Superiority of glycated albumin over glycated haemoglobin as indicator of glycaemic control and predictor of all-cause mortality in patients with type 2 diabetes mellitus receiving peritoneal dialysis

Abstract

Background

Glycated albumin, in contrast to glycated haemoglobin, precisely reflects glycaemic control and predicts all-cause mortality in haemodialysis patients with diabetes mellitus. However, whether those associations exist in diabetes mellitus patients receiving peritoneal dialysis remains unclear.

Methods

This was a retrospective cross-sectional and longitudinal observational study. We measured glycated albumin, glycated haemoglobin and casual plasma glucose for two months in diabetes mellitus-peritoneal dialysis (n = 44) and diabetes mellitus-haemodialysis (n = 88) patients (age-, gender-matched). The diabetes mellitus-peritoneal dialysis patients were followed for three years to monitor occurrence of all-cause mortality.

Results

Glycated albumin and glycated albumin/casual plasma glucose ratios, but not casual plasma glucose, glycated haemoglobin, or glycated haemoglobin/casual plasma glucose, were significantly lower in the diabetes mellitus-peritoneal dialysis as compared with the diabetes mellitus-haemodialysis patients. The regression lines between casual plasma glucose and glycated albumin showed a significant parallel shift downwards in diabetes mellitus-peritoneal dialysis as compared with diabetes mellitus-haemodialysis patients, while the slope did not differ significantly between the groups, resulting in underestimation of glycaemic control by 4.5%. Kapan-Meier analysis of the diabetes mellitus-peritoneal dialysis patients revealed that higher glycated albumin (median >18.0%), but not glycated haemoglobin (median >6.6%), indicated significantly elevated risk for all-cause mortality, which occurred in 15 patients (34.1%), as compared with those with a lower glycated albumin concentration. Higher glycated albumin concentration was also significantly and independently associated with all-cause mortality in multivariate Cox proportional hazards analysis.

Conclusions

Glycated albumin, in contrast to glycated haemoglobin, more precisely reflects glycaemic control in diabetes mellitus-peritoneal dialysis patients, based on its significant association with all-cause mortality. Furthermore, adjustment of the true glycated albumin concentration by adding 4.5% might provide a more precise measurement for determining glycaemic control in such patients.

Keywords

Peritoneal dialysis glycated albumin glycated haemoglobin glycaemic control all-cause mortality

Background

Glycated haemoglobin (HbA1c) is widely used as a marker of glycaemic control, although it is known to be influenced by the lifespan of red blood cells.¹ In patients with advanced chronic kidney disease (CKD) or receiving dialysis, it is becoming increasingly recognized that HbA1c measurement might underestimate glycaemic condition due to anaemia, use of erythropoiesis stimulating agents (ESA) and/or iron, independent of glycaemic control.² We previously demonstrated that glycated albumin (GA), which is not influenced by the lifespan of red blood cells, provides a significantly better estimate of glycaemic control in haemodialysis (HD) patients with diabetes mellitus (DM) as compared with HbA1c.^3,4 On the other hand, subjects with increased albumin metabolism are considered to have lower GA values, due to the shortened time of serum albumin exposure to glucose in plasma.⁵ Patients receiving peritoneal dialysis (PD) lose protein in peritoneal dialysate or urine; thus, it is possible that increased albumin turnover can falsely suppress GA in PD patients with DM, making it unclear whether GA and/or HbA1c are reliable markers of glycaemic control in DM-PD patients.

Strict glycaemic control has favourable effects not only on diabetes complications^6–8 but also cardiovascular disease, a leading cause of death in DM patients.^9–11 We previously found that GA, but not HbA1c, is independently and positively associated with pulse wave velocity and peripheral vascular calcification in DM-HD patients.^12,13 In addition, we reported that GA predicts all-cause mortality,¹⁴ which is confirmed by other studies that clearly demonstrated that GA, rather than HbA1c, precisely predicts all-cause mortality in DM-HD patients,¹⁵ likely because it accurately reflects glycaemic control status. Although a recent study showed that GA, but not HbA1c, predicts all-cause mortality in DM-PD patients,¹⁶ no comparisons of GA with HbA1c or plasma glucose were performed because of a lack of simultaneous measurements of the three markers. In the present investigation, those three glycaemic markers were simultaneously determined; thus, this is the first study to precisely assess the clinical significance of GA as a marker for glycaemic control as well as predictor of all-cause mortality in DM-PD patients.

Methods

Subjects and study design

This was a retrospective cross-sectional and longitudinal observational study. The entry criteria for this study included diagnosis of DM-PD in patients who had been treated while in a stable condition at the Inoue Hospital, and whose therapeutic regimen for DM had not been altered during the preceding six months and weekly dose of ESA had not been changing during the preceding three months prior to simultaneous determination of GA and HbA1c (n = 51). Those who had been receiving dialysis therapy for less than six months were excluded (n = 2). Furthermore, patients with liver cirrhosis, malignancy, infection, acute illness or history of peritonitis were also excluded from the analysis (n = 5). As a result, 44 DM-PD patients (32 males, 12 females) were enrolled as subjects in the present study. We previously performed a cross-sectional study (Osaka CKD Expert Research) that included 538 DM-HD patients who also underwent simultaneous measurements of GA and HbA1c, and whose therapeutic regimen for DM and weekly dose of ESA had not been altered during the preceding six and three months, respectively, prior to those measurements. That study aimed to assess whether GA might provide a better indication as compared with HbA1c of glycaemic control in DM-HD patients.³ Corresponding to the present 44 enrolled patients, we sequentially selected 88 age- and gender-matched DM-HD patients from the Osaka CKD Expert Research cohort for analysis. The diagnosis of DM was based on history of diabetes or criteria presented in the Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus.¹⁷

The present study was conducted in accordance with the principles of the Declaration of Helsinki and retrospectively approved by the Ethics Committee of Inoue Hospital (approval no. 219). Institutional approval for an opt-out consent method was given because of the observational nature (retrospective analysis) of this study.

Laboratory measurements

Blood was drawn without overnight fasting from DM-PD patients between 8:00 and 10:00 am, and from DM-HD patients just prior to starting their first dialysis session of the week. GA was measured with an enzymatic method using a Lucica GA-L kit (Asahi Kasei Pharma Corp., Tokyo, Japan), as previously reported.¹⁸ Briefly, GA was hydrolysed to amino acids by albumin-specific proteinase and then oxidized by ketoamine oxidase to produce hydrogen peroxide, which was quantitatively measured. The GA value was calculated as the percentage of GA relative to total albumin, which was measured using the new bromocresol purple method.¹⁸ GA assay findings are not influenced by the physiologic concentrations of ascorbic acid, bilirubin, or glucose up to 1000 mg/dL. HbA1c was measured by routine HPLC and latex agglutination immunoassay techniques, and was then standardized according to the method of the Japan Diabetes Society. HbA1c values were converted to National Glycohemoglobin Standardization Program equivalent values, according to the recommendations of the Japanese Diabetes Society and International Federation of Clinical Chemistry.¹⁹ The mean values of three measurements of casual plasma glucose (CPG) obtained during the two months prior to determination of serum GA and HbA1c were used in the present analyses. We calculated Geriatric Nutritional Risk Index (GNRI) values based on serum albumin and body weight, as follows^20–22

\begin{array}{l} GNRI = [14.89 \times albumin (g / dL)] \\ + [41.7 \times (body weight / ideal body weight) \end{array}

Outcome data collection

The DM-PD patients were followed for three years. At the end of the follow-up period, 18 patients were alive and receiving PD, 11 were transferred to HD and the remaining 15 had died.

Statistical analysis

Data are expressed as the mean ± standard deviation. A non-repeated t-test (continuous variables with normal distribution) and Fisher's exact test (categorical variables) were used to compare variables between groups. To compare variables between the three measured factors, we used one-way analysis of variance (ANOVA), followed by a Tukey-Kramer post hoc test. Pearson’s correlation test was used to determine correlations between continuous variables. Analysis of covariance was used to compare regression lines. Survival curves were estimated using the Kaplan-Meier analysis method with a log-rank test. A receiver operator characteristic (ROC) curve was constructed to identify the optimum cut-off value for predicting all-cause mortality. Prognostic variables for survival were examined using a Cox proportional hazards regression model. All statistical analyses were performed using the SPSS 22.0 statistical package for Windows (SPSS Inc., Chicago, IL, USA). All reported P values are two-tailed and were considered to be statistically significant at <0.05.

Results

Clinical characteristics of DM-PD and DM-HD patients

Table 1 shows the clinical characteristics of the 44 DM-PD and 88 DM-HD patients. The DM-PD group exhibited significantly lower concentrations of serum albumin and significantly shorter dialysis durations as compared with the DM-HD group, and also tended to have higher haemoglobin concentrations and body mass index (BMI). Although the average CPG concentration was not significantly different between the groups, GA was significantly lower and HbA1c tended to be higher in the DM-PD group. In addition, the GA/CPG and GA/HbA1c ratios were significantly lower in the DM-PD patients, while HbA1c/CPG ratio was not significantly different between the groups.

Table 1.

Clinical characteristics of the patients.

	DM-PD(n=44)	DM-HD(n=88)	P
Age (years)	62.2±9.8	62.2±10.0	1.000
Male	32 (72.7)	64 (72.7)	1.000
Dialysis duration (years)	2.4±2.2	5.2±4.1	<0.001
History of CVD events	21 (47.7)	–	–
Body weight (kg)	63.3±12.3	59.7±10.7	0.089
Body mass index (kg/m²)	23.9±3.6	22.8±3.2	0.087
Haemoglobin (g/dL)	10.3±1.2	9.9±1.2	0.066
Albumin (g/dL)	3.4±0.6	3.7±0.4	0.011
Creatinine (mg/dL)	9.7±2.6	10.8±2.6	0.016
GNRI	96.3±11.7	98.0±8.4	0.334
CPG (mg/dL)	157.6±37.1	149.6±51.4	0.359
HbA1c (%)	6.6±0.9	6.3±1.1	0.078
GA (%)	19.0±4.5	23.0±6.3	<0.001
GA/CPG ratio	0.125±0.032	0.163±0.050	<0.001
HbA1c/CPG ratio	0.044±0.012	0.045±0.012	0.657
GA/HbA1c ratio	2.879±0.579	3.655±0.697	<0.001
Dose of erythropoietin (U/week)	5054.0±2964.2	4721.6±2244.6	0.513

Note: Values are presented as the mean ± standard deviation or number (%) of dichotomous variables. P values are shown for comparisons of mean values between the groups (unrepeated t-test) or categorical variables (Fisher's exact test).

CVD: cardiovascular disease; CPG: casual plasma glucose; HbA1c: glycated haemoglobin; GA: glycated albumin; GNRI: Geriatric Nutritional Risk Index; DM: diabetes mellitus; PD: peritoneal dialysis.

CPG concentration variations during the two-month study period in DM-PD and DM-HD patients

CPG concentrations in the DM-PD patients obtained at two months and one month before and at the time (0 months) of the GA and HbA1c measurements were 162.5 ± 48.0, 165.3 ± 56.1 and 147.7 ± 40.6 mg/dL, respectively. Findings of one-way ANOVA did not indicate significant differences (P = 0.192), while those of the Tukey-Kramer post hoc test also showed that CPG in DM-PD patients was not significantly different among the three measurement time points (–2 vs. 0 months, P = 0.335; –1 vs. 0 months, P = 0.208; –2 vs. 1 month; P = 0.962). In the DM-HD patients, CPG concentrations were 152.6 ± 68.0, 144.0 ± 55.2 and 152.2 ± 63.3 mg/dL, respectively, at the three measurement time points. One-way ANOVA did not show significant differences (P = 0.537), and post hoc Tukey-Kramer test findings also showed that CPG in DM-HD patients was not significantly different among the three measurement time points (–2 vs. 0 months, P = 0.999; –1 vs. 0 months, P = 0.654; –2 vs. 1 month; P = 0.631). These results suggested that glycaemic control in both the DM-PD and DM-HD patients during the two-month examination period, which might have significant effects on HbA1c and GA, were stable enough to determine the association of the average CPG value during the prior two months with GA or HbA1c in such patients.

Correlation of CPG with GA and HbA1c in DM-PD and DM-HD patients

Figure 1 shows simple correlations of CPG with GA (Figure 1(a)) and HbA1c (Figure 1(b)) in DM-PD and DM-HD patients. Although the slope of the regression lines between CPG and GA did not differ significantly between the groups (P = 0.772), those lines for the DM-PD patients showed a significant downward direction as compared with those for the DM-HD patients (P < 0.001) (Figure 1(a)). On the other hand, the slope or position of the regression lines between CPG and HbA1c in both groups was nearly identical (Figure 1(b)).

Figure 1.

Correlations of CPG with GA and HbA1c in DM-PD and DM-HD patients. (a) Linear regression equations used. (1) Patients receiving peritoneal dialysis: y = 0.065x + 8.76, r = 0.538, P < 0.001; (2) Patients receiving haemodialysis: y = 0.071x + 12.39, r = 0.578, P < 0.001. Analysis of covariance (ANCOVA) findings indicated that the regression line was significantly different between the groups (P < 0.001). (b) Linear regression equations used: (1) Patients receiving peritoneal dialysis: y = 0.0095x + 5.107, r = 0.401, P = 0.007; (2) Patients receiving haemodialysis: y = 0.012x + 4.473, r = 0.582, P < 0.001. ANCOVA findings indicated that the regression line was not significantly different between the groups (P = 0.134).

Correlation of clinical parameters with GA in DM-PD patients

Among the clinical parameters examined, CPG (r = 0.538, P < 0.001) and HbA1c (r = 0.500, P = 0.001) were significantly and positively correlated with GA, whereas there was no correlation shown for age (r = 0.110, P = 0.478), gender (r = –0.039, P = 0.802), dialysis duration (r = –0.015, P = 0.925), serum albumin (r = –0.116, P = 0.455), serum creatinine (r = –0.055, P = 0.723), haemoglobin (r = –0.190, P = 0.216), BMI (r = –0.174, P = 0.260) or GNRI (r = –190, P = 0.217).

Association of clinical parameters with all-cause mortality in DM-PD patients

During the follow-up period of 36 months, 15 of 44 DM-PD patients died, including seven from cardiovascular causes (cardiovascular disease [n = 3], cerebrovascular disease [n = 4]) and eight from non-cardiovascular causes (infectious disease [n = 6], malignancy [n = 2]). When the DM-PD patients were divided into two subgroups based on the median levels of GA (18.0%), HbA1c (6.6%) and CPG (155 mg/dL), Kaplan-Meier analysis (Figure 2) showed that a higher GA value was positively associated with all-cause mortality, whereas higher HbA1c and CPG were not. ROC analysis also showed that GA had a greater AUC (AUC = 0.711, P = 0.023) to predict all-cause mortality as compared with that of HbA1c (AUC = 0.602, P = 0.271) and CPG (AUC = 0.632, P = 0.155) (Figure 3). The cut-off value for greatest sensitivity and specificity for GA was 18.0% (sensitivity 0.800, specificity 0.345), which was consistent with the median value of GA. To confirm the independent association of higher GA with all-cause mortality in DM-PD patients, multivariate Cox proportional hazards analyses were performed (Table 2). In model 2, which included higher GA and albumin as covariates, higher GA concentration and lower albumin concentration were significantly associated with all-cause mortality. When age, gender, dialysis duration, BMI, haemoglobin, past cardiovascular disease events, dose of erythropoietin, serum creatinine and GNRI were sequentially added to model 2 (models 3–11, respectively), higher GA remained significantly associated with greater mortality in all of the models.

Figure 2.

Findings of Kaplan-Meier analysis of associations of GA, HbA1c and CPG with all-cause mortality. The patients were divided into two groups according to the median values for GA (18.0%), CPG (155 mg/dL) and HbA1c (6.6%). Probability was analysed using a log-rank test.

Figure 3.

Receiver operating characteristic (ROC) curves for GA, HbA1c and CPG to predict all-cause mortality. (a) GA, (b) HbA1c, (c) CPG.

Table 2.

Independent association of higher GA concentration with all-cause mortality determined by multivariate Cox proportional analysis.

Model	Adjustment	HR	95% CI	P
1	GA (high=1, low=0), unadjusted	4.782	1.346–16.998	0.016
2	Model 1+ albumin	4.775	1.334–17.090	0.016
3	Model 2 + age	4.420	1.223–15.982	0.023
4	Model 2 + gender	4.788	1.337–17.147	0.016
5	Model 2 + dialysis duration	4.868	1.352–17.524	0.015
6	Model 2 + body mass index	4.632	1.294–16.576	0.018
7	Model 2 + haemoglobin	4.735	1.321–16.973	0.017
8	Model 2 + past CVD events	4.736	1.317–17.028	0.017
9	Model 2 + dose of erythropoietin	4.978	1.391–17.817	0.014
10	Model 2 + creatinine	4.462	1.229–16.196	0.023
11	Model 2 + GNRI	4.632	1.294–16.576	0.018

HR: hazard ratio; CI: confidence interval; GA: glycated albumin; CVD: cardiovascular disease; GNRI: Geriatric Nutritional Risk Index.

Discussion

In the present study, GA concentration and GA/CPG ratio were significantly lower in DM-PD as compared with DM-HD patients, while CPG was not significantly different between the groups. This suggests that GA was significantly low for CPG in the DM-PD patients as compared with the DM-HD patients. Interestingly, the regression lines between CPG and GA for the DM-PD and DM-HD patients were parallel, though shifted significantly downwards in the former as compared with the latter group. It seems that GA value may reflect glycaemic control in DM-PD patients as precisely as seen in DM-HD patients, although measured GA in DM-PD patients may lead to underestimation of glycaemic control (Figure 1). Furthermore, the present study showed that impaired glycaemic control in DM-PD patients determined by GA, but not HbA1c value is able to independently predict all-cause mortality (Table 2).

Other studies including ours^3,4 have accumulated results, indicating that GA provides a more precise measurement to indicate glycaemic control than HbA1c in DM-HD patients. This conclusion is based on findings showing that suppression of HbA1c by renal anaemia, as well as the use of ESA and/or iron independent of glycaemic control can lead to an underestimation of glycaemic control in DM-HD patients. On the other hand, a large-scale cohort study of 850 DM-PD patients reported that HbA1c values were influenced not only by mean CPG but also by haemoglobin.²³ Therefore, since HbA1c may be suppressed by anaemia independent of glycaemic control in DM-PD patients, glycaemic control might be underestimated in anaemic DM-PD patients as compared with non-anaemic DM patients without CKD when estimated based on HbA1c. In the present study, the regression lines between HbA1c and CPG did not differ between the DM-HD and DM-PD groups, while the slope of those lines was significantly shallower than a regression line previously reported for non-CKD DM patients,³ suggesting that HbA1c measurements might be falsely reduced by anaemia or ESA dosage in both DM-PD and DM-HD patients.

Previously, only a limited number of studies^24–26 have examined the clinical significance of GA measurement for evaluation of glycaemic control in DM-PD patients. Although GA is not affected by anaemia or haemoglobin concentration, albumin metabolism is known to have an influence.⁵ Low GA values have been observed in subjects with increased albumin metabolism, such as hyperthyroidism²⁷ and nephrotic syndrome,²⁸ due to the shortened exposure time of serum albumin to glucose in plasma. PD patients lose protein at a rate of approximately 5 g/day in peritoneal fluid and urine,²⁹ and albumin synthesis in the liver is stimulated to compensate for that loss,³⁰ resulting in a higher turnover of circulating albumin. Therefore, DM-PD patients are considered to exhibit low GA concentrations for their CPG as compared with DM non-PD patients, because of the shortened half-life of albumin in circulation. As expected, the present study found that GA and GA/CPG were significantly lower in DM-PD as compared with DM-HD patients (Table 1), as previously reported in investigations of DM-PD.^24,25 Interestingly, we found a downward shift of the regression line without a significant change in the DM-PD as compared with the DM-HD group (Figure 1), suggesting that GA measurement retained its significance as a glycaemic marker, except for underestimation of glycaemic control in those patients. The present findings also confirm those of previous studies showing that serum albumin concentration was not correlated with GA concentration in DM-PD patients.^24,26 Considering the parallel downward shift of the regression lines between GA and CPG in the DM-PD as compared with the DM-HD group, it seems that adjustment of the true GA concentration by adding 4.5% might provide more a precise measurement for determining glycaemic control in DM-PD patients.

We previously reported that GA, but not HbA1c, may be a significant measurement for determination of arterial wall stiffening¹² and peripheral vascular calcification¹³ in DM-HD patients. Furthermore, several studies including ours^14,15 have demonstrated that GA, rather than HbA1c, can precisely predict all-cause mortality in those patients. Consistent with a recent study showing that GA (≥20%) but not HbA1c concentration was associated with all-cause mortality in DM-PD patients,¹⁶ the present results also showed that those patients with elevated GA (>18.0%), but not elevated HbA1c have a greater risk for death, which was independent of serum creatinine concentration as a marker of sarcopenic status³¹ and GNRI as a marker of nutritional status.^20–22 Furthermore, the adjusted cut-off value increased to 22.5%, when the GA concentration was adjusted (measured GA concentration +4.5%). Our previous report¹⁴ as well as a large-scale epidemiological study¹⁵ both showed that DM-HD patients with a GA concentration >20–21% have significantly higher risk for all-cause mortality; thus, the cut-off GA value to determine increased risk for all-cause mortality might be similar in DM-PD and DM-HD patients.

This study has some important limitations. First, it was retrospective in design, and the number of subjects was small. Also, peritoneal transport function and albumin loss were not fully measured. In addition, the study population consisted of nearly exclusively Japanese patients; thus, it is unclear whether these findings can be generalized for other ethnic groups. Nevertheless, the present study clearly indicated that GA is superior than HbA1c as an indicator of glycaemic control and all-cause mortality in DM-PD patients.

In conclusion, GA, in contrast to HbA1c, more precisely reflects glycaemic control in DM-PD patients, although GA concentration is consistently lower than the CPG concentration in DM-PD patients as compared with DM-HD patients. GA might be a better predictor of all-cause mortality in cases of DM-PD.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

The present study was approved by the Ethics Committee of Inoue Hospital (Approval No. 219).

Guarantor

MK.

Contributorship

MM researched literature, conceived the study, researched data and contributed to the discussions and edited the manuscript. MK and MI researched literature, contributed to the discussions, wrote and edited the article. KM, Senji O, Shigeki O, ME and YT contributed to the discussion, and reviewed and edited the article. All authors reviewed and edited the article and approved the final version of the article.

References

Khera

Smith

Lindsell

, et al. Use of an oral stable isotope label to confirm variation in red blood cell mean age that influences HbA1c interpretation. Am J Hematol 2015; 90: 50–55.

Zheng

, et al. Glycated albumin in diabetic patients with chronic kidney disease. Clin Chim Acta 2012; 413: 1555–1561.

Inaba

Okuno

Kumeda

, et al. Glycated albumin is a better glycemic indicator than glycated hemoglobin values in hemodialysis patients with diabetes: effect of anemia and erythropoietin injection. JASN 2007; 18: 896–903.

Nagayama

Inaba

Okabe

, et al. Glycated albumin as an improved indicator of glycemic control in hemodialysis patients with type 2 diabetes based on fasting plasma glucose and oral glucose tolerance test. Biomed Pharmacother 2009; 63: 236–240.

Koga

Glycated albumin; clinical usefulness. Clin Chim Acta 2014; 433: 96–104.

Diabetes

Complications Trial Research

Nathan

, et al. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: 977–986.

Ohkubo

Kishikawa

Araki

, et al. Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract 1995; 28: 103–117.

Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet 1998; 352: 837–853.

Gaede

Vedel

Larsen

, et al. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med 2003; 348: 383–393.

10.

Chiasson

Josse

Gomis

, et al. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial. JAMA 2003; 290: 486–494.

11.

Nathan

Lachin

Cleary

, et al. Intensive diabetes therapy and carotid intima-media thickness in type 1 diabetes mellitus. N Engl J Med 2003; 348: 2294–2303.

12.

Kumeda

Inaba

Shoji

, et al. Significant correlation of glycated albumin, but not glycated haemoglobin, with arterial stiffening in haemodialysis patients with type 2 diabetes. Clin Endocrinol (Oxf) 2008; 69: 556–561.

13.

Yamada

Inaba

Shidara

, et al. Association of glycated albumin, but not glycated hemoglobin, with peripheral vascular calcification in hemodialysis patients with type 2 diabetes. Life Sci 2008; 83: 516–519.

14.

Inaba

Maekawa

Okuno

, et al. Impact of atherosclerosis on the relationship of glycemic control and mortality in diabetic patients on hemodialysis. Clin Nephrol 2012; 78: 273–280.

15.

Chen

Drechsler

Suntharalingam

, et al. High glycated albumin and mortality in persons with diabetes mellitus on hemodialysis. Clin Chem 2017; 63: 477–485.

16.

Abe

Hamano

Hoshino

, et al. Glycemic control and survival in peritoneal dialysis patients with diabetes: a 2-year nationwide cohort study. Sci Rep 2019; 9: 3320.

17.

Expert Committee on the D and Classification of Diabetes M. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 2003; 26: S5–S20.

18.

Kouzuma

Uemastu

Usami

, et al. Study of glycated amino acid elimination reaction for an improved enzymatic glycated albumin measurement method. Clin Chim Acta 2004; 346: 135–143.

19.

Committee of the Japan Diabetes Society on the Diagnostic Criteria of Diabetes M, Seino

Nanjo

, et al. Report of the committee on the classification and diagnostic criteria of diabetes mellitus. J Diab Invest 2010; 1: 212–228.

20.

Yamada

Furuya

Takita

, et al. Simplified nutritional screening tools for patients on maintenance hemodialysis. Am J Clin Nutr 2008; 87: 106–113.

21.

Kobayashi

Ishimura

Kato

, et al. Geriatric Nutritional Risk Index, a simplified nutritional screening index, is a significant predictor of mortality in chronic dialysis patients. Nephrol, Dial, Transplant 2010; 25: 3361–3365.

22.

Kang

Cho

Park

, et al. Geriatric Nutritional Risk Index as a prognostic factor in peritoneal dialysis patients. Perit Dial Int 2013; 33: 405–410.

23.

Hoshino

Mehrotra

Rhee

, et al. Using hemoglobin A1c to derive mean blood glucose in peritoneal dialysis patients. Am J Nephrol 2013; 37: 413–420.

24.

Kobayashi

Abe

Yoshida

, et al. Glycated albumin versus glycated hemoglobin as a glycemic indicator in diabetic patients on peritoneal dialysis. Int J Mol Sci 2016; 17: 619.

25.

Freedman

Shenoy

Planer

, et al. Comparison of glycated albumin and hemoglobin A1c concentrations in diabetic subjects on peritoneal and hemodialysis. Perit Dial Int 2010; 30: 72–79.

26.

Watanabe

Ohno

Inoue

, et al. Blood glucose levels in peritoneal dialysis are better reflected by HbA1c than by glycated albumin. Adv Perit Dial 2014; 30: 75–82.

27.

Koga

Murai

Saito

, et al. Effects of thyroid hormone on serum glycated albumin levels: study on non-diabetic subjects. Diabetes Res Clin Pract 2009; 84: 163–167.

28.

Okada

Nakao

Matsumoto

, et al. Influence of proteinuria on glycated albumin values in diabetic patients with chronic kidney disease. Intern Med 2011; 50: 23–29.

29.

Blumenkrantz

Gahl

Kopple

, et al. Protein losses during peritoneal dialysis. Kidney Int 1981; 19: 593–602.

30.

Kaysen

Schoenfeld

PY.

Albumin homeostasis in patients undergoing continuous ambulatory peritoneal dialysis. Kidney Int 1984; 25: 107–114.

31.

Inaba

Kurajoh

Okuno

, et al. Poor muscle quality rather than reduced lean body mass is responsible for the lower serum creatinine level in hemodialysis patients with diabetes mellitus. Clin Nephrol 2010; 74: 266–272.