The clinical significance of Δglucose as the increment of plasma glucose levels during GST has not yet been elucidated, although it is assumed mostly to reflect a part of HGO by glucagon. To the best of our knowledge, this is the first study to investigate the relationship between Δglucose during GST and various clinical parameters in subjects with T2DM.
First, we demonstrated that Δglucose was positively associated with FFA. FFA is known to increase HGO via stimulation of gluconeogenesis in the liver. The mechanisms how FFA stimulates gluconeogenesis can be attributed to the generation of 1) acetyl-CoA derived from FFA oxidation, which activates pyruvate carboxylase, 2) nicotinamide adenine dinucleotide (NADH), which is used for the formation of glyceraldehyde 3-phosphate from 1,3-bisphosphoglycerate, and 3) adenosine triphosphate (ATP), which is used as an energy source [7]. Chen and colleague showed that increment of plasma FFA levels results in increased gluconeogenesis, and FFA positively correlates to gluconeogenesis in humans [8]. Based on these previous findings, the positive relationship between Δglucose by exogenous glucagon and FFA in this study is assumed to reflect the increment of blood glucose levels by FFA-associated gluconeogenesis in the liver.
Second, we anticipated that Δglucose was associated with both liver function or the indices of residual liver function, since the liver is the most important organ supplying glucose into the circulation in the body and Δglucose is presumably to reflect a part of HGO [1]. However, we observed no relationship between Δglucose and liver function such as AST and ALT, or the indices of residual liver function such as serum albumin, ChE, bilirubin, PT, and cholesterol. Why was Δglucose not associated with both liver function or the indices of residual liver function? Indeed, AST and ALT are established as the clinical indices of liver function, but these parameters are valuable only when serum levels are elevated at the liver injury or inflammation. Thus, we presume that both AST and ALT may not manifest the liver functional capacity such as glucose production.
How about serum albumin, ChE, bilirubin, PT, and cholesterols? These parameters, reflecting the synthetic (albumin, ChE, prothrombin, and cholesterols) or elimination (bilirubin) functions of the liver [9], have been well entrenched in assessment of residual liver function. In particular, albumin, bilirubin, and PT constitute ‘Child-Pugh score,’ the indicator of both residual liver function and prognosis in the subjects with liver cirrhosis [9]. Although we speculate these parameters are associated with Δglucose during GST presumably reflecting HGO by glucagon, we failed to observe any relationship between them. Thus, we consider several possible reasons for our results. First, in the current study, we excluded the subjects with liver dysfunction or liver cirrhosis, who might be with impaired parameters such as albumin, bilirubin and PT. Second, these parameters are influenced by various factors other than the synthetic and elimination functions in the liver. For instance, serum albumin levels are generally altered in the presence of proteinuria, hypermetabolism, and malnutrition. Additionally, serum bilirubin levels are also influenced by the existence of renal insufficiency and hemolysis [9]. We consider these variables are the indices not independently reflecting the residual liver function simply represented by synthetic and elimination functions in the liver, thus the results were away from our speculations.
Interestingly, we observed that Δglucose showed positive relationship with eGFR, and inverse relationship with serum creatinine and cystatin C. The kidney also plays critical roles in glucose metabolism via gluconeogenesis, glucose utilization, and glucose reabsorption. Renal glucose production is only through gluconeogenesis, since the kidney cannot produce glucose via glycogenolysis because the kidney has much smaller amount of glycogen than the liver and renal cells can synthesize glycogen but lack glucose–6-phosphatase [10]. Indeed, liver and kidney are almost equivalently involved in glucose production via gluconeogenesis in the post-absorptive state of normal subjects [4, 10]. Moreover, patients with T2DM exhibit abnormal increase of glucose release into the circulation through gluconeogenesis in both liver and kidney [5]. Thus, renal gluconeogenesis not only plays an important role in maintaining homeostasis, but also is involved in pathogenesis of T2DM. These previous findings and our observation provide us the speculation that Δglucose might reflect the capacity for glucose release via renal gluconeogenesis in T2DM.
However, the liver, but not the kidney, has been considered to be the main organ of gluconeogenesis regulated by glucagon. Previous study using isolated perfused rat kidney showed that glucagon does not stimulate gluconeogenesis in the kidney [11]. Furthermore, glucagon stimulates gluconeogenesis from glutamine and increases glucose output not in the kidney but in the liver in human [12]. On the other hand, a recent report revealed that both kidney and intestine induce gluconeogenesis by glucagon in a mouse model for liver-specific deletion of glucose production, concluding that current dogma concerning the roles of the liver and the extrahepatic gluconeogenic organs in glucose homeostasis needs to be revisited [13]. Therefore, we speculate that these observations might explain why glucagon-derived glucose increment was related to renal function in our current study.
Based on our results, we further hypothesized that Δglucose could be the determinant of the future renal function. Then, we performed multiple regression analysis to clarify the variables associated with eGFR after a year. Multiple regression analysis revealed that Δglucose was an independent determinant for the eGFR after 1 year, in addition to BMI, HbA1c, serum lipids, and UAE. Previous several studies already showed that obesity, glycemic control, serum lipids, and proteinuria were respectively the predictors for the development of chronic kidney disease [14, 15]. Our results are very interesting since it suggests that Δglucose is not only related to the present renal function but also a determinant for the future renal function in addition to other metabolic parameters which has been already examined. The correlation between Δglucose and renal function could be linked to the view that Δglucose is a proxy of the functional activity of the nephrons.
There are several limitations in this study. First, our current study was conducted with retrospective cross-sectional design. Second, our results were based on the relatively small population without common selection criteria at single hospital. Third, we did not investigate mechanistic analysis that can explain our results in this study, especially in the relationship between Δglucose and renal function. Prospective multicenter cohort studies with sufficient number of participants are necessary to clarify the relationship between Δglucose and various clinical parameters especially in renal functions for further understanding.