Glucagon-like peptide 1 agonists have been found to modestly lower blood pressure in diabetic patients in most of the studies [8, 13]. Although the mechanisms leading to a blood pressure decrease during the treatment with GLP-1 agonists have not been fully elucidated it could be the result of an increased sodium and water excretion and/or inhibition of the RAAS activity [14]. In this study we investigated the effect of GLP-1R agonist liraglutide on blood pressure, hydration status, natriuresis and the activity of the renin-angiotensin-aldosterone axis in patients with diabetic kidney disease after a dose of liraglutide compared to placebo. The choice of liraglutide was substantiated by its favorable pharmacokinetic profile in patients with impaired kidney function and therefore by its increasing clinical use in this population [12]. Across studies and populations, liraglutide was found to be slowly absorbed following a subcutaneous injection, with a time to maximum plasma concentration of the drug (Tmax) of approximately 11–12 h [11]. The drug is completely degraded into several non-active metabolites within the body and as was shown in the study with radiolabeled drug, no intact liraglutide is excreted in either urine or feces [11]. What is more, no clear trend for a change in pharmacokinetics was evident across groups with increasing renal dysfunction [12] that allowed us to administer the same dose of the drug and the same blood sampling time points in patients with different degree of renal impairment who were included in our study.
The main finding of our study is the observation that a degree of renal impairment affects the magnitude of the effect of liraglutide on blood pressure and natriuresis. Contrary to our initial expectations, we did not observe a significant blood pressure lowering effect of liraglutide in patients with mild CKD, ie. with eGFR > 60 ml/min/1.73m2. Instead in patients with eGFR < 30 ml/min/1.73m2 a single injection of liraglutide led to a significant transient increase of blood pressure.
Seemingly the results of our study contrast with liraglutide in a range of clinical studies. In the metanalysis of 31 randomized controlled trials with GLP-1 agonists liraglutide and exenatide, Robinson et al. showed a reduction of systolic blood pressure by − 1.79 mm Hg [95%CI (-2.94)-(-0.64)] compared to placebo [15]. In the metanalysis of six LEAD trials, 26 weeks of liraglutide treatment with the same dose as in our study also led to a significant decrease of systolic blood pressure that reached 2.7 mmHg. A second metanalysis demonstrated that the treatment with 1.2 mg liraglutide reduced SBP compared with placebo with a mean difference of − 5.60 with 95% CI − 5.84 to − 5.36, p < 0.00001 [13]. The studies included in these metanalyses were performed however in patients with without or with only moderately impaired kidney function. In contrast to the hypotensive effect of liraglutide that was reported in earlier studies [16], our study showed that both systolic and diastolic blood pressure did not decrease. Instead we noticed a significant increase of systolic as well as diastolic blood pressure after liraglutide injection but this finding was limited to the patients with eGFR < 30 ml/min/1.73m2. Interestingly, in a few small randomized controlled trials that assessed the safety and efficacy of liraglutide in patients with impaired kidney excretory function, the effect of the drug on blood pressure differed from that observed in patients with normal kidney function or in early CKD. In LIRA-RENAL study, that included 279 patients with moderate renal impairment (eGFR 30–59 mL/min/1.73 m2), SBP reduction of 2.45 mmHg was seen but the change was not significantly different between liraglutide and placebo (p = 0.25) [17]. In the LIRA-RENAL trial a mean eGFR was however significantly lower than in our patients, the patients had a longer history of diabetes and liraglutide was administered in a 50% higher daily dose of 1.8 mg. Hiramatsu et al. studied a long-term effect of liraglutide in three subgroups of CKD patients with eGFR > 60 ml/min/1.73m2, 30–60, and < 30 ml/min/1.73m2 [18]. They found that a systolic blood pressure decreased significantly after 24 months of treatment with 0.9 mg/24h compared to baseline in all groups. A small improvement of blood pressure control after liraglutide treatment was also seen in patients with end-stage kidney disease treated by peritoneal dialysis [19]. In contrast to the blood pressure decrease reported by Hiramatsu et al. in patients with CKD including stage 4 with eGFR < 30 ml/min/1.73m2, Indorn et al. did not find any significant change of blood pressure in chronic dialysis patients after 12 weeks of liraglutide treatment, titrated to a maximum dose of 1.8 mg daily [20]. We may speculate that the difference between the results of our study and from most previous studies may primarily be dependent of treatment duration but also may be affected by different methods of blood pressure measurement. In the randomized placebo controlled study of Kumarathurai et al. in patients with type 2 diabetes mellitus and ischemic heart disease, in which ambulatory blood pressure monitoring was used liraglutide did not induce any significant change of mean 24h SBP, whereas the tendency towards blood pressure reduction was observed if office blood pressure was analyzed [21]. In another study there was a small but significant increase of 24-h and nighttime diastolic BP after 3 weeks of liraglutide treatment [22] similar to the effect of a single dose of the drug in our study. In contrast Liakos et al. found that liraglutide administered for 5 weeks had a favorable effect on 24h systolic but not diastolic blood pressure profile [23].
The acute effect of liraglutide on blood pressure was investigated in several studies. The studies of Bharucha et al. [24], Asmar et al. [25] and Mendis et al. [26] in healthy volunteers revealed a small but significant increase of blood pressure after a single injection of liraglutide. These results are similar to seen in our study but only in patients with eGFR < 30 ml/min/1.73m2. These results may show that the acute effect of liraglutide may be different that chronic. The difference of the acute and chronic effects of GLP-1R agonist was also demonstrated in a study of von Scholten et al. [27] that showed 24h SBP increase during the treatment as high as 10 mmHg after 3 days of treatment (p = 0.008). The effect decreased to 7 mmHg on day 7 (p = 0.033) and disappeared after 29 days of treatment. At that time 24h SBP decreased and was 7 mmHg lower compared with baseline value. That difference was however not significant (p = 0.11).
The cumulative evidence may support a role of GLP-1 in the regulation of fluid and electrolyte balance mediated by the reduction of thirst and increased natriuresis. GLP-1 infusion significantly increased the renal clearance of sodium by about 40% with a simultaneous 20% decrease of sodium reabsorption in the proximal tubule [28]. Gutzwiller et al. observed that an intravenous infusion of native GLP-1 increased sodium excretion from 59 to 96 mmol/180 min (p = 0.015) in obese males [29]. A similar significant effect of GLP-1R agonists on sodium excretion was observed in both experimental and clinical studies. The administration of exenatide led to an increase of both absolute and fractional sodium excretion in healthy men in a study of Muskiet et al. [30] and similar results were reported in the patients with diabetes [31]. Liraglutide significantly increased 24-h and nighttime urine sodium excretion in patients with type 2 diabetes mellitus and arterial hypertension after 3-week treatment [22] but similar effect was also reported after a single dose of this drug [32]. While our results confirmed the natriuretic effect of liraglutide, the increase of sodium excretion was significantly smaller in the patients with eGFR < 30 ml/min/1.73 m2 compared to those with eGFR > 60 ml/min/1.73 m2. These finding is not unexpected since GLP-1 may increase natriuresis through the inhibition of the sodium-hydrogen ion exchanger isoform 3 in the proximal tubule which expression decreases with the progression of CKD [33]. Although the natriuretic effect of liraglutide was small in the patients with eGFR < 30 ml/min/1.73 m2 the thoracic fluid index, a marker of hydration status assessed with impedance cardiography significantly decreased. The same effect of liraglutide was revealed in another study in patients with diabetes and hypertension but without renal impairment [27]. The different effect on the hydration status seen in our groups might have been caused by the different baseline hydration status as fluid retention tends to progress with renal impairment [34].
A significant decrease of plasma aldosterone was seen in our study after liraglutide compared with placebo only in patients with eGFR < 30 ml/min/1.73m2. Skov et al. did not find any significant changes of renin, aldosterone, and urinary angiotensinogen after a single GLP-1 infusion in healthy men [28] and the same result was reported in patients with diabetes without kidney function impairment [32]. In contrast the decrease of plasma aldosterone was observed by Baretic et al. [35] and Sedam et al. [36] in healthy participants. We may speculate that a larger reduction of plasma aldosterone in the patients with advanced CKD seen in our study could be caused by a higher local activity of RAAS due to kidney dysfunction.
After the injection of liraglutide serum atrial natriuretic peptide increased more than after placebo in both study groups. Such effect was not seen in previous studies but most of them were conducted in healthy men. Contrary to our results von Scholten et al. found that after 7 weeks of liraglutide administration serum ANP decreased by 20% (95% CI 12–28%, p < 0.001) [27]. A significant increase of ANP level after the administration of a GLP-1 agonist was found mainly in experimental models with hyperreninemia and fluid overload-dependent hypertension [8]. The tendency to fluid retention is highly prevalent in diabetic kidney disease [34]. We suppose that there might be an association between a grade of overhydration and the magnitude of the effect of GLP-1R agonists on atrial natriuretic peptide secretion, that may explain a significantly larger increase of serum ANP concentration in patients with eGFR < 30 ml/min/1.73m2 in our study. However we did not reveal any significant correlation between thoracic fluid index and its change and the change of ANP concentration. The small sample size could be one of possible reasons.
Despite the limitations like small sample size, the recruitment limited to two stages of DKD, the assessment of only an acute effect of the drug our study found a short-term hypertensive effect of GLP-1R agonist liraglutide in diabetic kidney disease that seems to be related to impaired renal sodium excretion. The natriuretic effect may not sufficiently counteract the initial increase of blood pressure in patients with severe kidney function impairment. We also postulate that increased natriuretic after liraglutide in severe diabetic kidney disease may result from increased atrial natriuretic peptide secretion and decreased aldosterone secretion. The mechanism of these changes will however require further research.