Most critically ill patients exhibit a myriad of symptoms with irregular glycaemic regulation and renal failure contributing to a large proportion of mortality and morbidity. Hence, the effect of dialysis on glycaemic regulation should be of interest to clinicians as it is a common therapy addressing renal failure. In this investigation, we measure transient changes in model-based insulin sensitivity during the commencement and end of dialysis periods in 51 critically ill patients with acute renal failure (ARF). The clinically validated model-based insulin sensitivity (SI) metric is a lumped parameter, which, in this case, can account for variance in insulin pharmacokinetics and production, as well as the efficiency of insulin mediated glucose uptake. Apparent SI is expected to be higher during renal failure as insulin will not be cleared as fast as the model assumed. Thus, it is hypothesized given model assumptions on steady, population levels of renal insulin clearance, that dialysis will cause a drop in model-based SI, and vice versa. This study is the first to investigate the effect of dialysis on insulin action of critically ill patients.

This investigation found a significant reduction in model-based SI after the commencement of dialysis, but an insignificant change when dialysis was stopped. As dialysis was considered to have little effect on true insulin sensitivity and the effect on insulin production would have been contrary to the observed behavior, it was concluded that commencing dialysis had a significant effect in increasing insulin clearance over the cohort. This effect was not reciprocated immediately following dialysis due to a slower return to complete renal failure following treatment.