快樂小藥師 Im pharmacist nichts glücklich

Introduction

In a recent observational study by Culver and colleagues that was part of the Women's Health Initiative (WHI), statin use was reported to be associated with increased risk for hyperglycemia/"diabetes" among 153,840 postmenopausal women aged 50-79.^[1] As expected, this report generated a frenzy of sensational misinterpretation of the data by news media and considerable concern and doubt about the safety of statins among healthcare providers and the public alike. A closer look at this report reveals numerous flaws in the design and data acquisition that render the conclusions highly questionable.

At the start of the study, 7% of the 153,840 women included in the observational study were on various doses of different statins. During follow-up, 10,242 (about 7%) of all women in the study developed hyperglycemia/"diabetes" (1076 had been on statins, 9166 had not). In an unadjusted risk model, statin use at baseline was associated with a 71% increased risk (95% confidence interval [CI]: 1.61-1.83), but after adjusting for some of the more obvious potential confounding variables the risk declined to 48%. The association was observed for all statins, with all doses or duration of statin therapy, and across a range of body mass indices (BMIs <25.0, 25.0-29.9, and >30.0 kg/m²). Interestingly, women with the lowest BMI (<25.0 kg/m²) appeared to be at higher risk for diabetes compared with obese women, a finding that the investigators speculated was related to phenotype or hormonal differences between the women. Quite surprisingly, among African American women, who are more prone to develop diabetes, there was no significant increase in the risk for diabetes associated with statin use at baseline.

Flaws of This Study

What were some of the more glaring flaws of this study?

1. This was a collaborative study that included 40 centers, each with numerous contributors. The justifications and indications for initiating statin therapy in only 7% of the 153,840 postmenopausal women observed, particularly during the early years of the study, were highly variable and far more stringent than they are now. It is quite likely that many of the women who were selected for statin therapy had various coexisting diabetogenic risk factors including family history, metabolic syndrome, sedentary lifestyle, atherogenic and diabetogenic dietary practices, abdominal obesity, or prediabetes, many or most of which continued during statin therapy.

2. Women not selected to receive statin therapy most likely did not have significant major cardiovascular risk factors, including metabolic syndrome, to justify statin therapy. Therefore they were healthier, hence less likely to develop diabetes.

3. The incidence of diabetes varies from state to state; those with the highest age-adjusted incidence are predominately located in the South, specifically Alabama, Florida, Georgia, Kentucky, Louisiana, South Carolina, Tennessee, Texas, and West Virginia. Thus the number of cases included from geographic locations that have a high incidence of diabetes also impacts the overall number of cases with "diabetes."

4. The diagnosis of "diabetes" is also problematic. Patients on niacin, hydrochlorothiazide, or some beta blockers may develop hyperglycemia, which, in the vast majority of cases, may spontaneously resolve even while on the drug or after the drug's discontinuation. Should we erroneously label these hyperglycemic responses as diabetes?

5. Women not selected to receive statin therapy did not need repeated blood tests and therefore would be far less likely to have an A1c or 2-hour glucose tolerance test. Women selected to receive statins invariably had repeated blood tests each year, creating more opportunities to detect hyperglycemia and therefore undergo additional tests, including an A1c or 2-hour glucose tolerance test to detect diabetes, thus contributing to the frequent testing/screening bias. As examples, screening colonoscopy is associated with a much greater likelihood of detecting dysplastic/precancerous polyps and colon cancers. Regular mammography is also associated with much higher detection rates of breast cancers compared with detection in women who do not have regular mammograms. Testing/screening for almost any disease is more likely to detect the disease in question compared with no testing/screening.

6. We know very little about insulin resistance or insulin insufficiency in cases selected for statin therapy or about the vigor or lack thereof among the various study contributors for diagnosing impaired fasting blood glucose, prediabetes, or diabetes before starting their patients on statins. Conversely, it is quite possible that prediabetes or "slightly elevated blood sugar" with or without other, coexisting risk factors may have biased or reinforced the justification for statin therapy. The fact that BMI and the dose, type, and duration of statin therapy were not contributing factors, but older age was, is consistent with such a bias. I frequently see patients for management of their dyslipidemia who have slightly elevated fasting blood glucose and, on further investigation with 2-hour glucose tolerance test, prediabetes or diabetes. Without pretesting, they, too, would be placed in the "statin-induced diabetes" category.

7. Long-term statin adherence outside of prospective, double-blind studies is highly variable and often poor, and patients remain off statins for months or longer, often without informing their healthcare providers. It is impossible to adjust for these noncompliant cases in an observational study. What happened to the glycemic status of those who stopped their statins (with or without the advice or knowledge of their healthcare providers) or took them intermittently? Did they remain "dysglycemic" or revert to normoglycemia?

These major confounders and individual contributors' practice patterns or judgment as to who would be prescribed statins simply cannot be and were not measured or adjusted for. Thus conclusions based on insufficient and lack of crucial data are neither informative nor reliable and are subject to misinterpretation or spinning. No matter how sincere the authors are, no amount of sincerity can be a substitute for flawed and inadequate data. The harm of such misinformation may be immeasurable if it misleads healthcare providers about the so-called "statin-induced diabetes," hence contributes to their reluctance to prescribe these agents and their patients' resistance to take them.

Other Supportive Studies

Even if we accept the notion that statins may increase the risk for dysglycemia or diabetes, there is no data to suggest that they increase the risk for cardiovascular disease and micro- or macrovascular diseases or that negate the benefits of statin therapy.

Yeboah and colleaguesused data from the Multi-Ethnic Study of Atherosclerosis (MESA) baseline and follow-up examinations to assess the associations of impaired fasting glucose (IFG) with type 2 diabetes and cardiovascular disease (CVD).^[2] During 7.5 years of follow-up, IFG was not associated with incident CV events in multivariate model compared with normal fasting glucose. However, IFG was strongly associated with progression to and development of type 2 diabetes. Thus the association between IFG and CVD risk is not due to IFG but its progression to diabetes and the coexistence of other major CVD risk factors in persons with IFG. Thus the development of diabetes in patients with IFG who are treated with statins may be misunderstood and mislabeled as "statin-induced diabetes." These findings also suggest that the threshold of fasting blood glucose level that is independently associated with CVD risk may be as high as the threshold for the diagnosis of diabetes.

What about previous reports of statin-associated hyperglycemia, dysglycemia, or "diabetes"? Waters and colleagues examined the incidence and clinical predictors of new-onset type 2 diabetes within 3 large randomized trials that used atorvastatin.^[3] They used a standard definition of diabetes (2 elevated fasting blood glucose levels) and excluded patients with prevalent diabetes, but not those with IFG, at baseline.

1. In the TNT (Treating to New Targets) trial,^[4] 351 of 3798 patients randomized to 80 mg of atorvastatin and 308 of 3797 randomized to 10 mg developed new-onset type 2 diabetes mellitus (T2DM) (9.24% vs 8.11%, adjusted hazard ratio [HR]: 1.10, 95% confidence interval [CI]: 0.94-1.29, P = .226).
2. In the IDEAL (Incremental Decrease in End Points Through Aggressive Lipid Lowering) trial,^[5] 239 of 3737 patients randomized to atorvastatin 80 mg/day and 208 of 3724 patients randomized to simvastatin 20 mg/day developed new-onset T2DM (6.40% vs 5.59%, adjusted HR: 1.19, 95% CI: 0.98-1.43, P = .072).
3. In the SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels) trial,^[6] new-onset T2DM developed in 166 of 1905 patients randomized to atorvastatin 80 mg/day and in 115 of 1898 patients in the placebo group (8.71% vs 6.06%, adjusted HR: 1.37, 95% CI: 1.08-1.75, P = .011). In each of the 3 trials, baseline fasting blood glucose, body mass index, hypertension, and fasting triglycerides were independent predictors of new-onset T2DM.

Across the 3 trials, there was no difference in the major cardiovascular events, which were 11.3% in patients with and 10.8% in patients without new-onset T2DM (adjusted HR: 1.02, 95% CI: 0.77-1.35, P = .69). The authors did find that high-dose atorvastatin treatment compared with placebo in the SPARCL trial was associated with a slightly increased risk for new-onset T2DM.

In a meta-analysis of 13 clinical trials with 91,140 participants, Sattar and colleagues examined the possible association of statins with development of diabetes.^[7] Among the 91,140 participants without diabetes, 4278 developed incident diabetes over a mean study follow-up of about 4 years. The rate of diabetes in individual trials varied substantially. In the combined data, 174 more cases of incident diabetes occurred in the groups assigned to statin treatment than in the placebo or standard-care groups, representing a 9% increase in the likelihood of development of diabetes during follow-up. The investigators estimated that this amounted to 1 additional case of diabetes per 255 patients treated with statins over 4 years. The results were nearly identical when the analyses were restricted to placebo-controlled trials. Although the association between statin therapy and risk for incident diabetes was stronger in trials with older participants, baseline BMI and percent change in LDL cholesterol did not seem to be important factors.

Final Comments

As yet, there is no clear or plausible reason(s) or mechanism(s) by which statins might raise blood glucose level: Is it upstream at pancreatic beta cell level or downstream due to insulin resistance at the liver or the muscle cells? Statins have an impact on muscle cell mitochondria and result in reduced coenzyme Q10 in myocytes. Leptin enters the hypothalamic appetite/satiety center thorough leptin-receptors, where it signals mitochondria to produce a host of hormones including melanocortins. Among other functions, melanocortins stimulate muscle cells to burn more glucose at a faster rate. Theoretically, statins may decrease the production or release of melanocortins from the hypothalamic appetite/satiety center, or decrease peripheral myocytes' mitochondrial response to melanocortins. If so, then such changes may result in delayed or slow glucose metabolism downstream from the pancreas, and contribute to dysglycemia, but not to insulin resistance or diabetes. This is only speculation on my part and I know of no study to support or refute this speculation.

Meta-analyses, like observational studies, cannot establish causation, because they cannot change or adjust for data deficiencies, unmeasured confounders, and flaws in the studies they put together. Furthermore, none of the trials included in the meta-analyses nor the current observational study was designed a priori to test the hypothesis of statin-associated diabetes; hence individually or in aggregate they do not establish cause and effect but only association. The key question is whether raising the fasting blood glucose by 15 to 20 mg/dL to reach the threshold for up-classifying patients to diabetes has any clinical significance.

Every study included in the meta-analyses referred to above showed significant adjudicated benefit of statin therapy even among the subgroup with dysglycemia/diabetes, suggesting that statin-induced dysglycemia, even if real, may well be a distraction and "much ado about nothing." Only large-scale, randomized, long-term prospective studies with adequate control for all the variables outlined above may answer whether statin-associated dysglycemia has significant micro- and macrovascular negative consequences.^[8] Such studies will be very difficult to organize and carry out over many years and therefore highly unlikely to be done anytime soon. For now, healthcare providers should feel comfortable to assure their patients that the well-established benefits of statin therapy far outweigh the slight risk for dysglycemia. Just as important is the clear need for placing much greater emphasis on dietary and lifestyle modifications than reliance on statin therapy alone.

References

1. Culver AL, Ockene IS, Balasubramanian R, et al. Statin use and risk of diabetes mellitus in postmenopausal women in the Women's Health Initiative. Arch Intern Med.2012;172:144-152.
2. Yeboah J, Bertoni AG, Herrington DM, Post WS, Burke GL. Impaired fasting glucose and the risk of incident diabetes mellitus and cardiovascular events in an adult population: MESA (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol. 2011;58:140-146.
3. Waters DD, Ho JE, DeMicco DA, et al. Predictors of new-onset diabetes in patients treated with atorvastatin: results from 3 large randomized clinical trials. J Am Coll Cardiol. 2011;57:1535-1545.
4. LaRosa JC, Grundy SM, Waters DD, et al; Treating to New Targets (TNT) Investigators. Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N Engl J Med. 2005;352:1425-1435.
5. Pedersen TR, Faergeman O, Kastelein JJ, et al. Incremental Decrease in End Points Through Aggressive Lipid Lowering (IDEAL) Study Group. High-dose atorvastatin vs usual-dose simvastatin for secondary prevention after myocardial infarction: the IDEAL study: a randomized controlled trial. JAMA. 2005;294:2437-2445.
6. Amarenco P, Bogousslavsky J, Callahan A 3rd, et al; Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Investigators. High-dose atorvastatin after stroke or transient ischemic attack. N Engl J Med. 2006;355:549-559.
7. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident diabetes: a collaborative meta-analysis of randomized statin trials. Lancet. 2010;375:735-742.
8. Hennekens CH, 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.