Introduction

Diabetic nephropathy (DN) is the most common single cause of end-stage renal disease (ESRD) in the western world and develops in approximately 30% of all type 1 diabetic patients [1]. Aggressive antihypertensive treatment with a special focus on blocking the renin–angiotensin–aldosterone system (RAAS) has been shown to be particularly effective in improving outcome, and ACE inhibitors (ACEI) are now considered the first-line therapy in type 1 diabetic patients with DN.

Dose titration studies of ACEIs are traditionally conducted according to BP levels in essential hypertension [2, 3]. The optimal BP-lowering dose is not, however, necessarily the same as the optimal dose for renoprotection. The aim of the current study was to determine the optimal renoprotective dose of lisinopril, as evaluated by short-term changes in albuminuria, in type 1 diabetic patients with DN.

Methods

Subjects

From the outpatient clinic at Steno Diabetes Center, 63 type 1 diabetic patients with DN and hypertension (>135/85 mmHg) were enrolled. Their DN was diagnosed clinically as previously described [4].

Main exclusion criteria were GFR <30 ml min−1 1.73 m−2, plasma potassium >4.8 mmol/l, severe hypertension (>180/100), known renal artery stenosis, age <18 or >70 years, allergy to ACEI, chronic heart failure, myocardial infarction, unstable angina or coronary bypass grafting within the previous 3 months.

Design

The study was a randomised, double-masked, crossover trial consisting of an initial 2 month washout period followed by three treatment periods, each lasting 2 months. During the washout period, all antihypertensive treatment was withdrawn and patients were put on slow-release furosemide in individual but fixed doses (median [range] 60 [0–360] mg daily) to prevent fluid retention, BP becoming too high and hyperkalaemia during the study. Thereafter patients received lisinopril 20, 40 or 60 mg once daily in random order, with an initial 2 week run-in period when patients received lisinopril 20 mg.

At the end of each treatment period, the primary endpoint, urinary albumin excretion rate (UAER), and the secondary endpoints, 24 h ambulatory BP (ABP) and estimated GFR (eGFR), were determined, and blood samples were drawn to allow measurement of the components of the RAAS. For safety reasons, BP, plasma potassium, sodium and creatinine were determined 3 weeks after the beginning of each treatment period.

The study was performed according to the principles of the Declaration of Helsinki and was approved by the ethical committee of Copenhagen County. All patients gave their informed consent.

Laboratory procedures

Urinary excretion of albumin was determined by turbidimetry in three consecutive 24 h urine collections completed immediately before the end of each treatment period (Hitachi 912 system; Roche Diagnostics, Mannheim, Germany).

Office BP and 24 h ABP were measured as described previously [4]. The eGFR was calculated using the re-expressed four-variable Modification of Diet in Renal Disease (MDRD) study equation [5].

From venous samples, haemoglobin, plasma potassium, sodium, creatinine and cholesterol concentrations and HbA1c was measured using standard methods [4]. Blood samples for plasma renin activity, angiotensin I, angiotensin II, ACE-activity and aldosterone concentrations were taken after 30 min of supine rest.

Statistical analysis

Normally distributed variables are expressed as means with either SD or SE in parentheses. Non-normally distributed variables were logarithmically transformed before statistical analysis and are given as geometric means (95% CI). Comparisons of UAER, ABP and eGFR between each treatment period were performed using linear mixed models.

Based on a previously calculated SD (log scale 0.1771) of the mean difference in UAER, a sample-size calculation showed a necessary minimum of 50 patients to detect a 15% difference in change in UAER between any two dose levels (α = 0.05, β = 0.80).

A value of p < 0.05 was considered significant (two-tailed test). Data were evaluated using SPSS version 14.0 (SPSS, Chicago, IL, USA).

Results

A total of 63 patients gave their informed consent to participate in the study. Seven patients were excluded during the washout period because of high BP. Fifty-six patients were randomised, but two patients were excluded because of a reversible increase in creatinine (one on lisinopril 20 mg and one on lisinopril 60 mg) and one because of high BP (lisinopril 60 mg). Four patients withdrew consent before completing the study because of minor complaints (see Electronic supplementary material [ESM] Fig. 1 for details of patient-flow and adverse events). Three patients had a transient increase in plasma potassium after 2 months of lisinopril 60 mg, which was reversed by standard treatment. No patients had plasma potassium >5.5 mmol/l on lisinopril 20 or 40 mg. No other adverse events were reported during the study.

Results are given for 49 patients who completed the study. Baseline variables are given in Table 1.

Table 1 Baseline characteristics in 49 type 1 diabetic patients with DN, after a 2 month washout period where all antihypertensive medication, except for slow-release furosemide, was withdrawn
Full size table

All doses of lisinopril significantly reduced UAER, ABP and eGFR from baseline (Table 2).

Table 2 Laboratory data during treatment with lisinopril 20, 40 and 60 mg in random order compared with baseline in 49 type 1 diabetic patients with DN
Full size table

Reductions in UAER, given as mean difference (95% CI) from baseline were 63% (55–69), 71% (66–76) and 70% (64–75) with increasing doses of lisinopril (p < 0.001; see ESM Fig. 2). Compared with the standard dose, i.e. lisinopril 20 mg, there was a significant further reduction in UAER with lisinopril 40 mg of 23% (8–35) and 60 mg 19% (3–31) p < 0.05. There was no difference in UAER between lisinopril 40 mg and 60 mg.

The reduction in ABP did not differ significantly between doses (except p < 0.05 for diastolic ABP 20 mg vs 40 mg). The effect on BP lasted through 24 h, with optimal effect from lisinopril 40 mg, as shown in ESM Fig. 2. Laboratory data are specified in Table 2.

Patients who responded better on increasing doses of lisinopril could not be predicted by levels of UAER, ABP, eGFR or plasma renin activity at baseline or during treatment with lisinopril 20 mg. There was a significant correlation between ΔUAER and ΔBP (changes from 20 to 40 mg, R = 0.49, p < 0.001), corresponding to a 2.8% reduction in UAER for every 1 mmHg reduction in systolic BP. After adjustment for changes in 24 h ABP, the difference in UAER between 20 mg and 40 mg lisinopril was still significant (p < 0.01).

Discussion

We evaluated the optimal renoprotective dose of the ACEI lisinopril in type 1 diabetic patients with DN. The overall reduction in albuminuria and BP was most pronounced during treatment with lisinopril 40 mg compared with the previously recommended dose of 20 mg daily. Our study shows that increasing the dose of lisinopril to 60 mg does not reduce albuminuria or BP further. The beneficial effects of lisinopril 40 mg were obtained without any additional adverse effects compared with 20 mg.

Our study has the strength of being a randomised, double-masked crossover trial conducted in one centre, ensuring a uniform setting and uniform treatment of patients. Long-term evaluation of the efficacy and safety of high dosing of lisinopril has not been performed and should be conducted in a controlled manner in the future.

The importance of finding the optimal renoprotective dose was previously acknowledged in other studies evaluating the optimal dose of angiotensin II receptor blockers in type 2 diabetic patients with micro- or macroalbuminuria [6, 7]. Although BP reductions were similar, further reduction in UAER was achieved by increasing dosage of angiotensin II receptor blocker above the recommended dose [6, 7]. Anti-proteinuric effects are associated with reno- and cardioprotective long-term effects.

Increasing RAAS blockade without further clinical benefit is costly and may be associated with more adverse events and even risk of renal impairment, as recently suggested by the ONTARGET study [8]. Detailed information from ONTARGET about albuminuria and causes of renal endpoints including dialysis, is so far lacking but may provide important information about the benefit or harm of aggressive RAAS blockade on renal disease, including DN.

Previous dose titration studies with lisinopril have been conducted in (1) non-diabetic patients with essential hypertension, using BP reductions as the main outcome [2, 3]; and (2) patients with non-diabetic nephropathies [9, 10]. Lisinopril 40 mg was shown to lower albuminuria and BP more effectively than lisinopril 20 mg. Higher doses have only been tested in essential hypertension studies. Overall, previous studies have provided evidence that suggests a better BP reduction with lisinopril doses higher than 20 mg daily and more efficient albuminuria lowering with lisinopril 40 mg daily (no higher doses tested); however, the recommended doses have been 20–40 mg lisinopril for hypertension and 10–20 mg for ‘renal complications caused by diabetes’.

Two patients had an increase in creatinine which was normalised after withdrawal of the study medication. This illustrates that it is important to evaluate kidney function during treatment with ACEI. In particular, this is important with ACEIs that are mainly (or exclusively) eliminated through the kidney, such as lisinopril. No patients had plasma potassium above 5.5 mmol/l on lisinopril 40 mg.

In conclusion, lisinopril 40 mg once daily was safe in our patients with DN and induced a further reduction in albuminuria and BP compared with the currently recommended dose of lisinopril 20 mg. Increasing the dose to 60 mg daily offered no beneficial effect on albuminuria or BP.