DIABETICMedicine DOI: 10.1111/dme.12432
Research: Epidemiology Establishing HbA1c–mean blood glucose formulae for patients on continuous ambulatory peritoneal dialysis C. Zhao*, Q. Luo*, F. He, F. Peng, X. Xia, F. Huang and X.Yu Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China Accepted 9 March 2014
Abstract To determine the relationship between HbA1c and mean blood glucose concentrations by using HbA1c–mean blood glucose formulae for people on continuous ambulatory peritoneal dialysis.
Aim
Methods A total of 305 people on continuous ambulatory peritoneal dialysis, including 13 people with Type 1 diabetes mellitus, 161 people with Type 2 diabetes mellitus and 131 people without diabetes, from a single peritoneal dialysis centre at the First Affiliated Hospital of Sun Yat-sen University, were enrolled between January 2006 and June 2011. Serum HbA1c concentration was measured quarterly and other laboratory variables, including blood glucose, were measured every month. The formulae were established using regression analysis and adjusted for other factors. The estimated blood glucose level calculated using our formulae was compared with that using previous formulae namely the Diabetes Control and Complications Trial and A1c-Derived Average Glucose formulae for people not on dialysis and the Hoshino formula for people on haemodialysis.
The HbA1c–mean blood glucose formulae obtained by linear regression analysis were: 1) mBGmmol/l = 0.107 9 HbA1c(mmol/mol) + 1.764 [adjusted R2 (R2adj ) = 0.494]; 2) mBGmmol/l = 0.101 9 HbA1c (mmol/mol) 0.001 9 Cr (lmol/l) + 2.850 (R2adj = 0.507); 3) mBGmmol/l = 0.102 9 HbA1c (mmol/mol) 0.095 9 Alb (g/l) + 5.394 (R2adj = 0.521); and 4) mBGmmol/l = 0.099 9 HbA1c (mmol/mol) 0.001 9 Cr (lmol/l) 0.084 9 Alb (g/l) + 5.754 (R2adj = 0.526), where mBG is mean blood glucose, Cr is serum creatinine and Alb is serum albumin. These new formulae performed as well as or better than previous formulae.
Results
The relationship between HbA1c and mean blood glucose for people on continuous ambulatory peritoneal dialysis differs from that for people not on dialysis or for those on haemodialysis. Clinicians and patients can determine glycaemic control targets by applying our formulae.
Conclusions
Diabet. Med. 31, 813–820 (2014)
Introduction Thanks to its convenience, low cost, minimal haemodynamic change and other advantages, people with end-stage renal disease are increasingly being treated with continuous ambulatory peritoneal dialysis (CAPD), especially in developing countries [1,2]. Some studies reported that intensive glucose control was associated with lower mortality rates in people with diabetes who are on CAPD, and poor glucose control led to higher mortality rates [3,4]; therefore, it is important to assess glucose control accurately in people on CAPD.
Correspondence to: Fengxian Huang, E-mail: [email protected] *These authors contributed equally to the work
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
HbA1c has been widely accepted as the preferred standard of blood glucose evaluation in people with diabetes without end-stage renal disease; however, it is important to translate HbA1c, which reflects the previous 2–3 months glucose control, into mean blood glucose for day-to-day monitoring of blood glucose [5,6]. Several studies have used formulae to establish close correlations between HbA1c and mean blood glucose; for example, the Diabetes Control and Complications Trial (DCCT) formula was mBGmmol/l = 1.98 9 HbA1c (%) 4.29 and the A1c-Derived Average Glucose (ADAG) study formula was mBGmmol/l = 1.59 9 HbA1c (%) 2.59 [5,6], where mBG is mean blood glucose. When it comes to people on dialysis, however, the relationship between HbA1c and mean blood glucose becomes very complicated. The formation of HbA1c is affected by various clinical factors, such as red blood cell survival, haemoglobin concentration,
813
DIABETICMedicine
HbA1c–mean blood glucose formulae for continuous ambulatory peritoneal dialysis C. Zhao et al.
What’s new? • The exact relationship between HbA1c and mean blood glucose concentrations remains unclear in clinical practice, especially in patients on dialysis. To our knowledge, this study is the first to provide formulae to determine the exact relationship between HbA1c and mean blood glucose in Asian populations on continuous ambulatory peritoneal dialysis • The variables serum albumin and creatinine were included in the formulae, but haemoglobin was not. • The estimated blood glucose levels calculated using our formulae were lower than those calculated using previous formulae.
CAPD or < 6 months follow-up. If a participant who was enrolled in the study for > 6 months and who had provided at least two HbA1c measurements died, transferred to haemodialysis treatment or was lost to follow-up, the follow-up period was terminated and their previous laboratory results were still used in the analysis. People with Type 1 diabetes, Type 2 diabetes and without diabetes were all included in the study. The diagnostic criterion for diabetes at the beginning of CAPD was based on the diagnostic criteria of the American Diabetes Association [12]. The research design was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University and written informed consent was obtained from all participants before their inclusion in the study.
Participant data
erythropoietin injection, PH, uraemia environment and the glucose exposure duration [7,8]. For people on maintenance haemodialysis, the formula was mBGmg/dl = 29.7 9 HbA1c (%) 18.4 9 Alb (g/dL) 4.7 9 Hb (g/dL) + 104.8 [9], where mBG is mean blood glucose, Alb is serum albumin and Hb is haemoglobin. This formula implies that, because of some important covariates such as serum albumin and haemoglobin, the relationship between HbA1c and mean blood glucose is quite different in people on haemodialysis from that in people with diabetes who are not on dialysis. For people on CAPD, HbA1c is still thought to be a reliable biomarker of glycaemic control, which is associated with mortality rates [3,10]. Compared with people on haemodialysis, those on CAPD have lower serum albumin levels and use the higher glucose-containing peritoneal dialysis solution with longer exposure to glucose [7,11]; therefore, the exact relationship between HbA1c and mean blood glucose for CAPD would be expected to be different from that for haemodialysis. In the present study, HbA1c–mean blood glucose formulae for people on CAPD were established and adjusted for other covariates. The estimated blood glucose was then calculated using both our formulae and the existing formulae, described above.
Baseline demographic and clinical data, such as age, gender, height, weight and systolic and diastolic blood pressure, were obtained 3 months after the beginning of CAPD. Most laboratory variables, for example, haemoglobin, serum albumin, high-sensitivity C-reactive protein, total triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol, serum creatinine and blood glucose, were assayed in serum samples monthly. The HbA1c values were assayed every 3 months using high-performance liquid chromatography (VARIANT II Hemoglobin A1c Testing System; Bio-Rad Corp., Hercules, CA, USA). The blood glucose values were fasting blood glucose levels and were assayed by an automatic biochemical analyser (Abbott AEROSET 2000; Abbott Corp., Chicago, IL, USA). All serological variables were assayed in the biochemical laboratory of the First Affiliated Hospital of Sun Yat-sen University before dialysis. Dialysis duration was measured from the day on which the patient first began CAPD. Kt/V values were calculated to assess CAPD adequacy, where K is dialyser clearance of urea, t is dialysis time, and V is volume of distribution of urea.
Statistical analysis
Patients and Methods Study design and participants
A total of 309 people on CAPD were recruited from a single peritoneal dialysis centre at the First Affiliated Hospital of Sun Yat-sen University in Guangzhou, Guangdong province, China, between 1 January 2006 and 30 June 2011. The follow-up period ended on 31 March 2012. Inclusion criteria included patients aged > 18 years who had received stable and regular CAPD for > 3 months and who did not have acute illness, malignant diseases or other serious complications at the time of recruitment. Exclusion criteria included recent haemodialysis treatment, kidney transplant before
814
Continuous quantitative variables were reported as mean (SD) or median (interquartile range) values and analysed using Student’s t-test or the non-parametric Mann–Whitney test, as appropriate. Categorical variables were reported as frequencies or percentages and were analysed using the chi-squared test or Fisher’s exact test, as appropriate. The values of all laboratory variables were averaged to obtain mean values. Linear regression analysis according to the least squares method was used to establish formulae which described the relationship between mean HbA1c and mean blood glucose. We then adjusted the formulae for other covariates such as serum albumin, serum creatinine, haemoglobin, BMI and age. Non-linear regression analyses, for example, quadratic, logarithms and exponential regression models, were also
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
DIABETICMedicine
tested. The adjusted coefficient of determination (R2adj) and root mean squared error were used to assess the fitting degree of the above formulae. The previous formulae were used as control formulae; these were mBGmmol/l = 1.98 9 HbA1c (%) 4.29, mBGmmol/l = 1.59 9 HbA1c (%) 2.59 and mBGmg/dl = 29.4 9 HbA1c (%) 20.8 9 Alb(g/dL) + 59.2 [5,6,9], where mBG is mean blood glucose. To convert from mg/dl units to mmol/l units for blood glucose, the value of the conventional unit was multiplied by 0.0555. All statistical analyses were performed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined as P < 0.05 with two-tailed tests.
Results Baseline characteristics
Of the 309 initially enrolled participants, four died or transferred to haemodialysis within the first 6 months of the study, and were excluded from the analysis; thus 305 participants with adequate data were included in the final analysis. These 305 participants provided a total of 9042 blood samples for assessing blood glucose levels and other serological variables, and 3011 HbA1c measurements with an average of ~10 measured values per patient. The baseline characteristics for the patients stratified by the presence (n = 174) or absence of diabetes (n = 131) are shown in Table 1. The mean (SD; range) participant age was 55.88 (13.87; 21–79) years and patients with diabetes were significantly older than patients without diabetes [mean (SD) age 60.11 (11.02) vs 50.26 (15.25) years, P < 0.001]. The mean (SD) blood glucose level for all participants was 6.71 (2.23) mmol/l, and participants with diabetes had a
significantly higher mean (SD) blood glucose level than participants wihout diabetes [7.86 (2.14) vs 5.19 (1.21) mmol/l, P < 0.001]. The mean (SD) HbA1c level for all participants was 46 (15) mmol/mol or 6.4 (1.4)%. People with diabetes had a significantly higher mean (SD) HbA1c level than those without diabetes [53 (14) mmol/mol or 7.0 (1.3)% vs 38 (11) mmol/mol or 5.6 (1.0)%, P < 0.001]. Meanwhile, people with diabetes had significantly lower mean (SD) diastolic blood pressure [74.87 (12.15) vs 83.57 (12.70) mmHg, P < 0.001], lower mean (SD) serum albumin [34.83 (3.62) vs 37.71 (3.91) g/l, P < 0.001), lower mean (SD) serum creatinine [752.08 (248.81) vs 898.90 (274.55) lmol/l, P < 0.001], higher mean (SD) BMI [22.64 (2.87) vs 21.74 (3.07) kg/m2, P = 0.008] and higher mean (SD) Kt/V values [2.44 (0.56) vs 2.26 (0.49), P = 0.017] than patients without diabetes. There was no significant difference in gender, systolic blood pressure, haemoglobin, high-sensitivity C-reactive protein, total triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol or dialysis duration between the two groups. A total of 68 participants had peritonitis infection events during the study period.
Blood glucose status in participants during the study period
The mean (SD) blood glucose level at first measurement for all participants was 6.05 (2.62) mmol/l when they were enrolled into the study and at the last measurement it was 6.97 (3.81) mmol/l, which was significantly higher (P < 0.001). Similarly, the mean (SD) HbA1c at first measurement for all patients was 45 (13) mmol/mol or 6.3 (1.2)%, which significantly increased to 46 (16) mmol/mol or 6.4(1.5)% at the last measurement (P = 0.043). The different degrees of glucose control, based on the mean HbA1c values,
Table 1 Baseline demographic and clinical characteristics of all the study participants
Characteristic Mean (SD) age, years Men, n (%) Mean (SD) BMI, kg/m2 Mean (SD) systolic blood pressure, mmHg Mean (SD) diastolic blood pressure, mmHg Mean (SD) haemoglobin, g/l Mean (SD) serum albumin, g/l Median (interquartile range) high-sensitivity C-reactive protein, mg/l Mean (SD) total triglycerides, mmol/l Mean (SD) total cholesterol, mmol/l Mean (SD) LDL cholesterol, mmol/l Mean (SD) HDL cholesterol, mmol/l Mean (SD) serum creatinine, lmol/l Mean (SD) blood glucose, mmol/l Mean (SD) HbA1c, mmol/mol Mean (SD) HbA1c, % Mean (SD) dialysis duration, months Kt/V
All participants N = 305
People without diabetes n = 131
People with diabetes n = 174
P*
55.88 178 22.25 135.59 78.58 101.59 36.07 4.90
(13.87) (58.4) (2.98) (20.56) (13.10) (14.66) (4.00) (2.14,8.61)
50.26 71 21.74 134.93 83.57 101.95 37.71 4.71
(15.25) (54.2) (3.07) (18.84) (12.70) (15.44) (3.91) (1.95,8.08)
60.11 107 22.64 136.08 74.87 101.32 34.83 5.18
(11.02) (61.5) (2.87) (21.79) (12.15) (14.08) (3.62) (2.22,9.67)
Research: Epidemiology Establishing HbA1c–mean blood glucose formulae for patients on continuous ambulatory peritoneal dialysis C. Zhao*, Q. Luo*, F. He, F. Peng, X. Xia, F. Huang and X.Yu Department of Nephrology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China Accepted 9 March 2014
Abstract To determine the relationship between HbA1c and mean blood glucose concentrations by using HbA1c–mean blood glucose formulae for people on continuous ambulatory peritoneal dialysis.
Aim
Methods A total of 305 people on continuous ambulatory peritoneal dialysis, including 13 people with Type 1 diabetes mellitus, 161 people with Type 2 diabetes mellitus and 131 people without diabetes, from a single peritoneal dialysis centre at the First Affiliated Hospital of Sun Yat-sen University, were enrolled between January 2006 and June 2011. Serum HbA1c concentration was measured quarterly and other laboratory variables, including blood glucose, were measured every month. The formulae were established using regression analysis and adjusted for other factors. The estimated blood glucose level calculated using our formulae was compared with that using previous formulae namely the Diabetes Control and Complications Trial and A1c-Derived Average Glucose formulae for people not on dialysis and the Hoshino formula for people on haemodialysis.
The HbA1c–mean blood glucose formulae obtained by linear regression analysis were: 1) mBGmmol/l = 0.107 9 HbA1c(mmol/mol) + 1.764 [adjusted R2 (R2adj ) = 0.494]; 2) mBGmmol/l = 0.101 9 HbA1c (mmol/mol) 0.001 9 Cr (lmol/l) + 2.850 (R2adj = 0.507); 3) mBGmmol/l = 0.102 9 HbA1c (mmol/mol) 0.095 9 Alb (g/l) + 5.394 (R2adj = 0.521); and 4) mBGmmol/l = 0.099 9 HbA1c (mmol/mol) 0.001 9 Cr (lmol/l) 0.084 9 Alb (g/l) + 5.754 (R2adj = 0.526), where mBG is mean blood glucose, Cr is serum creatinine and Alb is serum albumin. These new formulae performed as well as or better than previous formulae.
Results
The relationship between HbA1c and mean blood glucose for people on continuous ambulatory peritoneal dialysis differs from that for people not on dialysis or for those on haemodialysis. Clinicians and patients can determine glycaemic control targets by applying our formulae.
Conclusions
Diabet. Med. 31, 813–820 (2014)
Introduction Thanks to its convenience, low cost, minimal haemodynamic change and other advantages, people with end-stage renal disease are increasingly being treated with continuous ambulatory peritoneal dialysis (CAPD), especially in developing countries [1,2]. Some studies reported that intensive glucose control was associated with lower mortality rates in people with diabetes who are on CAPD, and poor glucose control led to higher mortality rates [3,4]; therefore, it is important to assess glucose control accurately in people on CAPD.
Correspondence to: Fengxian Huang, E-mail: [email protected] *These authors contributed equally to the work
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
HbA1c has been widely accepted as the preferred standard of blood glucose evaluation in people with diabetes without end-stage renal disease; however, it is important to translate HbA1c, which reflects the previous 2–3 months glucose control, into mean blood glucose for day-to-day monitoring of blood glucose [5,6]. Several studies have used formulae to establish close correlations between HbA1c and mean blood glucose; for example, the Diabetes Control and Complications Trial (DCCT) formula was mBGmmol/l = 1.98 9 HbA1c (%) 4.29 and the A1c-Derived Average Glucose (ADAG) study formula was mBGmmol/l = 1.59 9 HbA1c (%) 2.59 [5,6], where mBG is mean blood glucose. When it comes to people on dialysis, however, the relationship between HbA1c and mean blood glucose becomes very complicated. The formation of HbA1c is affected by various clinical factors, such as red blood cell survival, haemoglobin concentration,
813
DIABETICMedicine
HbA1c–mean blood glucose formulae for continuous ambulatory peritoneal dialysis C. Zhao et al.
What’s new? • The exact relationship between HbA1c and mean blood glucose concentrations remains unclear in clinical practice, especially in patients on dialysis. To our knowledge, this study is the first to provide formulae to determine the exact relationship between HbA1c and mean blood glucose in Asian populations on continuous ambulatory peritoneal dialysis • The variables serum albumin and creatinine were included in the formulae, but haemoglobin was not. • The estimated blood glucose levels calculated using our formulae were lower than those calculated using previous formulae.
CAPD or < 6 months follow-up. If a participant who was enrolled in the study for > 6 months and who had provided at least two HbA1c measurements died, transferred to haemodialysis treatment or was lost to follow-up, the follow-up period was terminated and their previous laboratory results were still used in the analysis. People with Type 1 diabetes, Type 2 diabetes and without diabetes were all included in the study. The diagnostic criterion for diabetes at the beginning of CAPD was based on the diagnostic criteria of the American Diabetes Association [12]. The research design was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University and written informed consent was obtained from all participants before their inclusion in the study.
Participant data
erythropoietin injection, PH, uraemia environment and the glucose exposure duration [7,8]. For people on maintenance haemodialysis, the formula was mBGmg/dl = 29.7 9 HbA1c (%) 18.4 9 Alb (g/dL) 4.7 9 Hb (g/dL) + 104.8 [9], where mBG is mean blood glucose, Alb is serum albumin and Hb is haemoglobin. This formula implies that, because of some important covariates such as serum albumin and haemoglobin, the relationship between HbA1c and mean blood glucose is quite different in people on haemodialysis from that in people with diabetes who are not on dialysis. For people on CAPD, HbA1c is still thought to be a reliable biomarker of glycaemic control, which is associated with mortality rates [3,10]. Compared with people on haemodialysis, those on CAPD have lower serum albumin levels and use the higher glucose-containing peritoneal dialysis solution with longer exposure to glucose [7,11]; therefore, the exact relationship between HbA1c and mean blood glucose for CAPD would be expected to be different from that for haemodialysis. In the present study, HbA1c–mean blood glucose formulae for people on CAPD were established and adjusted for other covariates. The estimated blood glucose was then calculated using both our formulae and the existing formulae, described above.
Baseline demographic and clinical data, such as age, gender, height, weight and systolic and diastolic blood pressure, were obtained 3 months after the beginning of CAPD. Most laboratory variables, for example, haemoglobin, serum albumin, high-sensitivity C-reactive protein, total triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol, serum creatinine and blood glucose, were assayed in serum samples monthly. The HbA1c values were assayed every 3 months using high-performance liquid chromatography (VARIANT II Hemoglobin A1c Testing System; Bio-Rad Corp., Hercules, CA, USA). The blood glucose values were fasting blood glucose levels and were assayed by an automatic biochemical analyser (Abbott AEROSET 2000; Abbott Corp., Chicago, IL, USA). All serological variables were assayed in the biochemical laboratory of the First Affiliated Hospital of Sun Yat-sen University before dialysis. Dialysis duration was measured from the day on which the patient first began CAPD. Kt/V values were calculated to assess CAPD adequacy, where K is dialyser clearance of urea, t is dialysis time, and V is volume of distribution of urea.
Statistical analysis
Patients and Methods Study design and participants
A total of 309 people on CAPD were recruited from a single peritoneal dialysis centre at the First Affiliated Hospital of Sun Yat-sen University in Guangzhou, Guangdong province, China, between 1 January 2006 and 30 June 2011. The follow-up period ended on 31 March 2012. Inclusion criteria included patients aged > 18 years who had received stable and regular CAPD for > 3 months and who did not have acute illness, malignant diseases or other serious complications at the time of recruitment. Exclusion criteria included recent haemodialysis treatment, kidney transplant before
814
Continuous quantitative variables were reported as mean (SD) or median (interquartile range) values and analysed using Student’s t-test or the non-parametric Mann–Whitney test, as appropriate. Categorical variables were reported as frequencies or percentages and were analysed using the chi-squared test or Fisher’s exact test, as appropriate. The values of all laboratory variables were averaged to obtain mean values. Linear regression analysis according to the least squares method was used to establish formulae which described the relationship between mean HbA1c and mean blood glucose. We then adjusted the formulae for other covariates such as serum albumin, serum creatinine, haemoglobin, BMI and age. Non-linear regression analyses, for example, quadratic, logarithms and exponential regression models, were also
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
DIABETICMedicine
tested. The adjusted coefficient of determination (R2adj) and root mean squared error were used to assess the fitting degree of the above formulae. The previous formulae were used as control formulae; these were mBGmmol/l = 1.98 9 HbA1c (%) 4.29, mBGmmol/l = 1.59 9 HbA1c (%) 2.59 and mBGmg/dl = 29.4 9 HbA1c (%) 20.8 9 Alb(g/dL) + 59.2 [5,6,9], where mBG is mean blood glucose. To convert from mg/dl units to mmol/l units for blood glucose, the value of the conventional unit was multiplied by 0.0555. All statistical analyses were performed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined as P < 0.05 with two-tailed tests.
Results Baseline characteristics
Of the 309 initially enrolled participants, four died or transferred to haemodialysis within the first 6 months of the study, and were excluded from the analysis; thus 305 participants with adequate data were included in the final analysis. These 305 participants provided a total of 9042 blood samples for assessing blood glucose levels and other serological variables, and 3011 HbA1c measurements with an average of ~10 measured values per patient. The baseline characteristics for the patients stratified by the presence (n = 174) or absence of diabetes (n = 131) are shown in Table 1. The mean (SD; range) participant age was 55.88 (13.87; 21–79) years and patients with diabetes were significantly older than patients without diabetes [mean (SD) age 60.11 (11.02) vs 50.26 (15.25) years, P < 0.001]. The mean (SD) blood glucose level for all participants was 6.71 (2.23) mmol/l, and participants with diabetes had a
significantly higher mean (SD) blood glucose level than participants wihout diabetes [7.86 (2.14) vs 5.19 (1.21) mmol/l, P < 0.001]. The mean (SD) HbA1c level for all participants was 46 (15) mmol/mol or 6.4 (1.4)%. People with diabetes had a significantly higher mean (SD) HbA1c level than those without diabetes [53 (14) mmol/mol or 7.0 (1.3)% vs 38 (11) mmol/mol or 5.6 (1.0)%, P < 0.001]. Meanwhile, people with diabetes had significantly lower mean (SD) diastolic blood pressure [74.87 (12.15) vs 83.57 (12.70) mmHg, P < 0.001], lower mean (SD) serum albumin [34.83 (3.62) vs 37.71 (3.91) g/l, P < 0.001), lower mean (SD) serum creatinine [752.08 (248.81) vs 898.90 (274.55) lmol/l, P < 0.001], higher mean (SD) BMI [22.64 (2.87) vs 21.74 (3.07) kg/m2, P = 0.008] and higher mean (SD) Kt/V values [2.44 (0.56) vs 2.26 (0.49), P = 0.017] than patients without diabetes. There was no significant difference in gender, systolic blood pressure, haemoglobin, high-sensitivity C-reactive protein, total triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol or dialysis duration between the two groups. A total of 68 participants had peritonitis infection events during the study period.
Blood glucose status in participants during the study period
The mean (SD) blood glucose level at first measurement for all participants was 6.05 (2.62) mmol/l when they were enrolled into the study and at the last measurement it was 6.97 (3.81) mmol/l, which was significantly higher (P < 0.001). Similarly, the mean (SD) HbA1c at first measurement for all patients was 45 (13) mmol/mol or 6.3 (1.2)%, which significantly increased to 46 (16) mmol/mol or 6.4(1.5)% at the last measurement (P = 0.043). The different degrees of glucose control, based on the mean HbA1c values,
Table 1 Baseline demographic and clinical characteristics of all the study participants
Characteristic Mean (SD) age, years Men, n (%) Mean (SD) BMI, kg/m2 Mean (SD) systolic blood pressure, mmHg Mean (SD) diastolic blood pressure, mmHg Mean (SD) haemoglobin, g/l Mean (SD) serum albumin, g/l Median (interquartile range) high-sensitivity C-reactive protein, mg/l Mean (SD) total triglycerides, mmol/l Mean (SD) total cholesterol, mmol/l Mean (SD) LDL cholesterol, mmol/l Mean (SD) HDL cholesterol, mmol/l Mean (SD) serum creatinine, lmol/l Mean (SD) blood glucose, mmol/l Mean (SD) HbA1c, mmol/mol Mean (SD) HbA1c, % Mean (SD) dialysis duration, months Kt/V
All participants N = 305
People without diabetes n = 131
People with diabetes n = 174
P*
55.88 178 22.25 135.59 78.58 101.59 36.07 4.90
(13.87) (58.4) (2.98) (20.56) (13.10) (14.66) (4.00) (2.14,8.61)
50.26 71 21.74 134.93 83.57 101.95 37.71 4.71
(15.25) (54.2) (3.07) (18.84) (12.70) (15.44) (3.91) (1.95,8.08)
60.11 107 22.64 136.08 74.87 101.32 34.83 5.18
(11.02) (61.5) (2.87) (21.79) (12.15) (14.08) (3.62) (2.22,9.67)
