Magnitude of gluconeogenesis and endogenous glucose production: are they predictable in clinical settings?

INTRODUCTION

Regulation of endogenous glucose production (EGP) is essential for glucose homeostasis. It includes gluconeogenesis (GNG) from non-carbohydrate substrates and hepatic glycogenolysis. Both these pathways are dysregulated in acute stress, but the magnitude of this deregulation cannot be assessed in clinical practice. The study aims at identifying clinically available variables predictive of EGP and GNG magnitude by modeling routinely available data.

METHODS

This exploratory study is based on the data from the Supplemental Parenteral Nutrition study 2 (SPN2), which measured EGP and GNG at days 4 and 10 in 23 critically ill patients. The correlation between EGP and GNG and 83 potential clinical indicators were explored, using single-stage and multivariate analysis.

RESULTS

On single-stage analysis, the strongest correlations were noradrenaline dose at day 4 with GNG (R = 0.71; P = 0.0004) and Nutrition risk screening score (NRS) with EGP (R = 0.42; P = 0.05). At day 10, VO2 (R = 0.59, P = 0.04) was correlated with GNG and VCO2 with EGP (R = 0.85, P = 0.00003). Cumulated insulin dose between days 5 and 9 was correlated to EGP at day 10 (R = 0.55, P = 0.03). Our multivariate model could predict EGP at day 4 (VCO2, glucose and energy intake) with an error coefficient (e.c.) between 7.8% and 23.4% (minimal and maximal error), and GNG at day 10 (age, mean and basal blood glucose), with an e.c. of 18.5% and 29.9%. GNG at day 4 and EGP at day 10 could not be predicted with an e.c. < 40%.

CONCLUSION

This preliminary exploratory study shows that GNG and EGP have different predictors on days 4 and 10; EGP is more correlated with the metabolic level, while GNG is dependent on external factors. Nevertheless, a bundle of variables could be identified to empirically assess the magnitude of both values. Our results suggest that a robust model might be built, but requires a prospective study including a larger number of patients.