Insulin resistance is a key factor in the pathogenesis of the metabolic syndrome and type 2 diabetes, but the underlying mechanisms remain poorly understood. Adipocytokines, including tumour necrosis factor α (TNFα), interleukin 6, leptin and adiponectin,1,2 are increasingly recognised as important regulators of both insulin sensitivity, as well as inflammation, and a dysregulation of their levels and/or functions has been shown in both obesity and rheumatoid arthritis.3 Further investigations have substantiated the important negative effects of TNFα on insulin‐mediated glucose uptake and the development of insulin resistance.4 In this study, the influence of therapeutic TNFα blockade on insulin sensitivity was investigated in regularly treated patients with rheumatic diseases.
A group of eight patients who were non‐diabetic and having various chronic inflammatory disorders (table 11)) were investigated in an open prospective study design. Patients were attending the outpatient clinic of the Sint Maartenskliniek in Nijmegen and entered the study after giving their written informed consent. The regional medical ethical committee approved the study. Patients started the treatment with infliximab (3 mg/kg) and were followed up over a period of 6 weeks. During this interval, the additional antirheumatic drugs remained unchanged. Table 11 gives the patient characteristics. Hyperinsulinaemic–euglycaemic clamps were performed as described in detail elsewhere5,6 before each anti‐TNF infusion, and the M value was used to quantify insulin sensitivity.5,6 At baseline, two patients with rheumatoid arthritis had both a higher fasting glucose concentration and a slight decrease in insulin sensitivity, and an additional four had low‐insulin sensitivity. This suggests that our group of patients were characterised by insulin resistance, which is in line with previous reports.3 Although, TNFα neutralisation was previously reported to fail to improve insulin sensitivity in obese non‐insulin dependent patients with diabetes,7 in this study, it improved insulin sensitivity in seven patients (5.09 (6.24) at baseline vs 6.56 (8.97) after 6 weeks, p=0.05), restored insulin sensitivity within the normal ranges in three patients, whereas TNF neutralisation slightly worsened insulin sensitivity in one patient (fig 1A1A)) who had ankylosing spondylitis and manifestations of peripheral arthritis (table 11).). Our findings are in line with previous studies, which reported that anti‐TNF agents corrected disturbances in glucose metabolism in patients with rheumatoid arthritis.8,9 To our knowledge, our study is the first to apply hyperinsulinaemic–euglycaemic clamps, the gold standard method for assessing insulin sensitivity, in documenting this effect.
During our study, except for one patient who later turned out to have Whipple's disease, TNFα neutralisation was accompanied by a considerable decrease in DAS28 and ESR (table 11).). We therefore hypothesised that the suppression of the inflammatory cascade might be responsible for the improvement of observed insulin sensitivity.4 However, no significant correlation between disease activity or other inflammatory markers and insulin sensitivity at both time points could be detected. Interestingly, the improvement in insulin sensitivity correlated negatively with the baseline BMI (r=−0.83, p=0.01) in our study (fig 1B1B).). This finding was unexpected and an explanation for that is currently under investigation in our laboratory.
In conclusion, using the hyperinsulinaemic–euglycaemic clamp technique, we found that TNFα blockade in patients with chronic inflammatory conditions improves insulin sensitivity, which is likely to be mainly due to the reduction of the inflammatory status. This change may decrease long‐term cardiovascular risk in rheumatoid arthritis.