The Relationship of Fasting Serum Radioimmune Insulin Levels to Incident Coronary Heart Disease in an Insulin-Treated Diabetic Cohort

P. Dandona, A. Chaudhuri and A. Aljada

Division of Endocrinology, Diabetes and Metabolism, State University of New York at Buffalo, Kaleida Health, Buffalo, New York 14209

Address correspondence to: P. Dandona, Millard Fillmore Hospital, 3 Gates Circle, Buffalo, New York 14209. E-mail: pdandona{at}kaleidahealth.org.

To the editor:

The paper by Kronmal et al. (1) raises an important and interesting issue related to the role of insulin in the pathogenesis of cardiovascular events in type 2 diabetes treated with insulin. After numerous statistical manipulations, the authors conclude that immune reactive plasma insulin concentrations, which have significant cross-reactivity with proinsulin, predict cardiovascular events and that insulin is, therefore, atherogenic. The unadjusted interquartile hazard ratio between patients with the highest immune reactive insulin concentrations and those with the lowest concentrations was approximately 3.6, with no significant increase in risk in the quartiles of human-specific insulin. The adjusted (for black race, internal carotid intimal-medial thickness, uric acid) hazard ratios for immune reactive and human-specific insulin were 5.5 and 4.0, respectively, whereas the hazard ratio for insulin antibodies was approximately 3.0. The plasma concentration of insulin antibodies and proinsulin in the quartile with the highest insulin concentration was 16 and 3.5 times that in patients with lowest insulin concentrations; insulin antibodies and proinsulin were measured in only 50% of the total cohort and 80% of the cohort with incident coronary heart disease (CHD).

This was a post hoc analysis of the data by defining the quartiles of insulin on the basis of equal distribution of incident CHD events, after realizing that most of the incident CHD events occurred at higher insulin levels. The fourth quartile was defined as a range of 150–400 µU/ml, with approximately 60% of the subjects in the 351–400 µU/ml range, in whom insulin antibodies could have made a significant contribution to the measured insulin levels. Plasma insulin concentration is directly related to insulin antibodies, and most of the circulating insulin may be bound to antibodies. In any such study, free insulin concentrations would be more indicative of the bioavailable and bioactive insulin.

In addition, it is of interest that in the group with the highest quartile of insulin (data only shown for the total cohort with no description of those with incident CHD alone), glucose and triglyceride concentrations were the highest and high-density lipoprotein (HDL) levels were the lowest. Remarkably, the data on glycosylated hemoglobin (HbA1c) are missing. Low-density lipoprotein concentrations in the two higher quartiles were also greater than those in the lower quartiles. The history of previous CHD was the highest in the fourth quartile (27 vs. 43%), as was the prevalence of significant major electrocardiogram changes (32.3 vs. 51.7%). This group also did 20% less exercise than the lowest quartile and had a higher body mass index..

Beyond the intrinsic deficiencies in the interpretation of these data, there is important evidence that the authors have ignored in their discussion. Insulin has an antioxidant and antiinflammatory effect in vitro and in vivo in human. Low-dose infusions of insulin suppress nuclear factor {kappa}B, induce inhibitor of {kappa}B, and suppress superoxide generation p47phox expression and Egr-1 expression in human mononuclear cell (2). They also suppress plasma concentration of soluble intercellular adhesion molecule-1, monocyte chemotactic protein-1, matrix metalloproteinase-9, tissue factor, plasminogen activator inhibitor-1 (PAI-1), and vascular endothelial growth factor in obese subjects (2). Low-dose infusion of insulin also has a suppressive effect on C-reactive protein, serum amyloid A, PAI-1, and p47phox in patients with acute myocardial infarction (AMI) (3). In addition, insulin reduces the magnitude of the increase in creatine kinase and creatine kinase-MB after AMI and appears to have a cardioprotective effect, as seen in animals (3). The infusion of insulin also causes a profound antiinflammatory effect in patients with critical, severe hyperglycemia as in diabetic ketoacidosis or nonketotic hyperosmolar hyperglycemic crisis (Kitabchi, A., personal communication). More recently, insulin administration in type 2 diabetes has been shown to reduce C-reactive protein and PAI-1 over a period of 2 wk (4). Insulin administration in endotoxin-challenged rats has been shown to reduce proinflammatory mediators significantly (5).

Finally, in two experimental approaches examining the role of insulin in atherogenesis, insulin has been shown to be antiatherogenic. First, the administration of insulin in apolipoprotein E-null mice suppresses atherogenesis (6). Second, interference in insulin signaling by creating insulin receptor substrate-1 and -2 null mice leads to atherogenesis (7).

On balance, therefore, insulin administration is acutely antiinflammatory in the human and endotoxin-challenged animal models and also chronically antiatherogenic in animal models. Furthermore, it is cardioprotective in AMI. It is unlikely that it contributed to the increase in cardiovascular events in the series of patients presented by Kronmal et al. (1). It is more likely that the patients with the highest insulin concentrations were the ones who received the highest doses of insulin, because they had the highest glucose (HbA1c data would have helped to resolve this), highest triglyceride, and lowest HDL concentrations or were the ones with the highest insulin antibodies, which led to falsely elevated values. They also had the greatest prevalence of electrocardiogram abnormalities and exercised less. The absence of data on HbA1c is unfortunate, because blood glucose concentrations were higher in the group with the highest insulin concentrations. Higher glucose concentrations are proinflammatory and may be atherogenic (2). Because triglyceride concentration was also the highest and HDL the lowest in this group, as was insulin antibody concentration, a multivariate analysis of all the data and a description of the characteristics of those with incident CHD alone is strongly suggested. We would be grateful if the authors can kindly provide this so that their data can be critically assessed.

Received July 22, 2004.

References

  1. Kronmal RA, Barzilay JI, Tracy RP, Savage PJ, Orchard TJ, Burke GL 2004 The relationship of fasting serum radioimmune insulin levels to incident coronary heart disease in an insulin-treated diabetic cohort. J Clin Endocrinol Metab 89:2852–2858[Abstract/Free Full Text]
  2. Dandona P, Aljada A, Mohanty P 2002 The anti-inflammatory and potential anti-atherogenic effect of insulin: a new paradigm. Diabetologia 45:924–930[CrossRef][Medline]
  3. Chaudhuri A, Janicke D, Wilson MF, Tripathy D, Garg R, Bandyopadhyay A, Calieri J, Hoffmeyer D, Tufail S, Ghanim H, Aljada A, Dandona P 2004 Anti-inflammatory and pro-fibrinolytic effect of insulin in acute ST-elevation myocardial infarction. Circulation 109:849–854[Abstract/Free Full Text]
  4. Takebayashi K, Aso Y, Inukai T 2004 Initiation of insulin therapy reduces serum concentrations of high-sensitivity C-reactive protein in patients with type 2 diabetes. Metabolism 53:693–699[CrossRef][Medline]
  5. Jeschke MG, Klein D, Bolder U, Einspanier R 2004 Insulin attenuates the systemic inflammatory response in endotoxemic rats. Endocrinology 145:4084–4093[Abstract/Free Full Text]
  6. Shamir R, Shehadeh N, Rosenblat M, Eshach-Adiv O, Coleman R, Kaplan M, Hamoud S, Lischinsky S, Hayek T 2003 Oral insulin supplementation attenuates atherosclerosis progression in apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol 23:104–110[Abstract/Free Full Text]
  7. Kubota T, Kubota N, Moroi M, Terauchi Y, Kobayashi T, Kamata K, Suzuki R, Tobe K, Namiki A, Aizawa S, Nagai R, Kadowaki T, Yamaguchi T 2003 Lack of insulin receptor substrate-2 causes progressive neointima formation in response to vessel injury. Circulation 107:3073–3080[Abstract/Free Full Text]