King et al. (1) recently suggested an association between glycemic control and systemic inflammation, i.e., between HbA1c levels and highly sensitive C-reactive protein (hsCRP) levels, based on data from 1,018 participants in the Third National Health and Nutrition Examination Survey. This report prompted us to search for a similar association in our clinical practice.
Since hsCRP levels can be lowered by statins, thiazolidinediones (TZDs), and anti-inflammatory drugs, we first looked at 64 C-peptide-negative type 1 diabetic patients whose only medication was insulin and found no association (r = 0.0748, P = 0.28) between HbA1c levels and hsCRP. With this negative association, we investigated 108 C-peptide-positive type 2 diabetic patients, all of whom were on a statin, an aspirin, and a TZD, to see whether there was an association between hsCRP and HbA1c in this homogenous group on maximal hsCRP-lowering therapy. The association was again negative, with an r value of 0.0424 and a P value of 0.78.
Why then did King et al. find an association of HbA1c with hsCRP and we did not? We believe that King et al.’s association was with insulin resistance and not hyperglycemia. An association of insulin resistance and hsCRP has been well documented, and theoretically at least, the greater the insulin resistance the worse the glycemic control and, conversely, the higher the glucose the greater the insulin resistance (glucotoxicity). In our group of type 2 diabetic patients who were all on a TZD, insulin resistance should be maximally treated so that if hyperglycemia did affect the hsCRP, its effects would not be confounded by the effects of insulin resistance.
That insulin resistance was not a factor in the King et al. study could be concluded from the inclusion of fasting insulin levels in the regression model. When diabetic subjects are treated with insulin, insulin secretagogues, or insulin sensitizers, the effectiveness of a fasting serum insulin level as a marker for insulin resistance is negated and the conclusion that insulin resistance was eliminated as a factor nullified.
To resolve this problem of differing conclusions from an epidemiological cross-sectional study and a retrospective cross-sectional clinical study, a prospective longitudinal study should be performed. An ideal study would be of type 1 diabetic patients at onset who are clinically free of infection, with measurements of hsCRP being performed before insulin therapy and 2 months later when they are well controlled in the honeymoon period. This is of clinical importance because if hsCRP levels are elevated due to hyperglycemia, then hsCRP levels should only be measured when glycemia is controlled to avoid unnecessary prescribing.
