Copeptin and renal function decline, cardiovascular events and mortality in type 1 diabetes

Abstract

Background

Plasma copeptin is a surrogate of arginine vasopressin (AVP) secretion and is associated with a risk of renal and cardiovascular disease. We investigated associations between copeptin and renal events, cardiovascular events and mortality in type 1 diabetes (T1D).

Methods

We conducted a prospective cohort study on 658 individuals with T1D from Steno Diabetes Center Copenhagen. Plasma copeptin concentrations and conventional risk factors were assessed at baseline. The five endpoints were traced through national registries and electronic laboratory records.

Results

Baseline mean age was 55 ± 13 years and estimated glomerular filtration rate (eGFR) was 81 ± 26 mL/min/1.73 m². The median follow-up was 6.2 years (interquartile range 5.8–6.7); 123 participants reached a combined renal endpoint [decline in eGFR ≥30%, end-stage kidney disease (ESKD) or all-cause mortality], 93 had a decrease in eGFR ≥30%, 21 developed ESKD, 94 experienced a combined cardiovascular endpoint and 58 died from all causes. Higher copeptin was associated with all endpoints in unadjusted Cox regression analyses. Upon adjustment for baseline eGFR, the associations were attenuated and remained significant only for the combined renal endpoint and decrease in eGFR ≥30%. Results were similar upon further adjustment for other risk factors, after which hazard ratios for the two renal endpoints were 2.27 (95% confidence interval 1.08–4.74) and 4.49 (1.77–11.4), respectively, for the highest versus the lowest quartile of copeptin.

Conclusions

Higher copeptin was an independent risk marker for a combined renal endpoint and decline in renal function. AVP may be a marker of renal damage or a factor whose contribution to renal and cardiovascular risk is partially mediated by renal damage.

INTRODUCTION

Arginine vasopressin (AVP), also called antidiuretic hormone, is crucial for fluid homeostasis. Upon osmotic stimulation, it is released from the posterior pituitary to allow water reabsorption in the collecting ducts of the kidneys and thus concentrates and reduces urine volume. AVP also contributes to vascular tone and the endocrine stress response. AVP is central to disorders of fluid homeostasis, such as diabetes insipidus and syndrome of inappropriate antidiuretic hormone secretion [1]. A growing body of evidence suggests a pathophysiologic role for AVP in the development of diabetes, kidney disease and cardiovascular disease [1, 2].

Several preanalytical- and assay-related factors make quantifying AVP in the blood difficult. In contrast, measuring copeptin in the blood has proven to be robust and accurate. Both derive from the same molecule and are cosecreted and their concentrations in the blood, though not equimolar, are highly correlated. Hence copeptin concentrations in the blood may be used to gauge the levels of AVP secretion [3].

Copeptin appears to be a prognostic marker in acute illness [4], for example, in heart failure [5], myocardial infarction [6, 7] and stroke [8]. Cross-sectionally, higher copeptin is associated with diabetes, albuminuria and lower renal function [9]. Moreover, higher copeptin is an independent risk marker for renal and cardiovascular events as well as mortality in individuals with type 2 diabetes (T2D) [10–15] and in the general population [16, 17]. However, little literature exists with respect to its role in type 1 diabetes (T1D). Only one study of individuals with T1D has found higher copeptin to be associated with incident end-stage kidney disease (ESKD) after adjustment for potential confounders, but not with all-cause mortality or coronary events [18].

In this prospective study we investigate the relationship between plasma copeptin concentration and renal events, cardiovascular events and all-cause mortality in individuals with T1D.

MATERIALS AND METHODS

Participants

Details on the cohort have previously been published [19]. A total of 667 participants with T1D were recruited from the outpatient clinic at Steno Diabetes Center (Copenhagen, Denmark). Baseline examination, including blood sampling, was performed between 2009 and 2011. The cohort was stratified by levels of albuminuria (normo-, micro- and macroalbuminuria). ESKD [chronic kidney disease Stage 5, chronic dialysis, renal transplantation or estimated glomerular filtration rate (eGFR) <15 mL/min/1.73 m²] was an exclusion criterion. Data on baseline plasma copeptin were available for 658 participants.

The study complied with the Declaration of Helsinki. The protocol was approved by the local ethics committee and all participants gave informed consent.

Follow-up

Details on the assessment of endpoints have previously been published [20]. Briefly, all participants were traced with no loss to follow-up in the Danish National Death and Health Registries on 31 December 2016. Information on cause of death was available until 31 December 2015, at which time follow-up ended for endpoints including cause-specific mortality. Participants were also traced with no loss to follow-up in the electronic laboratory records for data on eGFR and urine albumin:creatinine ratio (UACR) obtained at regular outpatient visits.

We considered five endpoints: a combined renal endpoint comprising a decrease in eGFR ≥30%, incident ESKD or all-cause mortality, whichever occurred first; a decrease in eGFR ≥30%; incident ESKD; a combined cardiovascular endpoint comprising cardiovascular mortality, ischemic heart disease, nonfatal myocardial infarction, nonfatal stroke, coronary interventions or peripheral arterial interventions including amputations, whichever occurred first; and all-cause mortality.

In addition, yearly changes in eGFR and UACR (slopes) were calculated among 509 and 507 participants, respectively, with at least two measurements and a minimum follow-up duration of 3 years. Calculations were based on all available measurements from outpatient visits during follow-up.

Laboratory analyses

Copeptin concentration was measured in baseline plasma samples stored at −80°C using an automated immunoassay (copeptin US KRYPTOR assay performed on the KRYPTOR compact PLUS; BRAHMS, Hennigsdorf, Germany) whose analytical performance has previously been deemed reliable [21].

Details on the measurement of other variables have previously been published [19]. Briefly, eGFR was calculated with the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [22]. At baseline, urine albumin excretion rate (UAER) was measured in 24-h urine collections by enzyme immunoassay. During follow-up, the UACR was measured by enzyme immunoassay. For classification into categories (normo-, micro- and macroalbuminuria), 24-h UAER and UACR have previously been demonstrated to be highly concordant [23].

Statistical analysis

The nonnormally distributed variables (copeptin and UAER) are summarized as medians with interquartile ranges (IQRs) and log transformed in all analyses. All other continuous variables are given as mean ± standard deviation (SD) and categorical variables are reported as percentages.

We divided the population according to sex-specific quartiles of copeptin. Baseline characteristics were compared between the four groups with analysis of variance (ANOVA) for continuous variables and the χ² test or Fisher’s exact test, as appropriate, for categorical variables.

Kaplan–Meier functions and log-rank test were applied to compare risks across sex-specific quartiles of copeptin. Associations between copeptin and the respective endpoints were analyzed with Cox proportional hazard regression models with copeptin analyzed in sex-specific quartiles and as a continuous variable per SD increase in log2-transformed copeptin (1.26 pmol/L). However, analyses in sex-specific quartiles were not possible for ESKD due to the small number of events (Q1 = 0, Q2 = 1, Q3 = 3 and Q4 = 19). For all endpoints, analyses were unadjusted (Model 1); adjusted for baseline eGFR (Model 2) and adjusted for baseline eGFR, age, diabetes duration, body mass index, low-density lipoprotein (LDL) cholesterol, smoking, hemoglobin A1c (HbA1c), systolic blood pressure, UAER, retinopathy status and history of cardiovascular disease (Model 3). Analyses of copeptin as a continuous variable also included adjustment for sex in Model 3. The Cox model assumption of proportional hazards was tested using cumulative sums of martingale residuals and was fulfilled for all outcomes. Moreover, the assumption of linearity of the logarithm of copeptin was fulfilled for all outcomes.

A general linear model was applied for calculating the eGFR and UACR slopes (yearly changes) for each individual using multiple follow-up measures. For the analyses of yearly changes in eGFR and UACR we applied a linear regression model and calculated the β estimates per log SD increase in copeptin using three models of adjustment as described above.

Statistical analyses were performed with SAS software (version 9.4; SAS Institute, Cary, NC, USA) except for the eGFR and UACR slope analyses and the scatter and Kaplan–Meier plots, which were executed in the R statistical platform (R Foundation for Statistical Computing, Vienna, Austria; https://www.r-project.org/). A two-tailed P-value <0.05 was considered significant.

RESULTS

Baseline characteristics are provided in Table 1. In the total population, the mean age was 55 ± 13 years, duration of diabetes was 33 ± 16 years, 56% were male, eGFR was 81 ± 26 mL/min/1.73 m² and median UAER was 17 mg/24 h (IQR 8–68). For the higher sex-specific quartiles of copeptin, as shown in Table 1, the baseline duration of diabetes, body mass index, HbA1c, UAER, history of cardiovascular disease, presence of retinopathy, use of antihypertensives and of renin–angiotensin–aldosterone system (RAAS) blockers was generally higher and eGFR was lower (P for trend between quartiles <0.05). Figure 1 illustrates the negative correlation between baseline copeptin and eGFR for men and women.

The median follow-up was 5.2 years (IQR 4.7–5.7) for the combined renal endpoint, 5.3 (2.7–6.2) for a decrease in eGFR ≥30%, 5.3 (4.8–5.7) for ESKD, 5.1 (4.7–5.6) for cardiovascular events and 6.2 (5.8–6.7) for all-cause mortality.

A total of 123 participants experienced the combined renal endpoint, 93 had a decrease in eGFR ≥30%, 21 developed ESKD, 94 reached the combined cardiovascular endpoint and 58 died from all causes. For the combined renal endpoint, the contribution of decline in eGFR ≥30%, ESKD and all-cause mortality was 76, 10 and 37 events, respectively. Kaplan–Meier plots (Figure 2A–D) illustrate across sex-specific quartiles of copeptin the cumulative survival probability for all endpoints, except ESKD due to the small number of events.

Tables 2 and 3 display the hazard ratios (HRs) per increase in sex-specific copeptin quartile and per log SD increase in copeptin, respectively, in all three models for all endpoints (though only per log SD increase for ESKD). Higher baseline copeptin was associated with all endpoints in Model 1. Upon adjustment for baseline eGFR in Model 2, however, the associations were attenuated and remained significant only for the combined renal endpoint and a decline in eGFR ≥30%. Results were similar after further adjustment for other risk factors (Model 3) in which HRs for these two renal endpoints were 2.27 [95% confidence interval (CI) 1.08–4.74] and 4.49 (1.77–11.4), respectively, for the highest versus the lowest quartile. As displayed in Table 4, higher copeptin was associated with a steeper yearly decline in eGFR in Models 1 and 2, but the association lost significance in Model 3. There were no associations with yearly change in UACR.

Additional analyses

Significant associations persisted when baseline treatment with RAAS blockers and diuretics was added to Model 3 for the combined renal endpoint and decline in eGFR ≥30%.

We also analyzed the association between copeptin as a continuous variable and incident hospitalization for nonfatal or fatal heart failure (International Classification of Diseases, Revision 10 code I50). Higher copeptin was associated with a higher risk of heart failure (27 events) in Model 1 (P < 0.001), but significance was lost after adjustment for eGFR (Model 2, P = 0.53). As an additional renal endpoint, we analyzed the association of copeptin with progression from normo- to microalbuminuria or from micro- to macroalbuminuria (defined as progression in classification based on two of three consecutive urine collections during follow-up). Copeptin was not associated with the risk of progression in albuminuria (36 events) in any of the models (P ≥ 0.73).

DISCUSSION

We investigated plasma copeptin as a risk marker for renal and cardiovascular complications in T1D. The key findings can be summarized as follows: copeptin was independently associated with a combined renal endpoint (decline in eGFR ≥30%, incident ESKD or all-cause mortality) and a decline in eGFR ≥30%; and copeptin was associated with incident ESKD, a combined cardiovascular endpoint and all-cause mortality, but not independent of baseline eGFR. This may reflect AVP as a marker of renal damage. It might also reflect AVP as a factor whose contribution to renal and cardiovascular risk is (partially) mediated by renal damage, though we cannot preclude that the associations between AVP or copeptin and the endpoints were confounded by unknown factors.

To our knowledge only one previous study has investigated the associations of copeptin with renal and cardiovascular events as well as mortality in T1D. In that study, participants with T1D (mean duration 27.2 years) from two separate cohorts were followed for a median of 10 (n = 218) and 5 (n = 518) years. Higher copeptin was associated with incident ESKD (n = 47), coronary events (n = 67) and all-cause mortality (n = 67) after adjustment for cohort, sex, age and diabetes duration. Baseline urinary albumin and eGFR were responsible for much of the attenuation upon further adjustment and a significant association remained only for incident ESKD. Also, higher copeptin was associated with a steeper yearly decrease in eGFR in adjusted analyses including baseline eGFR and urinary albumin [18]. This was slightly dissimilar to our findings in which higher copeptin was associated with a steeper yearly decrease in eGFR after adjustment for baseline eGFR but lost significance, albeit barely (P = 0.059), upon adjustment for several other risk factors.

In other populations, independent associations between copeptin and adverse renal outcomes have been reported. One recent study, including 13 597 individuals from three large community-based cohorts, demonstrated that higher copeptin was associated with incident chronic kidney disease, ‘certain drop in eGFR’ (Kidney Disease: Improving Global Outcomes criterion) and rapid kidney function decline. Of note, however, baseline albuminuria was not included in the adjusted analyses. In two of the cohorts, higher copeptin was also associated with incident UACR ≥30 mg/g [16].

Two cohorts of individuals with T2D (n = 1407 and n = 3098) were examined in another study. Higher copeptin was associated with a rapid decrease in eGFR as well as doubling of creatinine or incident ESKD [12]. In a third study, higher copeptin was associated with a faster decrease in eGFR in a subset of 756 individuals with T2D not treated with RAAS inhibition [10].

With respect to cardiovascular and mortality endpoints, independent associations with higher copeptin have been demonstrated in three population-based cohorts [14, 15, 17], though only in the subgroups with diabetes in two of these [14, 15]. Such associations were also found in the three cohorts of individuals with T2D described above [12, 13].

Thus independent associations between higher copeptin and renal function decline seem to be present in T1D, T2D and the general population. This accords with our findings. In our study, higher copeptin was also associated with cardiovascular events and mortality, albeit not independently of eGFR, as seems to be the case in T2D. Importantly, the lack of adjustment for baseline eGFR in two of the three population-based studies [15, 17] and albuminuria in all three population-based studies may constitute shortcomings. Whether it is important to adjust for eGFR and albuminuria in analyses of the general population is uncertain, but it is noteworthy that higher copeptin is also associated with lower renal function [14, 16, 24] and higher albuminuria [16, 24] in the general population.

The mechanisms for clearing copeptin and AVP from the circulation are uncertain. Plasma concentrations of both are known to be higher with lower renal function, and one study identified a breakpoint at an eGFR of 28 mL/min/1.73 m² below which the ratio of copeptin to AVP became increasingly higher. Thus renal clearance may play a role, and caution is warranted when interpreting copeptin as a surrogate of AVP at higher stages of chronic kidney disease [25, 26]. However, the inverse association between eGFR and copeptin may also be attributable to a diminished urine concentration capacity. This seems to accompany kidney disease, although it is not inevitably associated with renal function in persons without kidney disease [27] and might in turn elicit a compensatory higher secretion of copeptin and AVP [25, 28, 29]. One possible explanation, among others, for a diminished urine concentration capacity in kidney disease (and with advanced age) might be a blunted response to AVP in the collecting ducts, perhaps due to a lower abundance of the V2 receptors through which AVP exerts its effects on water reabsorption [29, 30].

Other mechanisms may mediate associations of copeptin with renal and cardiovascular risk. Experimental studies (primarily from murine models) have demonstrated that AVP can induce glomerular hyperfiltration and increase albuminuria [31]. AVP can affect glucose control unfavorably via receptors in the liver and pancreas and its stimulation of the cortisol axis [32]. AVP is also integral to the control of both systemic and local circulation, for example, in the kidneys, and closely linked to the RAAS [33].

Beyond being a risk marker, AVP may be a target for intervention, as its secretion can be modified by changes in water intake and its effects interrupted by vasopressin receptor antagonists [34]. Observational studies indicate that indices of higher hydration associate with better renal function [9]. Nevertheless, effects on renal function and albuminuria were not detected in a trial comparing high versus low water intake for 1 year in 631 individuals with an eGFR of 30–60 mL/min/1.73 m² and albuminuria. It was noted that 1 year of follow-up may not have been sufficient [35].

As to the strengths and limitations of our study, the main strengths are our well-characterized cohort permitting thorough adjustments and our complete follow-up ensured by the comprehensive registries in Denmark and frequent and regular outpatient visits.

A decrease in eGFR ≥30% may be considered a ‘soft’ renal endpoint but has been shown to be sturdily associated with the risk of ESKD and mortality [36]. Our combined renal endpoint may also be controversial, as it includes all-cause mortality, but this is in line with a five-point major adverse renal endpoint recently proposed [37]. Although cardiovascular endpoints were defined differently among the studies, it seems that cardiovascular events were less frequent in our study than in the T2D cohorts described, despite similar follow-up length. This may be ascribed in part to a higher baseline age and prevalence of cardiovascular disease in the T2D cohorts when compared with ours and the other described T1D cohorts [18]. The T2D cohorts were also generally larger. A lack of statistical power may therefore have been an issue, although in two of the population-based studies the subgroups with diabetes were smaller than our population [14, 17].

Another potential limitation is that copeptin was only measured once and at different times of the day. Although hydration status may affect copeptin levels [38], a substantial intraindividual variation from day to day has not been demonstrated and the circadian variation appears to be minor [39]. In one study, a high intraindividual reproducibility for plasma copeptin was demonstrated over a 3-week interval [40].

In conclusion, plasma copeptin seems to be an independent risk marker for adverse renal outcomes in T1D. It is also a risk marker for adverse cardiovascular outcomes and mortality, but not independent of baseline eGFR.

SUPPLEMENTARY DATA

Supplementary data are available at ndt online.

FUNDING

This work was supported by internal funding from the Steno Diabetes Center, Copenhagen, Denmark.

AUTHORS’ CONTRIBUTIONS

The contributions of each author are described using the Contributor Roles Taxonomy (CRediT). N.S.H. was responsible for writing the original draft. S.T., J.L.J. and F.P. were responsible for conceptualization. N.S.H. and S.T. were responsible for project administration. S.T., T.S.A., T.W.H. and J.P.G. were responsible for the methodology. J.L.J. and P.R. were responsible for funding acquisition and resources. S.T., F.P., T.W.H. and P.R. were responsible for supervision. N.S.H. and T.S.A. were responsible for software. N.S.H., S.A.W., N.T., T.S.A. and T.W.H. were responsible for the formal analysis. S.A.W. and N.T. were responsible for data curation and validation. S.T., S.A.W., N.T. and F.P. were responsible for the investigation. N.S.H. and T.W.H. were responsible for visualization. All authors were involved in the writing, review and editing the manuscript.

CONFLICT OF INTEREST STATEMENT

N.S.H., S.T., T.W.H., T.S.A. and P.R. own stock in Novo Nordisk A/S. N.S.H. owns stock in Akcea Therapeutics. T.S.A. owns shares in Zealand Pharma A/S. F.P. has received research grants from AstraZeneca, Novo Nordisk and Novartis; lecture fees from MSD, AstraZeneca, Novo Nordisk, Novartis, Eli Lilly and Boehringer Ingelheim; and served as a consultant for AstraZeneca, Novo Nordisk, Amgen and MSD. P.R. has received fees (to his institution) for consultancy and/or speaking from Astellas, AstraZeneca, Bayer, Boehringer Ingelheim, Gilead, Eli Lilly, MSD, Mundipharma, Novo Nordisk and Sanofi Aventis and research grants from AstraZeneca and Novo Nordisk. The other authors report no potential conflicts of interest.

DATA AVAILABILITY STATEMENT

The data underlying this article will be shared upon reasonable request to the corresponding author assuming relevant ethics and data protection agreements are in place.

REFERENCES