- Email: [email protected]
Discovery and validation of serum creatinine variability as novel biomarker for predicting onset of albuminuria in Type 2 diabetes mellitus
diabetes research and clinical practice
1 3 8 ( 2 0 1 8 ) 8 –1 5
Contents available at ScienceDirect
Diabetes Research and Clinical Practice journa l home page: www.e lse vier.com/locate/diabres
Discovery and validation of serum creatinine variability as novel biomarker for predicting onset of albuminuria in Type 2 diabetes mellitus Serena Low a, Xiao Zhang a, Keven Ang a, Su Jian Darren Yeo a, Guanyi Joel Lim a, Lee Ying Yeoh b, Yan Lun Liu b, Tavintharan Subramaniam c, Chee Fang Sum c, Su Chi Lim a,c,d,* a
Clinical Research Unit, Khoo Teck Puat Hospital, Singapore Department of Medicine, Khoo Teck Puat Hospital, Singapore c Diabetes Centre, Khoo Teck Puat Hospital, Singapore d Saw Swee Hock School of Public Health, National University of Singapore, Singapore b
A R T I C L E I N F O
A B S T R A C T
Article history:
Aim: We aim to study association serum creatinine(cr) variability and albuminuria progres-
Received 26 August 2017
sion.
Received in revised form
Methods: We conducted a retrospective cohort study on patients with Type 2 Diabetes Mel-
10 October 2017
litus at a Diabetes Centre in Singapore (‘‘discovery cohort”). Outcome is worsening of uri-
Accepted 7 November 2017
nary albumin-to-creatinine(ACR) across stages. Cr variability was expressed as adjusted
Available online 31 January 2018
cr-intrapersonal standard deviation(SD) and coefficient-of-variation(cr-CV). A separate cohort was used for validating association between cr variability and albuminuria progres-
Keywords: Creatinine Albuminuria Type 2 Diabetes Diabetic nephropathy
sion (‘‘validation cohort”). Results: Over median follow-up of 4.2 years, 38.4% of 636 patients had albuminuria progression in the discovery cohort. Increasing log-transformed adjusted cr-intrapersonal SD and cr-CV were significantly associated with albuminuria progression: HRs 1.43 (95%CI 1.11– 1.85) and 1.44 (1.11–1.87) respectively in the discovery cohort, and HRs 1.94 (1.09–3.45) and 1.91 (1.05–3.45) respectively in the validation cohort. When stratified by baseline urinary ACR, higher cr variability was significantly associated with albuminuria progression in patients with normoalbuminuria but not microalbuminuria. Conclusions: Cr variability independently predicts albuminuria onset. This is evident in patients with normoalbuminuria, suggesting that higher cr variability could herald albuminuria onset. Ó 2017 Published by Elsevier Ireland Ltd.
1.
Introduction
Kidney disease is a major complication of Type 2 Diabetes Mellitus (T2DM), affecting 25–40% of individuals with T2DM
[1]. Of note, Asians have a higher tendency to develop diabetic kidney disease (DKD) than other ethnicities such as Caucasians [2]. Given the substantial morbidity, mortality and economic costs imposed by DKD, it is imperative to identify
* Corresponding author at: Diabetes Centre, Khoo Teck Puat Hospital, 90 Yishun Central, 768828, Singapore. E-mail address: [email protected] (S.C. Lim). https://doi.org/10.1016/j.diabres.2017.11.003 0168-8227/Ó 2017 Published by Elsevier Ireland Ltd.
diabetes research and clinical practice
the risk factors of DKD onset and progression in order to support prevention efforts. There is accumulating literature on risk factors predictive of chronic kidney disease progression. These included age, gender, body mass index, systolic blood pressure, serum creatinine, urinary albumin-to-creatinine ratio (uACR), estimated glomerular filtration rate and even novel biomarkers [3]. Baseline serum GFR has long been recognized as a risk factor of kidney disease progression in T2DM. However, baseline GFR reveals little information on the functional reserve of the kidney. Functional renal mass can differ in different individuals even though their baseline GFR can be similar [4]. In the presence of reduced residual renal mass, more stress is exerted on the remaining nephrons. This in turn leads to hyperfiltration and kidney disease progression [4]. There is a need to look beyond baseline renal function to gain deeper insights into the progression of kidney disease. It was observed that reported patients with higher serum creatinine fluctuation during admission experienced a higher risk of mortality post discharge compared to those with more fluctuation [5]. One possible explanation was that rapid fluctuations in the GFR are reflective of acute kidney injury (AKI). Patients with AKI have a higher death risk during inpatient hospitalization and in the first year post hospital discharge [5,6]. To date, the role of stability of renal function on progression of chronic kidney disease remains under-explored. We therefore aim to study the association between serum creatinine variability over long-term and onset and progression of albuminuria in T2DM.
2.
Materials and methods
2.1.
Study population
We conducted a retrospective cohort study on patients with T2DM who enrolled at a Diabetes Centre in a regional hospital in Singapore in 2002–2014 (‘‘discovery cohort”). The exclusion criteria were: less than 21 years of age, pregnancy, active infections, active cancer, autoimmune disease, presence of other suspected causes of kidney disease (e.g. urinary tract infection, polycystic kidney disease, haematuria or history of glomerulonephritis). For the purpose of analysis, subjects were included if they had at least two years of follow-up for serum creatinine, absence of macroalbuminuria indicated by urinary albumin-to-creatinine ratio (ACR) >300 mg/g at baseline, at least two measurements of urinary ACR, four or more measurements of serum creatinine, and a follow-up period for urinary ACR of at least one month after the last serum creatinine measurement. There were altogether 636 patients eligible for the analysis. See Supplementary Fig. 1 for the flowchart.
9
size in the sitting position after a resting period of at least 10 min. One random day spot urine sample for ACR was collected for each patient at baseline and through follow-ups. Follow-up urinary ACR was collected on a yearly basis or more frequently if the ACR was high. Blood and spot urine samples were collected and sent to the hospital laboratory accredited by the College of the American Pathologists (CAP). Serum creatinine was quantitated with enzymatic colorimetric test (Roche cobasÒ c 501); Haemoglobin A1c (HbA1c) with Tinaquant Haemoglobin A1c Gen.3 (Roche cobasÒ c 501); and urinary albumin with immunoturibidmimetric assay (Roche cobasÒ c 501). Measurements were taken at recruitment and at multiple time points during follow-up. GFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation formula [7]. A separate cohort was also used for validating the association between serum creatinine variability and kidney disease progression. This comprised adult subjects attending the Diabetes Centre in the same regional hospital or a primary-care polyclinic in 2011–2016. The patients were included in the analysis if they had no macroalbuminuria at baseline, at least two urinary ACR measurements, at least four measurements of serum creatinine, and the last followup urinary ACR measurement taken after the last serum creatinine measurement. A total of 199 subjects met the inclusion criteria. See Supplementary Fig. 2 for the flowchart. Trained nurses measured BP using a standard sphygmomanometer using an appropriate cuff size in the sitting position after a resting period of at least 10 min. Two BP readings were taken and the average BP was derived. Clinical and laboratory measurements were taken at multiple time points till February 2016. One random day urine sample for ACR was taken at baseline and follow-ups on yearly basis or more frequently if the ACR was high. Blood and urine samples were quantitated at the same hospital laboratory using assays similar to those for the DN cohort. The study was approved by the National Healthcare Group Domain Specific Review Board in Singapore. All participants provided written informed consent.
2.3.
Outcome definition
The following stages of diabetic nephropathy were graded as follows: normoalbuminuria, microalbuminuria and macroalbuminuria indicated by uACR < 30 mg/g, 30–299 mg/g and 300 mg/g respectively. Patients who experienced worsening of these stages – from normo- to micro- or macroalbuminuria, or from micro- to macroalbuminuria were considered as having albuminuria progression (‘‘progressors”). Those who did not experience worsening of these stages were ‘‘nonprogressors”.
2.4. 2.2.
1 3 8 ( 2 0 1 8 ) 8 –1 5
Calculation of serum creatinine variability
Data collection
Clinical and demographic information were obtained from a standard questionnaire or extracted from patient’s medical records. Trained nurses measured Blood pressure (BP) using a standard sphygmomanometer using an appropriate cuff
The median number of serum creatinine measurements was 6.0 (IQR 5.0–8.0) in the discovery cohort and 5.0 (4.0–5.0) in the validation cohort. Intrapersonal mean and standard deviation (SD) of serum creatinine were calculated for each patient. Serum creatinine
10
diabetes research and clinical practice
1 3 8 ( 2 0 1 8 ) 8 –1 5
variability was expressed as adjusted cr-intrapersonal SD p using the formula: SD/ [n/(n1) where SD is the intrapersonal standard deviation of serum creatinine and n is the number of serum creatinine readings [8–10]. Furthermore, the coefficient of variation of serum creatinine variability (cr-CV) was calculated as intrapersonal SD of serum creatinine divided by intrapersonal mean (crintrapersonal mean) and expressed as a percentage [11]. CrCV was chosen as a normalized measure of serum creatinine variability so as to correct for larger SD attributed to higher absolute values of serum creatinine based on reference from past studies on HbA1c variability [12,13].
disease progression including age of entry, gender, ethnicity, duration of DM, HbA1c, SBP, HbA1c, estimated GFR, urinary ACR group and use of renin-angiotensin antagonist [3]. Assumption of proportional hazard was tested for all covariates with global test using scale Schoenfeld residuals. The assumption was not violated by the cox regression model in our analysis (p > .05). Multivariable logistic regression was used to estimate odds ratio for albuminuria progression in the validation cohort, adjusting for the same variables. All statistical tests were two-sided and considered significant if P < .05. Statistical analysis was done using STATA Version 14.0 (STATA Corporation, College Station, Texas).
2.5.
3.
Statistical analysis
Categorical variables were presented as number (percentage), and continuous variables as means (standard deviation) or median (interquartile range) as appropriate. Differences in patient characteristics stratified by progression were examined by Chi-Square test for categorical variables and student-t test or Mann-Whitney test for continuous variables. Multivariable cox proportional hazards regression model was used to estimate hazard ratios (HR) for albuminuria progression in the discovery cohort, adjusting for crintrapersonal mean and other known risk factors of kidney
Results
The overall median follow-up duration was 4.2 years (3.0–5.7) in the discovery cohort and 2.4 years (1.6–3.3) in the validation cohort. The patients were followed up for a median period of 8.4 months (IQR 3.6–12.0) and 22.8 months (IQR 13.2–32.4) after the last serum creatinine measurement in the discovery cohort and validation cohort respectively. Progression occurred in 38.4% in the discovery cohort anmd 29.6% in the validation cohort. The baseline characteristics in the discovery cohort were as follows: mean age 56.2 ± 11.5 years, 61.8% males, 70.2%
Table 1 – Baseline characteristics by progression of albuminuria in the discovery cohort. Variables
Number Age (years) Male (%) Ethnicity (%) Chinese Malay Indian DM duration (%) 10 years 11–20 years >20 years SBP 140 mmHg (%) HbA1c (%) (mmol/mol) <7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) eGFR (ml/min/1.73 m2) Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist (%) Use of Insulin (%) Cr-intrapersonal mean (umol/l) Adjusted Cr-intrapersonal SD (umol/l) Cr-coefficient of variation (%)
All
Progression
P-value
No
Yes
636 56.2 ± 11.5 393 (61.8)
392 55.2 ± 11.1 240 (61.2)
244 57.7 ± 11.9 153 (62.7)
446 (70.2) 90 (14.2) 99 (15.6)
279 (71.4) 44 (11.3) 68 (17.4)
167 (68.4) 46 (18.9) 31 (12.7)
390 (62.1) 177 (28.2) 61 (9.7) 224 (36.3)
251 (65.0) 106 (27.5) 29 (7.5) 119 (31.5)
139 (57.4) 71 (29.3) 32 (13.2) 105 (43.9)
173 (27.2)
112 (28.6)
61 (25.0)
174 (27.4)
106 (27.0)
68 (27.9)
131 (20.6)
79 (20.2)
52 (21.3)
158 (24.8)
95 (24.2)
63 (25.8)
88.0 (69.7–101.9)
91.6 (77.0–103.2)
78.9 (57.0–98.4)
350 (55.0) 286 (45.0) 363 (57.4) 191 (31.6) 79.4 (64.3–101.4) 6.7 (4.5–10.2) 9.6 (6.8–13.6)
253 (64.5) 139 (35.5) 188 (48.3) 110 (29.7) 75.4 (60.7–90.7) 6.2 (4.2–8.8) 8.9 (6.5–12.7)
97 (39.8) 147 (60.3) 175 (71.7) 81 (34.8) 89.2 (69.1–117.1) 8.6 (5.3–13.5) 10.8 (7.4–15.0)
0.008 0.709 0.015
0.038
0.002 0.803
<0.001 <0.001
<0.001 0.188 <0.001 <0.001 <0.001
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-tocreatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine.
diabetes research and clinical practice
Chinese, 14.2% Malay, 15.6% Indian, 36.3% SBP 140 mmHg, 45.4% HbA1c 8.0% (63.9 mmol/mol), 55.0% normoalbuminuria, 45.0% microalbuminuria, 57.4% used RAS antagonist, 31.6% used insulin, median adjusted cr-SD 6.7 (IQR 4.5–10.2), cr-CV 9.6% (6.8–13.6). There were no statistically significant differences between the discovery and validation cohorts in terms of age at entry, duration of DM and urinary ACR category. See Table 1 and Supplementary Table 1. In the discovery cohort, the progressors had poorer baseline clinical profile in terms of SBP, eGFR and urinary ACR (p < .05). They also had higher cr-intrapersonal mean, crintrapersonal SD and cr-CV (p < .001) than the nonprogressors. Similarly, the cr-intrapersonal SD and cr-CV were higher in progressors than non-progressors in the validation cohort (p < .05). See Table 1 and Supplementary Table 1. In the multivariable model which included only age, gender and ethnicity, increasing log-transformed adjusted crintrapersonal SD and log-transformed cr-CV were significantly associated with progression of albuminuria: adjusted hazards ratios (HRs) 1.45 (95% Confidence Interval (CI), 1.15– 1.85; and 1.45 (95%CI, 1.15–1.83) respectively in the discovery cohort, and HRs 1.71 (95%CI, 1.03–2.85) and 1.69 (95%CI, 1.01–2.82) respectively in the validation cohort. See Tables 2 and 3, and Supplementary Tables 2 and 3. After additional adjustment for duration of DM, SBP, HbA1c, eGFR, urinary ACR group, use of RAS antagonist and cr-intrapersonal mean, increasing log-transformed adjusted cr-intrapersonal SD and log-transformed cr-CV remained significantly associated with progression of albuminuria without much attenuation: HRs 1.43 (95%CI, 1.11–1.85) and 1.44 (95% CI, 1.11–1.87) respectively in the discovery cohorts, and HRs 1.94 (95%CI, 1.09–3.45) and 1.91 (95%CI, 1.05–3.45) respectively in the validation cohort. See Tables 2 and 3, and Supplementary Tables 2 and 3. In the multivariable models stratified by urinary ACR group in the discovery cohort, the association between increasing log-transformed adjusted cr-intrapersonal SD and log-transformed cr-CV with progression were strengthened in patients with normoalbuminuria with corresponding HRs 1.59 (95%CI, 1.06–2.38) and 1.61 (95%CI, 1.07–2.42). However, the associations were largely attenuated in patients with microalbuminuria, with corresponding HRs 1.33 (95%CI, 0.96–1.85) and 1.34(95%CI, 0.96–1.86). Similar findings were noted in the validation cohort. Although increasing logtransformed cr-intrapersonal mean was found associated with progression of albuminuria when it was included in the multivariable model relating cr-CV to outcome in the discovery cohort (Table 3), there was no association in the other multivariable models. See Table 2, and Supplementary Tables 2 and 3.
4.
Discussion
We have demonstrated that higher serum creatinine variability was associated with elevated risk of onset of albuminuria. This finding was consistent in two different cohorts with T2DM in which serum creatinine variability was expressed as adjusted cr-intrapersonal SD and logtransformed cr-CV.
1 3 8 ( 2 0 1 8 ) 8 –1 5
11
Research has shown that serum creatinine variability predicted mortality in inpatients [5]. Studies have also have investigated the relationship between eGFR variability and ESRD. [14,15] However, to the best of our knowledge, the role of serum creatinine variability has not been explored in the early phase of DKD (i.e. progression from normoalbuminuria to higher grade albuminuria). Albuminuria is present in many individuals who have normal eGFR. It has been linked to cardiovascular and renal outcomes independently of eGFR at baseline. There was suggestion that albuminuria reflects glomerular and tubular injury, thereby contributing to improvement in prediction of renal prognosis in addition to eGFR which only reflects glomerular injury [16]. Our study, for the first time, revealed the putative role of serum creatinine variability on onset of albuminuria. This corroborates previous findings on the association between variability or short-term alterations in eGFR, which is derived from serum creatinine, and end-stage renal disease or mortality [14,15,17,18]. Furthermore, there is an accumulating body of evidence showing the effect of variability in terms of HbA1c, BP, lipids and uric acid on kidney disease progression in recent years [10,19–23]. It is known that variability plays a vital role in maintenance of homeostasis in the body. It has also been shown that maintaining certain biological parameters in a restricted range confers protective effect in clinical outcome [19]. This supports the observation that higher HbA1c variability is related to deterioration of kidney function in our previous study and other studies [10,19–22]. Our current study adds serum creatinine to the current limited pool of biological markers known to affect outcomes in kidney disease in T2DM through fluctuations in their levels. The exact mechanisms of how serum creatinine variability putatively contributed to onset of albuminuria are not clear. There are a few possible mechanisms. Firstly, serum creatinine variability may be reflective of maladaption to hemodynamic changes in the background of reduced nephron mass [18,24]. It is known that haemodynamic pathway contributes to the pathophysiology of diabetes kidney disease through imbalance in resistance of afferent and efferent arterioles which then leads to elevated hydrostatic pressure in the glomerulus [25]. Elevations in renal plasma flow, intraglomerular pressure, and glomerular hyperfiltration and hypertrophy contribute to structural and functional abnormalities in the reduced renal mass [24]. Secondly, acute kidney injury (AKI) involves modest and possibly reversible elevation of serum creatinine levels [18,26,27]. It has been suggested that AKI predisposes individuals at higher risk of CKD by causing renal fibrosis [28–30]. Interestingly, the relationship between serum creatinine variability and progression of albuminuria was only evident in the presence of normoalbuminuria but not microalbuminuria at baseline. Serum creatinine variability may be an early marker of albuminuria progression. It is noted that many individuals with T2DM and normoalbuminuria experienced deterioration in kidney function i.e. decline in eGFR [31]. One possible reason is use of RAS anatgonists which may have controlled proteinuria and delayed but may not have prevented the continued deterioration of renal function [31]. Further studies are necessary to understand the putative role
Variable
Crude HR (95% CI) p-value
Model 1a HR (95% CI) p-value Model 2b HR (95% CI) p-value Model 3c HR (95% CI) p-value Model 4c HR (95% CI) p-value Normoalbuminuria
Microalbuminuria
1.01 (0.98–1.03) 0.517
0.99 (0.97–1.01) 0.424
1.00 (0.98–1.01) 0.517
Gender Female Male
1.00 0.91 (0.70–1.18) 0.480
1.00 0.71 (0.53–0.95) 0.022
1.00 0.79 (0.54–1.16) 0.233
Ethnicity Chinese Malay Indian
1.00 1.57 (1.13–2.17) 0.007 0.90 (0.61–1.31) 0.574
1.00 1.07 (0.75–1.53) 0.691 1.00 (0.68–1.48) 0.989
1.00 0.83 (0.56–1.22) 0.333 0.92 (0.61–1.38) 0.689
DM duration 10 years 11–20 years >20 years SBP 140 mmHg
1.00 1.32 (0.99–1.76) 0.058 2.15 (1.46–3.16) <0.001 1.63 (1.26–2.11) <0.001
1.00 0.90 (0.65–1.24) 0.531 1.47 (0.95–2.29) 0.085 1.38 (1.04–1.84) 0.025
1.52 (0.93–2.49) 0.098
1.45 (1.01–2.08) 0.042
1.00
1.00
1.00
1.00
1.10 (0.78–1.55) 0.601
1.25 (0.85–1.84) 0.250
0.83 (0.46–1.50) 0.546
1.81 (1.07–3.05) 0.026
1.02 (0.70–1.48) 0.916
1.36 (0.90–2.07) 0.148
1.04 (0.56–1.95) 0.898
1.70 (0.99–2.91) 0.056
1.27 (0.89–1.81) 0.182
1.50 (0.97–2.31) 0.066
1.07 (0.55–2.07) 0.850
1.66 (0.94–2.94) 0.080
0.27 (0.20–0.37) <0.001
0.86 (0.38–1.97) 0.727
0.62 (0.17–2.29) 0.472
0.70 (0.31–1.62) 0.407
1.00 2.10 2.02 1.40 4.82
1.00 1.45 1.46 0.97 2.48
0.016 0.023 0.860 0.064
1.60 (1.00–2.57) 0.050 1.35 (0.7902.32) 0.271 2.52 (0.68–9.38) 0.168
1.03 (0.66–1.60) 0.903 0.95 (0.62–1.43) 0.791 2.07 (0.79–5.43) 0.140
1.43 (1.11–1.85) 0.006
1.59 (1.06–2.38) 0.024
1.33 (0.96–1.85) 0.089
Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist Use of insulin Log-transformed Crintrapersonal mean (per umol/l) Log-transformed Adjusted Crintrapersonal SD (per umol/l)
(1.62–2.72) (1.53–2.68) (1.07–1.83) (3.30–7.04)
<0.001 <0.001 0.016 <0.001 3.16 (1.89–5.27) <0.001
2.08 (1.73–2.49) <0.001 1.45 (1.15–1.82) 0.002
(1.07–1.97) (1.05–2.01) (0.69–1.36) (0.95–6.47)
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-to-creatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine a Model 1 is adjusted for age at entry, gender and ethnicity. b Model 2 is adjusted for age at entry, gender, ethnicity, duration of DM, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, Urinary ACR group, use of RAS inhibitor(s) and use of insulin. c Models 3 and 4 are adjusted for age at entry, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, use of RAS inhibitor(s) and use of insulin.
1 3 8 ( 2 0 1 8 ) 8 –1 5
1.02 (1.01–1.03) <0.001 1.00 (0.99–1.01) 0.683
diabetes research and clinical practice
Age (per year)
HbA1c (per%) < 7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) Log-transformed eGFR (per ml/min/1.73 m2)
12
Table 2 – Cox Regression Models Relating Adjusted Creatinine-Intrapersonal Standard Deviation to Progression of Albuminuria in the Discovery Cohort.
Table 3 – Cox Regression Models Relating to Creatinine Coefficient of Variation to Progression of Albuminuria in the Discovery Cohort. Variable
Crude HR (95% CI) p-value
Model 5a HR (95% CI) p-value Model 6b HR (95% CI) p-value Model 7c HR (95% CI) p-value Model 8c HR (95% CI) p-value Normoalbuminuria
Microalbuminuria
1.01 (0.98–1.03) 0.525
0.99 (0.97–1.01) 0.416
1.00 (1.48–1.01) 0.504
Gender Female Male
1.00 0.91 (0.70–1.18) 0.480
1.00 0.71 (0.53–0.95) 0.022
1.00 0.80 (0.54–1.17) 0.241
Ethnicity Chinese Malay Indian
1.00 1.57 (1.13–2.17) 0.007 0.90 (0.61–1.31) 0.574
1.00 1.07 (0.75–1.53) 0.690 1.00 (0.68–1.48) 0.986
1.00 0.82 (0.56–1.21) 0.327 0.92 (0.61–1.38) 0.690
DM duration 10 years 11–20 years >20 years SBP 140 mmHg
1.00 1.32 (0.99–1.76) 0.058 2.15 (1.46–3.16) <0.001 1.63 (1.26–2.11) <0.001
1.00 0.91 (0.65–1.24) 0.529 1.48 (0.95–2.29) 0.083 1.38 (1.04–1.84) 0.025
1.52 (0.93–2.49) 0.097
1.45 (1.01–2.08) 0.042
1.00
1.00
1.00
1.00
1.10 (0.78–1.55) 0.601
1.25 (0.86–1.84) 0.246
0.84 (0.46–1.50) 0.548
1.81 (1.07–3.05) 0.026
1.02 (0.70–1.48) 0.916
1.36 (0.89–2.07) 0.150
1.04 (0.55–1.95) 0.909
1.69 (0.99–2.91) 0.056
1.27 (0.89–1.81) 0.182
1.51 (0.98–2.32) 0.063
1.06 (0.55–2.07) 0.854
1.67 (0.94–2.95) 0.078
0.27 (0.20–0.37) <0.001
0.86 (0.38–1.96) 0.718
0.63 (0.17–2.32) 0.485
0.70 (0.30–1.61) 0.401
1.00 2.10 2.02 1.40 4.82
1.00 1.45 1.46 0.97 3.54
1.60 (1.00–2.57) 0.051 1.36 (0.79–2.32) 0.268 4.10 (1.11–15.18) 0.035
1.03 (0.66–1.60) 0.901 0.94 (0.62–1.43) 0.789 2.75 (1.19–6.37) 0.018
1.61 (1.07–2.42) 0.023
1.34 (0.96–1.86) 0.088
HbA1c (per%) < 7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) Log-transformed eGFR (per ml/min/1.73 m2) Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist Use of insulin Log-transformed Crintrapersonal mean (per umol/l) Log-transformed cr coefficient of variation (per%)
(1.62–2.72) (1.53–2.68) (1.07–1.83) (3.30–7.04)
<0.001 <0.001 0.016 <0.001 4.60 (2.98–7.08) <0.001
1.65 (1.31–2.07) <0.001 1.45 (1.1501.83) 0.002
(1.07–1.97) (1.05–2.02) (0.69–1.36) (1.48–8.48)
0.016 0.023 0.854 0.005
1.44 (1.11–1.87) 0.006
13
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-to-creatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine a Model 5 is adjusted for age at entry, gender and ethnicity b Model 6 is adjusted for age at entry, gender, ethnicity, duration of DM, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, Urinary ACR group, use of RAS inhibitor(s) and use of insulin. c Models 7 and 8 are adjusted for age at entry, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, use of RAS inhibitor(s) and use of insulin.
1 3 8 ( 2 0 1 8 ) 8 –1 5
1.02 (1.01–1.03) <0.001 1.00 (0.99–1.01) 0.695
diabetes research and clinical practice
Age (per year)
14
diabetes research and clinical practice
of serum creatinine variability in the presence of normal urinary ACR. The unique features of our study include (1) availability of two separate longitudinal cohorts to establish the relationship between serum creatinine variability and albuminuria progression; (2) use of two measures of variability to consistently demonstrate association between serum creatinine variability and albuminuria progression; (3) availability of rich clinical information, including medications such as RAS and insulin, which could have altered the course of albuminuria progression; (4) sampling of patients from healthcare facilities which provided insights into the ‘‘real-world” relevant to healthcare providers and policy makers; and (5) availability of the follow-up serum creatinine values in the medical records for both inpatient and outpatient settings, even though the study was conducted in the diabetes centre. Such information allowed us to take into account the greater creatinine variability from more creatinine measurements in some patients with acute illness, especially the hospitalized patients, compared to those with a relatively stable disease. However, there are limitations in this study. Firstly, our study was confined to patients attending diabetes centre in a regional hospital. Hence the findings could not be generalized to the entire diabetes population. Secondly, urinary ACR was measured once at baseline and each follow-up using spot urine samples. We were unable to confirm the transition of albuminuria with a second measurement. Hence the outcome measure was prone to variability. However, it was reported that spot urine protein-creatinine ratio correlated well with 24 h urinary protein in patients with T2DM, suggesting that spot ACR measurement could be utilized as a reliable and fast alternative for assessing albuminuria [32,33]. Another limitation was that serum creatinine fluctuations may be attributed to changes in muscle mass, acute sickness, generation of creatinine in the muscle and malnutrition [5,34,35]. It may be necessary to exclude these factors when studying the impact of the serum creatinine variability. In addition, there could still be residual confounding due to socio-economic status and behavioural factors which may influence the progression of albuminuria. Nevertheless, our findings may have potential clinical implications. The measures of serum creatinine variability could be used in the routine clinical practice by incorporating them in the electronic clinical system. These measures would help clinicians to track the changes in serum creatinine and institute appropriate management to control the fluctuations. There is potential for therapeutic interventions to be developed to target at fluctuations of serum creatinine or to avoid stressors leading to fluctuating renal function e.g. contrast nephropathy, nonsteroidal anti-inflammatory drugs.
5.
Conclusions
In conclusion, higher serum creatinine variability is an independent predictor of albuminuria onset. The relationship is most evident in patients with normoalbuminuria, suggesting that higher serum creatinine variability could herald albuminuria of onset.
1 3 8 ( 2 0 1 8 ) 8 –1 5
Conflict of interest statement The authors declare that they have no conflict of interest.
Acknowledgements This study was supported by the Alexandra Health Grant (AHPL SIGII/11001; SIG/11029 and SIG/12024), National Medical Research Council Grant (NMRC/PPG/AH(KTPH)/2011) and National Medical Research Council Grant (NMRC/ CIRG/1398/2014).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.diabres.2017. 11.003.
R E F E R E N C E S
[1] Remuzzi G, Schieppati A, Ruggenenti P. Clinical practice. Nephropathy in patients with type 2 diabetes. N Engl J Med 2002;346:1145–51. [2] Kong AP, Xu G, Brown N, So WY, Ma RC, Chan JC. Diabetes and its comorbidities–where east meets west. Nat Rev Endocrinol 2013;9:537–47. [3] Echouffo-Tcheugui JB, Kengne AP. Risk models to predict chronic kidney disease and its progression: a systematic review. PLoS Med 2012;9:e1001344. [4] Ronco C, Rosner MH. Acute kidney injury and residual renal function. Crit Care 2012;16:144. [5] Kao SS, Kim SW, Horwood CM, Hakendorf P, Li JY, Thompson CH. Variability in inpatient serum creatinine: its impact upon short- and long-term mortality. QJM 2015;108:781–7. [6] McAlister FA, Ezekowitz J, Tonelli M, Armstrong PW. Renal insufficiency and heart failure: prognostic and therapeutic implications from a prospective cohort study. Circulation 2004;109:1004–9. [7] Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro 3rd AF, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–12. [8] Kilpatrick ES, Rigby AS, Atkin SL. A1c variability and the risk of microvascular complications in type 1 diabetes: data from the diabetes control and complications trial. Diabetes Care 2008;31:2198–202. [9] Marcovecchio ML, Dalton RN, Chiarelli F, Dunger DB. A1c variability as an independent risk factor for microalbuminuria in young people with type 1 diabetes. Diabetes Care 2011;34:1011–3. [10] Penno G, Solini A, Bonora E, Fondelli C, Orsi E, Zerbini G, et al. Hba1c variability as an independent correlate of nephropathy, but not retinopathy, in patients with type 2 diabetes: The renal insufficiency and cardiovascular events (riace) italian multicenter study. Diabetes Care 2013;36:2301–10. [11] Mosca A, Lapolla A, Gillery P. Glycemic control in the clinical management of diabetic patients. Clin Chem Lab Med 2013;51:753–66. [12] Rodriguez-Segade S, Rodriguez J, Garcia Lopez JM, Casanueva FF, Camina F. Intrapersonal hba(1c) variability and the risk of progression of nephropathy in patients with type 2 diabetes. Diabet Med 2012;29:1562–6.
diabetes research and clinical practice
[13] Waden J, Forsblom C, Thorn LM, Gordin D, Saraheimo M, Groop PH, et al. A1c variability predicts incident cardiovascular events, microalbuminuria, and overt diabetic nephropathy in patients with type 1 diabetes. Diabetes 2009;58:2649–55. [14] Tseng CL, Lafrance JP, Lu SE, Soroka O, Miller DR, Maney M, et al. Variability in estimated glomerular filtration rate values is a risk factor in chronic kidney disease progression among patients with diabetes. BMC Nephrol 2015;16:34. [15] Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, et al. Short-term change in kidney function and risk of end-stage renal disease. Nephrol Dial Transplant 2012;27:3835–43. [16] Gansevoort RT, Nauta FL, Bakker SJ. Albuminuria: all you need to predict outcomes in chronic kidney disease? Curr Opin Nephrol Hypertens 2010;19:513–8. [17] Perkins RM, Tang X, Bengier AC, Kirchner HL, Bucaloiu ID. Variability in estimated glomerular filtration rate is an independent risk factor for death among patients with stage 3 chronic kidney disease. Kidney Int 2012;82:1332–8. [18] Uehara K, Yasuda T, Shibagaki Y, Kimura K. Estimated glomerular filtration rate variability independently predicts renal prognosis in advanced chronic kidney disease patients. Nephron 2015;130:256–62. [19] Ceriello A, De Cosmo S, Rossi MC, Lucisano G, Genovese S, Pontremoli R, et al. Variability in hba1c, blood pressure, lipid parameters and serum uric acid, and risk of development of chronic kidney disease in type 2 diabetes. Diabetes Obes Metab 2017. [20] Low S, Lim SC, Yeoh LY, Liu YL, Liu JJ, Fun S, et al. Effect of long-term glycemic variability on estimated glomerular filtration rate decline among patients with type 2 diabetes mellitus: Insights from the diabetic nephropathy cohort in singapore. J Diabetes 2016. [21] Sugawara A, Kawai K, Motohashi S, Saito K, Kodama S, Yachi Y, et al. Hba(1c) variability and the development of microalbuminuria in type 2 diabetes: Tsukuba kawai diabetes registry 2. Diabetologia 2012;55:2128–31. [22] Tsuboi A, Takenouchi A, Kurata M, Fukuo K, Kazumi T. Postmeal triglyceridemia and variability of hba1c and postmeal glycemia were predictors of annual decline in estimated glomerular filtration rate in type 2 diabetic patients with different stages of nephropathy. J Diabetes Metab Disord 2017;16:1. [23] Yeh CH, Yu HC, Huang TY, Huang PF, Wang YC, Chen TP, et al. The risk of diabetic renal function impairment in the first
[24]
[25] [26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
1 3 8 ( 2 0 1 8 ) 8 –1 5
15
decade after diagnosed of diabetes mellitus is correlated with high variability of visit-to-visit systolic and diastolic blood pressure: a case control study. BMC Nephrol 2017;18:99. Gluhovschi G, Gadalean F, Gluhovschi C, Petrica L, Velciov S, Gluhovschi A, et al. The solitary kidney–a nephrological perspective. Rom J Intern Med 2013;51:80–8. Lim A. Diabetic nephropathy – complications and treatment. Int J Nephrol Renovasc Dis 2014;7:361–81. Bucaloiu ID, Kirchner HL, Norfolk ER, Hartle 2nd JE, Perkins RM. Increased risk of death and de novo chronic kidney disease following reversible acute kidney injury. Kidney Int 2012;81:477–85. Goldenberg I, Chonchol M, Guetta V. Reversible acute kidney injury following contrast exposure and the risk of long-term mortality. Am J Nephrol 2009;29:136–44. Basile DP. Rarefaction of peritubular capillaries following ischemic acute renal failure: a potential factor predisposing to progressive nephropathy. Curr Opin Nephrol Hypertens 2004;13:1–7. Basile DP, Donohoe D, Roethe K, Osborn JL. Renal ischemic injury results in permanent damage to peritubular capillaries and influences long-term function. Am J Physiol Renal Physiol 2001;281:F887–99. Coca SG, Singanamala S, Parikh CR. Chronic kidney disease after acute kidney injury: a systematic review and metaanalysis. Kidney Int 2012;81:442–8. Dwyer JP, Parving HH, Hunsicker LG, Ravid M, Remuzzi G, Lewis JB. Renal dysfunction in the presence of normoalbuminuria in type 2 diabetes: results from the demand study. Cardiorenal Med 2012;2:1–10. Biradar SB, Kallaganad GS, Rangappa M, Kashinakunti SV, Retnakaran R. Correlation of spot urine protein-creatinine ratio with 24-hour urinary protein in type 2 diabetes mellitus patients: a cross sectional study. J Res Med Sci 2011;16:634–9. Zhang X, Low S, Sum CF, Tavintharan S, Yeoh LY, Liu J, et al. Arterial stiffness is an independent predictor for albuminuria progression among asians with type 2 diabetes-a prospective cohort study. J Diabetes Complications 2017;31:933–8. Kovesdy CP, George SM, Anderson JE, Kalantar-Zadeh K. Outcome predictability of biomarkers of protein-energy wasting and inflammation in moderate and advanced chronic kidney disease. Am J Clin Nutr 2009;90:407–14. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, et al. Change in the estimated glomerular filtration rate over time and risk of all-cause mortality. Kidney Int 2013;83:684–91.
1 3 8 ( 2 0 1 8 ) 8 –1 5
Contents available at ScienceDirect
Diabetes Research and Clinical Practice journa l home page: www.e lse vier.com/locate/diabres
Discovery and validation of serum creatinine variability as novel biomarker for predicting onset of albuminuria in Type 2 diabetes mellitus Serena Low a, Xiao Zhang a, Keven Ang a, Su Jian Darren Yeo a, Guanyi Joel Lim a, Lee Ying Yeoh b, Yan Lun Liu b, Tavintharan Subramaniam c, Chee Fang Sum c, Su Chi Lim a,c,d,* a
Clinical Research Unit, Khoo Teck Puat Hospital, Singapore Department of Medicine, Khoo Teck Puat Hospital, Singapore c Diabetes Centre, Khoo Teck Puat Hospital, Singapore d Saw Swee Hock School of Public Health, National University of Singapore, Singapore b
A R T I C L E I N F O
A B S T R A C T
Article history:
Aim: We aim to study association serum creatinine(cr) variability and albuminuria progres-
Received 26 August 2017
sion.
Received in revised form
Methods: We conducted a retrospective cohort study on patients with Type 2 Diabetes Mel-
10 October 2017
litus at a Diabetes Centre in Singapore (‘‘discovery cohort”). Outcome is worsening of uri-
Accepted 7 November 2017
nary albumin-to-creatinine(ACR) across stages. Cr variability was expressed as adjusted
Available online 31 January 2018
cr-intrapersonal standard deviation(SD) and coefficient-of-variation(cr-CV). A separate cohort was used for validating association between cr variability and albuminuria progres-
Keywords: Creatinine Albuminuria Type 2 Diabetes Diabetic nephropathy
sion (‘‘validation cohort”). Results: Over median follow-up of 4.2 years, 38.4% of 636 patients had albuminuria progression in the discovery cohort. Increasing log-transformed adjusted cr-intrapersonal SD and cr-CV were significantly associated with albuminuria progression: HRs 1.43 (95%CI 1.11– 1.85) and 1.44 (1.11–1.87) respectively in the discovery cohort, and HRs 1.94 (1.09–3.45) and 1.91 (1.05–3.45) respectively in the validation cohort. When stratified by baseline urinary ACR, higher cr variability was significantly associated with albuminuria progression in patients with normoalbuminuria but not microalbuminuria. Conclusions: Cr variability independently predicts albuminuria onset. This is evident in patients with normoalbuminuria, suggesting that higher cr variability could herald albuminuria onset. Ó 2017 Published by Elsevier Ireland Ltd.
1.
Introduction
Kidney disease is a major complication of Type 2 Diabetes Mellitus (T2DM), affecting 25–40% of individuals with T2DM
[1]. Of note, Asians have a higher tendency to develop diabetic kidney disease (DKD) than other ethnicities such as Caucasians [2]. Given the substantial morbidity, mortality and economic costs imposed by DKD, it is imperative to identify
* Corresponding author at: Diabetes Centre, Khoo Teck Puat Hospital, 90 Yishun Central, 768828, Singapore. E-mail address: [email protected] (S.C. Lim). https://doi.org/10.1016/j.diabres.2017.11.003 0168-8227/Ó 2017 Published by Elsevier Ireland Ltd.
diabetes research and clinical practice
the risk factors of DKD onset and progression in order to support prevention efforts. There is accumulating literature on risk factors predictive of chronic kidney disease progression. These included age, gender, body mass index, systolic blood pressure, serum creatinine, urinary albumin-to-creatinine ratio (uACR), estimated glomerular filtration rate and even novel biomarkers [3]. Baseline serum GFR has long been recognized as a risk factor of kidney disease progression in T2DM. However, baseline GFR reveals little information on the functional reserve of the kidney. Functional renal mass can differ in different individuals even though their baseline GFR can be similar [4]. In the presence of reduced residual renal mass, more stress is exerted on the remaining nephrons. This in turn leads to hyperfiltration and kidney disease progression [4]. There is a need to look beyond baseline renal function to gain deeper insights into the progression of kidney disease. It was observed that reported patients with higher serum creatinine fluctuation during admission experienced a higher risk of mortality post discharge compared to those with more fluctuation [5]. One possible explanation was that rapid fluctuations in the GFR are reflective of acute kidney injury (AKI). Patients with AKI have a higher death risk during inpatient hospitalization and in the first year post hospital discharge [5,6]. To date, the role of stability of renal function on progression of chronic kidney disease remains under-explored. We therefore aim to study the association between serum creatinine variability over long-term and onset and progression of albuminuria in T2DM.
2.
Materials and methods
2.1.
Study population
We conducted a retrospective cohort study on patients with T2DM who enrolled at a Diabetes Centre in a regional hospital in Singapore in 2002–2014 (‘‘discovery cohort”). The exclusion criteria were: less than 21 years of age, pregnancy, active infections, active cancer, autoimmune disease, presence of other suspected causes of kidney disease (e.g. urinary tract infection, polycystic kidney disease, haematuria or history of glomerulonephritis). For the purpose of analysis, subjects were included if they had at least two years of follow-up for serum creatinine, absence of macroalbuminuria indicated by urinary albumin-to-creatinine ratio (ACR) >300 mg/g at baseline, at least two measurements of urinary ACR, four or more measurements of serum creatinine, and a follow-up period for urinary ACR of at least one month after the last serum creatinine measurement. There were altogether 636 patients eligible for the analysis. See Supplementary Fig. 1 for the flowchart.
9
size in the sitting position after a resting period of at least 10 min. One random day spot urine sample for ACR was collected for each patient at baseline and through follow-ups. Follow-up urinary ACR was collected on a yearly basis or more frequently if the ACR was high. Blood and spot urine samples were collected and sent to the hospital laboratory accredited by the College of the American Pathologists (CAP). Serum creatinine was quantitated with enzymatic colorimetric test (Roche cobasÒ c 501); Haemoglobin A1c (HbA1c) with Tinaquant Haemoglobin A1c Gen.3 (Roche cobasÒ c 501); and urinary albumin with immunoturibidmimetric assay (Roche cobasÒ c 501). Measurements were taken at recruitment and at multiple time points during follow-up. GFR was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation formula [7]. A separate cohort was also used for validating the association between serum creatinine variability and kidney disease progression. This comprised adult subjects attending the Diabetes Centre in the same regional hospital or a primary-care polyclinic in 2011–2016. The patients were included in the analysis if they had no macroalbuminuria at baseline, at least two urinary ACR measurements, at least four measurements of serum creatinine, and the last followup urinary ACR measurement taken after the last serum creatinine measurement. A total of 199 subjects met the inclusion criteria. See Supplementary Fig. 2 for the flowchart. Trained nurses measured BP using a standard sphygmomanometer using an appropriate cuff size in the sitting position after a resting period of at least 10 min. Two BP readings were taken and the average BP was derived. Clinical and laboratory measurements were taken at multiple time points till February 2016. One random day urine sample for ACR was taken at baseline and follow-ups on yearly basis or more frequently if the ACR was high. Blood and urine samples were quantitated at the same hospital laboratory using assays similar to those for the DN cohort. The study was approved by the National Healthcare Group Domain Specific Review Board in Singapore. All participants provided written informed consent.
2.3.
Outcome definition
The following stages of diabetic nephropathy were graded as follows: normoalbuminuria, microalbuminuria and macroalbuminuria indicated by uACR < 30 mg/g, 30–299 mg/g and 300 mg/g respectively. Patients who experienced worsening of these stages – from normo- to micro- or macroalbuminuria, or from micro- to macroalbuminuria were considered as having albuminuria progression (‘‘progressors”). Those who did not experience worsening of these stages were ‘‘nonprogressors”.
2.4. 2.2.
1 3 8 ( 2 0 1 8 ) 8 –1 5
Calculation of serum creatinine variability
Data collection
Clinical and demographic information were obtained from a standard questionnaire or extracted from patient’s medical records. Trained nurses measured Blood pressure (BP) using a standard sphygmomanometer using an appropriate cuff
The median number of serum creatinine measurements was 6.0 (IQR 5.0–8.0) in the discovery cohort and 5.0 (4.0–5.0) in the validation cohort. Intrapersonal mean and standard deviation (SD) of serum creatinine were calculated for each patient. Serum creatinine
10
diabetes research and clinical practice
1 3 8 ( 2 0 1 8 ) 8 –1 5
variability was expressed as adjusted cr-intrapersonal SD p using the formula: SD/ [n/(n1) where SD is the intrapersonal standard deviation of serum creatinine and n is the number of serum creatinine readings [8–10]. Furthermore, the coefficient of variation of serum creatinine variability (cr-CV) was calculated as intrapersonal SD of serum creatinine divided by intrapersonal mean (crintrapersonal mean) and expressed as a percentage [11]. CrCV was chosen as a normalized measure of serum creatinine variability so as to correct for larger SD attributed to higher absolute values of serum creatinine based on reference from past studies on HbA1c variability [12,13].
disease progression including age of entry, gender, ethnicity, duration of DM, HbA1c, SBP, HbA1c, estimated GFR, urinary ACR group and use of renin-angiotensin antagonist [3]. Assumption of proportional hazard was tested for all covariates with global test using scale Schoenfeld residuals. The assumption was not violated by the cox regression model in our analysis (p > .05). Multivariable logistic regression was used to estimate odds ratio for albuminuria progression in the validation cohort, adjusting for the same variables. All statistical tests were two-sided and considered significant if P < .05. Statistical analysis was done using STATA Version 14.0 (STATA Corporation, College Station, Texas).
2.5.
3.
Statistical analysis
Categorical variables were presented as number (percentage), and continuous variables as means (standard deviation) or median (interquartile range) as appropriate. Differences in patient characteristics stratified by progression were examined by Chi-Square test for categorical variables and student-t test or Mann-Whitney test for continuous variables. Multivariable cox proportional hazards regression model was used to estimate hazard ratios (HR) for albuminuria progression in the discovery cohort, adjusting for crintrapersonal mean and other known risk factors of kidney
Results
The overall median follow-up duration was 4.2 years (3.0–5.7) in the discovery cohort and 2.4 years (1.6–3.3) in the validation cohort. The patients were followed up for a median period of 8.4 months (IQR 3.6–12.0) and 22.8 months (IQR 13.2–32.4) after the last serum creatinine measurement in the discovery cohort and validation cohort respectively. Progression occurred in 38.4% in the discovery cohort anmd 29.6% in the validation cohort. The baseline characteristics in the discovery cohort were as follows: mean age 56.2 ± 11.5 years, 61.8% males, 70.2%
Table 1 – Baseline characteristics by progression of albuminuria in the discovery cohort. Variables
Number Age (years) Male (%) Ethnicity (%) Chinese Malay Indian DM duration (%) 10 years 11–20 years >20 years SBP 140 mmHg (%) HbA1c (%) (mmol/mol) <7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) eGFR (ml/min/1.73 m2) Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist (%) Use of Insulin (%) Cr-intrapersonal mean (umol/l) Adjusted Cr-intrapersonal SD (umol/l) Cr-coefficient of variation (%)
All
Progression
P-value
No
Yes
636 56.2 ± 11.5 393 (61.8)
392 55.2 ± 11.1 240 (61.2)
244 57.7 ± 11.9 153 (62.7)
446 (70.2) 90 (14.2) 99 (15.6)
279 (71.4) 44 (11.3) 68 (17.4)
167 (68.4) 46 (18.9) 31 (12.7)
390 (62.1) 177 (28.2) 61 (9.7) 224 (36.3)
251 (65.0) 106 (27.5) 29 (7.5) 119 (31.5)
139 (57.4) 71 (29.3) 32 (13.2) 105 (43.9)
173 (27.2)
112 (28.6)
61 (25.0)
174 (27.4)
106 (27.0)
68 (27.9)
131 (20.6)
79 (20.2)
52 (21.3)
158 (24.8)
95 (24.2)
63 (25.8)
88.0 (69.7–101.9)
91.6 (77.0–103.2)
78.9 (57.0–98.4)
350 (55.0) 286 (45.0) 363 (57.4) 191 (31.6) 79.4 (64.3–101.4) 6.7 (4.5–10.2) 9.6 (6.8–13.6)
253 (64.5) 139 (35.5) 188 (48.3) 110 (29.7) 75.4 (60.7–90.7) 6.2 (4.2–8.8) 8.9 (6.5–12.7)
97 (39.8) 147 (60.3) 175 (71.7) 81 (34.8) 89.2 (69.1–117.1) 8.6 (5.3–13.5) 10.8 (7.4–15.0)
0.008 0.709 0.015
0.038
0.002 0.803
<0.001 <0.001
<0.001 0.188 <0.001 <0.001 <0.001
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-tocreatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine.
diabetes research and clinical practice
Chinese, 14.2% Malay, 15.6% Indian, 36.3% SBP 140 mmHg, 45.4% HbA1c 8.0% (63.9 mmol/mol), 55.0% normoalbuminuria, 45.0% microalbuminuria, 57.4% used RAS antagonist, 31.6% used insulin, median adjusted cr-SD 6.7 (IQR 4.5–10.2), cr-CV 9.6% (6.8–13.6). There were no statistically significant differences between the discovery and validation cohorts in terms of age at entry, duration of DM and urinary ACR category. See Table 1 and Supplementary Table 1. In the discovery cohort, the progressors had poorer baseline clinical profile in terms of SBP, eGFR and urinary ACR (p < .05). They also had higher cr-intrapersonal mean, crintrapersonal SD and cr-CV (p < .001) than the nonprogressors. Similarly, the cr-intrapersonal SD and cr-CV were higher in progressors than non-progressors in the validation cohort (p < .05). See Table 1 and Supplementary Table 1. In the multivariable model which included only age, gender and ethnicity, increasing log-transformed adjusted crintrapersonal SD and log-transformed cr-CV were significantly associated with progression of albuminuria: adjusted hazards ratios (HRs) 1.45 (95% Confidence Interval (CI), 1.15– 1.85; and 1.45 (95%CI, 1.15–1.83) respectively in the discovery cohort, and HRs 1.71 (95%CI, 1.03–2.85) and 1.69 (95%CI, 1.01–2.82) respectively in the validation cohort. See Tables 2 and 3, and Supplementary Tables 2 and 3. After additional adjustment for duration of DM, SBP, HbA1c, eGFR, urinary ACR group, use of RAS antagonist and cr-intrapersonal mean, increasing log-transformed adjusted cr-intrapersonal SD and log-transformed cr-CV remained significantly associated with progression of albuminuria without much attenuation: HRs 1.43 (95%CI, 1.11–1.85) and 1.44 (95% CI, 1.11–1.87) respectively in the discovery cohorts, and HRs 1.94 (95%CI, 1.09–3.45) and 1.91 (95%CI, 1.05–3.45) respectively in the validation cohort. See Tables 2 and 3, and Supplementary Tables 2 and 3. In the multivariable models stratified by urinary ACR group in the discovery cohort, the association between increasing log-transformed adjusted cr-intrapersonal SD and log-transformed cr-CV with progression were strengthened in patients with normoalbuminuria with corresponding HRs 1.59 (95%CI, 1.06–2.38) and 1.61 (95%CI, 1.07–2.42). However, the associations were largely attenuated in patients with microalbuminuria, with corresponding HRs 1.33 (95%CI, 0.96–1.85) and 1.34(95%CI, 0.96–1.86). Similar findings were noted in the validation cohort. Although increasing logtransformed cr-intrapersonal mean was found associated with progression of albuminuria when it was included in the multivariable model relating cr-CV to outcome in the discovery cohort (Table 3), there was no association in the other multivariable models. See Table 2, and Supplementary Tables 2 and 3.
4.
Discussion
We have demonstrated that higher serum creatinine variability was associated with elevated risk of onset of albuminuria. This finding was consistent in two different cohorts with T2DM in which serum creatinine variability was expressed as adjusted cr-intrapersonal SD and logtransformed cr-CV.
1 3 8 ( 2 0 1 8 ) 8 –1 5
11
Research has shown that serum creatinine variability predicted mortality in inpatients [5]. Studies have also have investigated the relationship between eGFR variability and ESRD. [14,15] However, to the best of our knowledge, the role of serum creatinine variability has not been explored in the early phase of DKD (i.e. progression from normoalbuminuria to higher grade albuminuria). Albuminuria is present in many individuals who have normal eGFR. It has been linked to cardiovascular and renal outcomes independently of eGFR at baseline. There was suggestion that albuminuria reflects glomerular and tubular injury, thereby contributing to improvement in prediction of renal prognosis in addition to eGFR which only reflects glomerular injury [16]. Our study, for the first time, revealed the putative role of serum creatinine variability on onset of albuminuria. This corroborates previous findings on the association between variability or short-term alterations in eGFR, which is derived from serum creatinine, and end-stage renal disease or mortality [14,15,17,18]. Furthermore, there is an accumulating body of evidence showing the effect of variability in terms of HbA1c, BP, lipids and uric acid on kidney disease progression in recent years [10,19–23]. It is known that variability plays a vital role in maintenance of homeostasis in the body. It has also been shown that maintaining certain biological parameters in a restricted range confers protective effect in clinical outcome [19]. This supports the observation that higher HbA1c variability is related to deterioration of kidney function in our previous study and other studies [10,19–22]. Our current study adds serum creatinine to the current limited pool of biological markers known to affect outcomes in kidney disease in T2DM through fluctuations in their levels. The exact mechanisms of how serum creatinine variability putatively contributed to onset of albuminuria are not clear. There are a few possible mechanisms. Firstly, serum creatinine variability may be reflective of maladaption to hemodynamic changes in the background of reduced nephron mass [18,24]. It is known that haemodynamic pathway contributes to the pathophysiology of diabetes kidney disease through imbalance in resistance of afferent and efferent arterioles which then leads to elevated hydrostatic pressure in the glomerulus [25]. Elevations in renal plasma flow, intraglomerular pressure, and glomerular hyperfiltration and hypertrophy contribute to structural and functional abnormalities in the reduced renal mass [24]. Secondly, acute kidney injury (AKI) involves modest and possibly reversible elevation of serum creatinine levels [18,26,27]. It has been suggested that AKI predisposes individuals at higher risk of CKD by causing renal fibrosis [28–30]. Interestingly, the relationship between serum creatinine variability and progression of albuminuria was only evident in the presence of normoalbuminuria but not microalbuminuria at baseline. Serum creatinine variability may be an early marker of albuminuria progression. It is noted that many individuals with T2DM and normoalbuminuria experienced deterioration in kidney function i.e. decline in eGFR [31]. One possible reason is use of RAS anatgonists which may have controlled proteinuria and delayed but may not have prevented the continued deterioration of renal function [31]. Further studies are necessary to understand the putative role
Variable
Crude HR (95% CI) p-value
Model 1a HR (95% CI) p-value Model 2b HR (95% CI) p-value Model 3c HR (95% CI) p-value Model 4c HR (95% CI) p-value Normoalbuminuria
Microalbuminuria
1.01 (0.98–1.03) 0.517
0.99 (0.97–1.01) 0.424
1.00 (0.98–1.01) 0.517
Gender Female Male
1.00 0.91 (0.70–1.18) 0.480
1.00 0.71 (0.53–0.95) 0.022
1.00 0.79 (0.54–1.16) 0.233
Ethnicity Chinese Malay Indian
1.00 1.57 (1.13–2.17) 0.007 0.90 (0.61–1.31) 0.574
1.00 1.07 (0.75–1.53) 0.691 1.00 (0.68–1.48) 0.989
1.00 0.83 (0.56–1.22) 0.333 0.92 (0.61–1.38) 0.689
DM duration 10 years 11–20 years >20 years SBP 140 mmHg
1.00 1.32 (0.99–1.76) 0.058 2.15 (1.46–3.16) <0.001 1.63 (1.26–2.11) <0.001
1.00 0.90 (0.65–1.24) 0.531 1.47 (0.95–2.29) 0.085 1.38 (1.04–1.84) 0.025
1.52 (0.93–2.49) 0.098
1.45 (1.01–2.08) 0.042
1.00
1.00
1.00
1.00
1.10 (0.78–1.55) 0.601
1.25 (0.85–1.84) 0.250
0.83 (0.46–1.50) 0.546
1.81 (1.07–3.05) 0.026
1.02 (0.70–1.48) 0.916
1.36 (0.90–2.07) 0.148
1.04 (0.56–1.95) 0.898
1.70 (0.99–2.91) 0.056
1.27 (0.89–1.81) 0.182
1.50 (0.97–2.31) 0.066
1.07 (0.55–2.07) 0.850
1.66 (0.94–2.94) 0.080
0.27 (0.20–0.37) <0.001
0.86 (0.38–1.97) 0.727
0.62 (0.17–2.29) 0.472
0.70 (0.31–1.62) 0.407
1.00 2.10 2.02 1.40 4.82
1.00 1.45 1.46 0.97 2.48
0.016 0.023 0.860 0.064
1.60 (1.00–2.57) 0.050 1.35 (0.7902.32) 0.271 2.52 (0.68–9.38) 0.168
1.03 (0.66–1.60) 0.903 0.95 (0.62–1.43) 0.791 2.07 (0.79–5.43) 0.140
1.43 (1.11–1.85) 0.006
1.59 (1.06–2.38) 0.024
1.33 (0.96–1.85) 0.089
Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist Use of insulin Log-transformed Crintrapersonal mean (per umol/l) Log-transformed Adjusted Crintrapersonal SD (per umol/l)
(1.62–2.72) (1.53–2.68) (1.07–1.83) (3.30–7.04)
<0.001 <0.001 0.016 <0.001 3.16 (1.89–5.27) <0.001
2.08 (1.73–2.49) <0.001 1.45 (1.15–1.82) 0.002
(1.07–1.97) (1.05–2.01) (0.69–1.36) (0.95–6.47)
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-to-creatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine a Model 1 is adjusted for age at entry, gender and ethnicity. b Model 2 is adjusted for age at entry, gender, ethnicity, duration of DM, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, Urinary ACR group, use of RAS inhibitor(s) and use of insulin. c Models 3 and 4 are adjusted for age at entry, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, use of RAS inhibitor(s) and use of insulin.
1 3 8 ( 2 0 1 8 ) 8 –1 5
1.02 (1.01–1.03) <0.001 1.00 (0.99–1.01) 0.683
diabetes research and clinical practice
Age (per year)
HbA1c (per%) < 7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) Log-transformed eGFR (per ml/min/1.73 m2)
12
Table 2 – Cox Regression Models Relating Adjusted Creatinine-Intrapersonal Standard Deviation to Progression of Albuminuria in the Discovery Cohort.
Table 3 – Cox Regression Models Relating to Creatinine Coefficient of Variation to Progression of Albuminuria in the Discovery Cohort. Variable
Crude HR (95% CI) p-value
Model 5a HR (95% CI) p-value Model 6b HR (95% CI) p-value Model 7c HR (95% CI) p-value Model 8c HR (95% CI) p-value Normoalbuminuria
Microalbuminuria
1.01 (0.98–1.03) 0.525
0.99 (0.97–1.01) 0.416
1.00 (1.48–1.01) 0.504
Gender Female Male
1.00 0.91 (0.70–1.18) 0.480
1.00 0.71 (0.53–0.95) 0.022
1.00 0.80 (0.54–1.17) 0.241
Ethnicity Chinese Malay Indian
1.00 1.57 (1.13–2.17) 0.007 0.90 (0.61–1.31) 0.574
1.00 1.07 (0.75–1.53) 0.690 1.00 (0.68–1.48) 0.986
1.00 0.82 (0.56–1.21) 0.327 0.92 (0.61–1.38) 0.690
DM duration 10 years 11–20 years >20 years SBP 140 mmHg
1.00 1.32 (0.99–1.76) 0.058 2.15 (1.46–3.16) <0.001 1.63 (1.26–2.11) <0.001
1.00 0.91 (0.65–1.24) 0.529 1.48 (0.95–2.29) 0.083 1.38 (1.04–1.84) 0.025
1.52 (0.93–2.49) 0.097
1.45 (1.01–2.08) 0.042
1.00
1.00
1.00
1.00
1.10 (0.78–1.55) 0.601
1.25 (0.86–1.84) 0.246
0.84 (0.46–1.50) 0.548
1.81 (1.07–3.05) 0.026
1.02 (0.70–1.48) 0.916
1.36 (0.89–2.07) 0.150
1.04 (0.55–1.95) 0.909
1.69 (0.99–2.91) 0.056
1.27 (0.89–1.81) 0.182
1.51 (0.98–2.32) 0.063
1.06 (0.55–2.07) 0.854
1.67 (0.94–2.95) 0.078
0.27 (0.20–0.37) <0.001
0.86 (0.38–1.96) 0.718
0.63 (0.17–2.32) 0.485
0.70 (0.30–1.61) 0.401
1.00 2.10 2.02 1.40 4.82
1.00 1.45 1.46 0.97 3.54
1.60 (1.00–2.57) 0.051 1.36 (0.79–2.32) 0.268 4.10 (1.11–15.18) 0.035
1.03 (0.66–1.60) 0.901 0.94 (0.62–1.43) 0.789 2.75 (1.19–6.37) 0.018
1.61 (1.07–2.42) 0.023
1.34 (0.96–1.86) 0.088
HbA1c (per%) < 7.0% (53.0 mmol/mol) 7.1–7.9% (54.1–62.8 mmol/mol) 8.0–8.9% (63.9–73.8 mmol/mol) 9.0% (74.9 mmol/mol) Log-transformed eGFR (per ml/min/1.73 m2) Urinary ACR group Normoalbuminuria Microalbuminuria Use of RAS-antagonist Use of insulin Log-transformed Crintrapersonal mean (per umol/l) Log-transformed cr coefficient of variation (per%)
(1.62–2.72) (1.53–2.68) (1.07–1.83) (3.30–7.04)
<0.001 <0.001 0.016 <0.001 4.60 (2.98–7.08) <0.001
1.65 (1.31–2.07) <0.001 1.45 (1.1501.83) 0.002
(1.07–1.97) (1.05–2.02) (0.69–1.36) (1.48–8.48)
0.016 0.023 0.854 0.005
1.44 (1.11–1.87) 0.006
13
DM, diabetes mellitus; SBP, systolic blood pressure; HbA1c, haemoglobin A1c; eGFR, estimated glomerular filtration rate; ACR, albumin-to-creatinine ratio; RAS-antagonist, renin-angiotensin system antagonist; SD, standard deviation; cr, creatinine a Model 5 is adjusted for age at entry, gender and ethnicity b Model 6 is adjusted for age at entry, gender, ethnicity, duration of DM, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, Urinary ACR group, use of RAS inhibitor(s) and use of insulin. c Models 7 and 8 are adjusted for age at entry, SBP 140 mmHg, HbA1c group, log-transformed baseline eGFR, use of RAS inhibitor(s) and use of insulin.
1 3 8 ( 2 0 1 8 ) 8 –1 5
1.02 (1.01–1.03) <0.001 1.00 (0.99–1.01) 0.695
diabetes research and clinical practice
Age (per year)
14
diabetes research and clinical practice
of serum creatinine variability in the presence of normal urinary ACR. The unique features of our study include (1) availability of two separate longitudinal cohorts to establish the relationship between serum creatinine variability and albuminuria progression; (2) use of two measures of variability to consistently demonstrate association between serum creatinine variability and albuminuria progression; (3) availability of rich clinical information, including medications such as RAS and insulin, which could have altered the course of albuminuria progression; (4) sampling of patients from healthcare facilities which provided insights into the ‘‘real-world” relevant to healthcare providers and policy makers; and (5) availability of the follow-up serum creatinine values in the medical records for both inpatient and outpatient settings, even though the study was conducted in the diabetes centre. Such information allowed us to take into account the greater creatinine variability from more creatinine measurements in some patients with acute illness, especially the hospitalized patients, compared to those with a relatively stable disease. However, there are limitations in this study. Firstly, our study was confined to patients attending diabetes centre in a regional hospital. Hence the findings could not be generalized to the entire diabetes population. Secondly, urinary ACR was measured once at baseline and each follow-up using spot urine samples. We were unable to confirm the transition of albuminuria with a second measurement. Hence the outcome measure was prone to variability. However, it was reported that spot urine protein-creatinine ratio correlated well with 24 h urinary protein in patients with T2DM, suggesting that spot ACR measurement could be utilized as a reliable and fast alternative for assessing albuminuria [32,33]. Another limitation was that serum creatinine fluctuations may be attributed to changes in muscle mass, acute sickness, generation of creatinine in the muscle and malnutrition [5,34,35]. It may be necessary to exclude these factors when studying the impact of the serum creatinine variability. In addition, there could still be residual confounding due to socio-economic status and behavioural factors which may influence the progression of albuminuria. Nevertheless, our findings may have potential clinical implications. The measures of serum creatinine variability could be used in the routine clinical practice by incorporating them in the electronic clinical system. These measures would help clinicians to track the changes in serum creatinine and institute appropriate management to control the fluctuations. There is potential for therapeutic interventions to be developed to target at fluctuations of serum creatinine or to avoid stressors leading to fluctuating renal function e.g. contrast nephropathy, nonsteroidal anti-inflammatory drugs.
5.
Conclusions
In conclusion, higher serum creatinine variability is an independent predictor of albuminuria onset. The relationship is most evident in patients with normoalbuminuria, suggesting that higher serum creatinine variability could herald albuminuria of onset.
1 3 8 ( 2 0 1 8 ) 8 –1 5
Conflict of interest statement The authors declare that they have no conflict of interest.
Acknowledgements This study was supported by the Alexandra Health Grant (AHPL SIGII/11001; SIG/11029 and SIG/12024), National Medical Research Council Grant (NMRC/PPG/AH(KTPH)/2011) and National Medical Research Council Grant (NMRC/ CIRG/1398/2014).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.diabres.2017. 11.003.
R E F E R E N C E S
[1] Remuzzi G, Schieppati A, Ruggenenti P. Clinical practice. Nephropathy in patients with type 2 diabetes. N Engl J Med 2002;346:1145–51. [2] Kong AP, Xu G, Brown N, So WY, Ma RC, Chan JC. Diabetes and its comorbidities–where east meets west. Nat Rev Endocrinol 2013;9:537–47. [3] Echouffo-Tcheugui JB, Kengne AP. Risk models to predict chronic kidney disease and its progression: a systematic review. PLoS Med 2012;9:e1001344. [4] Ronco C, Rosner MH. Acute kidney injury and residual renal function. Crit Care 2012;16:144. [5] Kao SS, Kim SW, Horwood CM, Hakendorf P, Li JY, Thompson CH. Variability in inpatient serum creatinine: its impact upon short- and long-term mortality. QJM 2015;108:781–7. [6] McAlister FA, Ezekowitz J, Tonelli M, Armstrong PW. Renal insufficiency and heart failure: prognostic and therapeutic implications from a prospective cohort study. Circulation 2004;109:1004–9. [7] Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro 3rd AF, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–12. [8] Kilpatrick ES, Rigby AS, Atkin SL. A1c variability and the risk of microvascular complications in type 1 diabetes: data from the diabetes control and complications trial. Diabetes Care 2008;31:2198–202. [9] Marcovecchio ML, Dalton RN, Chiarelli F, Dunger DB. A1c variability as an independent risk factor for microalbuminuria in young people with type 1 diabetes. Diabetes Care 2011;34:1011–3. [10] Penno G, Solini A, Bonora E, Fondelli C, Orsi E, Zerbini G, et al. Hba1c variability as an independent correlate of nephropathy, but not retinopathy, in patients with type 2 diabetes: The renal insufficiency and cardiovascular events (riace) italian multicenter study. Diabetes Care 2013;36:2301–10. [11] Mosca A, Lapolla A, Gillery P. Glycemic control in the clinical management of diabetic patients. Clin Chem Lab Med 2013;51:753–66. [12] Rodriguez-Segade S, Rodriguez J, Garcia Lopez JM, Casanueva FF, Camina F. Intrapersonal hba(1c) variability and the risk of progression of nephropathy in patients with type 2 diabetes. Diabet Med 2012;29:1562–6.
diabetes research and clinical practice
[13] Waden J, Forsblom C, Thorn LM, Gordin D, Saraheimo M, Groop PH, et al. A1c variability predicts incident cardiovascular events, microalbuminuria, and overt diabetic nephropathy in patients with type 1 diabetes. Diabetes 2009;58:2649–55. [14] Tseng CL, Lafrance JP, Lu SE, Soroka O, Miller DR, Maney M, et al. Variability in estimated glomerular filtration rate values is a risk factor in chronic kidney disease progression among patients with diabetes. BMC Nephrol 2015;16:34. [15] Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, et al. Short-term change in kidney function and risk of end-stage renal disease. Nephrol Dial Transplant 2012;27:3835–43. [16] Gansevoort RT, Nauta FL, Bakker SJ. Albuminuria: all you need to predict outcomes in chronic kidney disease? Curr Opin Nephrol Hypertens 2010;19:513–8. [17] Perkins RM, Tang X, Bengier AC, Kirchner HL, Bucaloiu ID. Variability in estimated glomerular filtration rate is an independent risk factor for death among patients with stage 3 chronic kidney disease. Kidney Int 2012;82:1332–8. [18] Uehara K, Yasuda T, Shibagaki Y, Kimura K. Estimated glomerular filtration rate variability independently predicts renal prognosis in advanced chronic kidney disease patients. Nephron 2015;130:256–62. [19] Ceriello A, De Cosmo S, Rossi MC, Lucisano G, Genovese S, Pontremoli R, et al. Variability in hba1c, blood pressure, lipid parameters and serum uric acid, and risk of development of chronic kidney disease in type 2 diabetes. Diabetes Obes Metab 2017. [20] Low S, Lim SC, Yeoh LY, Liu YL, Liu JJ, Fun S, et al. Effect of long-term glycemic variability on estimated glomerular filtration rate decline among patients with type 2 diabetes mellitus: Insights from the diabetic nephropathy cohort in singapore. J Diabetes 2016. [21] Sugawara A, Kawai K, Motohashi S, Saito K, Kodama S, Yachi Y, et al. Hba(1c) variability and the development of microalbuminuria in type 2 diabetes: Tsukuba kawai diabetes registry 2. Diabetologia 2012;55:2128–31. [22] Tsuboi A, Takenouchi A, Kurata M, Fukuo K, Kazumi T. Postmeal triglyceridemia and variability of hba1c and postmeal glycemia were predictors of annual decline in estimated glomerular filtration rate in type 2 diabetic patients with different stages of nephropathy. J Diabetes Metab Disord 2017;16:1. [23] Yeh CH, Yu HC, Huang TY, Huang PF, Wang YC, Chen TP, et al. The risk of diabetic renal function impairment in the first
[24]
[25] [26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
1 3 8 ( 2 0 1 8 ) 8 –1 5
15
decade after diagnosed of diabetes mellitus is correlated with high variability of visit-to-visit systolic and diastolic blood pressure: a case control study. BMC Nephrol 2017;18:99. Gluhovschi G, Gadalean F, Gluhovschi C, Petrica L, Velciov S, Gluhovschi A, et al. The solitary kidney–a nephrological perspective. Rom J Intern Med 2013;51:80–8. Lim A. Diabetic nephropathy – complications and treatment. Int J Nephrol Renovasc Dis 2014;7:361–81. Bucaloiu ID, Kirchner HL, Norfolk ER, Hartle 2nd JE, Perkins RM. Increased risk of death and de novo chronic kidney disease following reversible acute kidney injury. Kidney Int 2012;81:477–85. Goldenberg I, Chonchol M, Guetta V. Reversible acute kidney injury following contrast exposure and the risk of long-term mortality. Am J Nephrol 2009;29:136–44. Basile DP. Rarefaction of peritubular capillaries following ischemic acute renal failure: a potential factor predisposing to progressive nephropathy. Curr Opin Nephrol Hypertens 2004;13:1–7. Basile DP, Donohoe D, Roethe K, Osborn JL. Renal ischemic injury results in permanent damage to peritubular capillaries and influences long-term function. Am J Physiol Renal Physiol 2001;281:F887–99. Coca SG, Singanamala S, Parikh CR. Chronic kidney disease after acute kidney injury: a systematic review and metaanalysis. Kidney Int 2012;81:442–8. Dwyer JP, Parving HH, Hunsicker LG, Ravid M, Remuzzi G, Lewis JB. Renal dysfunction in the presence of normoalbuminuria in type 2 diabetes: results from the demand study. Cardiorenal Med 2012;2:1–10. Biradar SB, Kallaganad GS, Rangappa M, Kashinakunti SV, Retnakaran R. Correlation of spot urine protein-creatinine ratio with 24-hour urinary protein in type 2 diabetes mellitus patients: a cross sectional study. J Res Med Sci 2011;16:634–9. Zhang X, Low S, Sum CF, Tavintharan S, Yeoh LY, Liu J, et al. Arterial stiffness is an independent predictor for albuminuria progression among asians with type 2 diabetes-a prospective cohort study. J Diabetes Complications 2017;31:933–8. Kovesdy CP, George SM, Anderson JE, Kalantar-Zadeh K. Outcome predictability of biomarkers of protein-energy wasting and inflammation in moderate and advanced chronic kidney disease. Am J Clin Nutr 2009;90:407–14. Turin TC, Coresh J, Tonelli M, Stevens PE, de Jong PE, Farmer CK, et al. Change in the estimated glomerular filtration rate over time and risk of all-cause mortality. Kidney Int 2013;83:684–91.