Increment of plasma glucose by exogenous glucagon is associated with present and future renal function in type 2 diabetes:a retrospective study from glucagon stimulation test

We included a total of 209 (108 male and 101 female) Japanese patients with T2DM who underwent GST during admission in Asahikawa Medical University Hospital from 2013 to 2016, in this cross-sectional study. There was no participant who had symptomatic cerebrovascular disorders, severe heart disease, hepatitis and liver cirrhosis, renal failure (exclusion criteria: eGFR below 30 ml min^− 1 1.73 m^− 2 or serum creatinine above 2.0 mg/dl), and malignant neoplasm, who was in the pre- or post-operative period, or who was positive for diabetes-related autoantibodies. The protocol of the current study was approved by the ethics committee of Asahikawa Medical University (Approval number: 17104).

Prior to GST, all inpatient participants were treated with appropriate diet and insulin-based glucose lowering treatments after administration, resulting in fasting glucose levels and pre-meal glucose levels of 140 mg/dl or less to avoid glucotoxicity. No subjects received rapid-acting insulin after 6 pm on the 1 day before GST. Some participants received long acting insulin such as insulin glargine, insulin detemir or insulin degludec subcutaneously at 8 pm the day before GST to maintain adequate fasting blood glucose levels. Then, GST was carried out at 8 am after an overnight fast by measuring serum CPR at fasting and 6 min after single intravenous injection of 1 mg glucagon (Novo Nordisk, Tokyo, Japan). We defined the increment of serum CPR after glucagon injection as ΔCPR, the difference between plasma glucose at fasting and 6 min after glucagon injection as Δglucose. Blood and urine samples for laboratory measurements were collected at overnight fast. Height and body weight (BW) were measured, and body mass index (BMI) was calculated by dividing BW (kg) by height squared (m²). Plasma glucose levels were measured by automatic analyzer (GA09, A&T, Fujisawa, Japan) using glucose oxidase method. HbA1c levels were measured by the automatic analyzer (HLC-723G8, TOSOH Bioscience, Tokyo, Japan) using high performance liquid chromatography (HPLC) method. Both plasma insulin concentrations as immunoreactive insulin (IRI) and serum CPR were measured using an electrochemiluminescence method (cobas 6000, Roche Diagnostics Japan, Tokyo, Japan). Other biochemical values were measured by the automatic analyzer (LABOSPECT 008, HITACHI High-Technologies, Tokyo, Japan).

We assessed correlations between Δglucose and clinical parameters including BMI, duration of diabetes, HbA1c, serum CPR during GST, urinary CPR, serum lipids, serum FFA, liver function, the indices of residual liver function, renal function, and urinary albumin excretion (UAE). As renal function, we employed serum creatinine, cystatin C and eGFR. We calculated eGFR using the previously established equation for eGFR in Japanese subjects, which was calculated from serum creatinine, gender, and age in individual subjects [10].

The clinical characteristics and parameters of the total 209 participants are shown in Table 1. The mean age was 60.7 ± 1.0 years, mean BMI was 27.5 ± 0.4 kg/m², mean diabetic duration was 13.0 ± 0.8 years, and mean HbA1c level was 9.7 ± 0.1% at the time of administration. For the GST, the mean plasma glucose at fasting and 6 min after glucagon injection were 122.2 ± 1.6 and 140.4 ± 1.6 mg/dl, respectively, and the mean Δglucose was 18.2 ± 0.5 mg/dl. Additionally, there was no significant difference in Δglucose among subjects with treated by glucose lowering agents, especially with or without biguanides or dipeptidyl peptidase-4 (DPP-4) inhibitors, which may influence the value of Δglucose (data not shown).

First, we examined the correlations between Δglucose and clinical parameters as follows: age, BMI, duration of diabetes, HbA1c, serum CPR during GST, daily urinary CPR, serum lipids, FFA, liver function and residual liver function, renal function, and UAE (Table 2). We observed positive correlation between Δglucose and FFA (r = 0.2117, p = 0.0025; Fig. 1a), which is known to stimulate gluconeogenesis. Interestingly, Δglucose showed positive correlation with eGFR at baseline (r = 0.2108, p = 0.0025; Fig. 1b), and inverse correlation with serum creatinine (r = − 0.2166, p = 0.0017) and cystatin C (r = − 0.1515, p = 0.0302). On the other hand, we observed no relationship between Δglucose and liver function such as AST and ALT, and the indices of residual liver function such as serum albumin, ChE, bilirubin, PT, total cholesterol, triglyceride, HDL cholesterol, and LDL cholesterol.

Since we observed that Δglucose correlated with present renal function, we hypothesized that Δglucose is also related to the future renal function and could be a determinant of the future renal function. Then, we further performed multiple regression analysis to identify the variables associated with eGFR after 1 year. We observed positive relationship between Δglucose and eGFR 1 year later (r = 0.2412, p = 0.0082; Fig. 1c) similarly to eGFR at baseline. Intriguingly, multiple regression analysis revealed that BMI (β = 0.1830, p = 0.0337), HbA1c (β = 0.2909, p = 0.0008), total cholesterol (β = − 1.4550, p = 0.0119), HDL-cholesterol (β = 0.5937, p = 0.0053), LDL-cholesterol (β = 1.3343, p = 0.0073), UAE (β = − 0.3141, p = 0.0040), and Δglucose (β = 0.2308, p = 0.0034) were independent determinants of the eGFR after 1 year (Table 3).