Statin-induced diabetes: perhaps, it’s the tip of the icebergThe meta-analysis by Mills et al.1 involving 170 255 patients randomized in 76 trials reported on the efficacy and safety of statin therapy for the prevention of cardiovascular disease (CVD) and found a relative 9% increased risk in the development of incident diabetes (P = 0.001) among subjects randomized to statins compared with placebo in the 17 trials reporting on diabetes development. It is noteworthy that the average age of the subjects in the meta-analysis was 59.6 years, average follow-up was 2.7 years and more than half of the subjects were randomized for the primary prevention of CVD. We feel that the implications of statin-induced diabetes are not trivial, but of major concern, particularly in the primary prevention of CVD when statin therapy might be used for decades in individuals at relatively low risk2; many questions need answering before statin therapy can be safely recommended across broad populations.
Interestingly, a recently published meta-analysis involving 91 140 patients randomized in 13 trials3 specifically looking at the risk of incident diabetes from statin therapy also revealed a significant 9% increased relative risk of the development of diabetes over a mean overall trial period of 4 years. Disturbingly, 2 of the 13 trials demonstrated very high incidence of the development of diabetes among the statin-treated subjects. The Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER),4 a primary prevention trial of 1.9-year duration in subjects with a mean age of 66 years, demonstrated a significant relative increase in diabetes incidence of 26% among subjects randomized to rosuvastatin; the absolute rate of incident diabetes expressed in events per 1000 patient-years was 13 and 16 among the placebo and rosuvastatin subjects, respectively. Low-density lipoprotein (LDL) cholesterol was decreased robustly by 50% in the rosuvastatin subjects and the median LDL cholesterol at the end of follow-up was 55 mg/dl. The PROspective Study of Pravastatin in the Elderly at Risk (PROSPER),5 a combined primary and secondary prevention trial of 3.2-year duration in subjects with a mean age of 76 years, demonstrated a significant relative increase in diabetes incidence of 32% among subjects randomized to pravastatin; the absolute rate of incident diabetes expressed in events per 1000 patient-years was 16 and 21 among the placebo and pravastatin subjects, respectively. LDL cholesterol was decreased by 31% in the pravastatin subjects. Therefore, it appears that the risk of statin-induced diabetes is more prominent with aggressive LDL cholesterol lowering and among the elderly subjects. It is of concern that thought leaders in the cardiovascular arena strongly suggest that statin use should be increased from 16 to 100 million people in the USA and LDL cholesterol should be aggressively lowered.6 This issue takes even more relevance given that the prevalence of diabetes is rapidly increasing in the USA7 and worldwide8; alarmingly, three-quarters of the elderly in the USA have diabetes or pre-diabetes.7
In vivo studies have demonstrated that despite lowering LDL cholesterol levels, some9–11 but not all statins11 significantly increase fasting plasma insulin levels and significantly decrease insulin sensitivity in hypercholesterolemic patients in a dose-dependent manner. Statins can significantly increase fasting plasma insulin levels and glycated hemoglobin levels in the absence of significant changes in fasting glucose.9–11 Additionally, some statins have been shown to significantly decrease plasma adiponectin levels.10,11
In vitro and animal studies12,13 have shown that statins can significantly decrease the expression of the insulin-responsive glucose transporter 4 (GLUT4) in adipocytes. GLUT4 is distributed in the intracellular compartment in the basal state and relocates to the cell membrane in response to insulin signaling. Moreover, statins increase the expression of GLUT112 in adipocytes; GLUT1 is localized in the cell membrane. It is unclear how statins change the expression of GLUT1 and GLUT4; perhaps, it is related to an inhibition of isoprenoid biosynthesis by statins12 or cholesterol lowering, leading to a change in membrane lipid raft structure resulting in decreased insulin signaling.14 Since GLUT4 concentrations are not reduced in skeletal muscle in obese subjects and subjects with diabetes, and skeletal muscle is the primary source of insulin-stimulated glucose disposal, it has been argued that whole-body insulin sensitivity cannot be explained by a decrease in the production of GLUT415; however, it has been shown that the downregulation of GLUT4 and resulting glucose transport in adipose tissue can cause insulin resistance.16 It is notable that GLUT4 concentrations are decreased in skeletal muscle in elderly compared with younger subjects,15 which might explain why the elderly are more sensitive to the diabetes promoting effects of statin therapy. Furthermore, dysregulation of cellular cholesterol may attenuate pancreatic β-cell function, since cholesterol maintains normal function of voltage gated calcium channels and is vital in the mobilization and fusion of insulin granules with the cell membrane.17 In summation, there are many ways by which statin therapy might lead to hyperinsulinemia, insulin resistance, prediabetes and diabetes.
In addition to the classic complications of diabetes such as CVD, renal failure, blindness and neuropathy, epidemiological studies demonstrate that diabetes is related to the increased risk of many cancers.18 These include liver, pancreas, kidney, endometrial, colorectal, bladder and breast cancer and non-Hodgkin’s lymphoma. A large European population study with a median follow-up of 15.8 years has shown that compared with individuals with normal glucose tolerance, men and women with prediabetes or diabetes had a significant increase in cancer mortality, irrespective of the body mass index.19 There is epidemiological evidence that insulin resistance is associated with cancer in Eastern populations.20 Interestingly visceral fat mass, assessed by computed tomography, but not subcutaneous fat mass, correlates positively with cancer21; indeed, visceral fat is a strong determinant of insulin resistance and hyperinsulinemia.
There are many ways by which hyperinsulinemia can promote cancer.18,22 Hyperinsulinemia results in an increase in the biologically active free circulating insulin-like growth factor-1 (IGF-1) by increasing hepatic IGF-1 production22 and decreasing IGF-1 binding proteins.18 Tumor cells are replete with IGF-1 receptors and two isoforms of insulin receptors (IR-A and IR-B).23 IGF-1 primarily signals through the IGF-1 receptor resulting in mitogenic effects and, not surprisingly, higher IGF-1 blood levels have been associated with an increased risk of several cancers.24,25 Insulin signaling through the IR-A and IR-B results in mitogenic and metabolic effects, respectively.22 Hyperinsulinemia can persist for decades in prediabetic states and it is certainly conceivable that this prolonged mitogenic stimulus increases cancer promotion as has been seen in epidemiological studies.
Statin therapy might affect tumor metabolism by insulin independent means. As previously mentioned, some statins decrease adiponectin levels.10,11 This is potentially problematic over the long-term since adiponectin is anti-proliferative and anti-angiogenic and has other oncostatic properties.26 Furthermore, obesity is associated with lower circulating adiponectin levels and this might partially explain the association of obesity and various cancers. Additionally, the fact that statin therapy might increase GLUT1 expression12 is of concern since GLUT1 is already overexpressed and is the main glucose transporter in cancer cells.27 Glucose uptake by cancer cells is extremely avid and up to 30 times that of normal cells and utilized by glycolysis for energy and supplying important metabolites for rapid cellular proliferation.28 Indeed, in human studies, increased expression of GLUT1 in cancer cells has been associated with poor prognosis of many cancers.27,29
The Western diet is permissive to the diabetogenic effects of statin therapy. The prevalence of obesity has been steadily increasing in the USA and more than two-thirds of adults are overweight or obese.30 As mentioned, the prevalence of diabetes has been increasing in the USA and a majority of the elderly subjects in the USA now have pre-diabetes or diabetes.7 Interestingly, total cholesterol and LDL cholesterol have been decreasing in the USA likely due to cholesterol awareness and the increased use of lipid-lowering medication, and more than half of the elderly subjects in the USA have reported using lipid-lowering medications.31 However, blood triglyceride levels have been steadily increasing despite the increasing use of lipid-lowering therapy.31 Intriguingly, it is now believed that abnormalities in fatty acid metabolism are at the root of diabetes, and ectopic lipid accumulation in muscle, liver and pancreatic β-cells leads to the development of insulin resistance by interfering with insulin signaling.32,33 The increase in blood triglyceride levels is driven by a high carbohydrate diet and partially fueled by the increase in dietary sweetener consumption.34 Unfortunately, fructose consumption, largely from sweetened beverages, has escalated drastically in North America over the past three decades35 and excessive fructose intake leads to increased hepatic de novo lipogenesis resulting in hepatic steatosis, visceral fat accumulation and ectopic lipid deposition in skeletal muscle, thereby all leading to insulin resistance.
We are living in times when there seems to be a much stronger emphasis on the use of drugs over lifestyle change to prevent disease. The food industry has been uncooperative and blames personal responsibility as a cause of the obesity problem.36 There is a belief among many patients that they can eat whatever they want as long as they are on statin therapy.37 This has been amplified by a proposal to offer powdered statin in packets to be sprinkled on hamburgers at fast-food restaurants in order to neutralize the detrimental effect of the food choice.38 Plant-based diets have been shown to decrease both CVD and cancer risk and even result in a rapid change in gene expression in neoplastic tissue, and they are not diabetogenic.39–41 Moreover, a Mediterranean diet has been shown to counter the insulin raising effects of simvastatin therapy.42 It is extremely troubling that a goal has been proposed for decreasing the LDL cholesterol levels of all subjects worldwide to below 100 mg/dl and ideally below 60 mg/dl by statin therapy.43
In conclusion, many important questions need answering before expanding the use of statin therapy, particularly for the primary prevention of CVD. In what proportion of subjects do statins increase plasma insulin levels, even if there is no progression to diabetes? Are some statins more likely than others to cause hyperinsulinemia because of physiochemical differences? Will prolonged statin therapy result in chronic hyperinsulinemia and potentially increase prediabetes, diabetes and/or cancer? Will this risk outweigh any perceived benefits, particularly in the elderly or in aggressively treated patients? Will the Western diet and lifestyle encourage the use of more statin therapy and provide a metabolic substrate to further perpetuate hyperinsulinemia and its subsequent complications? Is the increase in diabetes prevalence in the elderly subjects fueled partially by the increasing use of statins in this age group? What statins decrease blood adiponectin levels, and is it continuous, and, if so, what are the long-term clinical implications? How should physicians monitor patients for the adverse metabolic effects of statin therapy? Should subjects have a plasma insulin level measured prior to initiating and during statin therapy? What diet or diets will mitigate the hyperinsulinemic effects of statin therapy? Will statin therapy used by subjects with a history of cancer increase the chance of hyperinsulinemia increasing the promotion of occult micrometastatic disease? Finally, physicians should realize that statin-induced diabetes, as seen in the relatively short-term clinical trials, might be just the tip of the iceberg, and properly designed clinical trials must be done to determine what else lurks beneath the water in order to ensure the safety of patients on long-term treatment with these drugs.
- © The Author 2010. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved. For Permissions, please email: firstname.lastname@example.org