Copeptin, a surrogate marker for vasopressin, is associated with kidney function decline in subjects with autosomal dominant polycystic kidney disease

Abstract

Background

Experimental studies have suggested that vasopressin plays a detrimental role in autosomal dominant polycystic kidney disease (ADPKD). It is, however, unknown whether endogenous vasopressin concentration is associated with kidney function decline in subjects with ADPKD.

Methods

We measured plasma copeptin (a marker of vasopressin) in 79 ADPKD subjects with renal function assessed during short-term follow-up by inulin clearance measured glomerular filtration rate (mGFR) and during long-term follow-up by Modification of Diet in Renal Disease (MDRD) equation estimated GFR (eGFR).

Results

In these subjects (43% male, age 36.8 ± 10.1 years, GFR 96.8 ± 18.2 mL/min/1.73 m²), median copeptin concentration at baseline was 2.71 [interquartile ranges (IQR) 1.63–5.46] pmol/L. Baseline copeptin concentration was inversely associated both with change in mGFR during follow-up for 3.3 (3.1–3.5) years, (R = −0.300, P = 0.01), as well as with change in eGFR during follow-up for 11.2 (4.5–14.3) years, (R = −0.302, P < 0.01). These associations were independent of age, gender and baseline GFR. Nine subjects started renal replacement therapy during follow-up of which eight had at baseline a copeptin concentration above the median in this population.

Conclusion

In ADPKD subjects, a higher copeptin concentration is associated with kidney function decline during follow-up, suggesting that copeptin may be a new marker to predict kidney outcome in ADPKD.

ADPKD, copeptin, GFR, vasopressin

Introduction

Autosomal dominant polycystic kidney disease (ADPKD) is the most common hereditary kidney disease. It is characterized by the development of kidney cysts and progressive kidney function loss, often leading to end-stage renal disease [1, 2]. Mutations in two genes, PKD1 and PKD2 on chromosome 16 and 4, respectively, can lead to this disease. There is no proven treatment available yet, which can attenuate the rate of cyst formation and kidney function decline.

Vasopressin, also known as antidiuretic hormone, regulates water homeostasis in the body. It is secreted by the neurohypophysis in response to an increase in plasma osmolarity or a decrease in plasma volume. Vasopressin is known to bind to three receptors. The V_1A receptor is found in several organs, among others the kidney, where binding of vasopressin to this receptor results in a decrease in blood flow to the inner medulla and stimulation of prostaglandin synthesis [3]. The V_1B receptor stimulates the release of adrenocorticotropin from the anterior pituitary [4]. The V₂ receptor is found in the kidney and is located at the interstitial side of the principal cells of the collecting ducts. Upon stimulation, these V₂ receptors induce insertion of the water channel aquaporin-2 in the luminal membrane of the principal cells, which leads to reabsorption of water from the lumen of the collecting duct into the blood, thereby reducing water excretion [3].

In ADPKD, vasopressin is hypothesized to play a role in the pathogenesis of cyst formation in the kidneys. In experimental studies, it has been shown that when this hormone binds to the V₂ receptors, production of cyclic adenosine monophosphate (cAMP) is stimulated in the principal cells [2, 5]. In turn, cAMP stimulates cyst production by promotion of fluid secretion and activation and proliferation of cyst-derived cells [6]. Furthermore, in animal models for ADPKD, it has been shown that blockade of the effect of vasopressin leads to a reduction of cyst formation [7–11].

Vasopressin is difficult to measure because of its binding to platelets [12], short half-life time [13] and instability in isolated plasma [14]. Copeptin is a part of the precursor of vasopressin (preprovasopressin) and can be measured by a recently developed sensitive sandwich immunoassay [15]. Copeptin has been shown to be stable in isolated plasma [16] and to be a reliable substitute for circulating vasopressin concentration [17].

Recently, we found an association between copeptin concentration and various markers of disease severity in subjects with ADPKD [18], supporting the results of the above-mentioned experimental studies that suggest vasopressin to have a deleterious effect in ADPKD. However, this study was cross-sectional in design, thus limiting firm conclusions on temporal relationships and a possible causal role for vasopressin in kidney function loss. In the present study, we therefore aimed to investigate prospectively the association between plasma copeptin concentration at baseline and the prognosis with respect to kidney function during follow-up in subjects with ADPKD. A priori, we hypothesized that copeptin concentration is associated with the rate of decline of kidney function in these subjects.

Materials and methods

Subjects

We included ADPKD subjects who participated in a trial investigating the efficacy of angiotensin-converting enzyme (ACE)-inhibition to preserve kidney function that was performed between 1994 and 1999 [19]. In this study, normotensive subjects were randomized to enalapril once daily (OD) (5 or 10 mg) or placebo. Hypertensive subjects were randomized to a step-up dosage regime with a maximum of 20 mg enalapril OD or 100 mg atenolol OD. Inclusion criteria for this trial were ADPKD as defined by the ultrasonographic criteria formulated by Ravine [20], age 18–70 years and plasma creatinine <2.5 mg/dL. Exclusion criteria were a history of myocardial infarction, cerebrovascular accident, presence of other kidney disease, diabetes mellitus, congestive heart failure, peripheral vascular disease, hepatic dysfunction, chronic use of immunosuppressants, NSAIDS, uricosurics and levodopa, adverse reactions to ACE inhibitors or pregnancy. This study found no beneficial effect of ACE inhibition; there was no difference in rate of kidney disease progression in both groups. Inclusion and exclusion criteria for the present study conform with those in the original trial. In addition, we excluded subjects with missing data.

Measurements and definitions

Baseline

At baseline, subjects visited an outpatient department and their medical history was taken. Blood pressure was measured three times, of which the average was taken. Body mass index (BMI) was calculated using the standard formula: weight (kilogram)/square of height (metre). Blood was drawn at baseline while patients were fasting; they were allowed to drink water ad libitum. Plasma creatinine was measured using the Jaffe method. Total cholesterol and glucose concentrations were measured by standard methods. Plasma samples were stored at −80°C. All samples were collected between 1994 and 1996 and have been stored for a similar period of time. Morgenthaler et al. [15] showed that frozen storage did not have an effect on concentration of copeptin with recovery values of around 100%. Copeptin was measured in these samples by a sandwich immunoassay (CT-proAVP LIA; ThermoFisher Scientific, B.R.A.H.M.S. Biomarkers, Hennigsdorf/Berlin, Germany) as described previously [15, 21], with a modification insofar that the capture antibody was replaced by a murine monoclonal antibody directed to amino acids 137–144 of proAVP. This modification improved the sensitivity of the assay. The lower limit of detection was 0.4 pmol/L [22]. Copeptin measurements were carried out by an employee of B.R.A.H.M.S. (the manufacturer of the copeptin assay), who had no access to patient files and was therefore blinded for outcome.

Follow-up

Kidney disease progression during follow-up was assessed in three different ways.

• (i) Short-term follow-up: change in measured glomerular filtration rate (mGFR). At baseline and at the end of the 3-year study, GFR was measured after an overnight fast by inulin clearance. A loading dose of inulin was given in 10 min, followed by 3 h of continuous infusion. During the infusion, subjects stayed in the resting position and maintained hydration by oral water intake ad libitum. After 1.5 h, three urine samples were obtained over 30-min periods, with blood samples before and at the end of each collection period for determination of inulin concentrations. For each urine portion and corresponding blood sample, the mGFR was calculated and then averaged. mGFR was measured twice at the beginning of the study with a median time between the two measurements of 14 [7–26] days. The mean of the two measurements was used. mGFR was corrected for body surface area calculated by the Mosteller formula [23]. Change in mGFR was calculated as the difference between the baseline and last available mGFR value divided by follow-up time in years.

• (ii) Long-term follow-up: change in estimated GFR (eGFR). Plasma creatinine was measured at baseline, and this value was used to estimate GFR using the abbreviated Modification of Diet in Renal Disease (MDRD) equation [24]. The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was not used for our primary analysis because the creatinine values obtained at baseline were not standardized to IDMS-traceable values. Subjects were followed over time and the last available plasma creatinine value was obtained for assessment of the most recent eGFR. For subjects who started renal replacement therapy (RRT) during follow-up or who died, the last available creatinine value before the start date of RRT or death was used as the most recent value. The date at which creatinine was measured was used as end of follow-up. Change in eGFR was calculated as the difference between the baseline and last available eGFR value divided by follow-up time in years.

• (iii) Long-term follow-up: incident RRT. If applicable, the start date of RRT was reported and we investigated the association between baseline plasma copeptin concentration and the hazard ratio for start of RRT. The hazard ratio should be interpreted as the increase of risk in case of a 10-fold higher copeptin level (due to the log transformation of copeptin).

Statistical analyses

Analyses were performed with SPSS version 18.0 (SPSS Inc., Chicago, IL). Normality was tested with the Kolmogorov–Smirnov test. Parametric variables are expressed as means with the standard deviation (±SD). Non-parametric variables are expressed as medians with interquartile ranges (IQR). A two-sided P-value of <0.05 was considered to indicate statistical significance.

For the analyses of copeptin as predictor of change in renal function during follow-up, we performed continuous analyses by performing univariate regression analyses with baseline log copeptin as the independent variable and change in renal function (mGFR or eGFR) as the dependent variable. To visualize these associations, scatter plots were made and a regression line was drawn. Using a multivariable regression model, these associations were adjusted for covariates that could potentially be confounders in this association. We built multivariable models stepwise. Firstly, our association was adjusted for gender and age (Model 2), and subsequently also for baseline GFR, use of diuretics, hypertension (defined as systolic blood pressure >140 or diastolic blood pressure >90 or use of anti-hypertensive drugs) and treatment group (placebo/atenolol versus enalapril) (Model 3). We adjusted for use of diuretics because these drugs influence volume status, and therefore vasopressin concentration.

To investigate the association between plasma copeptin and incident RRT (Analysis 3), Cox regression analysis was performed. This analysis was performed similarly, first crude and subsequently stepwise with adjustment for gender, age, baseline eGFR, use of diuretics, hypertension (defined as systolic blood pressure >140 or diastolic blood pressure >90 or use of anti-hypertensive drugs) and treatment group (placebo/atenolol versus enalapril). Subjects were censored at the date they deceased, lost to follow-up, started RRT or at the date of the last available serum creatinine value.

For regression analyses, logarithmic transformation (Lg10) of copeptin was applied to fulfil the requirement of equal distribution of the residuals. Interactions between log copeptin concentration and age and gender were tested for change in mGFR, eGFR and start of RRT as dependent variables.

Results

Of 108 available subjects, 29 subjects were excluded because of missing plasma samples at baseline or missing data during follow-up (Figure 1). Baseline characteristics of the remaining 79 subjects are given (Table 1). Forty-three per cent of these subjects were male and mean age was 36.8 ± 10.1 years. Median baseline plasma copeptin concentration was 2.71 (1.63–5.46) pmol/L. The 29 subjects that were excluded for the present analyses did not differ significantly from the 79 subjects that were included with respect to any of the characteristics listed in Table 1, except for baseline GFR (mGFR 71.4 ± 32.5 mL/min/1.73 m² versus 96.8 ± 18.2 mL/min/1.73 m², respectively, P < 0.01; eGFR 59.4 ± 22.7 mL/min/1.73 m² versus 75.9 ± 18.1 mL/min/1.73 m², respectively, P < 0.01) and systolic blood pressure (145.0 ± 16.7 mmHg versus 136.6 ± 16.4 mmHg, respectively, P = 0.03). In these 79 patients, 43 patients were randomized to enalapril, 5 to atenolol and 31 to placebo. The limited size of the subgroup receiving atenolol did not allow formal analyses of the efficacy of atenolol versus enalapril.

Fig. 1.

Flow diagram with in/exclusion of subjects for the three analyses: baseline plasma copeptin concentration versus (i) change in mGFR (inulin clearance) during short-term follow-up, (ii) change in eGFR (MDRD) during long-term follow-up and (iii) start of RRT during long-term follow-up.

Table 1.

Baseline characteristics of all 79 subjects analysed^a

Variables
Male gender, n (%)	34 (43)
Age (years)	36.8 ± 10.1
BMI (kg/m2)	24.4 (21.9–26.2)
Systolic blood pressure (mmHg)	136.5 ± 16.4
Diastolic blood pressure (mmHg)	86.5 ± 8.5
Serum glucose (mg/dL)	85 (81–92)
Serum total cholesterol (mg/dL)	196 ± 34
Serum creatinine (mg/dL)	1.0 (0.87–1.9)
mGFR (by inulin clearance in mL/min/1.73 m2)	96.8 ± 18.2
eGFR (by MDRD in mL/min/1.73 m2)	75.9 ± 18.1
Copeptin (pmol/L)	2.71 (1.63–5.46)

^aResults are given as means ± SD or as median (IQR) in case of non-normal distribution. Conversion factors for units: serum glucose in mg/dL to mmol/L, ×0.05551, serum total cholesterol in mg/dL to mmol/L, ×0.02586 and serum creatinine in mg/dL to μmol/L, ×88.4.

Baseline associations

Baseline mGFR and baseline eGFR were significantly associated with each other (std B 0.776, P < 0.001). Plasma copeptin concentrations were inversely associated with mGFR (std B−0.258, P = 0.02) and with eGFR (std B−0.207, P = 0.06). Age was not associated with plasma copeptin concentration (std B 0.052, P = 0.64).

Short-term follow-up: change in mGFR

For the first analysis of the association between copeptin and kidney outcome in ADPKD, inulin clearances assessed at baseline and at the end of the clinical trial in which the subjects under analysis participated were used. At the beginning of the trial, mGFR was assessed twice, with a coefficient of variation of 4.9%. In eight subjects, mGFR was not assessed at the end of the trial. These subjects were therefore excluded from this analysis, leaving 71 subjects, who had a median follow-up of 3.3 (3.1–3.5) years with a mean change in mGFR of −2.5 ± 3.7 mL/min/1.73 m²/year (Table 2). Baseline copeptin was significantly associated with change in mGFR during the trial (std B−0.300, P = 0.01) (Figure 2). When adjusted for gender, age, use of diuretics, baseline mGFR, hypertension and treatment group, a significant association remained between baseline plasma copeptin concentration and change in mGFR during short-term follow-up (std B−0.345, P < 0.01, Table 3).

Fig. 2.

Baseline plasma copeptin concentration versus change in mGFR (inulin clearance) during short-term follow-up [n = 71, follow-up 3.3 (IQR 3.1–3.5) years, standardized B−0.300, P = 0.01]. Triangles represent patients who started RRT during follow-up.

Table 2.

Results during short- and long-term follow-up, where mGFR is measured as inulin clearance (Analysis 1) and eGFR is estimated by the MDRD equation (Analyses 2)^a

Variables	mGFR (inulin clearance)	eGFR (MDRD)
N	71	77
Baseline GFR (mL/min/1.73 m2)	97.1 ± 19.2	75.4 ± 18.1
Follow-up (years)	3.3 (3.1–3.5)	11.2 (4.5–14.3)
GFR end of study (mL/min/1.73 m2)	90.0 ± 23.3	57.7 ± 29.5
Annual change in GFR (mL/min/1.73 m2/year)	−2.5 ± 3.7	−1.7 ± 1.8
Died (all cause), n (%)		4 (5.2)
Start of RRT, n (%)		9 (11.7)

^aResults are given as means ± SD or as median (IQR) in case of non-normal distribution.

Table 3.

Association between baseline plasma copeptin concentration and change in mGFR (n = 71), change in eGFR (n = 77) and start of RRT (n = 79) during follow-up (3.3 and 11.2 years, respectively)^a

	Model 1		Model 2		Model 3
	Std B	P-value	Std B	P-value	Std B	P-value
Change in mGFR (inulin clearance)
Lg10[Copeptin] (pmol/L)	−0.300	0.01	−0.341	0.004	−0.345	0.006
Gender (female)			−0.242	0.04	−0.222	0.07
Age (year)			−0.244	0.03	−0.175	0.2
Baseline mGFR (mL/min/1.73 m2)					0.073	0.6
Use of diuretics (yes)					NA	NA
Hypertension (yes)					−0.180	0.1
Treatment group (enalapril)					−0.087	0.5
Change in eGFR (MDRD)
Lg10[Copeptin] (pmol/L)	−0.302	0.008	−0.272	0.02	−0.254	0.05
Gender (female)			0.114	0.3	0.158	0.2
Age (year)			−0.007	0.9	0.090	0.6
Baseline eGFR (mL/min/1.73 m2)					0.105	0.6
Use of diuretics (yes)					0.098	0.4
Hypertension (yes)					−0.141	0.2
Treatment group (enalapril)					−0.137	0.3
Incidence RRT	HR	P-value	HR	P-value	HR	P-value
Lg10[Copeptin] (pmol/L)	9.88	0.01	10.20	0.02	5.73	0.1
Gender (female)			0.45	0.3	0.02	0.02
Age (year)			1.02	0.5	0.99	0.9
Baseline eGFR (mL/min/1.73 m2)					0.81	0.02
Use of diuretics (yes)					0.00	0.9
Hypertension (yes)					0.12	0.2
Treatment group (enalapril)					15.75	0.09

^aHR, hazard ratio; Std B, standardized beta; NA, not applicable (because none of the patients included in this analysis used diuretics); Lg10[copeptin], 10log transformed value of copeptin was used. The hazard ratio should be interpreted as the risk increase with a ten-fold higher copeptin level.

Long-term follow-up: change in eGFR

For the second analysis, the baseline and last available creatinine values were used to estimate GFR with the abbreviated MDRD equation [24]. Two subjects were lost to follow-up, leaving 77 subjects for analysis, who had a median follow-up of 11.2 (4.5–14.3) years with a mean change in eGFR during long-term follow-up of −1.7 ± 1.8 mL/min/1.73 m²/year (Table 2). Baseline plasma copeptin concentration was significantly associated with the decline in eGFR during long-term follow-up (std B−0.302, P < 0.01, Figure 3). Results of the multivariable regression analyses for this end point are shown (Table 3). When adjusted for gender, age, use of diuretics, baseline eGFR, hypertension and treatment group, the significant association remained between baseline copeptin concentration and change in eGFR during long-term follow-up (std B−0.254, P = 0.05).

Fig. 3.

Baseline plasma copeptin concentration versus change in eGFR (MDRD) during long-term follow-up [n = 77, follow-up 11.2 (IQR 4.5–14.3) years, standardized B−0.302, P < 0.01]. Triangles represent patients who started RRT during follow-up.

Long-term follow-up: incident RRT

For the third analysis, the association between baseline copeptin concentration and start of RRT during follow-up was investigated. Of the 79 subjects with long-term follow-up, 4 subjects died, whereas 9 subjects started RRT (11.4%). Eight of those 9 patients had a copeptin concentration above the median value of 2.71 pmol/L. Subjects who started RRT were compared to subjects who did not start RRT for all characteristics listed in Table 1. Those who started RRT had similar characteristics as those who did not, except for plasma copeptin concentration [4.10 (3.27–17.6) versus 2.27 (1.55–5.19) pmol/L, P = 0.01], BMI [22.6 (19.9–24.0) versus 24.7 (22.1–26.2) kg/m², P = 0.03] and eGFR (60.0 ± 19.9 versus 77.5 ± 16.9 mL/min/1.73 m², P = 0.03). Results of the Cox regression analyses are given (Table 3). In the crude analyses, a 10-fold higher plasma copeptin concentration was associated with a hazard ratio for start of RRT of 9.88 (95% confidence interval: 1.72–56.88, P = 0.01). When adjusted for age and gender, the hazard ratio was 10.20 and remained significant (P = 0.02). When additionally adjusted for baseline eGFR, use of diuretics, hypertension and treatment group, the hazard ratio for incident RRT was lowered and of borderline statistical significance.

Sensitivity analyses

Several sensitivity analyses were performed. Firstly, we investigated in the short-term follow-up study the association between baseline copeptin concentration and change in eGFR (MDRD) instead of mGFR: crude std B−0.305 (P = 0.01) and Model 3 std B−0.259 (P = 0.07). Secondly, we used in the long-term follow-up study the CKD-EPI equation instead of the MDRD equation to estimate GFR: crude std B−0.312 (P < 0.01) and Model 3 std B−0.261, P = 0.05. Thirdly, change in the reciprocal of serum creatinine concentration over time was used as outcome measure for the short-term follow-up study crude std B−0.325 (P = 0.007) and Model 3 std B−0.285 (P = 0.05) as well as the long-term follow-up study crude std B−0.335 (P = 0.003) and Model 3 std B−0.298 (P = 0.02). All these analyses showed essentially similar results when compared with our primary analyses. In addition, we tested whether there was interaction between baseline copeptin and age or gender in the association with risk. No significant interactions were found for all the three analyses.

Discussion

This study shows that in subjects with ADPKD, plasma concentration of copeptin is associated with a decline in kidney function, assessed as either change in inulin clearance (mGFR) during short-term follow-up or as change in eGFR during long-term follow-up. These associations remained significant after adjustment for age, gender, baseline GFR, diuretics, hypertension and treatment group. Furthermore, plasma concentration of copeptin was found to be associated with need for RRT during long-term follow-up, independent of age and sex.

As far as we know, the present study is the first to prospectively assess the association between baseline copeptin concentrations and renal outcome in subjects with ADPKD. Interestingly, we previously found a cross-sectional association between copeptin concentrations and disease severity in ADPKD [18], copeptin concentration and albuminuria in healthy subjects [25] and also an association between copeptin and accelerated renal function decline during follow-up in kidney transplant recipients [26]. Both studies showed a positive association, the higher copeptin the worse renal outcome. These studies, in combination with the present one, suggest that copeptin is associated with renal outcome in ADPKD, but possibly also in other chronic kidney diseases.

From other studies, it is known that copeptin values can decrease very rapidly [15] suggesting extrarenal clearance as predominant clearance mechanism. Nevertheless, we took the possibility into account that lower renal function may lead to less clearance of copeptin and adjusted therefore for baseline GFR in our multivariable models. This adjustment led to only a minor decrease in the regression coefficient of the association between baseline copeptin and renal outcome. Consequently, this association remained significant. Furthermore, we recently found that under standardized conditions, copeptin levels were significantly elevated in young ADPKD patients when compared to age- and sex-matched healthy controls. Importantly, these young ADPKD patients and healthy controls appeared to have similar kidney function (eGFR CKD-EPI 100 versus 104 mL/min/1.73 m² and 24 h creatinine clearance 116 versus 117 mL/min/1.73 m², respectively) [27]. These data suggest that in ADPKD patients, the association between copeptin level and renal outcome is unlikely to be an effect of a lower renal clearance of copeptin at baseline, and indicate that a rise in copeptin precedes kidney function decline. We propose that relatively early in the course of ADPKD, due to a urinary concentrating defect, vasopressin levels are elevated in order to maintain plasma osmolality within the normal range. Unfortunately, plasma osmolality was not measured at the start of this study.

Our results also agree with the possible detrimental role of vasopressin in ADPKD, since the rise in vasopressin (measured as copeptin) precedes a decline in GFR. Consequently, it may be hypothesized that lowering vasopressin can lead to renoprotection in human ADPKD. To lower vasopressin concentration, one of the options is to achieve ample hydration. Another way to suppress the effect of vasopressin is to block the V₂ receptor in the kidney with medication. This option has been tested in animal experiments, which showed that vasopressin antagonism indeed prevented cyst growth and kidney function decline [9]. At this moment, a large-scale randomized controlled trial is ongoing that investigates whether these vasopressin V₂ receptor antagonists are renoprotective in ADPKD patients [28].

Our data suggest that plasma copeptin concentration may be a promising new, relatively easy to measure marker to predict kidney function decline in subjects with ADPKD. Nowadays, it is difficult to predict the prognosis of a patient with ADPKD. There are risk factors known, such as the type of genetic mutation (PKD1 or PKD2) and family history with respect to age at time of need for RRT. Both are, however, not very specific at an individual level [29] and the genetic mutation is relatively difficult and time-consuming to measure for routine diagnostics. The same holds true for renal blood flow [30] and magnetic resonance imaging-assessed kidney (or cyst) volume [31]. Measurement of GFR is of limited value because it remains for a prolonged period near normal [32]. Plasma copeptin concentration may therefore be a new prognostic factor to help distinguish between patients with a low risk of reaching end-stage kidney disease and patients with a higher risk, alone or in combination with other prognostic factors. Replication of our findings is, however, necessary before it can be used as such.

Some limitations of this study should be addressed. Firstly, not all subjects who participated in the original study were included in the present analyses. The reason to exclude these subjects was, however, due to a random process (missing plasma samples) and no bias is therefore to be expected. Secondly, kidney volume was not assessed as outcome because at the time of initiation of this study, kidney volume measurements were not assessed routinely, nor in the context of clinical trials. Thirdly, baseline creatinine measurements were performed in one centre and standardized. During follow-up, however, the most recent creatinine was measured at various sites. Therefore, differences between assays for creatinine may be present. However, this is expected to result in effect dilution bias and therefore in an under- rather than an overestimation of the associations that were found. Fourthly, baseline copeptin concentration in these patients was measured only once. However, this is again expected to result in an under- rather than overestimation of the true effect size. Fifthly, only a limited number of incident RRT cases were observed during follow-up. The results of the analyses studying this outcome should therefore be interpreted with caution, and be regarded primarily as supporting the findings that were obtained in the analyses studying the association between baseline copeptin and changes in mGFR and eGFR. Lastly, from this study, it cannot be concluded that copeptin predicts outcome because of a pathophysiological role of vasopressin in ADPKD specifically or that it is a marker for kidney disease progression in general.

Strengths of this study are that this study has a relatively long duration of follow-up and that renal outcome was measured in three different ways, among which the gold standard for measuring GFR, being inulin clearance. Our findings, showing that baseline copeptin concentrations are associated with all three renal outcome measures during follow-up, make our data robust and likely to be valid.

In conclusion, plasma copeptin concentration, a surrogate marker for vasopressin, is associated with the rate of kidney function decline in subjects with ADPKD. These results suggest that in such subjects, copeptin may be a promising, relatively easy to measure new marker to predict renal outcome, alone or in combination with other markers.

Acknowledgements

We thank Leoni van der Plas and Inger Kunnekes for secretarial work.

Conflict of interest statement. J.S. is an employee of ThermoFisher Scientific, B.R.A.H.M.S. Biomarkers, the company that manufactures and holds patent rights on the copeptin assay. None of the other authors have anything to declare.

(See related article by Fenske and Wanner. Copeptin: a marker for ADPKD progression?. Nephrol Dial Transplant 2012; 27: 3985–3987.)