Background:
In persons with type 2 diabetes, glycated haemoglobin A1c (HbA1c) is used as an indirect measure of the mean glucose over the past 3-4 months. The normal range of HbA1c and the correlation to the mean glucose has been determined from studies in subjects without severe chronic kidney disease.
In persons with end-stage renal disease (ESRD) and type 2 diabetes, HbA1c has been shown in small studies to underestimate mean glucose when compared with continuous glucose monitoring (CGM) or alternative glycaemic markers such as glycated albumin or fructosamine. The cause of a falsely low HbA1c in persons with ESRD and diabetes is not clear, but adjuvant erythropoietin treatment and iron infusion increase erythropoiesis, thus increasing the number of new non-glycated erythrocytes that are thought to lower HbA1c. Furthermore, the erythrocyte life span was found to be decreased in individuals receiving hemodialysis and peritoneal dialysis, which may also contribute to a reduction in HbA1c.
CGM is an alternative to HbA1c in glycaemic monitoring and measures interstitial glucose that calibrates with blood glucose within minutes. CGM measures interstitial glucose every five minutes for up to eight days (Medtronic's Ipro2). Studies of persons with type 1 diabetes and CGM measurements performed over three months showed that approximately two-weeks of CGM measurements correlated to HbA1c. When CGM was performed beyond 14 days, the correlation between the mean glucose from CGM and HbA1c improved only modestly.
Values of CGM measurements also provide a basis for monitoring glycaemic variability, time-in-range and hypoglycaemic episodes. For persons with diabetes undergoing peritoneal dialysis, there are only a few studies validating HbA1c and none have been prospective with a control group over a long period of time. A retrospective study of 60 persons with diabetes and in peritoneal dialysis (without control group) found a significant correlation between HbA1c and mean interstitial glucose as measured by CGM (r = 0.48, P≤0.0001). A prospective study of 25 persons with diabetes and in peritoneal dialysis (without control group) also found a significant correlation (r = 0.51, P≤0.01). A study of three persons with diabetes and in peritoneal dialysis (without control group) found that HbA1c compared to CGM underestimated mean glucose in one, but overestimated mean glucose in the other two. However, all three studies did not include a control group and had few days of CGM measurements, which collectively mean that the evidence for the use of HbA1c in this patient group is sparse.
In a study of peritoneal dialysis subjects with diabetic nephropathy (n = 1296), 17.7% had normal HbA1c and were without antidiabetic treatment. When the same study included glycated albumin, 10.9% were found to have normal HbA1c, normal glycated albumin (defined as less than 16%) and were without antidiabetic medication. The study concluded that normal HbA1c alone overestimates the prevalence of burnt-out diabetes.
The uncertainty associated with HbA1c in persons with type 2 diabetes undergoing peritoneal dialysis is a problem both in terms of diagnosing and treating diabetes appropriately. CGM, glycated albumin and fructosamine are independent of erythrocyte turnover thus allowing for a validation of HbA1c in this group of patients.
Objective:
To investigate the correlation between mean glucose measured by continuous glucose monitoring (CGM) and mean glucose estimated from glycated haemoglobin A1c (HbA1c) in persons with type 2 diabetes undergoing peritoneal dialysis compared with a control group of persons with type 2 diabetes and normal renal function.
Method:
Prospective case-control study over 16 days with 26 persons in each group. The case group consist of persons with type 2 diabetes undergoing peritoneal dialysis and the control group consists of persons with type 2 diabetes with normal renal function (defined as an estimated glomerular filtration rate (eGFR) above 60 ml/min and urine albumin-creatinine ratio below 300mg/g). CGM is performed for a maximum of 16 days. On visit 1 (day 0), CGM is performed for 8 days and visit 2 (day 8) (+7/-4 days), CGM is performed for another 8 days. Glycaemic markers are measured at baseline and on visit 3 (final day of the study). At least 12 days of CGM must be completed for each person. Data is collected using REDcap.
Statistics:
The null hypothesis is that there is no difference in the ratio between the two groups measured as mean glucose concentration measured over 16 days at CGM (mmol / L) divided by the estimated glucose concentration from HbA1c (mmol / L) measured at day 16.
The alternative hypothesis is that there is a 17% difference between groups for the ratio of mean glucose concentration measured by CGM (mmol / L) divided by the mean glucose concentration from HbA1c (mmol / L). In a pilot study of persons with type 2 diabetes and dialysis dependent ESRD compared with persons with normal renal function, a 17% difference was found in the ratio between the two groups. In a two-sided unpaired t-test with α = 0.05, SD = 1.17 and power of 90%, a sample size of 24 individuals in each group will show a significant difference. A drop of approximately 10% is expected and a total of 26 are included in each group.
