Associations Between Continuous Glucose Monitoring-derived Metrics and Diabetic Retinopathy and Albuminuria in Patients With Type 2 Diabetes

doi:10.21203/rs.3.rs-72340/v1

Download PDF

Original investigation

Associations Between Continuous Glucose Monitoring-derived Metrics and Diabetic Retinopathy and Albuminuria in Patients With Type 2 Diabetes

https://doi.org/10.21203/rs.3.rs-72340/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Apr, 2021

Read the published version in BMJ Open Diabetes Research & Care →

Version 1

posted

You are reading this latest preprint version

Background

Preventing the development and progression of diabetic microvascular complications through optimal blood glucose control remains an important challenge. Whether metrics based on continuous glucose monitoring are useful for the management of diabetic microvascular complications is not entirely clear.

Methods

This is an exploratory analysis of an ongoing prospective, multicenter, 5-year follow-up observational study. Study participants included 999 outpatients with type 2 diabetes who underwent continuous glucose monitoring at baseline. Associations between continuous glucose monitoring-derived metrics and the severity of diabetic retinopathy or albuminuria were investigated using multivariable proportional odds models.

Results

The overall prevalence of diabetic retinopathy was 22.2%. Multivariate analysis with proportional odds models demonstrated that continuous glucose monitoring-derived metrics related to intra-day and inter-day glucose variability are significantly associated with the severity of diabetic retinopathy, even after adjusting for various possible risk factors. However, significant relationships were not observed after adjusting for HbA1c levels. The prevalence of microalbuminuria and macroalbuminuria were 20.3% and 6.7%, respectively. Similarly, multivariate analysis demonstrated that those metrics are significantly associated with the severity of albuminuria. These relationships remained significant even after further adjusting for HbA1c levels.

Conclusions

Continuous glucose monitoring-derived metrics related to intra-day and inter-day glucose variability are significantly associated with the severity of diabetic retinopathy or albuminuria in patients with type 2 diabetes. Thus, evaluating these metrics might possibly be useful for risk assessment of diabetic microvascular complications.

Cardiac & Cardiovascular Systems

glucose variability

continuous glucose monitoring

diabetic retinopathy

albuminuria

type 2 diabetes mellitus

Diabetic retinopathy (DR) is a main cause of visual impairment and blindness [1]. Diabetic nephropathy (DN) is the main cause of end-stage renal disease [2]. Accordingly, taking preventive measures against the development and progression of diabetic microvascular complications in patients with type 2 diabetes is an important task necessary for maintaining daily quality of life, extending healthy lifespan, and reducing healthcare costs. Given that the main cause of microvascular complications is damage to tissues and organs caused by persistent hyperglycemia [3], optimal glycemic control is the best way to prevent the development and progression of microvascular complications.

HbA1c is recognized as a gold standard for assessment of glycemic management. Several studies have demonstrated strong associations between HbA1c levels and diabetic microvascular complications [4, 5]. In addition, previous studies have indicated that improvement of HbA1c levels is associated with risk reduction in the incidence and progression of diabetic complications in patients with type 2 diabetes [6, 7]. Based on those data, current guidelines recommend a target HbA1c level of 7% or less [8, 9]. On the other hand, another study failed to show glycemic control to HbA1c < 7% has a beneficial effect on the prevention of microvascular complications [10]. One possible explanation for the discrepant findings may be that frequent episodes of severe hypoglycemia counterbalanced the beneficial effects of glycemic control [11]. This hypothesis is based on the fact that HbA1c reflects average glucose over the last few months, but it provides no information on intra-day and inter-day glucose variability and hypoglycemia, both of which may play an important role in the development of macrovascular and microvascular complications [11–13]. In addition, HbA1c has another limitation. HbA1c is affected by factors such as anemia, hemoglobinopathy, chronic kidney disease, and ethnicity [14]. Thus, new metrics reflecting various aspects of glycemic status are needed for the management of diabetic microvascular complications.

Continuous glucose monitoring (CGM) has emerged as an optimal method to obtain a more comprehensive glycemic profile, including data on intra-day and inter-day glucose variability and patterns of hyperglycemia and hyperglycemia. In particular, the Advanced Technologies & Treatments for Diabetes (ATTD) Congress recommend using 10 core CGM metrics that may be most useful in clinical practice [15]. The CGM metrics includes three key CGM measurements: 1) time in range (TIR), defined as the percentage of the time spent within the target glucose range; 2) time below range (TBR), and 3) time above range (TAR) [15]. These new metrics assessed with CGM could help improve clinical management by providing more information than HbA1c.

Nevertheless, until now, only limited data from cross-sectional studies investigating the relationship between CGM metrics and diabetic microvascular complications have been available. Intriguingly, two recent cross-sectional studies conducted by the same group demonstrated that standard deviation (SD) and TIR are each significantly associated with the presence of DR in inpatients with type 2 diabetes [16, 17]. Another study showed that TIR is associated with the presence of albuminuria in patients with type 2 diabetes, but this relationship did not reach statistical significance after adjusting for HbA1c levels [18]. On the other hand, a small retrospective cross-sectional study demonstrated that TAR is associated with the presence of DR, but other metrics of glucose variability was not associated with the presence of DR or DN in inpatients with type 2 diabetes [19]. Thus, the association between metrics from CGM and the presence or severity of diabetic microvascular complications in patients with type 2 diabetes has not been fully elucidated yet.

In this exploratory cross-sectional study, we investigated the relationship between CGM-derived metrics and the severity of DR and albuminuria in 999 outpatients with type 2 diabetes.

Study design

This study is an exploratory sub-analysis of an ongoing, observational, prospective cohort study that aims to investigate the relationships between glucose fluctuations evaluated with CGM and the incidence of composite cardiovascular events over a 5-year follow-up period as described previously [20]. This study used baseline study from the cohort study. This study has been registered in the University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR), which is a non-profit organization in Japan that meets the requirements of the International Committee of Medical Journal Editors (UMIN000032325).

Study Population

The study population consists of Japanese patients with type 2 diabetes who regularly attend the outpatient diabetes clinics of 34 institutions across Japan (with investigator names in parentheses) shown in supplementary Table 1. The inclusion criteria were as follows: 1) age ≥ 30 years and ≤ 80 years, regardless of gender; 2) receiving treatment for type 2 diabetes at one of the participating outpatient clinics; 3) informed consent for study participation; 4) no changes (including new prescriptions) in antidiabetic medications for 6 months before written informed consent was obtained (insulin dosage changes were allowed); and 5) no anticipated changes in antidiabetic medications from the time of enrolment until a CGM device was applied on the back of the upper arm (insulin dosage changes were allowed). The following exclusion criteria were also applied: 1) type 1 or secondary diabetes; 2) presence of severe infectious disease preoperatively, postoperatively, or associated with severe trauma; 3) history of myocardial infarction, angina pectoris, cerebral stroke, cerebral infarction, or arteriosclerosis obliterans; 4) current treatment with artificial dialysis; 5) moderate liver dysfunction defined as aspartate aminotransferase ≥ 100 IU/L; 6) moderate or severe heart failure (New York Heart Association stage III or worse); 7) pregnancy, lactation, possible pregnancy, or plans to become pregnant during the study period; 8) present or past history of a malignant tumor; 9) use of a sensor-augmented insulin pump; 10) type 2 diabetes diagnosis within the past year; and 11) judged as ineligible by the clinical investigators. Patients not currently receiving medication for a malignant tumor, with no disease recurrence to date, and without recurrence risks during the study period were allowed to participate.

Consecutive subjects were screened. Patients who meet the eligibility criteria were asked to participate in the present study. A total of 1,000 patients who met the eligibility criteria were recruited between May 2018 and March 2019. One patient withdrew consent. The protocol was approved by the institutional review board of each participating institution in compliance with the Declaration of Helsinki and current legal regulations in Japan. Written informed consent was obtained from all participants after a full explanation of the study.

Biochemical Tests

Blood samples were obtained at visits after overnight fasting. Renal function tests, lipid levels, and HbA1c (National Glycohemoglobin Standardization Program) were measured with standard techniques. Urinary albumin excretion (UAE) was measured using a latex agglutination assay on a spot urine sample. The estimated glomerular filtration rate (eGFR) was calculated using a formula [21].

Dr And Dn Assessment

The presence and severity of DR were determined by trained ophthalmologists. The patients were grouped into four groups based on medical records: no diabetic retinopathy (NDR), simple diabetic retinopathy (SDR), pre-proliferative diabetic retinopathy (PPDR), or proliferative diabetic retinopathy (PDR). DN was defined according to the level of UAE: values < 30 mg/g creatinine were defined as normoalbuminuria, 30–300 mg/g creatinine was defined as microalbuminuria, and values ≥ 300 mg/g creatinine was defined as macroalbuminuria.

Cgm With The Freestyle Libre Pro Device

The FreeStyle Libre Pro (FLP) (Abbott Japan, Tokyo, Japan) CGM (FLP-CGM) device, which measures glucose levels every 15 minutes for up to 14 days, was used in this study as previously reported [20]. Other than wearing FLP-CGM, there were no restrictions on participants’ daily lives. Downloaded data sets were further analyzed. Glucose variability was assessed based on mean amplitude of glycemic excursion (MAGE) [22], SD, and glucose coefficient of variation (CV). MAGE was calculated as the arithmetic mean of the differences between consecutive peaks and nadirs, provided that the differences are greater than one SD of the mean glucose value. CV (%) was calculated by dividing SD by the mean of the corresponding glucose readings. The original statistical analysis plan (SAP) for this study was reported in the initial study protocol [20]. We added some CGM-derived metrics in this study since the ATTD Congress proposed some CGM-derived metrics as useful clinical targets that complement HbA1c [15]. Thus, we updated the SAP prior to database lock. Mean glucose was measured from data collected during FLP-CGM. TIR was defined as the percentage of the time spent in the target range between 3.9 and 10.0 mmol/L (TIR^{3.9–10 mmol/L}), time above target glucose range (TAR^{> 10 mmol/L}, TAR^{> 13.9 mmol/L}), and time below target glucose range (TBR^{< 3.9 mmol/L}, TBR^{< 3.0 mmol/L}). Low blood glucose index (LBGI) and high blood glucose index (HBGI) formulae were implemented by converting glucose values into risk scores [23]. In addition, mean of daily differences (MODD) [24] in glucose levels and interquartile range (IQR) were calculated to assess inter-day glucose variability. MODD was calculated as the mean of the absolute difference between glucose levels measured at the same time on 2 consecutive days. IQR was calculated using values from sthe same time of day during the monitoring period. Since a previous study demonstrated that FLP-CGM was less accurate during the first 24 hours (from the first day to the second day) after insertion and during the last four days of its 14-day lifetime [25], we analyzed FLP-CGM data over the middle 8-day period.

Statistical Analysis

Results are presented as means ± SD for continuous variables or number (proportion) of patients for categorical variables. Several parameters were logarithmically transformed to approximate the normal distribution. Continuous data were compared using analysis of variance and categorical data were compared using the chi-square test or Fisher's exact test as appropriate. Multivariate analysis with proportional odds models was performed to investigate whether FLP-CGM–derived metrics are associated with the severity of diabetic microvascular complications. Conventional possible risk factors evaluated by clinical, biochemical, metabolic tests based on clinical judgment were included in the models. All statistical tests were two-sided with a 5% significance level. All analyses were performed using SAS software version 9.4 or above (SAS Institute, Cary, NC).

Relationship Between FLP-CGM–derived Metrics and DR severity

The baseline clinical characteristics of the 999 patients with type 2 diabetes are summarized in Table 1. The mean age was 64.6 ± 9.6 years, 60.9% were male, mean HbA1c was 7.1 ± 0.8% (53.7 ± 8.8 mmol/mol), and estimated duration of type 2 diabetes was 12.9 ± 8.5 years.

Table 1

Patient demographic and background characteristics
Parameter
Age (years)	64.6 ± 9.6 (n = 999)
Male gender (%)	608 (60.9)
BMI (kg/m²)	24.6 ± 3.9 (n = 999)
Estimated duration of diabetes (years)	12.9 ± 8.5 (n = 999)
Systolic blood pressure (mmHg)	131.2 ± 14.8 (n = 999)
Diastolic blood pressure (mmHg)	75.5 ± 11.0 (n = 999)
HbA1c (%)	7.1 ± 0.8 (n = 999)
HbA1c (mmol/mol)	53.7 ± 8.8 (n = 999)
Total cholesterol (mmol/L)	4.81 ± 0.82 (n = 964)
LDL cholesterol (mmol/L)	2.67 ± 0.69 (n = 990)
HDL cholesterol (mmol/L)	1.56 ± 0.41 (n = 998)
Triglycerides (mmol/L)	1.4 ± 0.9(n = 999)
Uric acid (µmol/L)	307.4 ± 73.0 (n = 994)
Estimated glomerular filtration rate (mL/min/1.73 m²)	73.4 ± 20.6 (n = 999)
Use of oral glucose-lowering agents (%)	894 (89.5)
Metformin (%)	543 (54.4)
Sulfonylureas (%)	127 (12.7)
Glinides (%)	68 (6.8)
Dipeptidyl peptidase-4 inhibitors (%)	577 (57.8)
Sodium-glucose cotransporter 2 inhibitors (%)	231 (23.1)
Thiazolidinediones (%)	143 (14.3)
α-glucosidase inhibitors (%)	172 (17.2)
Glucagon-like peptide-1 antagonists (%)	74 (7.4)
Insulin (%)	158 (15.8)
Use of anti-hypertensive drugs	483 (48.3)
ACE inhibitors (%)	28 (2.8)
Angiotensin II receptor blockers (%)	390 (39.0)
Calcium channel blockers (%)	273 (27.3)
Diuretic drugs (%)	57 (5.7)
α-adrenergic receptor antagonists (%)	19 (1.9)
β-adrenergic receptor antagonists (%)	33 (3.3)
Use of lipid-lowering agents (%)	595 (59.7)
Statins (%)	508 (51.0)
Ezetimibe (%)	107 (10.7)
Fibrates (%)	41 (4.1)
Use of anti-thrombotic agents (%)	64 (6.4)
Antiplatelet agents (%)	50 (5.0)
Anticoagulants (%)	15 (1.5)
FLP-CGM–derived metrics
Mean glucose (mmol/L)	7.80 ± 1.79 (n = 999)
SD (mmol/L)	2.04 ± 0.63 (n = 999)
CV (%)	26.2 ± 5.79 (n = 999)
MAGE (mmol/L)	5.46 ± 2.00 (n = 999)
TIR^{3.9–10 mmol/L} (%)	78.9 ± 18.6 (n = 999)
TAR^{> 10 mmol/L} (%)	19.0 ± 19.2 (n = 999)
TAR^{> 13.9 mmol/L} (%)	3.85 ± 9.31 (n = 999)
TBR^{< 3.9 mmol/L} (%)	2.16 ± 4.71 (n = 999)
TBR^{< 3.0 mmol/L} (%)	0.33 ± 1.53 (n = 999)
LBGI	1.56 ± 1.67 (n = 999)
HBGI	5.58 ± 4.64 (n = 999)
MODD (mmol/L)	1.73 ± 0.64 (n = 999)
IQR (mmol/L)	2.14 ± 0.81 (n = 999)
Data are means ± SD or number of patients (%).
CV, coefficient of variation; FLP-CGM, FreeStyle Libre Pro continuous glucose monitoring device; HBGI, high blood glucose index; IQR, interquartile range; LBGI, low blood glucose index; MAGE, mean amplitude of glycemic excursions; MODD, mean of daily differences; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range.

In this study, 222 of 999 (22.2%) were diagnosed as having DR. Subject characteristics by DR stage are presented in Table 2. SDR was observed in 133 subjects (13.3%), PPDR in 50 (5.0%), and PDR in 39 (3.9%). Subjects with more severe DR were more likely to be older, have longer duration of DM, higher HbA1c levels, higher uric acid levels, higher UAE, and lower eGFR. All FLP-CGM–derived metrics except TBR^{< 3.9 mmol/L} and TBR^{< 3.0 mmol/L} were significantly different among the groups. Subjects with more severe DR were more likely to be treated with oral antidiabetic drugs and anti-hypertensive drugs, respectively.

Table 2

Patient demographic and background characteristics by diabetic retinopathy stage
Parameter	NDR (n = 777)	SDR (n = 133)	PPDR (n = 50)	PDR (n = 39)	P value
Age (years)	64.1 ± 9.8	65.3 ± 9.6	67.3 ± 8.5	66.6 ± 7.2	0.039
Male gender (%)	482 (62.0)	84 (63.2)	23 (46.0)	19 (48.7)	0.050
BMI (kg/m²)	24.6 ± 3.8	24.7 ± 3.8	24.4 ± 4.1	25.1 ± 4.3	0.838
Estimated duration of diabetes (years)	11.8 ± 8.0	15.6 ± 8.9	17.5 ± 9.6	19.1 ± 9.6	< 0.001
Systolic blood pressure (mmHg)	130.9 ± 15.1	133.1 ± 14.0	132.3 ± 13.9	130.3 ± 14.4	0.403
Diastolic blood pressure (mmHg)	75.6 ± 11.3	76.5 ± 10.2	71.9 ± 9.7	75.5 ± 10.4	0.089
HbA1c (%)	7.0 ± 0.8	7.2 ± 0.9	7.6 ± 0.9	7.6 ± 1.1	< 0.001
HbA1c (mmol/mol)	52.8 ± 8.2	55.2 ± 9.4	59.1 ± 9.6	59.8 ± 11.8	< 0.001
Total cholesterol (mmol/L)	4.87 ± 0.81	4.55 ± 0.79	4.51 ± 0.63	4.81 ± 0.96	< 0.001
LDL cholesterol (mmol/L)	2.71 ± 0.68	2.54 ± 0.67	2.45 ± 0.78	2.57 ± 0.65	0.005
HDL cholesterol (mmol/L)	1.57 ± 0.41	1.51 ± 0.33	1.57 ± 0.37	1.57 ± 0.49	0.450
Triglycerides (mmol/L)	1.4 ± 0.9	1.1 ± 0.6	1.5 ± 1.2	1.5 ± 1.0	0.029
Uric acid (µmol/L)	306.6 ± 72.3	315.4 ± 72.7	278.1 ± 71.3	333.9 ± 78.1	0.002
Estimated glomerular filtration rate (mL/min/1.73 m²)	75 ± 20	72 ± 24	68 ± 18	58 ± 25	< 0.001
Urinary albumin excretion (mg/g creatinine)	67.9 ± 254	73.6 ± 206	212 ± 464	607 ± 1261	< 0.001
Use of oral glucose-lowering agents	677 (87.1)	129 (97)	49 (98)	39 (100)	< 0.001
Metformin (%)	402 (51.7)	93 (69.9)	34 (68.0)	14 (35.9)	< 0.001
Sulfonylureas (%)	89 (11.5)	20 (15.0)	8 (16.0)	10 (25.6)	0.044
Glinides (%)	40 (5.1)	12 (9.0)	9 (18.0)	7 (17.9)	< 0.001
Dipeptidyl peptidase-4 inhibitors (%)	435(56.0)	87 (65.4)	31 (62.0)	24 (61.5)	0.188
Sodium-glucose cotransporter 2 inhibitors (%)	167 (21.5)	34 (25.6)	19 (38.0)	11 (28.2)	0.038
Thiazolidinediones (%)	100 (12.9)	23 (17.3)	13 (26.0)	7 (17.9)	0.041
α-glucosidase inhibitors (%)	117 (15.1)	39 (29.3)	10 (20.0)	6 (15.4)	< 0.001
Glucagon-like peptide-1 receptor agonists (%)	36 (4.6)	18 (13.5)	11 (22.0)	9 (23.1)	< 0.001
Insulin (%)	94 (12.1)	31 (23.3)	13 (26.0)	20 (51.3)	< 0.001
Use of anti-hypertensive drugs	346 (44.5)	79 (59.4)	29 (58.0)	29 (74.4)	< 0.001
ACE inhibitors (%)	19 (2.4)	5 (3.8)	2 (4.0)	2 (5.1)	0.323
Angiotensin II receptor blockers (%)	278 (35.8)	67 (50.4)	22 (44.0)	23 (59.0)	< 0.001
Calcium channel blockers (%)	193 (24.8)	44 (33.1)	19 (38.0)	17 (43.6)	0.005
Diuretic drugs (%)	35 (4.5)	13 (9.8)	3 (6.0)	6 (15.4)	0.006
α-adrenergic receptor antagonists (%)	13 (1.7)	4 (3.0)	1 (2.0)	1 (2.6)	0.455
β-adrenergic receptor antagonists (%)	26 (3.3)	4 (3.0)	1 (2.0)	2 (5.1)	1.000
Use of lipid-lowering agents (%)	450 (58.1)	93 (69.9)	35 (70.0)	17 (43.6)	0.005
Statins (%)	386 (49.8)	73 (54.9)	32 (64.0)	17 (43.6)	0.140
Ezetimibe (%)	76 (9.8)	23 (17.3)	5 (10.0)	3 (7.7)	0.089
Fibrates (%)	31 (4)	6 (4.5)	4 (8.0)	0 (0.0)	0.298
Use of anti-thrombotic agents (%)	48 (6.2)	8 (6.0)	5 (10.0)	3 (7.7)	0.623
Antiplatelet agents (%)	36 (4.6)	8 (6.0)	5 (10.0)	1 (2.6)	0.297
Anticoagulants (%)	13 (1.7)	0 (0.0)	0 (0.0)	2 (5.1)	0.110
FLP-CGM–derived metrics
Mean glucose (mmol/L)	7.66 ± 1.71	8.09 ± 1.77	8.55 ± 2.06	8.61 ± 2.48	< 0.001
SD (mmol/L)	2.00 ± 0.60	2.07 ± 0.61	2.20 ± 0.72	2.46 ± 0.80	< 0.001
CV (%)	26.2 ± 5.84	25.5 ± 5.38	25.9 ± 5.72	28.8 ± 5.66	0.019
MAGE (mmol/L)	5.36 ± 1.93	5.43 ± 1.76	5.93 ± 2.39	6.82 ± 2.99	< 0.001
TIR^{3.9–10 mmol/L} (%)	80.4 ± 17.6	76.3 ± 19.8	70.3 ± 22.8	68.9 ± 21.9	< 0.001
TAR^{> 10 mmol/L} (%)	17.4 ± 18.0	22.2 ± 20.7	27.6 ± 24.5	27.8 ± 22.8	< 0.001
TAR^{> 13.9 mmol/L} (%)	3.28 ± 8.64	4.41 ± 9.20	7.55 ± 10.9	8.51 ± 16.0	< 0.001
TBR^{< 3.9 mmol/L} (%)	2.21 ± 4.93	1.56 ± 3.33	2.08 ± 4.28	3.29 ± 4.72	0.210
TBR^{< 3.0 mmol/L} (%)	0.350 ± 1.67	0.11 ± 0.36	0.27 ± 0.77	0.68 ± 1.59	0.178
LBGI	1.60 ± 1.72	1.25 ± 1.20	1.31 ± 1.44	2.16 ± 1.99	0.010
HBGI	5.26 ± 4.35	5.99 ± 4.50	7.26 ± 5.29	8.38 ± 7.58	< 0.001
MODD (mmol/L)	1.69 ± 0.61	1.78 ± 0.56	1.95 ± 0.66	2.34 ± 1.00	< 0.001
IQR (mmol/L)	2.09 ± 0.79	2.16 ± 0.72	2.33 ± 0.76	2.95 ± 1.20	< 0.001
Data are means ± SD or number of patients (%). Continuous data were compared using analysis of variance. Categorical data were compared using the chi-square test or Fisher's exact test as appropriate.
CV, coefficient of variation; FLP-CGM, FreeStyle Libre Pro continuous glucose monitoring device; HBGI, high blood glucose index; HDL-C, IQR, interquartile range; LBGI, low blood glucose index; MAGE, mean amplitude of glycemic excursions; MODD, mean of daily differences; NDR, no diabetic retinopathy; PDR, proliferative diabetic retinopathy; PPDR, pre-proliferative diabetic retinopathy; SD, standard deviation; SDR, simple retinopathy; TAR, time above range; TBR, time below range; TIR, time in range.

Next, we investigated the relationship between FLP-CGM–derived metrics and DR severity in patients with type 2 diabetes. In a proportional odds model with the patients with NDR as the reference group, HbA1c and FLP-CGM–derived metrics except for CV, TBR^{< 3.9 mmol/L}, TBR^{< 3.0 mmol/L}, and LBGI were significantly associated with DR severity (Model 1 in Table 3). In Models 2 and 3, the associations remained significant after adjusting for age, gender, BMI, duration of diabetes, systolic blood pressure, lipid parameters, uric acid, eGFR, UAE, smoking, alcohol consumption, use of insulin therapy, use of ACE inhibitors and/or angiotensin II receptor blockers (ARBs), use of statins, and use of anti-platelet agents (Table 2). However, we did not find any significant associations between FLP-CGM–derived metrics and DR severity after adjusting for HbA1c in addition to parameters included in Model 3 (Model 4 in Table 3).

Table 3

Associations between FLP-CGM–derived metrics and diabetic retinopathy severity
Parameter	Odds ratio (95% CI)	P value
Mean glucose (1 mmol/L increase)
Model 1	1.21 (1.12–1.30)	< 0.001
Model 2	1.20 (1.10–1.29)	< 0.001
Model 3	1.22 (1.11–1.33)	< 0.001
Model 4	1.07 (0.94–1.22)	0.311
SD (mmol/L) (1 mmol/L increase)
Model 1	1.57 (1.26–1.96)	< 0.001
Model 2	1.36 (1.08–1.71)	0.010
Model 3	1.30 (1.00-1.69)	0.049
Model 4	0.97 (0.72–1.31)	0.842
CV (%) (1% increase)
Model 1	1.00 (0.98–1.03)	0.747
Model 2	0.99 (0.96–1.01)	0.365
Model 3	0.98 (0.95-1.00)	0.089
Model 4	0.98 (0.95–1.01)	0.170
MAGE (1 mmol/L increase)
Model 1	1.13 (1.05–1.21)	< 0.001
Model 2	1.10 (1.03–1.18)	0.007
Model 3	1.10 (1.02–1.19)	0.015
Model 4	1.03 (0.94–1.12)	0.531
TIR^{3.9–10 mmol/L} (10% increase)
Model 1	0.83 (0.77–0.89)	< 0.001
Model 2	0.85 (0.79–0.91)	< 0.001
Model 3	0.85 (0.78–0.93)	< 0.001
Model 4	0.97 (0.86–1.09)	0.616
TAR^{> 10 mmol/L} (1% increase)
Model 1	1.02 (1.01–1.04)	< 0.001
Model 2	1.02 (1.01–1.02)	< 0.001
Model 3	1.02 (1.01–1.03)	< 0.001
Model 4	1.01 (0.99–1.02)	0.371
TAR^{> 13.9 mmol/L} (1% increase)
Model 1	1.03 (1.01–1.04)	< 0.001
Model 2	1.03 (1.01–1.04)	< 0.001
Model 3	1.03 (1.01–1.05)	< 0.001
Model 4	1.01 (0.99–1.03)	0.541
TBR^{< 3.9 mmol/L} (1% increase)
Model 1	0.99 (0.96–1.03)	0.677
Model 2	0.99 (0.95–1.02)	0.397
Model 3	0.96 (0.93–0.99)	0.028
Model 4	0.98 (0.94–1.01)	0.238
TBR^{< 3.0 mmol/L} (1% increase)
Model 1	0.96 (0.85–1.08)	0.499
Model 2	0.94 (0.83–1.06)	0.288
Model 3	0.84 (0.72–0.97)	0.020
Model 4	0.88 (0.77–1.01)	0.071
LBGI (1unit increase)
Model 1	0.95 (0.86–1.04)	0.268
Model 2	0.92 (0.84–1.02)	0.097
Model 3	0.84 (0.76–0.94)	0.002
Model 4	0.91 (0.82–1.01)	0.069
HBGI (1unit increase)
Model 1	1.07 (1.04–1.10)	< 0.001
Model 2	1.06 (1.03–1.09)	< 0.001
Model 3	1.06 (1.03–1.10)	< 0.001
Model 4	1.01 (0.96–1.06)	0.745
MODD ( 1 mmol/L increase)
Model 1	1.84 (1.49–2.26)	< 0.001
Model 2	1.75 (1.41–2.18)	< 0.001
Model 3	1.67 (1.28–2.19)	< 0.001
Model 4	1.27 (0.93–1.74)	0.126
IQR (1 mmol/L increase)
Model 1	1.50 (1.28–1.77)	< 0.001
Model 2	1.48 (1.24–1.75)	< 0.001
Model 3	1.42 (1.15–1.76)	0.001
Model 4	1.11 (0.87–1.43)	0.399
HbA1c (1% increase)
Model 1	1.76 (1.48–2.08)	< 0.001
Model 2	1.62 (1.35–1.94)	< 0.001
Model 3	1.66 (1.35–2.04)	< 0.001
Model 1: crude
Model 2: adjusted for age and gender, BMI, and duration of diabetes
Model 3: adjusted for variables in Model 2 plus systolic blood pressure, total cholesterol, HDL cholesterol, logarithm of triglycerides, serum uric acid, estimated glomerular filtration rate, logarithm of urinary albumin excretion, smoker, alcohol consumption, use of insulin therapy, use of ACE inhibitors and/or angiotensin II receptor blockers, use of stains, and use anti-platelet agents
Model 4: adjusted for variables in Model 3 plus HbA1c
CV, coefficient of variation; HBGI, high blood glucose index; IQR, interquartile range; LBGI, low blood glucose index; MAGE, mean amplitude of glycemic excursions; MODD, mean of daily differences; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range.

Relationship Between Flp-cgm–derived Metrics And Albuminuria Severity

Of 999 subjects, 729 (73.0%) were classified as having normoalbuminuria, 203 (20.3%) were classified as having microalbuminuria, and 67 (6.7%) were classified as having macroalbuminuria. The clinical characteristics of the study participants stratified by albuminuria status are summarized in Table 4. Subjects with more severe albuminuria were more likely to have longer duration of DM, higher BMI, higher prevalence of DR, higher HbA1c levels, higher triglyceride levels, higher uric acid levels, lower HDL levels, and lower eGFR. Among the groups, there were significant differences in most FLP-CGM–derived metrics, except for CV, TBR^{< 3.9 mmol/L}, TBR^{< 3.0 mmol/L}, and LBGI. Subjects with more severe albuminuria were more likely to be treated with sodium-glucose cotransporter 2 (SGLT-2) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, and insulin, as well as anti-hypertensive drugs such as calcium channel blockers, ACE inhibitors, ARBs.

Table 4

Patient demographic and background characteristics stratified by albuminuria status
Parameter	Normoalbuminuria (n = 729)	Microalbuminuria (n = 203)	Macroalbuminuria (n = 67)	P value
Age (years)	64.1 ± 9.5	66.2 ± 9.8	64.6 ± 10.0	0.026
Male gender (%)	441 (60.5)	120 (59.1)	47 (70.1)	0.256
BMI (kg/m²)	24.3 ± 3.9	25.2 ± 3.6	25.9 ± 4.0	< 0.001
Estimated duration of diabetes (years)	12.0 ± 8.2	14.6 ± 8.7	17.2 ± 9.4	< 0.001
Systolic blood pressure (mmHg)	130.7 ± 14.7	132.5 ± 14.6	133.4 ± 17.1	0.147
Diastolic blood pressure (mmHg)	75.5 ± 10.8	74.8 ± 11.7	77.6 ± 11.9	0.198
HbA1c (%)	7.0 ± 0.7	7.3 ± 0.9	7.5 ± 1.1	< 0.001
HbA1c (mmol/mol)	52.6 ± 7.8	55.8 ± 10.0	58.5 ± 12.3	< 0.001
Total cholesterol (mmol/L)	4.83 ± 0.81	4.76 ± 0.74	4.72 ± 1.05	0.437
LDL cholesterol (mmol/L)	2.69 ± 0.68	2.64 ± 0.64	2.43 ± 0.81	0.010
HDL cholesterol (mmol/L)	1.59 ± 0.41	1.49 ± 0.41	1.49 ± 0.37	0.004
Triglycerides (mmol/L)	1.3 ± 0.8	1.4 ± 0.9	2.0 ± 1.8	< 0.001
Uric acid (µmol/L)	300.6 ± 69.9	314.8 ± 79.0	358.7 ± 65.0	< 0.001
Estimated glomerular filtration rate (mL/min/1.73 m²)	76 ± 19	69 ± 21	58 ± 25	< 0.001
Diabetic retinopathy (%)	123 (16.9)	63 (31.0)	36 (53.7)	< 0.001
Use of oral glucose-lowering agents (%)	639 (88)	188 (93)	67 (100)	0.002
Metformin (%)	387 (53)	120 (59)	36 (54)	0.311
Sulfonylureas (%)	84 (12)	32 (16)	11 (16)	0.177
Glinides (%)	44 (6)	17 (8)	7 (10)	0.238
Dipeptidyl peptidase-4 inhibitors (%)	412 (57)	118 (58)	47 (70)	0.096
Sodium-glucose cotransporter 2 inhibitors (%)	149 (20)	59 (29)	23 (34)	0.003
Thiazolidinediones (%)	99 (14)	30 (15)	14 (21)	0.256
α-glucosidase inhibitors (%)	125 (17)	32 (16)	15 (22)	0.458
Glucagon-like peptide-1 receptor agonists (%)	40 (6)	22 (11)	12 (18)	< 0.001
Insulin (%)	92 (13)	44 (22)	22 (33)	< 0.001
Use of anti-hypertensive drugs (%)	298 (41)	130 (64)	55 (82)	< 0.001
ACE inhibitors (%)	13 (2)	11 (5)	4 (6)	0.006
Angiotensin II receptor blockers (%)	245 (34)	101 (50)	44 (66)	< 0.001
Calcium channel blockers (%)	154 (21)	86 (42)	33 (49)	< 0.001
Diuretic drugs (%)	33 (5)	17 (8)	7 (10)	0.025
α-adrenergic receptor antagonists (%)	9 (1)	6 (3)	4 (6)	0.013
β-adrenergic receptor antagonists (%)	15 (2)	12 (6)	6 (9)	< 0.001
Use of lipid-lowering agents (%)	426 (59)	121 (60)	48 (72)	0.114
Statins (%)	360 (50)	106 (52)	42 (63)	0.110
Ezetimibe (%)	75 (10)	21 (10)	11 (16)	0.298
Fibrates (%)	29 (4)	9 (4)	3 (5)	0.949
Use of anti-thrombotic agents (%)	46 (6)	14 (7)	4 (6)	0.948
Antiplatelet agents (%)	36 (5)	11 (5)	3 (5)	0.935
Anticoagulants (%)	11 (2)	3 (2)	1 (2)	1.000
FLP-CGM–derived metrics
Mean glucose (mmol/L)	7.59 ± 1.59	8.25 ± 1.97	8.70 ± 2.62	< 0.001
SD (mmol/L)	1.97 ± 0.58	2.20 ± 0.68	2.29 ± 0.77	< 0.001
CV (%)	26.0 ± 5.72	26.8 ± 5.90	26.7 ± 6.11	0.142
MAGE (mmol/L)	5.26 ± 1.85	5.86 ± 2.08	6.34 ± 2.79	< 0.001
TIR^{3.9–10 mmol/L} (%)	81.1 ± 17.1	74.3 ± 19.5	68.0 ± 24.7	< 0.001
TAR^{> 10 mmol/L} (%)	16.7 ± 17.5	23.8 ± 20.4	28.7 ± 26.3	< 0.001
TAR^{> 13.9 mmol/L} (%)	2.87 ± 7.4	5.71 ± 11.8	8.86 ± 15.7	< 0.001
TBR^{< 3.9 mmol/L} (%)	2.12 ± 4.58	1.91 ± 4.06	3.35 ± 7.20	0.086
TBR^{< 3.0 mmol/L} (%)	0.33 ± 1.64	0.20 ± 0.77	0.65 ± 1.85	0.117
LBGI	1.58 ± 1.68	1.41 ± 1.44	1.79 ± 2.07	0.226
HBGI	5.04 ± 3.86	6.70 ± 5.48	8.08 ± 7.40	< 0.001
MODD (mmol/L)	1.66 ± 0.59	1.88 ± 0.64	2.13 ± 0.90	< 0.001
IQR (mmol/L)	2.05 ± 0.76	2.31 ± 0.82	2.65 ± 1.11	< 0.001
Data are means ± SD or number of patients (%). Continuous data were compared using analysis of variance. Categorical data were compared using the chi-square test or Fisher's exact test as appropriate. CV, coefficient of variation; FLP-CGM, FreeStyle Libre Pro continuous glucose monitoring device; HBGI, high blood glucose index; IQR, interquartile range; LBGI, low blood glucose index; MAGE, mean amplitude of glycemic excursions; MODD, mean of daily differences; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range.

Next, we investigated the relationship between FLP-CGM derived–metrics and albuminuria severity. In a proportional odds model with the patients with normoalbuminuria as the reference group, most FLP-CGM–derived metrics except for CV, TBR^{< 3.9 mmol/L}, TBR^{< 3.0 mmol/L}, and LBGI were significantly associated with albuminuria severity (Model 1), as shown in Table 5. In Models 2 and 3, most FLP-CGM–derived metrics, except for CV, TBR^{< 3.9 mmol/L}, TBR^{< 3.0 mmol/L}, and LBGI were significantly associated with albuminuria severity after adjusting for age, gender, BMI, duration of diabetes, systolic blood pressure, HbA1c levels, lipid parameters, uric acid, eGFR, smoking, alcohol consumption, use of insulin therapy, use of ACE inhibitors and/or ARBs, use of statins, use of anti-platelet agents, and presence of DR. Thus, these metrics are predictive factors for the severity of albuminuria independent of HbA1c levels.

Table 5

Associations between FLP-CGM–derived metrics and albuminuria severity
Parameter	Odds ratio (95% CI)	P value
Mean glucose (1 mmol/L increase)
Model 1	1.26 (1.17–1.35)	< 0.001
Model 2	1.24 (1.15–1.34)	< 0.001
Model 3	1.28 (1.14–1.45)	< 0.001
SD (1 mmol/L increase)
Model 1	1.86 (1.50–2.31)	< 0.001
Model 2	1.81 (1.45–2.27)	< 0.001
Model 3	1.57 (1.18–2.08)	0.002
CV (1% increase)
Model 1	1.02 (1.00-1.05)	0.053
Model 2	1.02 (1.00-1.05)	0.066
Model 3	1.02 (0.99–1.05)	0.251
MAGE (1 mmol/L increase)
Model 1	1.20 (1.12–1.28)	< 0.001
Model 2	1.19 (1.11–1.27)	< 0.001
Model 3	1.14 (1.05–1.24)	0.001
TIR^{3.9–10 mmol/L} (10% increase)
Model 1	0.97 (0.74–0.85)	< 0.001
Model 2	0.80 (0.75–0.86)	< 0.001
Model 3	0.81 (0.72–0.90)	< 0.001
TAR^{> 10 mmol/L} (1% increase)
Model 1	1.02 (1.02–1.03)	< 0.001
Model 2	1.02 (1.01–1.03)	< 0.001
Model 3	1.02 (1.01–1.03)	< 0.001
TAR^{> 13.9 mmol/L} (1% increase)
Model 1	1.04 (1.02–1.05)	< 0.001
Model 2	1.04 (1.02–1.05)	< 0.001
Model 3	1.03 (1.01–1.05)	0.002
TBR^{< 3.9 mmol/L} (1% increase)
Model 1	1.01 (0.98–1.04)	0.436
Model 2	1.01 (0.98–1.04)	0.523
Model 3	1.01 (0.97–1.03)	0.751
TBR^{< 3.0 mmol/L} (1% increase)
Model 1	1.01 (0.92–1.10)	0.914
Model 2	1.00 (0.91–1.09)	0.927
Model 3	0.98 (0.89–1.08)	0.663
LBGI (1unit increase)
Model 1	0.98 (0.90–1.07)	0.675
Model 2	0.98 (0.90–1.07)	0.623
Model 3	0.98 (0.89–1.08)	0.683
HBGI (1 unit increase)
Model 1	1.09 (1.06–1.13)	< 0.001
Model 2	1.08 (1.05–1.11)	< 0.001
Model 3	1.09 (1.04–1.13)	< 0.001
MODD (1 mmol/L increase)
Model 1	1.98 (1.62–2.43)	< 0.001
Model 2	1.97 (1.59–2.44)	< 0.001
Model 3	1.70 (1.26–2.29)	< 0.001
IQR (1 mmol/L increase)
Model 1	1.68 (1.44–1.98)	< 0.001
Model 2	1.70 (1.44–2.01)	< 0.001
Model 3	1.53 (1.21–1.93)	< 0.001
HbA1c (1% increase)
Model 1	1.68 (1.43–1.98)	< 0.001
Model 2	1.56 (1.31–1.85)	< 0.001
Model 3 (excluding HbA1c)	1.39 (1.14–1.69)	0.001
Model 1: crude
Model 2: adjusted for age and gender, BMI, and duration of diabetes
Model 3: adjusted for variables in Model 2 plus systolic blood pressure, HbA1c, total cholesterol, HDL cholesterol, logarithm of triglycerides, serum uric acid, estimated glomerular filtration rate, smoker, alcohol consumption, presence of diabetic retinopathy, use of insulin therapy, use of ACE inhibitors and/or angiotensin II receptor blockers, use of stains ,and use anti-platelet agents
CV, coefficient of variation; HBGI, High blood glucose index; IQR, interquartile range; LBGI, low blood glucose index; MAGE, mean amplitude of glycemic excursions; MODD, mean of daily differences; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range.

In this study, we demonstrated that most FLP-CGM–derived metrics related to intra-day and inter-day glucose variability are significantly associated with the severity of DR or albuminuria, even after adjusting for various risk factors in 999 outpatients with type 2 diabetes. Notably, these metrics remain predictive factors for determining the severity of albuminuria after adjusting for HbA1c levels.

A pooled sub-analysis of population-based studies demonstrated that the prevalence of any DR in patients with diabetes during 2000–2008 was substantially lower than the prevalence observed before 2000 [26]. The relative reduction in the prevalence of any DR from 49.6–24.8% may reflect improvements in medical care and management of diabetes and DR-related risk factors, including blood pressure, as well as early disease identification and medical provider awareness. In that study [26], the prevalence of any DR and PDR in patients with type 2 diabetes during 1980–2008 was 27.2% and 2.6%, respectively. Our study demonstrated that the prevalence of any DR was 22.2% and PDR was 3.9%. Given that risk factors for DR such as blood glucose and blood pressure were relatively well controlled in our study, the prevalence of DR in our study is reasonable.

The Japan Diabetes Complications Study demonstrated that HbA_1c is the strongest risk factor for development and progression of DR, while longer duration of diabetes, systolic blood pressure, and BMI are positively associated with incident DR [27]. In fact, previous studies have demonstrated that improvement in HbA1c levels is associated with reduced risk of DR development and progression in patients with type 2 diabetes [6, 7]. In accordance with those findings, our study indicated that HbA1c is positively associated with DR severity, even after adjusting for several risk factors. On the other hand, a recent study demonstrated that TIR^{3.9–10 mmol/L} based on seven-point glucose testing is inversely associated with the risk of DR progression in patients with type 1 diabetes [28]. Similarly, another cross-sectional study demonstrated that CGM-derived TIR^{3.9–10 mmol/L} is inversely associated with DR severity, independent of HbA1c levels in patients with type 2 diabetes [17]. Our study also demonstrated that FLP-CGM–derived TIR^{3.9–10.0 mmol/L} is inversely associated with DR severity, even after adjusting for several possible risk factors in patients with type 2 diabetes. However, associations between TIR^{3.9–10 mmol/L} as well as other FLP-CGM–derived metrics including SD, MAGE, TAR^{> 10 mmol/L}, TAR^{> 13.9 mmol/L}, HBGI, MODD, and IQR and DR severity did not reach statistical significance after adjusting for HbA1c levels.

A possible explanation for the discrepant findings may be differences in characteristics of patients between our studies and the other two previous studies. First, the subjects of a prior study [28] were patients with type 1 diabetes treated with insulin, which is completely different from our subjects. In addition, TIR^{3.9–10mmol/L} derived from seven-point blood glucose testing reflects one daytime values; it does not reflect the overnight period and has limited ability to assess intra-day and inter-glycemic variability. The subjects of the other prior study [17] were hospitalized for the treatment of diabetes. They were mainly treated with insulin and had hospital meals during a few days of CGM measurement. Thus, these patients are anticipated to have lower TIR^{3.9–10mmol/L} despite substantially higher HbA1c levels. Accordingly, that data may not be generalizable to outpatients with type 2 diabetes under their usual living conditions.

Recently, SGLT-2 inhibitors, DPP-4 inhibitors, and GLP-1 receptor agonists have been frequently used for patients with type 2 diabetes. These drugs are reported to decrease glucose fluctuations without increasing the risk of hypoglycemia [29]. Not surprisingly, our subjects had lower HbA1c levels and were more likely to have lower SD and MAGE and higher TIR^{3.9–10mmol/L}, because a higher proportion used DPP-4 inhibitors and SGLT-2 inhibitors than those in a previous study [29]. Thus, it could be difficult to detect the impact of glycemic variability on DR severity independent of HbA1c levels in our study subjects. However, it should be noted that FLP-CGM–derived metrics related to intra-day and inter-day glucose variability were still significantly associated with DR severity after adjusting for possible risk factors other than HbA1c. These findings highlight the substantial role of intra-day and inter-day glucose variability in the pathogenesis of DR and the important role of FLP-CGM–derived metrics to complement HbA1c.

According to serial cross-sectional studies of patients with diabetes who participated in National Health and Nutrition Examination Surveys, the prevalence of albuminuria declined progressively from 20.8% in 1988–1994 to 15.9% in 2009–2014 [30]. Higher prevalence of treatment with anti-diabetic agents, renin-angiotensin system inhibitors, and statins may account for the reduction in the prevalence of albuminuria. On the other hand, an in vitro study showed that intermittent treatment of high blood glucose levels increases apoptosis of mesangial cells by increasing levels of inflammatory cytokines and oxidative stress, leading to the development of DN [31]. However, there is limited data about the clinical impact of intra-day and inter-day glucose variability on the presence or progression of albuminuria in patients with type 2 diabetes. Two previous studies conducted by the same group did not show a significant association between CGM–derived metrics of intra-day and inter-day glucose variability and the presence of albuminuria independent of HbA1c levels in patients with type 2 diabetes with HbA1c levels of more than 8% [32, 33]. In contrast, our study clearly showed close relationships between FLP-CGM–derived metrics related to intra-day and inter-day glucose variability and the severity of albuminuria, even after adjusting for various possible risk factors including HbA1c, in patients with type 2 diabetes. The reason that those studies yielded conflicting results is not clear, but it may be due to differences in the characteristics of subjects or use of anti-diabetic agents among studies. Previous studies enrolled patients with inadequately controlled type 2 diabetes; thus, the prevalence of albuminuria was relatively high, approximately 40%. In contrast, our subjects had a substantially lower prevalence of albuminuria (27%) and were more likely to be taking SGLT-2 inhibitors, DPP-4 inhibitors, and GLP-1 receptor agonists. Taken together, our data suggest that intra-day and inter-day glucose variability are important targets in terms of reducing the risk of albuminuria in patients treated according to the current consensus about diabetes treatment [34].

HbA1c is recognized as a gold standard for treatment target. A few studies have demonstrated strong associations between HbA1c levels and diabetic microvascular complications [4, 5]. However, HbA1c alone may not adequately reflect an individual’s glycemic variation and risk of hyperglycemia and hypoglycemia. In this regard, the ATTD Congress recommended TIR as a key metric of glycemic management in clinical practice [15]. Our results showed that TIR^{3.9–10 mmol/L}, TAR^{> 10 mmol/L}, and TAR^{> 13.9 mmol/L} are significantly associated with the severity of albuminuria, even after adjusting for possible risk factors including HbA1c levels. On the other hand, a recent study demonstrated that severe hypoglycemia is a predictor of worsening renal dysfunction in patients with type 2 diabetes [35]. However, in addition to LBGI, TBR^{< 3.9 mmol/L}, and TBR^{< 3.0 mmol/L} were not associated with the severity of albuminuria in our study. Mild hypoglycemia, low frequency of hypoglycemic events, or both may not be involved in the development of albuminuria. Alternatively, the relatively low frequency of hypoglycemic events and short duration of hypoglycemia observed in our study may account for this finding. Taken together, based on FLP-CGM–derived metrics, focusing on improving hyperglycemia may be important to reduce the risk of albuminuria development. However, HbA1c alone does not provide enough information. Indeed, a previous study demonstrated that lower renal function (eGFR < 60 mL/min/1.73 m²) is strongly correlated with a higher prevalence of anemia in the general population [36]. Modest reductions in hemoglobin due to a shorter erythrocyte lifespan may affect the accuracy of HbA1c. In patients with more advanced DN, evaluating FLP-CGM–derived metrics could serve as a therapeutic target complementary to HbA1c.

Our study found that FLP-CGM–derived metrics related to glucose variability are associated with the severity of albuminuria, which is different from DR, even after adjusting for HbA1c levels. Although the exact reason for these findings is not clear, we postulated one scenario. Atherosclerosis of the intrarenal and extrarenal arteries and microangiopathy of the glomerular capillaries, afferent arterioles, and efferent arterioles are considered to contribute to the progression of glomerular lesions in DN [37, 38]. Previous studies have demonstrated that glucose fluctuation is more significantly associated with atherosclerotic-related diseases than the degree of hyperglycemic exposure as indicated by HbA1c levels in patients with type 2 diabetes [39, 40]. Thus, it is possible that atherosclerosis of the intrarenal and extrarenal arteries caused by glucose variability may also accelerate renal damage. Therefore, glucose variability is more likely to be associated with the pathogenesis of DN than DR.

The strengths of this study included its relatively large sample size and multicenter study design. Our study had certain limitations. First, the cross-sectional study design made it impossible to evaluate whether FLP-CGM–derived metrics had a causal relationship with diabetic microvascular complications. In this regard, we are currently conducting a long-term follow-up study in the same cohort that focuses on FLP-CGM–derived metrics and onset of outcomes such as primary cardiovascular disease and diabetic microvascular complications. Second, FLP-CGM–derived metrics were evaluated based FLP-CGM measurements during a limited time. Thus, FLP-CGM–derived metrics may not represent overall glycemic control of subjects. In order to attain the best measurements of glucose fluctuations with FLP-CGM at baseline, we only recruited patients with stable control. In addition, we employed a blind CGM system that prevented subjects from altering their lifestyle behaviors based on the results of glucose readings. Third, we only recruited Japanese patients with type 2 diabetes. These constraints may limit the generalizability of our results.

In conclusion, we demonstrated that FLP-CGM–derived metrics related to intra-day and inter-day glucose variability are significantly associated with the severity of DR and albuminuria, even after adjusting for various risk factors in patients with type 2 diabetes. Thus, these derived metrics could provide medical professionals with useful information for assessing the risk of severe diabetic microvascular complications. CGM might identified treatment targets in addition to those based on HbA1c.

Acknowledgments

The authors wish to thank the study investigators listed in Supplementary Table 1 and the participants for their contributions to this study. The authors also wish to acknowledge the assistance of D. Takayama and H. Yamada (Soiken Holdings Inc., Tokyo, Japan).

Funding.

This study was financially supported by the Japan Agency for Medical Research and Development (AMED) under Grant Number JP20ek0210105 (to H.W.) and the Manpei Suzuki Diabetes Foundation (to H.W.).

Ethics approval and consent to participate.

The protocol was approved by the institutional review board of each participating institution in compliance with the Declaration of Helsinki and current legal regulations in Japan. Written informed consent was obtained from all participants after a full explanation of the study.

Availability of data and materials.

The analyzed datasets are available from the corresponding author on reasonable request.

Consent for publication.

Not applicable.

Competing interests.

T. O. and H.W. have received research funds from Abbott Japan. H.W. is a member of the advisory board of Abbott Japan. All other authors (S.W., T.M., N.K., Y.O., H.Y., K.N., T.S., K.T., A.K., M.G., and I.S.) declare no conflicts of interest.

Author Contributions

All authors contributed to the study design and were involved in all stages of manuscript development. S.W. and T.M. drafted the manuscript. M.G., a statistician, was primarily responsible for data analysis. S.W, T.M., N.K., Y.O., T.O., H.Y., K.N., T.S., K.T., A.K., M.G., I.S., and H.W. also collected, analyzed, and interpreted the data; reviewed and edited the manuscript; and approved the final manuscript. H.W. is the principal guarantor of this work; he has full access to all study data and takes responsibility for the integrity of the data and the accuracy of data analysis. All authors have read and agreed to the publication of the manuscript.

Klein BE. Overview of epidemiologic studies of diabetic retinopathy. Ophthalmic Epidemiol. 2007;14(4):179–83.
Webster AC, Nagler EV, Morton RL, Masson P. Chronic Kidney Disease. Lancet. 2017;389(10075):1238–52.
Klein R, Klein BE, Moss SE, Cruickshanks KJ. Relationship of hyperglycemia to the long-term incidence and progression of diabetic retinopathy. Arch Intern Med. 1994;154(19):2169–78.
Stratton IM, Adler AI, Neil HA, Matthews DR, Manley SE, Cull CA, Hadden D, Turner RC, Holman RR. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. Bmj. 2000;321(7258):405–12.
The relationship of glycemic exposure (HbA1c) to the risk of development and progression of retinopathy in the diabetes control and complications trial. Diabetes 1995, 44(8):968–983.
Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet 1998, 352(9131):837–853.
Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M. Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract. 1995;28(2):103–17.
American Diabetes A. 6. Glycemic Targets: Standards of Medical Care in Diabetes-2020. Diabetes Care. 2020;43(Suppl 1):66–76.
Haneda M, Noda M, Origasa H, Noto H, Yabe D, Fujita Y, Goto A, Kondo T, Araki E: Japanese Clinical Practice Guideline for Diabetes 2016. Diabetology international 2018, 9(1):1–45.
Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, Reaven PD, Zieve FJ, Marks J, Davis SN, Hayward R, et al. Glucose control and vascular complications in veterans with type 2 diabetes. N Engl J Med. 2009;360(2):129–39.
Boussageon R, Bejan-Angoulvant T, Saadatian-Elahi M, Lafont S, Bergeonneau C, Kassai B, Erpeldinger S, Wright JM, Gueyffier F, Cornu C. Effect of intensive glucose lowering treatment on all cause mortality, cardiovascular death, and microvascular events in type 2 diabetes: meta-analysis of randomised controlled trials. Bmj. 2011;343:d4169.
Ceriello A, Esposito K, Piconi L, Ihnat MA, Thorpe JE, Testa R, Boemi M, Giugliano D. Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients. Diabetes. 2008;57(5):1349–54.
Quagliaro L, Piconi L, Assaloni R, Martinelli L, Motz E, Ceriello A. Intermittent high glucose enhances apoptosis related to oxidative stress in human umbilical vein endothelial cells: the role of protein kinase C and NAD(P)H-oxidase activation. Diabetes. 2003;52(11):2795–804.
Welsh KJ, Kirkman MS, Sacks DB. Role of Glycated Proteins in the Diagnosis and Management of Diabetes: Research Gaps and Future Directions. Diabetes Care. 2016;39(8):1299–306.
Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, Bosi E, Buckingham BA, Cefalu WT, Close KL, et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range. Diabetes Care. 2019;42(8):1593–603.
Lu J, Ma X, Zhang L, Mo Y, Ying L, Lu W, Zhu W, Bao Y, Zhou J. Glycemic variability assessed by continuous glucose monitoring and the risk of diabetic retinopathy in latent autoimmune diabetes of the adult and type 2 diabetes. Journal of diabetes investigation. 2019;10(3):753–9.
Lu J, Ma X, Zhou J, Zhang L, Mo Y, Ying L, Lu W, Zhu W, Bao Y, Vigersky RA, et al. Association of Time in Range, as Assessed by Continuous Glucose Monitoring, With Diabetic Retinopathy in Type 2 Diabetes. Diabetes Care. 2018;41(11):2370–6.
Xin Z, Zhu Y, Wang S, Liu S, Xu M, Wang T, Lu J, Chen Y, Zhao Z, Wang W, et al. Associations of subclinical atherosclerosis with nonalcoholic fatty liver disease and fibrosis assessed by non-invasive score. Liver international: official journal of the International Association for the Study of the Liver. 2020;40(4):806–14.
Sonoda S, Okada Y, Mori H, Uemura F, Sugai K, Hajime M, Tanaka K, Kurozumi A, Narisawa M, Torimoto K, et al. Association Between Diabetic Microangiopathies and Glycemic Variability Assessed by Continuous Glucose Monitoring. J UOEH. 2018;40(1):11–8.
Mita T, Katakami N, Okada Y, Yoshii H, Osonoi T, Nishida K, Shiraiwa T, Torimoto K, Kurozumi A, Wakasugi S, et al. Protocol of a Prospective Observational Study on the Relationship Between Glucose Fluctuation and Cardiovascular Events in Patients with Type 2 Diabetes. Diabetes therapy: research treatment education of diabetes related disorders. 2019;10(5):1565–75.
Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, Yamagata K, Tomino Y, Yokoyama H, Hishida A. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis. 2009;53(6):982–92.
Service FJ, Molnar GD, Rosevear JW, Ackerman E, Gatewood LC, Taylor WF. Mean amplitude of glycemic excursions, a measure of diabetic instability. Diabetes. 1970;19(9):644–55.
Kovatchev BP, Cox DJ, Kumar A, Gonder-Frederick L, Clarke WL: Algorithmic evaluation of metabolic control and risk of severe hypoglycemia in type 1 and type 2 diabetes using self-monitoring blood glucose data. Diabetes technology & therapeutics 2003, 5(5):817–828.
Hill NR, Oliver NS, Choudhary P, Levy JC, Hindmarsh P, Matthews DR. Normal reference range for mean tissue glucose and glycemic variability derived from continuous glucose monitoring for subjects without diabetes in different ethnic groups. Diabetes technology therapeutics. 2011;13(9):921–8.
Boscari F, Galasso S, Acciaroli G, Facchinetti A, Marescotti MC, Avogaro A, Bruttomesso D. Head-to-head comparison of the accuracy of Abbott FreeStyle Libre and Dexcom G5 mobile. Nutrition, metabolism, and cardiovascular diseases: NMCD 2018, 28(4):425–427.
Yau JW, Rogers SL, Kawasaki R, Lamoureux EL, Kowalski JW, Bek T, Chen SJ, Dekker JM, Fletcher A, Grauslund J, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care. 2012;35(3):556–64.
Kawasaki R, Tanaka S, Tanaka S, Yamamoto T, Sone H, Ohashi Y, Akanuma Y, Yamada N, Yamashita H. Japan Diabetes Complications Study G: Incidence and progression of diabetic retinopathy in Japanese adults with type 2 diabetes: 8 year follow-up study of the Japan Diabetes Complications Study (JDCS). Diabetologia. 2011;54(9):2288–94.
Beck RW, Bergenstal RM, Riddlesworth TD, Kollman C, Li Z, Brown AS, Close KL. Validation of Time in Range as an Outcome Measure for Diabetes Clinical Trials. Diabetes Care. 2019;42(3):400–5.
Zhou Z, Sun B, Huang S, Zhu C, Bian M. Glycemic variability: adverse clinical outcomes and how to improve it? Cardiovascular diabetology. 2020;19(1):102.
Afkarian M, Zelnick LR, Hall YN, Heagerty PJ, Tuttle K, Weiss NS, de Boer IH. Clinical Manifestations of Kidney Disease Among US Adults With Diabetes, 1988–2014. Jama. 2016;316(6):602–10.
Ying C, Wang S, Lu Y, Chen L, Mao Y, Ling H, Cheng X, Zhou X. Glucose fluctuation increased mesangial cell apoptosis related to AKT signal pathway. Archives of medical science: AMS. 2019;15(3):730–7.
Yoo JH, Choi MS, Ahn J, Park SW, Kim Y, Hur KY, Jin SM, Kim G, Kim JH: Association Between Continuous Glucose Monitoring-Derived Time in Range, Other Core Metrics, and Albuminuria in Type 2 Diabetes. Diabetes technology & therapeutics 2020.
Jin SM, Kim TH, Oh S, Baek J, Joung JY, Park SM, Cho YY, Sohn SY, Hur KY, Lee MS, et al. Association between the extent of urinary albumin excretion and glycaemic variability indices measured by continuous glucose monitoring. Diabetic medicine: a journal of the British Diabetic Association. 2015;32(2):274–9.
Davies MJ, D'Alessio DA, Fradkin J, Kernan WN, Mathieu C, Mingrone G, Rossing P, Tsapas A, Wexler DJ, Buse JB. Management of hyperglycaemia in type 2 diabetes, 2018. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia. 2018;61(12):2461–98.
Lee YL, Yen SJ, Shin SJ, Huang YC, He JS, Lin KD. Severe Hypoglycemia as a Predictor of End-Stage Renal Disease in Type 2 Diabetes: A National Cohort Study. International journal of environmental research and public health 2019, 16(5).
Astor BC, Muntner P, Levin A, Eustace JA, Coresh J. Association of kidney function with anemia: the Third National Health and Nutrition Examination Survey (1988–1994). Arch Intern Med. 2002;162(12):1401–8.
Bohle A, Wehrmann M, Bogenschutz O, Batz C, Muller CA, Muller GA. The pathogenesis of chronic renal failure in diabetic nephropathy. Investigation of 488 cases of diabetic glomerulosclerosis. Pathol Res Pract. 1991;187(2–3):251–9.
Shimizu M, Furuichi K, Toyama T, Funamoto T, Kitajima S, Hara A, Iwata Y, Sakai N, Takamura T, Kitagawa K, et al. Association of renal arteriosclerosis and hypertension with renal and cardiovascular outcomes in Japanese type 2 diabetes patients with diabetic nephropathy. Journal of diabetes investigation. 2019;10(4):1041–9.
Torimoto K, Okada Y, Mori H, Tanaka Y. Relationship between fluctuations in glucose levels measured by continuous glucose monitoring and vascular endothelial dysfunction in type 2 diabetes mellitus. Cardiovascular diabetology. 2013;12:1.
Su G, Mi S, Tao H, Li Z, Yang H, Zheng H, Zhou Y, Ma C. Association of glycemic variability and the presence and severity of coronary artery disease in patients with type 2 diabetes. Cardiovascular diabetology. 2011;10:19.

supTablewakasugietalCD.doc

Download PDF

Journal Publication

published 01 Apr, 2021

Read the published version in BMJ Open Diabetes Research & Care →

Version 1

posted

You are reading this latest preprint version

Associations Between Continuous Glucose Monitoring-derived Metrics and Diabetic Retinopathy and Albuminuria in Patients With Type 2 Diabetes

Status:

Journal Publication

Version 1

Abstract

Background

Methods

Study design

Study Population

Biochemical Tests

Dr And Dn Assessment

Cgm With The Freestyle Libre Pro Device

Statistical Analysis

Results

Relationship Between FLP-CGM–derived Metrics and DR severity

Relationship Between Flp-cgm–derived Metrics And Albuminuria Severity

Discussion

Conclusion

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1