Introduction
According to the Diabetes Control and Complications Trial (DCCT) and the UK Prospective Diabetes Study (UKPDS) [
1,
2], hemoglobin A1C is the gold standard for assessing glycaemic control for decades and is strongly associated with the risk of long-term diabetic complications. However, as an indirect measure of average blood glucose levels over three months, HbA1c is considered to have some limitations [
3]. HbA1c does not provide details about hypoglycaemia or hyperglycaemia and does not reflect glycemic variability [
4]. HbA1c also has limitations in interpreting the risk of chronic complications of diabetes. For instance, HbA1c accounted for only 11% of the variation in risk of diabetic retinopathy observed in the Diabetes Control and Complications Trial (DCCT) [
5].
Continuous glucose monitoring (CGM) provides an accurate reflection of an individual’s blood glucose status throughout the day. Compared to glycosylated haemoglobin, CGM technology can better reflect blood glucose variability. As a CGM-derived indicator, ‘time in range’ (TIR) is simple, intuitive and responsive to treatment and lifestyle changes, and has become a key indicator for assessing glycaemic control. It is negatively correlated with glycated haemoglobin and the American Diabetes Association 2021 guidelines [
6] stated that it can be used to assess glycaemic control and might be an acceptable endpoint for future clinical trials. Several different studies [
7] have reported that time range (TIR) is associated with the risk of microvascular complications and can predict the risk of future diabetic complications. Furthermore, according to the study [
8], average blood glucose levels at night, but not daytime blood glucose values or glucose variability, were independently associated with the degree of vascular remodelling. This suggests that nighttime blood glucose may be a better indicator of the risk of diabetic complications than full-day blood glucose status. Circadian rhythms have received increasing attention in recent years with the award of the 2017 Nobel Prize to Young MW et al. [
9] for their discoveries of molecular mechanisms controlling the circadian rhythm. Circadian rhythms also exist for various hypoglycemic hormones and baseline blood glucose levels, and studies have shown that baseline pre-meal glucose levels in animals and healthy humans show circadian rhythms under a regular light/dark cycle, with a trough during sleep and a peak during wakefulness [
10‐
12]. Nevertheless, the clinical significance of circadian rhythms in blood glucose is not clear.
Diabetic kidney disease (DKD) is a common chronic microvascular complication of diabetes that is now a major cause of CKD and end-stage renal disease, manifested mainly by a urinary albumin/creatinine ratio (UACR) ≥ 30 mg/g and/or an estimated glomerular filtration rate (eGFR) < 60 ml-min -¹-(1.73 m²)-¹ that persists for more than 3 months. In early screening, a random urine measurement of UACR is recommended to reflect urinary albumin excretion. TIR was found to be strongly associated with albuminuria in type 2 diabetes [
13]. Furthermore, respective study found that each 10% treatment-induced increase in TIR was associated with 18% reduction in albuminuria in patients with T1D [
14]. However, there is little report on the impact of nocturnal TIR and albuminuria in T2D.
The aim of this work was to investigate whether TIR measured by CGM, especially nocturnal TIR and hypoglycaemic events is related to the presence and severity of albuminuria and decrease of eGFR in type 2 patients with diabetes.
Discussion
In our cohort of 823 T2D patients, the prevalence of albuminuria decreased with increasing TIR quartiles. Binary logistic regression revealed that TIR as well as nocturnal TIR was obviously related to the presence of albuminuria after adjusting for age, diabetes duration, sex, BMI, lipid profile, blood pressure, HbA1c (%), SD, MAGE, CV and ADDR. However, after adjustment of M-value, the link between albuminuria and TIR was weakened, the association between the presence of albuminuria nocturnal TIR was still significant. Multiple regression found the relationship between TIR and severity of albuminuria. However, nocturnal TIR was obviously related to the severe stage of albuminuria regardless of SD, MAGE, CV and ADDR. In our study, eGFR was not significantly associated with TIR, but was significantly associated with the number of hypoglycemic events and the number of nocturnal hypoglycemic events.
As the gold standard for assessing glycaemic management, hemoglobin A1c (HbA1c) is considered to have a strong correlation with the microvascular complications of diabetes, including diabetic nephropathy [
16,
17]. However, HbA1c does not reflect information on hyperglycaemia, hypoglycaemia and fluctuations in blood glucose, nor does it reflect the magnitude and frequency of intra- and inter-day glucose changes [
18,
19]. In addition, the measurement of glycated haemoglobin can be affected by specific conditions including kidney insufficiency, anaemia, pregnancy, haemoglobinopathies and iron deficiency [
20‐
23]. Due to the shortcomings of HbA1c, CGM-derived indicators, especially TIR has become an alternative marker of glycemic control over the last few years [
24,
25].
There is insufficient evidence in previous studies to demonstrate the association of glycaemic variability with diabetic nephropathy in T2DM. The study by S.-M. Jin et al. did not find an independent association between glycaemic variability and the degree of proteinuria in patients with type 2 diabetes [
26]. Subramanian S et al. suggest that glycaemic variability may be a factor in the development of DKD, but clear evidence is lacking [
27]. Wakasugi S et al. demonstrated that FLP-CGMderived metrics related to intraday and interday glucose variability, including TIR, SD, MAGE and MODD, were significantly associated with albuminuria severity, and these associations remained significant after adjustment for HbA1c [
28]. Our study also found that TIR was obviously related to the presence of albuminuria after adjusting for glycaemic variability Indicators including SD, MAGE, CV and ADDR. However, in contrast to their results, multiple regression analysis found a weaker relationship between TIR and the stage of albuminuria severity.
A series of studies have demonstrated a correlation between CGM indicators, represented by TIR, and the development or progression of albuminuria in T2DM. Yoo JH et al. demonstrated that CGM-derived TIR significantly associated with the risk of albuminuria, even after adjusting for various confounding factors including CV. With a 10% increase in TIR, the risk of albuminuria was reduced by 6% [
9]. However, after further adjustment for HbA1c, the study did not show a significant association between TIR and albuminuria. In contrast to their results, we found that TIR as well as nocturnal TIR was obviously related to the presence of albuminuria even after adjustment for HbA1c. In the study by Varghese JS, participants were divided into three groups based on glycemic profile (‘TIR profile’, ‘hyper’ and ‘hypo’). Compared with ‘TIR profile’, both ‘hyper’ and ‘hypo’ profiles had higher odds of macroalbuminuria and higher odds of diabetic kidney disease [
29]. TIR was demonstrated to be associated with UACR, DPN and T2DM duration in the multicentre, prospective cohort study by Kuroda N et al. [
30]. Consistent with those fndings, we found that the prevalence of albuminuria decreased as the TIR quartiles increased. The independently negative association between TIR and the presence of albuminuria exists even after adjusting for several risk factors including HbA1c.
There is little previous research on the relationship between nocturnal blood glucose and diabetic complications. A study of T1DM showed that reduced nocturnal TIR was more closely associated with function of sudomotor nerves function of sudomotor nerves [
31]. In our study, nocturnal TIR showed a better correlation with albuminuria compared to TIR. Binary logistic regression analysis showed that TIR as well as nocturnal TIR were significantly correlated with the presence of albuminuria, after adjustment for a range of indicators. However, after further correction for M-value, only nocturnal TIR remained significant. Furthermore, multiple regression analysis showed that TIR did not correlate significantly with the severity of albuminuria, while nocturnal TIR correlated significantly, even after adjustment for a number of indicators including HbA1c. One possible explanation is attributed to the effect of growth hormone secreted at night on the kidneys. Growth hormone (GH) is widely used to treat short stature in children, including children with chronic kidney disease (CKD). GH-excess can affect kidney health by causing glomerular hyperfiltration, hypertrophy, and glomerulosclerosis [
32]. GH-excess is also an important promoter of diabetes nephropathy in T1DM patients. Studies in patients with T1DM showed that urinary GH and IGF-1 levels were related to microalbuminuria in patients [
33]. In addition, GH also has a renal protective effect on humans. Studies have shown that GH therapy can protect cisplatin-induced nephropathy in rats [
34]. To sum up, GH secreted at night has an impact on blood glucose and kidneys, and the nocturnal TIR can better include its impact on blood glucose, which may explain the better correlation between nocturnal TIR and albuminuria. In addition, cortisol as a circadian hormone is not easy to ignore. a study by Roy et al. [
35] found a tendency for elevated cortisol secretion in patients with diabetic retinopathy or diabetic cardiovascular complications. Chiodini I et al. [
36] showed that the degree of midnight cortisol secretion was directly related to the presence and number of complications of type 2 diabetes, including DKD. In contrast, other parameters of cortisol secretion were not significantly correlated with the presence or number of diabetic complications. We can therefore speculate that nocturnal TIR could incorporate the effect of nocturnal cortisol levels on blood glucose and therefore correlate better with diabetic nephropathy. In addition, diabetes significantly affects melatonin secretion levels at night. Hikichi et al. [
37] compared melatonin secretion at night and during the day in non-diabetic and diabetic subjects and found that melatonin levels were lower at night in patients with diabetes, but not during the day as affected by diabetes. The study by Baris Afsar et al. [
38] suggests that melatonin activates cardiovascular system and renal receptors to protect DN in preclinical models. Considering that elevated blood glucose at night may inhibit melatonin secretion, thus depriving the kidneys of the protective effect of melatonin, the association between nocturnal TIR and proteinuria is more significant compared to full-day TIR.
Unlike albuminuria and TIR, our study found no significant correlation between eGFR and TIR. In recent years, the role of the renal tubules has received increasing attention in studies on the pathogenesis of DKD. As a traditional indicator of kidney disease, the increase of microalbuminuria or creatinine based glomerular filtration rate (eGFR) may occur long after the decline of renal function [
39]. Some biomarkers of proximal renal tubular injury, such as kidney injury molecule 1 (KIM-1), N-acetyl-b-D-glucosaminidase (NAG), and liver fat acid binding protein (L-FABP), appear abnormal before proteinuria, and may become new biomarkers for DKD prediction [
40]. At present, it is believed that the mechanism of albuminuria is the excessive inflammatory reaction of proximate tubular epithelial cells (PTECs) under the condition of diabetes [
41,
42]. Considering the central role of renal tubules rather than glomeruli in the occurrence and development of early stage DKD, it can be explained why TIR, as a blood glucose control indicator, has a good correlation with albuminuria but not with eGFR. A notable result was that hypoglycaemic and nocturnal hypoglycaemic events were negatively correlated with eGFR. According to the study by Khanimov I [
43], eGFR was strongly associated with an increased incidence of hypoglycaemia during hospitalization in non-critically ill patients. However, this study also showed that renal function was a strong predictor of hypoglycaemia regardless of the presence of diabetes, implying that hypoglycaemic events may not be specific for the prediction of diabetic nephropathy.
Our study had several limitations that should be noted. Firstly, this is a retrospective, single-centre study and it can only describe the correlation, not the causal relationship. Secondly, all subjects received CGM for 72 h, which may not be representative of overall glucose status. Some glucose-raising hormones that may play a role in the development of diabetic complications such as cortisol, growth hormone and melatonin were not measured.
In conclusion, our study revealed that TIR and nocturnal TIR is associated with the presence and severity of albuminuria independent of HbA1c and GV metrics. Nocturnal TIR shows better correlation than TIR. Hypoglycemia and nocturnal hypoglycemia events were negatively correlated with eGFR, while TIR was not correlated with eGFR. The role of TIR and nocturnal TIR in the evaluation of diabetes kidney disease should be emphasized.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.