Plasma copeptin, a surrogate for vasopressin, was associated with albuminuria in population-based studies. These associations are consistent with the effect of vasopressin on albuminuria observed in humans and rodents. The objective of this study was to determine whether plasma copeptin is an independent marker of risk of renal events in people with type 2 diabetes and albuminuria.
We studied 3,101 participants of the DIABHYCAR trial (6-year follow-up) with type 2 diabetes and albuminuria. A renal event was defined as doubling of serum creatinine or development of end-stage renal disease.
During follow-up, 86 renal events occurred in 76 subjects (2.45%). Incidences by tertiles of baseline plasma copeptin were 1.06% (T1), 1.45% (T2), and 4.84% (T3). They were 2.43% (T1), 5.11% (T2), and 11.81% (T3) for the subset of subjects with macroalbuminuria at baseline (n = 729). Hazard ratio for plasma copeptin tertiles as a risk for renal events was 4.79 (95% CI, 2.48–9.24; P < 0.0001; for T3 vs. T1). In a stepwise regression analysis, urinary albumin excretion and plasma copeptin remained positively associated and HDL cholesterol and estimated glomerular filtration rate were inversely associated with the incidence of renal events. These independent predictors explained ∼18% of the variance of the outcome. The yearly variations of estimated glomerular filtration rate by copeptin tertiles were −1.43 ± 0.51 (T1), −2.29 ± 0.49 (T2), and −3.52 ± 0.44 mL/min/1.73 m2 per year (T3) (P = 0.005) in subjects with macroalbuminuria.
Plasma copeptin may help to identify subjects with diabetic chronic kidney disease who are at high risk for renal function decline.
Type 2 diabetes is a major cause of kidney and cardiovascular diseases, and screening for albuminuria and proteinuria is recommended for stratification of risk of these comorbidities (1). Care of patients with diabetes and albuminuria improved markedly in the past decades. However, as shown in recent randomized trials, even with optimal treatment up to 20% of patients with diabetes and proteinuria develop end-stage renal disease (ESRD) within a 3-year follow-up (2,3). There is still an important unmet therapeutic need and many innovative therapeutic options have been or are being evaluated. Some of these new strategies were disappointing because of severe side effects (4). Because not all patients with type 2 diabetes and proteinuria will progress to ESRD, the identification of those at higher risk would improve the benefit-to-risk ratio of intensification of treatment.
Copeptin, the stable COOH-terminal portion of the precursor of vasopressin, is an easily measurable surrogate marker of vasopressin. Studies in healthy subjects have shown plasma copeptin and vasopressin concentrations to correlate strongly over a wide range of osmolalities (5,6). We hypothesized that vasopressin might contribute to the progression of kidney damage and, thus, that plasma copeptin could be a good candidate for the identification of subjects at high risk for progression of nephropathy. Vasopressin is elevated in diabetes (7), and copeptin was associated with albuminuria in a recent, cross-sectional, population-based study (8). Moreover, infusion of DDAVP (a selective vasopressin V2 receptor agonist) was shown to increase urinary albumin excretion (UAE) in healthy subjects and in normal rats (9), and treatment with a selective vasopressin V2-receptor antagonist prevented the increase in albuminuria in rats with streptozotocin-induced diabetes (10).
The aim of the present investigation was to evaluate whether vasopressin could represent a risk factor for the decline of kidney function in patients with diabetic nephropathy. We examined the association of plasma copeptin with the risk of doubling plasma creatinine concentration and the risk of ESRD in subjects with type 2 diabetes and microalbuminuria or macroalbuminuria from the prospective DIABHYCAR study (11,12).
RESEARCH DESIGN AND METHODS
We studied 3,101 unrelated French type 2 diabetic patients from the DIABHYCAR cohort. DIABHYCAR was a 6-year clinical trial conducted in men and women with type 2 diabetes selected on the basis of persistent microalbuminuria (UAE, 20–200 mg/L) or macroalbuminuria (UAE >200 mg/L) without renal failure (plasma creatinine <150 µmol/L) at baseline. The trial tested whether a low dose of ramipril, an ACE inhibitor able to reduce UAE, also would reduce cardiovascular and/or renal events such as myocardial infarction, stroke, acute heart failure, ESRD, and cardiovascular death. For the purpose of the original DIABHYCAR trial, a renal event was defined as the doubling of serum creatinine levels or the requirement of hemodialysis or renal transplantation (ESRD) during follow-up. An independent committee reviewed all case records to validate selection criteria, to grade the renal involvement of each patient, and to adjudicate the clinical events during follow-up (11). Results were negative regarding the drug effect and were published previously (11,12). Participants gave written informed consent and study protocols were approved by the Ethics Committee of Angers University Hospital.
Copeptin concentration was measured in fasting plasma-EDTA samples, collected at baseline, and kept frozen at −80°C. An automated immunofluorescent sandwich immunoassay was used (BRAHMS Copeptin US KRYPTOR CT-proAVP; Thermofisher Scientific, Hennigsdorf, Germany) (13). Urinary albumin was measured by nephelometry (14). Estimated glomerular filtration rate (eGFR) was calculated according to the Modification of Diet in Renal Disease formula (15). Variation of eGFR during the study was computed as the difference between values at the end of follow-up and at baseline divided by the duration of follow-up.
Results are expressed as mean ± SD unless otherwise stated. Tertiles of plasma copeptin concentration were computed separately for women and men to take into account sex-related differences in copeptin levels. Differences between groups were assessed by Pearson χ2 test, Wilcoxon/Kruskal-Wallis test, ANOVA, and ANCOVA. When ANOVA or ANCOVA were significant, comparisons between pairs were made using the Tukey-Kramer honestly significant difference test. Kaplan-Meier curves were used to plot survival (renal event-free) rates over time by sex-specific tertiles of plasma copeptin. Cox proportional hazards survival regression analyses were used to examine the effect of explanatory variables on time-related survival (renal event-free) rates in prospective analyses. Hazard ratios with their 95% CIs were computed in these analyses. Competing risk regression analysis (Fine and Gray model) was performed to estimate subhazard ratios of risk factors assuming death as a competing risk (16). Adjustments for clinical and biological parameters were performed by including these parameters as covariates in the regression models. Interaction between plasma copeptin tertiles and sex or study treatment (randomization group in the original DIABHYCAR study: ramipril vs. placebo) on the risk of renal events or in the comparison of eGFR decline were assessed by including in the regression model (Cox or ANCOVA) a “crossed” compound covariate (copeptin tertiles/sex or copeptin tertiles/study treatment). C-statistics were used to compare areas of receiver operating characteristic (ROC) curves. Data were log-transformed for the analyses when the normality of the distribution was rejected by the Kolmogorov-Smirnov-Lilliefors goodness-of-fit test. Statistics were performed with JMP (SAS Institute, Cary, NC) and Stata (StataCorp, College Station, TX) software. P < 0.05 was considered significant.
Clinical characteristics at baseline
Characteristics of participants at baseline by plasma copeptin tertiles and by sex are shown in Table 1. For both men and women, subjects in the highest tertile were older, had higher HbA1c levels, had lower eGFR, and more often had arterial hypertension and proteinuria. Renal events during follow-up comprised 67 cases of doubling of serum creatinine and 19 cases of ESRD in 76 subjects (2.45%). Individuals who presented a renal event, as compared with those who did not, at baseline had a longer duration of diabetes, and higher copeptin, HbA1c, and triglyceride levels, and had lower HDL cholesterol levels. They had lower eGFR, higher UAE, more often had arterial hypertension and proteinuria, and more often were treated by diuretics. A history of myocardial infarction was more frequent in individuals who presented a renal event (Supplementary Table 1).
Plasma copeptin and renal events during follow-up
The incidences of renal events during follow-up by sex-specific tertiles of plasma copeptin were 1.06% (T1), 1.45% (T2), and 4.84% (T3) for the whole population (n = 3,101). They were 2.43% (T1), 5.11% (T2), and 11.81% (T3) for the subset of subjects with macroalbuminuria at baseline (n = 729). Cox proportional hazards survival regression analyses showed a positive association of the highest tertile of plasma copeptin with the incidence of renal events in the whole population, as well as in the subset of subjects with macroalbuminuria at baseline (Fig. 1 and Table 2, model 1). This association remained significant when adjusted for baseline UAE and eGFR (Table 2, model 2). The highest tertile of plasma copeptin was associated with the incidence of both outcomes comprising the renal events (doubling of plasma creatinine and ESRD) in subset analyses stratified by outcomes (Supplementary Table 2). Despite the smaller number of renal events in the subset of subjects with microalbuminuria at baseline (25 events in 2,372 subjects), the highest tertile of plasma copeptin also was associated with the outcome (data not shown). Plasma copeptin levels were, on average, 15% higher in subjects treated with diuretics as compared with those who were not (mean ± SEM, 7.55 ± 0.3 vs. 6.58 ± 0.02 pmol/L; P < 0.0001, adjusted for age and sex). However, associations of copeptin tertiles with renal events were similar in subjects treated or not treated by diuretics, and no interaction copeptin/use of diuretics was observed (data not shown).
A stepwise multiple regression analysis was performed to evaluate the independence of the association of plasma copeptin levels with renal events from other potentially confounding covariates (Table 3). The renal event status during follow-up (yes or no) was entered in the model as the dependent variable. Sex, age, duration of diabetes, arterial hypertension status (yes or no), use of diuretics (yes or no), HbA1c, triglycerides, HDL cholesterol, eGFR, UAE, and copeptin (tertiles) at baseline were entered as independent covariates. UAE and copeptin remained positively associated and HDL cholesterol and eGFR were inversely associated with the incidence of renal events. These independent predictors explained ∼18% of the variance of the outcome. Another potential confounder, smoking behavior, was not associated with plasma copeptin levels and had no impact on the association of copeptin with renal events (data not shown). In an additional analysis, the inclusion in the regression model of blood glucose levels and blood pressure at baseline and changes in blood glucose levels and blood pressure during follow-up had no impact on the results. We constructed ROC curves with UAE, plasma copeptin levels, or both as markers of renal events (data not shown). ROC areas were similar for UAE and plasma copeptin (0.78 vs. 0.73; C-statistics P = 0.16) adjusted for age, sex, and study treatment. We observed an added effect for the combined markers (UAE plus copeptin ROC area, 0.82) as compared with the UAE effect (P = 0.03).
Renal events and death as competing risks
Death occurred in 454 subjects (14.65%) during follow-up, including 19 subjects who had presented a renal event. The association between tertiles of plasma copeptin and the incidence of renal events evaluated with the Cox model might be biased if many patients died of other causes related to or associated with copeptin levels before achieving the renal end point (17). Consequently, we performed competing risk regression analyses to estimate subhazard ratios of plasma copeptin tertiles as a risk for renal events assuming death as a competing risk. Subhazard ratios and hazard ratios from the Cox model were similar, indicating that death was not a competing risk in the association of plasma copeptin levels with renal events (Table 2, model 3). Moreover, we observed no independent association of plasma copeptin levels with mortality during follow-up in our cohort (Supplementary Table 3).
Plasma copeptin and eGFR
The yearly variations of eGFR during follow-up by tertiles of plasma copeptin were −0.65 ± 0.24 (T1), −0.77 ± 0.24 (T2), and −1.91 ± 0.24 mL/min/1.73 m2 per year (T3) (mean ± SEM, ANCOVA P = 0.0001, adjusted for sex, age, and study treatment in the original DIABHYCAR trial) for the whole population. They were −1.43 ± 0.51 (T1), −2.29 ± 0.49 (T2), and −3.52 ± 0.44 mL/min/1.73 m2 per year (T3) (P = 0.005) for the subset of subjects with macroalbuminuria at baseline (Supplementary Fig. 1). These differences remained significant after additional adjustment for the duration of diabetes, HbA1c, triglycerides, HDL cholesterol, UAE, arterial hypertension status (yes or no), and use of diuretics (yes or no) at baseline (P = 0.003 for the whole population and P = 0.03 for the subjects with microalbuminuria). In all comparisons, either of the risk of renal events or of the variation of eGFR during follow-up, we observed no interaction between copeptin tertiles and sex or between copeptin tertiles and study treatment (ramipril vs. placebo) in the original DIABHYCAR study (data not shown).
In this study, we demonstrated that high plasma copeptin concentration was strongly associated with the risk of severe renal outcomes (doubling of plasma creatinine concentration and/or ESRD) in patients with type 2 diabetes and albuminuria. This association was independent of relevant covariates such as age, duration of diabetes, blood pressure, and baseline levels of HbA1c, UAE, and eGFR. Plasma copeptin and UAE similarly predicted the incidence of renal outcomes, and a 4% added effect was observed for the combined markers.
To our knowledge, this is the first clinical study that investigates the association between vasopressin or copeptin concentration and accelerated renal function decline in type 2 diabetic patients during a long-term follow-up. Our study confirms and extends previous results on copeptin concentration and kidney diseases. Plasma copeptin has been shown to be positively associated with the prevalence of microalbuminuria in cross-sectional observational and long-term follow-up studies in the general population (8,18). In autosomal-dominant polycystic kidney disease, copeptin is associated with various markers of disease severity including albuminuria and GFR (19–21). In renal transplant recipients, higher copeptin level at baseline was associated with a more rapid decline in effective GFR during 3- to 4-year follow-up (22). Moreover, studies in Australian and Canadian community-based cohorts of the general population showed, respectively, an inverse linear relationship between fluid intake and prevalence of chronic kidney disease (CKD) (23) and a faster decline of eGFR over time in people with lower 24-h urine volume (24). The results of those population studies are in agreement with an adverse effect of vasopressin on kidney function.
As in any epidemiological study, the association we observed may be confounded by other factors. Among them, age and sex have been shown to influence urinary concentrating activity and vasopressin levels (25). We observed a positive association of age with plasma copeptin levels in men and women. We also found copeptin levels to be higher in men than in women as previously observed (8,18,26). However, we observed no interaction between copeptin levels and sex on the risk of renal events. Moreover, when adjusted for confounding covariates, neither age nor sex was associated with renal events. In a recent study of 1,241 hemodialysis patients with type 2 diabetes in Germany, high levels of copeptin were found to be associated with increased risk for all-cause mortality (17). In our study, death was not a competing risk in the association of copeptin levels with renal events and we observed no association of plasma copeptin levels with mortality. However, it is noteworthy that all subjects in the German cohort had ESRD and were more severely ill than our patients. It is possible to speculate that the association between copeptin and renal events could be explained, at least in part, by vasopressin-mediated changes in blood pressure. Vasopressin may contribute to hypertension through activation of V1a receptors on the vascular smooth muscle or by V2-receptor–dependent tubular effects on sodium reabsorption (27). However, the association between copeptin and renal outcomes was independent of blood pressure as previously noted for the association of copeptin and UAE (8).
Experimental evidence strongly supports a causal role of vasopressin in aggravation of diabetic CKD through V2-receptor activation. Besides well-known antidiuretic effects at the collecting duct level, a V2-receptor agonist was shown to induce glomerular hyperfiltration and to increase UAE in normal rats (9,28). Moreover vasopressin has been shown to participate in progression to renal failure in rats with five-sixths reduction in renal mass (29,30) and in rodent models of type 1 diabetes (10,31). Acute administration of DDAVP has been shown to increase UAE in healthy humans but not in patients with diabetes insipidus caused by mutations in the V2-receptor (9). The mechanisms of these deleterious effects of vasopressin are not clear. They may involve changes in the composition of the tubular fluid at the macula densa that influence tubuloglomerular feedback control of GFR, as well as an increase in intraglomerular pressure subsequent to afferent arteriole vasodilatation (32). The latter hemodynamic effect is supposed to be one of the main drivers in the decline of renal function in diabetic CKD since the pioneering works by Brenner and colleagues (33,34). However, our study cannot preclude or confirm any of these mechanisms because of its observational nature. Further experimental studies are needed in that regard.
Our study has limitations. First, the design did not allow any conclusion regarding the possible causal relationship between copeptin and disease progression. Second, we have not measured the true GFR, but we used an estimation index based on plasma creatinine levels. Finally, half of the patients of DIABHYCAR received ramipril during the follow-up and this drug may influence blood pressure. However, we observed no interaction between copeptin and study treatment (ramipril vs. placebo) in any of our results, and adjustment for ramipril treatment had no impact on the results.
The strong and independent associations of copeptin with baseline albumin excretion rate and with progression to CKD offer new tools to identify patients with highest risk of progression. Several novel therapeutic strategies have been proposed for slowing the progression of CKD (4). Some of them, like double-blockade of the renin-angiotensin system (VA-nephron-D study, NCT00555217) or antibodies against tumor growth factor-β1 (NCT01113801), are still undergoing evaluation but their use will induce a higher risk of side effects, as suggested by the ONTARGET trial (35). We propose that plasma copeptin can help target patients with the highest risk of CKD progression to optimize the benefit/risk ratio. Our observations also raise exciting hypotheses of new therapeutic options in diabetic CKD. Increasing water intake may lower vasopressin secretion; such “medicinal use of water” has been discussed recently for kidney diseases other than diabetic nephropathy (36). Also, vaptans (vasopressin receptor antagonists) could selectively target the V2 receptor–mediated deleterious effects of vasopressin. Advantages and disadvantages of both options are debatable (37). If the increase of water intake or V2 receptor antagonists were shown to be beneficial, then this could open the way to individualized therapy. These interventions then could be used specifically in subjects with high copeptin levels who, presumably, would benefit the most from the reduction of vasopressin effects. Further studies are needed to test these hypotheses.
In conclusion, our study is the first investigation studying the association between endogenous vasopressin and kidney dysfunction in type 2 diabetic patients. It highlights the role of copeptin as a risk marker for long-term renal outcomes in these patients. Furthermore, it provides a rationale for intervention studies aiming to reduce vasopressin secretion or action.
The original DIABHYCAR trial was supported by grants from Sanofi (Paris), the French Ministry of Health (Programme Hospitalier de Recherche Clinique–PHRC–Angers 1996), and the Association Francaise des Diabétiques (research grant 2004). The present work was supported by grants from Association Diabète et Risque Vasculaire, Association Robert Debré, DHU Fire (Fibrosis, Inflammation, Remodeling in Cardiovascular, Respiratory, and Renal Diseases), and the Société Francophone du Diabète. Measurement of plasma copeptin was performed by Thermofisher Scientific (Hennigsdorf, Germany), in anonymized tubes, blinded to characteristics and outcomes of patients. The analysis and interpretation of the data have been performed without the participation of these organizations and companies.
R.R. has received research grants, honoraria, or consulting fees from Sanofi, MSD Chibret, Gambro, Servier, Danone Research, Eli Lilly, Novo Nordisk, Medtronic, Novartis, Bristol-Myers Squibb, Janssen-Cilag, and LifeScan. M.M. is a consultant for Association Diabète et Risque Vasculaire, and is a speaker or serves on boards for Merck, Sanofi-Aventis, Novo Nordisk, Servier, and Abbott Diagnostics. No other potential conflicts of interest relevant to this article were reported.
G.V. and N.B. researched data and wrote the manuscript. S.H. contributed to discussion and reviewed and edited the manuscript. N.M. and K.M. researched data. F.F. contributed to discussion and reviewed and edited the manuscript. L.P., N.B.-M., and C.T. researched data. F.A.-G., L.B., and M.M. contributed to discussion and reviewed and edited the manuscript. R.R. researched data and wrote the manuscript. R.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.