implications of all these findings for clinicians.
      By 2012, of 27 published trials, none of which was designed a priori to test the hypothesis, only 6 (22%) even mentioned newly diagnosed diabetes. Thus, any findings on statins and diabetes from any such individual trials or their meta-analyses should be considered hypothesis generating, not testing.
      • Hennekens C.H.
      • DeMets D.
      The need for large-scale randomized evidence without undue emphasis on small trials, meta-analyses, or subgroup analyses.
      Two of the individual trials reported statistically significant findings and 4 did not. Of the 2, the results were in opposite directions. The first, the West of Scotland Coronary Prevention Study (WOSCOPS), concluded that there was “evidence for a protective treatment effect” of pravastatin on diabetes.
      • Freeman D.J.
      • Norrie J.
      • Sattar N.
      • et al.
      Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland Coronary Prevention Study.
      Specifically, there was a statistically significant 30% decreased risk of newly diagnosed diabetes (relative risk [RR] 0.70; 95% confidence interval [CI], 0.50-0.98; P = .036) among those assigned at random to the statin. The second, the Justification for the Use of Statins in Primary Prevention: an Intervention Trial with Rosuvastatin (JUPITER) trial,
      • Ridker P.M.
      • Danielson E.
      • Fonseca F.A.
      • et al.
      Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein.
      reported an unexpected statistically significant 25% positive association between rosuvastatin and newly diagnosed diabetes (RR 1.25; 95% CI, 1.05-1.49; P = .01). This finding was unexpected because, based on the WOSCOPS findings, a benefit had been hypothesized a priori. In addition, the trial was terminated early, after 1.9 years, due to the emergence of a statistically extreme 44% benefit on the primary combined cardiovascular disease endpoint and a statistically significant 20% benefit on total mortality, so any findings on secondary endpoints, such as diabetes, are less reliable. In contrast, the recent and large Heart Outcomes Prevention Evaluation (HOPE-3) trial continued to its scheduled termination of 5.6 years and reported no statistically significant association between rosuvastatin and diabetes.
      • Yusuf S.
      • Bosch J.
      • Dagenais G.
      • et al.
      Cholesterol lowering in intermediate-risk persons without cardiovascular disease.
      In 2010, 13 individual trials, none designed a priori to test the hypothesis of statins and diabetes, were included in a meta-analysis. Of the 13, 11 reported no statistically significant association, and 4 were in the direction of a protective effect. In this meta-analysis, the authors reported a small but statistically significant 9% association between statins and diabetes (RR 1.09; 95% CI, 1.02-1.17). The range of the CI suggests that any risk is likely to be as low as 2% and no higher than 17%. The authors noted that, even assuming that the findings were real, 255 patients would have to be treated with a statin for 4 years before even one patient would be diagnosed with diabetes. In these same 255 patients, 9 major nonfatal and fatal vascular events would have been prevented. Thus, the authors concluded that, even assuming causality, the benefits of statins far exceed the potential risk of diabetes.
      • Sattar N.
      • Preiss D.
      • Murray H.M.
      • et al.
      Statins and risk of incident diabetes: a collaborative meta-analysis of randomized trials.
      In 2011, 5 individual trials were included in a subsequent meta-analysis of more vs less intensive statin therapy, which had been designed a priori to test clinical cardiovascular events, but not diabetes. Of the 5, 4 did not achieve statistical significance. In the meta-analysis, the authors reported a small but statistically significant 12% increase in risk (RR 1.12; 95% CI, 1.04-1.22).
      • Preiss D.
      • Seshasai S.R.
      • Welsh P.
      • et al.
      Risk of incident diabetes with intensive-dose compared with moderate-dose statin therapy – a meta-analysis.
      The Cholesterol Treatment Trialists' Collaboration published several worldwide comprehensive meta-analyses of trials designed a priori to test the hypothesis of benefits of statins on vascular events, but not diabetes, using individual patient data. The Cholesterol Treatment Trialist investigators concluded that, on the assumption of causality, the net absolute benefit observed with statin therapy in such individuals is more than 50 times larger than any putative effect on diabetes. Specifically, there are 11 fewer major vascular events per 1000 treated over 5 years per 1.0 mmol/L, or 39 mg/dL reduction in low-density lipoprotein cholesterol. Moreover, long-term follow-up of statin trials has shown that the absolute reductions in major vascular events increase while the statin treatment is continued, and that these benefits persist for at least 5 years after the treatment has stopped, with no evidence of any adverse effects emerging with extended follow-up.
      We believe that the totality of evidence about statins and newly diagnosed diabetes should be viewed as hypothesis formulating, not hypothesis testing. First, there should be substantial differences between how trial evidence is interpreted for any prespecified main effects in contrast to any somewhat unexpected side effects. Every adverse event that reaches a level of statistical significance of P ≤.01, especially those generated from small trials not designed a priori to test a hypothesis or their meta-analyses, should not be accepted as real. If that were to occur, then many drugs of lifesaving benefit would be mistakenly labeled as hazardous, and any real side effects might be obscured by a mass of unreal ones. Second, it is also plausible that differential follow-up times between the statin-treated patients who achieve clinical vascular disease endpoints and the comparison groups who achieve these endpoints at higher rates earlier, and thus have shorter durations of follow-up, may lead to higher but biased ascertainment rates of newly diagnosed diabetes in the statin group. Third, exclusion of chance as a plausible alternative explanation is a necessary, but not sufficient, basis upon which to conclude the presence of a valid statistical association, let alone a judgment of causality.
      • Hennekens C.H.
      • DeMets D.
      Statistical association and causation: contributions of different types of evidence.
      • Hennekens C.H.
      • DeMets D.
      The need for large-scale randomized evidence without undue emphasis on small trials, meta-analyses, or subgroup analyses.
      Clinicians should also consider that, even if the current insufficient totality of evidence were judged to be causal, the possible but unproven small risks of statins and newly diagnosed diabetes pale in comparison with the clear and conclusive benefits of statins on vascular disease outcomes, especially in subjects with diabetes or at high risk of being diagnosed with diabetes. The totality of evidence is large and persuasive that clinicians should more widely prescribe statins in the treatment and prevention of cardiovascular disease in women and men in secondary and primary prevention, including those at high, as well as low, risk. This is because the large and persuasive current totality of evidence indicates the need for more widespread and appropriate utilization of statins, as first-line drugs of choice as adjuncts, not alternatives, to therapeutic lifestyle changes. Further, there is evidence for benefits of statins even among subjects who are unwilling or unable to adopt therapeutic lifestyle changes.
      From the perspective of the health of the general public, there is underutilization of statins in the US in secondary and primary prevention.
      • Hennekens C.H.
      • Pfeffer M.
      Guidelines and guidance in lipid modification.
      Many premature deaths will occur needlessly if patients for whom statins should be prescribed do not agree to take the drug or if patients prescribed statins stop taking the drug as a result of misplaced concerns about the risk of diabetes. These public health issues are particularly alarming in women, for whom cardiovascular disease is also, far and away, the leading cause of death, and for whom there is more underutilization of statins than for men.
      • Hennekens C.H.
      • Lieberman E.
      • Rubenstein M.
      • Hebert P.
      • DeMets D.
      • Pfeffer M.
      Lipid modification in the treatment and prevention of cardiovascular diseases: emerging clinical and public health challenges.

      Acknowledgment

      We are indebted to Angela Moscaritolo for her expert technical assistance.

      References

        • Collins R.
        • Reith C.
        • Emberson J.
        • et al.
        Interpretation of the evidence for the efficacy and safety of statin therapy.
        Lancet. 2016; 388: 2532-2561
        • Hennekens C.H.
        • Lieberman E.
        • Rubenstein M.
        • Hebert P.
        • DeMets D.
        • Pfeffer M.
        Lipid modification in the treatment and prevention of cardiovascular diseases: emerging clinical and public health challenges.
        in: Watson R.R. Handbook of Cholesterol. AOCS Press, Urbana, IL2016: 155-181
        • Hennekens C.H.
        • DeMets D.
        Statistical association and causation: contributions of different types of evidence.
        JAMA. 2011; 305: 1134-1135
        • Hennekens C.H.
        • DeMets D.
        The need for large-scale randomized evidence without undue emphasis on small trials, meta-analyses, or subgroup analyses.
        JAMA. 2009; 302: 2361-2362
        • Freeman D.J.
        • Norrie J.
        • Sattar N.
        • et al.
        Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland Coronary Prevention Study.
        Circulation. 2001; 103: 357-362
        • Ridker P.M.
        • Danielson E.
        • Fonseca F.A.
        • et al.
        Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein.
        N Engl J Med. 2008; 359: 2195-2207
        • Yusuf S.
        • Bosch J.
        • Dagenais G.
        • et al.
        Cholesterol lowering in intermediate-risk persons without cardiovascular disease.
        N Engl J Med. 2016; 374: 2021-2031
        • Sattar N.
        • Preiss D.
        • Murray H.M.
        • et al.
        Statins and risk of incident diabetes: a collaborative meta-analysis of randomized trials.
        Lancet. 2010; 375: 735-742
        • Preiss D.
        • Seshasai S.R.
        • Welsh P.
        • et al.
        Risk of incident diabetes with intensive-dose compared with moderate-dose statin therapy – a meta-analysis.
        JAMA. 2011; 305: 2556-2564
        • Hennekens C.H.
        • Pfeffer M.
        Guidelines and guidance in lipid modification.
        Trends Cardiovasc Med. 2015; 25: 348-350

      Linked Article

      • Statins and Diabetes: Wider Utilization Is Needed in Treatment and Prevention
        The American Journal of MedicineVol. 130Issue 5
        • Preview
          In this issue of The American Journal of Medicine, Hennekens et al1 address many cogent methodologic issues concerning whether there is a valid statistical association between statins and the development of diabetes. While they conclude that the current totality of evidence is insufficient to confirm a valid association, if such an association were present, the most plausible magnitude of increased risk would be about 9%-12% of treated patients. So, even if a clinician believes that statins can increase the risk for developing diabetes, these data pale in comparison with the well-described benefits of statins in both secondary and primary prevention of clinical cardiovascular disease endpoints.
        • Full-Text
        • PDF
      • When and Why Statins Fail to Save Lives
        The American Journal of MedicineVol. 130Issue 9
        • Preview
          The Commentary by Hennekens et al1 about statins and diabetes makes suggestions about mortality that are not supported by most placebo-controlled trials. The opening sentence suggests that statins lower mortality in primary prevention and in diabetics, without being more group specific.
        • Full-Text
        • PDF