Elevated Copeptin, Arterial Stiffness and Elevated Albumin Excretion in Adolescents with Type 1 Diabetes

Abstract

Objective:

We sought to evaluate copeptin concentrations in adolescents with and without type 1 diabetes (T1D) and examine the associations between copeptin and measures of arterial stiffness and kidney dysfunction.

Research Design and Methods:

This analysis included 169 adolescents with T1D (12-19 years of age, 59% girls, mean HbA1c 9.0±1.5% and diabetes duration of 8.6±2.9 years), in addition to 61 controls without T1D. Arterial stiffness including carotid-femoral pulse wave velocity (CF-PWV), carotid-radial PWV (CR-PWV), augmentation index normalized to heart rate of 75bpm (AIx@HR75) and brachial artery distensibility (BAD). Serum copeptin, urinary albumin-to-creatinine ratio (UACR) and estimated glomerular filtration rate (eGFR) by serum creatinine and cystatin C were also assessed.

Results:

Compared to controls, adolescents with T1D had higher median (Q1-Q3) copeptin (7.5 [5.2-11.3] vs. 6.4 [4.8-8.3] pmol/l, p=0.01), mean±SD eGFR (121±23 vs. 112±16 ml/min/1.73m², p=0.002) and lower BAD (7.1±1.3 vs. 7.2±1.2%, p=0.02). Adolescents with T1D in the in high tertile copeptin group (>9.1 pmol/l) had higher AIx@HR75 (10.7±1.2 vs. 5±1.2, p=0.001), CR-PWV (5.30±1.0 vs. 5.18±1.0 m/s, p=0.04) and UACR (12±1 vs. 8±1 mg/g, p=0.025) compared to those in low tertile (<5.8 pmol/l) after adjusting for age, sex and eGFR. Copeptin inversely associated with CF-PWV independent of age, sex, eGFR, SBP and HbA1c in T1D adolescents.

Conclusions:

Our data demonstrate that elevated copeptin was associated with worse arterial stiffness in adolescents with T1D. These findings suggest that copeptin could improve CVD risk stratification in adolescents with T1D.

Introduction

Cardiovascular disease (CVD) and diabetic kidney disease (DKD) are the leading causes of mortality in type 1 diabetes (T1D).¹ Although overall morbidity and mortality in T1D have improved with therapeutic advancements, people with T1D still have higher rates of CVD events and death compared to the general population.² Children and adolescents (particularly age <10 years) exhibit higher CVD risk, compared to those in whom T1D are diagnosed later in life, and lose 14–18 years of their life expectancy.³ Our group and others have shown that adolescents with T1D have higher arterial stiffness, a strong predictor for future CVD, compared with their peers without diabetes.^4-6 Moreover, glomerular hyperfiltration, an early marker of kidney dysfunction, is common in adolescents with T1D and predicts rapid glomerular filtration rate (GFR) decline and abnormal circadian rhythms of blood pressure, important risk factors of CVD.^7,8 There is a need for modifiable biomarkers in adolescents with T1D to improve risk prediction and serve as novel therapeutic targets to impede the development and progression of CVD and DKD.

A growing body of epidemiological research has demonstrated that arginine vasopressin (AVP) concentrations are increased in adults with T1D and T2D, and may represent a biomarker of cardiorenal risk.^9-11 AVP is co-secreted with copeptin, the c-terminal part of AVP precursor, into the blood in an equimolar ratio.¹² In contrast to AVP, copeptin is stable for days, easily measured and mimics fluctuations in AVP concentration.¹² Several studies have demonstrated relationships between increased copeptin and the development or worsening of CVD and DKD in cohorts of adults with diabetes.^10,11,13 In addition to water reabsorption in the collecting ducts via vasopressin-2 (V2) receptor activity, AVP stimulates V1a receptor at arterial smooth muscles and the liver as well as V1b receptor at hypothalamus, resulting in vasoconstriction, hepatic glucose production and cortisol secretion, thereby hastening cardiorenal complications in diabetes.¹³ Increased AVP also stimulates the renin-angiotensin-aldosterone system, attenuates natriuresis and thereby increases renal oxygen consumption in experimental animal and human models.^14,15 Despite this body of evidence in adult cohorts of patients with diabetes, there are limited data describing the relationships between elevated vasopressin and risk for early CVD and marker of renal dysfunction in adolescents with T1D. Accordingly, we hypothesized that elevated copeptin, as a marker of vasopressin, would associate with markers of CVD and renal dysfunction in adolescents with T1D.

2. Methods

2.1. Participants

In the current analysis, 169 adolescents with T1D and 61 nondiabetic controls who had available serum samples and data on arterial stiffness and markers of nephropathy were included from the Determinants of Macrovascular Disease in Adolescents with Type 1 Diabetes study.^16,17 Briefly, the study was conducted during 2008–2014 to investigate atherosclerotic disease risk in adolescents with T1D. The inclusion criteria relevant to this study’s analysis were 1) ages 12–19 years, 2) diagnosed with diabetes per ADA guideline, 3) positive for islet autoantibody, 4) diabetes duration >5 years and 5) received care at the Barbara Davis Center for Childhood Diabetes. Participants were excluded if they had abnormal cardiac anatomy or arrhythmia. The study was approved by the Colorado Multiple Institution Review Board, and informed consent and assent (for subjects <18 years) were obtained from all subjects.

Physical examination Participants’ Tanner stage were evaluated by the study investigators who are trained pediatric endocrinologists. Height was measured to the nearest 0.1 cm using a stadiometer and standard procedures. Body weight was measured to the nearest 0.1 kg using a Detecto scale (Detecto, Webb City, Missouri). BMI percentile was calculated using the 2000 Centers for Disease Control standard growth charts. After participants laid supine for a minimum of 5 minutes, blood pressure was measured three times using a Dynapulse 5200A (Pulse Metric, San Diego, California), and the 3 measurements were averaged.

2.2. Fasting biochemical analysis

After an 8 hours overnight fast, blood was drawn for serum glucose, HbA1c (DCCT-calibrated ion-exchange HPLC, DCA Vantage, Siemens, Princeton, NJ), BUN, creatinine, total cholesterol, HDL-c, and triglycerides concentrations (enzymatically, Beckman Coulter AU system, Beckman Coulter Inc, Brea, CA). LDL-c was calculated using Friedewald formula. Cystatin C was measured in the University of Colorado Hospital clinical lab using the commercially available Dade-Behring assay following package insert instructions on a BNII or Prospec instrument, as previously described in detail.¹⁸

Copeptin was measured on participants’ stored serum samples using automated ultra-sensitive immunoluminometric assay (B.R.A.H.M.S Copeptin proAVP; Thermo Scientific, Hennigsdorf, Germany) on a KRYPTOR compact PLUS system.¹⁹ The detection limit was 0.9 pmol/l. The intra-assay coefficient variation was <15% and <8% for concentrations ranging from 2.0–4.0 pmol/L and 4.0–15.0 pmol/L, respectively. The interassay coefficient variation was <18% and <10%, respectively, for the lower and higher ranges of copeptin concentrations. Urine samples were collected for creatinine and albumin concentrations and measured with Hitachi 7600 using a turbidimetric immunoassay and an enzymatic method, respectively.

2.3. Arterial stiffness

All participants were asked to refrain from caffeine intake and smoking within 8 hours prior to the study visit. Brachial artery distensibility (BAD) was obtained with a DynaPulse Pathway instrument (Pulse Metric Inc., San Diego, CA). Pulse wave velocity was measured in the carotid-radial (CR-PWV) and carotid-femoral segment (CF-PWV) using applanation tonometry (Sphygmocor Vx, AtCor Medical, Lisle, IL).¹⁶ Augmentation Index was also collected over the right radial artery and corrected for individual heights and heart rates of 75 beats per minute (AIx@HR75). Detailed description of the methods have been previously published.^16,17

2.4. Markers of kidney function

Estimated GFR (eGFR) and urinary albumin to creatinine ratio (UACR) were measured to determine kidney fucntion (i.e. DKD risk). A spot urine sample was collected during the study visit and UACR was measured in duplicate. Albuminuria was measured using double-antibody radioimmunoassay (Diagnostic Products Corp., Los Angeles, CA) in duplicate. eGFR was calculated using the Zappitelli combined serum creatinine and cystatin C equation. Elevated albumin excretion was defined as UACR ≥30 mg/g.

2.5. Statistical analysis

Analyses were performed in SAS (version 9.4). Data distribution was evaluated before statistical analyses. Descriptive statistics were presented as mean±SD or median (Q1, Q3) as appropriate. Participants were also stratified into 3 groups according to tertiles of copeptin concentrations [low (<5.8 pmol/l), mid (5.8–9.1 pmol/l) and high tertile (>9.1 pmol/l)]. Comparisons between group means were performed using student t-test or one-way ANOVA and Chi-square test were used to compare the proportions. Age, sex and eGFR were considered as covariates for tertile comparison. Skewed data were natural log (ln)-transformed prior to regression analyses. Correlations were evaluated using Pearson’s method. Multivariable regression models were fit to evaluate the association between SUMOD and arterial stiffness (PWV, BAD and AIx@HR75) and markers of nephropathy (eGFR and UACR) adjusted for age, sex, SBP, HbA1c and eGFR. We did not perform the multiple comparison adjustment. Therefore, our statistical analysis should be considered hypothesis generating. A logistic regression model was fitted to evaluate the odd ratios of elevated UACR in adolescents with T1D in high vs. low copeptin tertile groups. P-values of <0.05 were considered significant.

3. Results

3.1. Participant characteristics and biochemical profile

Adolescents with and without T1D were similar in age, sex and ethnic distributions and pubertal status (Table 1). Adolescents with T1D had a mean ± SD diabetes duration of 8.6±2.9 years and HbA1c of 9.0±1.5%. Adolescents with T1D had higher BMI percentile (68.0±23.1 vs. 56.2±27.7 percentile, p=0.002), SBP (113±8 vs. 107±8 mmHg, p<0.001) and DBP (69±6 vs. 64±6 mmHg, p<0.001) than controls. Further, adolescents with T1D exhibited higher total cholesterol (161±35 vs. 145±26 mg/dl, p<0.001), LDL-c concentrations (92±28 vs. 80±20 mg/dl, p<0.001) and HDL-c concentration (52±11 vs. 48±9 mg/dl, p=0.02) compared to their peers without diabetes. However, triglyceride concentrations were not different between groups. Median (Q1, Q3) serum copeptin concentrations were significantly higher in the T1D group than controls (7.5 [5.2, 11.3] vs. 6.4 [4.8, 8.3] pmol/l, p=0.01).

Table 1.

Participant characteristics, fasting biochemical profile and measures of arterial stiffness and kidney function

Variables	T1D N=169	Controls N=61	p-value
Age (years)	15.4±2.2	15.3±2.2	0.85
Sex (% female)	59	57	0.76
Race (% non-Hispanic White)	63%	62.3%	0.88
Diabetes duration (years)	8.6±2.9	-	-
Tanner stage	5 (4, 5)	5 (3, 5)	0.83
BMI percentile	68.0±23.1	56.2±27.7	0.002
SBP (mmHg)	113±8	107±8	<0.001
DBP (mmHg)	69±6	64±6	<0.001
HbA1c (%)	9.0±1.5	5.3±0.3	<0.001
HbA1c (mmol/mol)	75±16	34±3	<0.001
Total cholesterol (mg/dL)	161±35	145±26	<0.001
LDL-c (mg/dL)	92±28	80±20	<0.001
HDL-c (mg/dL)	52±11	48±9	0.02
Triglycerides (mg/dL)	77 (72, 82)	75 (68, 83)	0.75
Creatinine (mg/dL)	0.64±0.14	0.70±0.15	0.02
Cystatin C (mg/L)	0.68±0.13	0.70±0.08	0.23
Zappitelli eGFR (ml/min/1.73m2)	121±23	112±16	0.002
UACR (mg/g)	10.1±3.3	9.2±2.7	0.86
Copeptin (pmol/L)	7.5 (5.2, 11.3)	6.4 (4.8, 8.3)	0.01
CF-PWV (m/s)	5.2±0.7	5.1±0.6	0.32
CR-PWV (m/s)	6.9±1.0	6.8±1.0	0.39
AIx@ HR75 (%)	2.0 (−2.3, 10)	1.8 (−7.4, 10)	0.21
BAD (%/mm Hg)	7.1±1.3	7.2±1.2	0.02

Data are presented as mean ± SD and median (Q1, Q3). P-value <0.05 indicates statistical significance.

T1D, type 1 diabetes; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; CF-PWV, carotid-femoral pulse wave velocity; CR-PWV, carotid-radial pulse wave velocity; BAD, brachial artery distensibility; AIx@HR75, augmentation index at heart rate of 75 beat per minute; UACR, urinary albumin to creatinine ratio; eGFR, estimated glomerular filtration rate

3.2. Arterial stiffness and measures of nephropathy

Adolescents with T1D had lower BAD (7.1±1.3 vs. 7.2±1.2 %mmHg, p=0.02) compared with controls, whereas CF-PWV, CR-PWV and AIx@HR75 were not significantly different between the two groups (Table 1). Adolescents with T1D also exhibited higher eGFR (110±20 vs 101±18 ml/min/1.73m², p=0.002), yet no differences were observed in UACR (10.1±3.3 vs 9.2±2.7 mg/g, p=0.86).

3.3. Arterial stiffness and markers of nephropathy in adolescents with T1D regarding copeptin tertiles

Adolescents with T1D were stratified according to tertiles of copeptin concentrations as following; low (<5.8 pmol/l), mid (5.8–9.1 pmol/l) and high tertile (>9.1 pmol/l). Participants in the high tertile had higher least square mean ± SEM AIx@HR75 (10.7±1.2 vs. 5.0±1.2, p=0.001, Figure 1A) and CR-PWV (5.30±1 vs. 5.18±1, p=0.04, Figure 1C) compared with those in the low tertile group after adjusting for age, sex and eGFR. BAD (Figure 1B) and CF-PWV (Figure 1D) were comparable across tertiles. In addition, participants in the high tertile group had higher UACR (12±1 vs. 8±1, p=0.025, Figure 1E) compared to those in the low tertile group, adjusting for age, sex and eGFR. Finally, participants in the high tertile group had 4.29 fold greater odds of having elevated UACR (≥30 mg/g) [95% CI 1.03–17.86, p=0.046) after multivariable adjustments for age, sex and eGFR (Figure 2).

Figure 1: Measures of vascular stiffness and urinary albumin excretion in adolescents with type 1 diabetes regarding tertiles of copeptin concentrations.

Data are shown by groups according to tertiles of copeptin concentrations. Bar graphs and whiskers indicate least-square means (adjusting for age, sex and eGFR) and standard errors of the mean, respectively.

CF-PWV, carotid-femoral pulse wave velocity; CR-PWV, carotid-radial pulse wave velocity; BAD, brachial artery distensibility; AIx@HR75, augmentation index at heart rate of 75 beats per minute; UACR, urinary albumin to creatinine ratio

Figure 2: Logistic regression model predicting elevated albumin excretion in T1D adolescents in high vs. low copeptin tertile groups.

Data are presented as forest plots. Black box indicates reference odds ratio of low copeptin tertile group. Black circle and lines indicate odds ratio and 95% confidence interval of elevated albuminuria in high vs. with low copeptin tertile group.

3.4. Associations between copeptin and clinical characteristics, fasting biochemical parameters and measures of arterial stiffness and kidney function

Copeptin was positively correlated with HbA1c (r=0.26, p=0.001) and CF-PWV (r=0.16, p=0.01) in adolescents with T1D (Table 2). In multivariable regression models, the associations between copeptin and CF-PWV (β±SE: 0.022±0.009, p=0.013) remained significant after adjusting for age, sex, SBP, HbA1c and eGFR (Table 3). However, copeptin did not significantly associated with other measures of vascular stiffness or UACR.

Table 2.

Correlation analysis between serum copeptin concentrations and clinical characteristics, fasting biochemical parameters, markers of arterial stiffness and kidney function

Variables	T1D		Controls
	r	p-value	r	p-value
Significant correlations
HbA1c	0.257	0.001	−0.112	0.40
CF-PWV	0.162	0.01	−0.094	0.48
BAD	−0.081	0.31	−0.391	0.005
eGFR	−0.205	0.008	−0.345	0.007
Non-significant correlations
Age	0.000	0.99	0.015	0.91
Diabetic duration	−0.028	0.71	−	−
BMI	−0.047	0.54	0.016	0.90
SBP	0.109	0.16	0.050	0.70
DBP	0.140	0.07	−0.245	0.06
CR-PWV	0.076	0.35	−0.167	0.23
AIx@HR75	0.105	0.21	−0.059	0.71
UACR	0.086	0.28	−0.168	0.18
HDL-c	0.027	0.73	0.016	0.90
LDL-c	0.071	0.36	−0.165	0.20
Triglycerides	0.100	0.19	−0.005	0.97
TG/HDL	0.097	0.21	−0.020	0.88
CRP	0.035	0.65	0.065	0.62

Table 3.

Associations of natural log (ln)-transformation of copeptin concentrations and markers of arterial stiffness and kidney function in adolescents with type 1 diabetes

Models	β±SE	p-value
Type 1 diabetes
CF-PWV	0.022±0.009	0.010
CR-PWV	0.009±0.015	0.58
AIx@HR75	0.265±0.151	0.08
BAD	−0.002±0.017	0.90
Ln UACR	0.010±0.015	0.49
Controls
CF-PWV	−0.008±0.033	0.81
CR-PWV	0.001±0.051	0.98
AIx@HR75	0.198±0.656	0.76
BAD	−0.070±0.049	0.16
Ln UACR	0.016±0.058	0.78

Age, sex, eGFR, SBP, HbA1c and copeptin were included in the prediction models.

4. Discussion

Our study demonstrated higher copeptin concentrations in adolescents with T1D compared to their peers without diabetes. Furthermore, elevated copeptin concentrations associated with worse central arterial stiffness independent of renal function in adolescents with T1D. These data are consistent with the adult literature in T1D and T2D.

To date, there has only been one study by Schiel et al. that measured copeptin concentrations in adolescents with T1D and found, in contrast to our study, that adolescents with T1D had no differences in copeptin concentrations compared to those without T1D.²⁰ However, the study by Schiel et al. had a smaller sample size and participants’ sex and BMI were not similar across the groups. Furthermore, their participants were younger than ours, and age and pubertal status are well-known factors affecting copeptin concentrations.²¹ In adults, several studies have demonstrated that copeptin concentrations are elevated in people with T1D,¹⁰ T2D,¹¹ and DKD and CVD.²² A growing body of epidemiological research also indicates that high copeptin concentrations predict future CVD and all-cause mortality in adults with diabetes, yet the data are less consistent in the non-diabetes population.^10,23,24 Previous work in adults with T1D has shown that the association of copeptin and CVD event was at least partially dependent on eGFR, UACR and blood pressure.¹⁰ In contrast, our study found that copeptin was associated with arterial stiffness regardless of those factors which were in normal range. These findings suggest that relationships between copeptin, kidney function and CVD may differ in adolescents and adults with T1D.^10,25 It is also important to note that the glycemic control in our cohort was suboptimal with a mean HbA1c of 9.0±1.5%. Poor glycemic control is a well-known risk factor of arterial stiffness in adolescents and adults with diabetes.^26-28 Although our adolescents with T1D had higher copeptin concentrations compared to controls, the study by Shiel et al found no significant differences in copeptin concentrations between participants with and without T1D. This discrepancy may relate to the better glycemic control in their participants with T1D. Therefore, our findings may not be generalized to those with better glycemic control.

In this study, we used multiple measures to assess arterial stiffness because atherosclerosis develops in an unpredictable fashion.²⁹ We found that adolescents with T1D had elevated measures of BAD but no other measures were significantly different. Among adolescents with T1D, worse AIx@HR75 and CR-PWV were observed across copeptin tertiles. However, copeptin was associated predominantly with CF-PWV, which is a marker for central (aortic) arterial stiffness and a robust predictor of CVD.^30,31 These findings may support the role of copeptin as a biomarker for central arterial stiffness and coronary artery disease.

We also found an inverse relationship between copeptin and eGFR in our cohort, which is congruent with previous studies in adolescents and adults with T1D.^10,20,32 This relationship has also been observed in people with chronic kidney disease as well as in general population.^33,34 Accordingly, we included eGFR as a confounder in all our multivariable models. Although we did not observe a significant linear relationship between copeptin and urinary albumin excretion as reported in adults with and without T1D,^10,35 adolescents with T1D whose copeptin levels were in the high tertile had approximately four times higher odds of having elevated albumin excretion (UACR ≥30mg/g). These data are consistent with observations we and others have made in adults.^35,36

The exact pathogenesis underlying the relationships between copeptin and cardiorenal health in adolescents with T1D remains unclear. However, supraphysiological concentrations of AVP are thought to have deleterious impacts on cardiometabolic health.³⁷ T1D complications are impacted by several mechanisms that increase tubular Na⁺ transport and thereby exacerbate renal hypoxia, including hyperfiltration, glucosuria and supraphysiological insulin exposure, and neurohormonal changes including increased AVP and RAAS activity.^8,38-41 Beyond stimulating free water reabsorption, AVP has antinatriuretic properties through its multitude of effects on Na⁺ transport mechanisms including the sodium-potassium-chloride cotransporter (NKCC2), the thiazide-sensitive sodium-chloride cotransporter (NCC), and ENaC.^32,42-45 Further, AVP increases the mRNA expression and subsequent protein abundance of the Na⁺/K⁺ ATPase.^46,47 The net effect of increased Na⁺ transport is increased renal O₂ consumption, since the majority of oxygen consumed by the kidney is used to support the active reabsorption of Na⁺. Circulating AVP can also stimulate systemic/renal vasoconstriction, platelet aggregation and hepatic glucose production (V1a receptor) and cortisol secretion (V1b receptor), resulting in hypertension, microalbuminuria, hyperglycemia and insulin resistance, typical features attributed to the development and progression of CVD and worse kidney function.^13,48 As such, AVP is a potential modifiable risk factor and novel target for future diabetes treatment. Whether increases in daily water intake can impede the development of vascular complications remains controversial. Recent short-term clinical trials have shown that increased water intake can reduce concentrations of plasma copeptin, fasting blood glucose and glucagon in healthy participants.^13,49 Another study also found that increased water consumption resulted in decreased copeptin concentrations in participants with chronic kidney disease, however, no differences in rates of eGFR decline were observed in participants with or without increased water intake.⁵⁰

Mechanisms leading to increased AVP, and thus copeptin, in diabetes are also incompletely elucidated. Our results showed that copeptin positively correlated with HbA1c in adolescents with T1D suggesting that hyperglycemia may contribute to increased copeptin possibly by increasing plasma osmolality or by inducing hypovolemia via glucosuria. However, under hyperglycemic conditions, increased serum sodium, but not osmolarity, is a factor precipitating AVP hyper-secretion.⁵¹ Robert et al. conducted an experimental study in adults with T1D and demonstrated that AVP secretion was increased in response to hypertonic saline infusion compared to healthy participants. On the other hand, hypertonic dextrose infusion resulted in a blunted response of AVP in both groups.⁵² Understanding the mechanism underlying diabetes-induced AVP secretion might be a target for treatment to prevent cardiorenal complication in diabetes.

Our study has important limitations worth mentioning. First, most participants had normal eGFR and UACR which may have limited our ability to identify a relationship between copeptin and kidney dysfunction. Second, we did not adjust for multiple comparison for which reason our data should be considered hypothesis generating.

In conclusion, adolescents with T1D have elevated serum copeptin concentrations compared to their peers without T1D. Additionally, we demonstrate for the first time that high copeptin concentrations were associated with increased arterial stiffness in adolescents with T1D. These data suggest that copeptin could potentially help improve CVD risk stratification in adolescents with T1D. Further long-term studies are needed to confirm the deleterious effect of high copeptin in adolescents with T1D.