All patients were included after written informed consent. The study was approved by The Ethical Committee of Copenhagen County and the Danish Medicines Agency. The study was carried out according to the Helsinki Declaration.

The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S1.

The study was carried out in a randomized, open-label, cross-over design comparing an 8-week control period with an 8-week period of once-daily administration of eplerenone. Randomization was done by the principal investigator drawing sealed opaque envelopes. Patients were either allocated to start in the intervention period followed by the control period or vice versa with no wash-out period in between.

Inclusion criteria were: age >18 years, persistent 24 hour proteinuria, initially planned as >2000 mg, but after inclusion of the first patient changed to >500 mg or albuminuria >300 mg, BP>130/80 mmHg or ongoing stable antihypertensive treatment, including RAS-blockade. There was no demand for ongoing RAS-blocking therapy.

Exclusion criteria were: diabetic nephropathy, creatinine clearance <20 mL/min, plasma (p-) potassium >5.0 mEq/l, allergy to aldosterone antagonists, chronic liver insufficiency, ongoing treatment with CYP3A4-inhibitors, lithium or immunosuppressive agents including steroids, invalidating psychiatric disorders, other severe non-renal disease, woman of childbearing potential not using safe contraception, pregnancy or breast-feeding. Patients were recruited from and followed in the outpatient clinics of the two participating departments. All patients were seen by the principal investigator.

Eplerenone treatment was initiated by a once daily oral dose of 25 mg administered as add-on treatment to ongoing therapy. The dose was doubled after one week to 50 mg once daily for seven weeks. Patients were seen at weeks 0, 1, 2, 4, 8, 9, 10, 12 and 16. The BP goal was <130/80 mmHg. In case of symptomatic hypotension, reductions were primarily made in non-RAS-blocking antihypertensive agents and in case of BP above target non-RAS-blocking agents were added. Main outcome variables were albuminuria based on single 24 hour urine samples collected for every visit, fractional excretion of albumin, BP, p-potassium, and creatinine clearance, which were measured at each visit.

Potassium supplements were withdrawn before initiation of eplerenone treatment. There were no dietary restrictions at trial entry. Withdrawal criteria were any CV event, serious non-renal disease, a persistent rise in p-creatinine above 30% of baseline value, persistent hyperkalaemia, defined as p-potassium >5.5 mEq/L at two successive visits, or pregnancy despite safe contraception.

BP was measured using the mercury sphygmomanometer auscultatory method. Cuff width was selected according to arm circumference. BP measurements were done after 5–10 minutes of rest in the sitting position. Analysis of urinary albumin was done by immunoturbidimetry with antibodies from Dako A/S (Konelab PRIME 60), inter-series coefficient of variation (CV) 2% at Herlev Hospital and Tinaquant from Roche (Roche Modular), inter-series CV 10%, at Rigshospitalet.

P-creatinine was analyzed by the enzymatic method, MDRD-IDMS calibrated (Vitros 5.1), inter-series CV 2% at Herlev Hospital and by Roche Modular Enzymatic method, MDRD-IDMS calibrated, inter-series CV 2.1% at Rigshospitalet.

Fractional albumin excretion was calculated according to the following formula:

A reduction of albuminuria of 25% was considered as the minimal clinically relevant response to eplerenone treatment. Sample size was estimated to be a total of 40 patients, (power 0.8 and level of significance 0.05). Data are presented as mean and standard deviations (SD). Where data were not normally distributed, logarithmic transformations were made and these data are presented as well as geometric means with 95% confidence intervals [CI]. Changes in logarithmic transformed parameters were calculated as ratios and additionally shown in percent.

The treatment effects were calculated by the paired samples t-test, using the mean value of the variable over all four visits in each period – i.e. control and treatment. Analyses of values at individual visits were done as a secondary end-point. Before analyzing treatment effect, analyses of carry-over and time effect were made. To test for carry-over effect, we conducted a two-sample t-test between the two randomization groups using the mean value of each subject. For time effect a paired samples t-test was conducted comparing values for each patient at the two different time periods. Two-way ANOVA was used for testing the treatment effect on urinary albumin excretion when adjusting for creatinine clearance and systolic BP. One-way ANOVA was used for testing differences in response to treatment, when level of RAS-blockade or baseline creatinine clearance was taken into account. Evaluation of significance over multiple periods was done by repeated measure ANOVA. Comparisons of data from individual visits were corrected for multiple comparisons by the Bonferroni method. Since only baseline data were tested against data from each of the four following visits, correction was made by multiplying the P-values by four. When applying correction for multiple testing the corrected P-values are shown. All P-values were two-sided and P-values<0.05 were regarded as significant. All confidence intervals are two-sided 95% CI. The statistical analyses were done using the SPSS statistical software, version 17.0 and R statistical software version 2.12.1, 2011, R Foundation for Statistical Software, Vienna, Austria.

Renal diagnoses were based on previous renal biopsies in chronic glomerulonephritis (n = 26), specified in table 1, and in vascular disease (n = 2). Autosomal dominant polycystic kidney disease (n = 1) was diagnosed by ultrasound scan. The remaining patients were categorized as having chronic nephropathy of unknown aetiology (n = 12). Patients with diabetic nephropathy were excluded.

Creatinine clearance ranged from 24 to 195 mL/min at the time of inclusion and patients were classified as CKD stages 1 (n = 18), 2 (n = 10), 3 (n = 10) and 4 (n = 2). All patients but two were on RAS-blockade. One patient received vitamin D supplements. Baseline patient characteristics of all patients are presented in table 1 and 2. The median baseline number of antihypertensive drugs in the individual patient was two (range 0 to 5).

Albuminuria was significantly lower during the add-on eplerenone period as compared with the control period with a 22% [CI: 14,28], P<0.001, lower excretion, as shown in table 3. The mean 24 hour excretion was 1481 mg [CI: 1192,1840] during the control period and 1163 mg [CI: 921, 1468] during add-on eplerenone. No significant carry-over, P = 0.3 or time effect, P = 0.3, was detected for the urinary albumin excretion. Reduction in albuminuria was observed in 31 of 40 patients. Mean changes in urinary excretion of albumin over time are shown in figure 1A, and the difference in albumin excretion during the two periods is illustrated in figure 1B. A significant decrease was seen after two weeks of add-on eplerenone, P = 0.003. Three patients had membranous nephropathy, a glomerular disease where resistance to aldosterone antagonists has been reported ^[18]. In two of these patients, a decreased urinary albumin excretion was observed during add-on eplerenone, while it was unchanged in the third patient.

Fractional albumin excretion was significantly lower in the eplerenone period, see table 3, P<0.001.

When systolic BP reduction was taken into account, the urinary albumin excretion was 17% [CI: 8, 25] lower during eplerenone treatment, P<0.001. When corrected for the lower creatinine clearance during the treatment period, the reduction in albumin excretion was 19% [CI: 10,27], P = 0.0003. When both BP and creatinine clearance were included in the statistical model, the total difference in 24 hour urinary albumin excretion was 14% [CI: 4,24], P = 0.008.

As there was no wash-out period between study periods, additional analysis was done excluding the first measurements in each block (i.e. visit 2 and 6). Excluding these measurements the difference in urinary albumin excretion was 25% [CI: 17,32], P<0.001. Comparing only visits 2 and 6, there was no difference in albumin excretion between treatment and control periods, as well as there was no detection of any significant carry-over effect, P = 0.3.

To study the influence of baseline RAS-blockade on the antialbuminuric effect of eplerenone baseline RAS-blockade was separated into three levels for post hoc analysis: Low level blockade when the daily dose of ACEi/ARB was <¼ of the maximum dose for the individual drug (n = 11), moderate level blockade when the ACEi/ARB dose was between ¼ and ½ of the maximum dose (n = 18) and high level blockade when the ACEi/ARB dose was >½ the maximum dose (n = 2) or in case of combined ACEi and ARB (n = 7). There was a non-significant tendency for patients on moderate RAS-blockade to have a greater response on eplerenone than the other groups. Patients on combined ACEi and ARB had no significant reduction in albuminuria during eplerenone treatment.

Systolic and diastolic BP was significantly lower during add-on eplerenone treatment when compared to the control period (table 3). There was a significant reduction of systolic BP after two weeks of eplerenone treatment, P = 0.003. At later visits, there were no further significant changes in systolic BP. The diastolic BP was significantly reduced after four weeks of eplerenone treatment, P = 0.002, and there was a significant difference in diastolic BP between the treatment period and control period at the same time point, P = 0.004.

There were no significant differences between diastolic BP at the end of the two periods. There were no significant differences in BP reduction between groups of low, moderate or high baseline RAS-blockade.

There were no significant carry-over, P = 0.4 and P = 0.9, or time effects, P = 0.5 and P = 0.2 for systolic or diastolic BP.

As shown in table 3 creatinine clearance was 5% ^[2], ^[8], P = 0.005 lower in the eplerenone period compared to the control period. Mean creatinine clearance was 86 mL/min [CI: 75,97] during control and 81 mL/min in the eplerenone period [CI: 71,93]. The time course analyses showed that, the difference was only significant at week four of treatment versus week four of the control period, P = 0.03. There were modest increases in p-creatinine during add-on eplerenone, although no patients had an increase above 30%. Patients on baseline combined ACEi and ARB tolerated eplerenone as well as the other patients.

Four patients reported increased dizziness during the treatment period of the study compared to their usual state without any change in BP. None required reduction in dose of eplerenone due to dizziness alone.

No patients were withdrawn due to hyperkalaemia or other adverse events. The mean difference in p-potassium was 0.1 mEq/L, P<0.001 between the eplerenone and control periods, as shown in table 3.

Four percent of all p-potassium values measured were above 5.0 mEq/L and none were above 6.0 mEq/L. Only one patient had a single p-potassium value above 5.6 mEq/L at a planned visit. Consequently the patient was seen at an extra control, where p-potassium had increased to 5.8 mEq/L. Because of this, as well as symptomatic hypotension, nausea and general discomfort the dose of eplerenone was reduced to 25 mg daily. The patient completed the study without further adverse effects.

There were no significant carry-over effects, P = 0.3 and P = 0.1, or time effects, P = 0.7 and P = 0.7, for either creatinine clearance or p-potassium.

Five patients had hypoalbuminemia, with a p-albumin <3.6 g/dL. There was a significantly higher concentration of albumin in plasma during treatment with eplerenone, P = 0.004. When the hypoalbuminemic patients were taken out of the statistical analysis, the P-value changed to 0.006. Twenty-nine patients had higher mean values of p-albumin during the treatment period, and 10 patients had lower values. Only one of the hypoalbuminemic patients had normalization of p- albumin in the treatment period and a 17% reduced urinary albumin excretion in the treatment period. Four patients were continuously hypoalbuminemic throughout the study. Two of these patients had a reduced excretion of albumin in urine and two had a higher excretion during treatment with eplerenone. Along with the analysis of changes in p-albumin it was also investigated whether there were any changes in p-cholesterol (total), p-LDL and p-HDL cholesterol and triglycerides. Neither of these changed significantly during the study. No patients were started or altered in lipid lowering treatment during the study period.

There were also no significant differences between study periods in the following parameters: weight, BMI, p-ionized calcium, p-uric acid, p-C-reactive-protein, blood haemoglobin, nor in the urinary excretion of creatinine, potassium and sodium.

The group of patients studied were not homogeneous with respect to baseline albuminuria and renal diagnoses. There were no requirements to baseline therapy or run-in period. The study was not designed to detect a difference in response to eplerenone based on the baseline RAS-blocking treatment. It was an open label design and a relatively small sample size, which limits the extent to which results should be generalized. The statistical analyses were based on samples from all visits which were given equal weight in the analyses.

Eplerenone in a dose of 25–50 mg daily is a safe and relevant add-on therapy in patients with non-diabetic CKD when aiming at a reduction in urinary albumin excretion. The potential risk of hyperkalaemia, especially in daily practice outside the setting of a controlled trial, should be kept in mind. Long-term studies need to be designed in order to assess the question of whether the effect of eplerenone on the reduction in urinary albumin excretion is persistent, and whether the progression rate of CKD is reduced.