Liraglutide improves metabolic parameters and carotid intima-media thickness in diabetic patients with the metabolic syndrome: an 18-month prospective study

Liraglutide, a glucagon-like peptide-1 (GLP-1) analogue, is widely used for the treatment of type 2 diabetes mellitus (T2DM). It exerts several glycemic and non-glycemic effects, including the regulation of glucose levels by stimulating glucose-dependent insulin secretion and the suppression of glucagon secretion [1]. Liraglutide is also favorable for body weight by delaying gastric emptying and promoting satiety [2]. In addition, liraglutide has beneficial effects on waist circumference and numerous metabolic parameters including inflammatory markers, blood pressure (BP) and plasma lipids [3], thus improving the cardio-metabolic profile of patients with T2DM [4, 5]. Similar beneficial effects of liraglutide on metabolic parameters were also seen in obese individuals with prediabetes who were at high risk for progression to T2DM and the development of cardiovascular disease (CVD) [6], as well as in subjects who had hypothalamic pituitary injury followed by hyperphagia, in whom several features of the metabolic syndrome (MetS) were present [7].

Liraglutide decreases the prevalence of both MetS and prediabetes in obese, nondiabetic individuals [2, 8], suggesting that this antidiabetic drug might be beneficial in patients with the MetS. It was observed that patients with the MetS responded better to exenatide therapy compared to those without the MetS [9]. Although several clinical trials (reviewed in [3]) suggest that liraglutide, and GLP-1 analogues in general, can beneficially modulate the distinctive features of the MetS, studies performed in patients with diagnosed MetS are lacking. Therefore, we performed an 18-month prospective study in order to assess the effect of liraglutide on cardio-metabolic parameters in subjects with the MetS, collecting real-world data.

Statistical analysis was performed using SPSS software (V.17.0 for Windows, SPSS Inc., Chicago, USA). All the investigated parameters were assessed for normal vs non-normal distribution by the Kolmogorov–Smirnov normality test. Univariate analysis was performed using paired t test to estimate the difference between baseline and the final measurement at each 6 months. Chi squared test was used to assess if there was any association between cIMT changes and the disappearance of the MetS at the end of the study. ANOVA was used to evaluate changes in cardio-metabolic parameters from baseline to 6, 12, and 18 months of liraglutide treatment (p for trend), while correlation analysis was performed using the Spearman rank correlation method. Multivariate analysis (by multiple regression model) was performed in order to determine the independent effect of clinical and laboratory parameters on changes in cIMT.

Baseline characteristics of the patients are shown in Table 1. The effects of adding liraglutide on all evaluated parameters are shown in Table 2. We found a significant reduction in waist circumference (p = 0.0052), BMI (p = 0.0303), fasting glycemia (p < 0.0001) and HbA1c (p < 0.0001). Regarding plasma lipids, liraglutide therapy significantly decreased TC, TG and LDL-C (p = 0.0017, p = 0.0061 and p = 0.0039, respectively), while HDL-C increased, although the difference did not reach the statistical significance (p = 0.0853). Carotid IMT also reduced over the time (p < 0.0001). The trends in these metabolic parameters over the course of the study are shown in Fig. 1.

Age (years)	62 ± 9
Women, n (%)	50 (41)
Diabetes duration (years)	9 ± 8
Smoking habit, n (%)	30 (25)
Past history of cardiovascular diseases n (%)	17 (14)
Past history of cerebro-cardiovascular diseases n (%)	10 (8)
Family history of cardiovascular diseases, n (%)	74 (61)
Systolic blood pressure (mmHg)	128 ± 11
Diastolic blood pressure (mmHg)	79 ± 8
Hypertension, n (%)	93 (77)
Obesity, n (%)	68 (56)
Dyslipidemia, n (%)	93 (77)
Use of anti-hypertensive therapies
Beta-blockers, n (%)	18 (15)
Angiotensin-converting enzyme inhibitors, n (%)	29 (24)
Calcium entry blockers, n (%)	12 (10)
Diuretics, n (%)	28 (23)
Use of lipid-lowering drugs
Statins, n (%)	57 (47)
Fibrates, n (%)	3 (3)
Omega-3 fatty acids, n (%)	14 (13)
Aspirin use, n (%)	43 (35)

n number

	Baseline	6 months	12 months	18 months	p
Body weight (kg)	86 ± 19	83 ± 18c	82 ± 18c	82 ± 17c	0.2142
CV (cm)	109 ± 15	106 ± 12c	105 ± 14c	102 ± 16c	0.0052
BMI (kg/m2)	32 ± 9	30 ± 5b	30 ± 5b	30 ± 5b	0.0303
Glycemia (mmol/l)	9.38 ± 3.55	7.62 ± 2.50c	7.25 ± 2.08c	7.21 ± 2.66c	<0.0001
HbA1c (%)	8.76 ± 0.93	6.94 ± 1.38c	6.79 ± 1.02c	6.79 ± 1.13c	<0.0001
TC (mmol/l)	4.67 ± 1.37	4.25 ± 1.00b	4.21 ± 0.97c	4.21 ± 0.96c	0.0017
TG (mmol/l)	1.87 ± 1.20	1.58 ± 0.73b	1.57 ± 0.75b	1.51 ± 0.72c	0.0061
LDL-C (mmol/l)	2.71 ± 1.30	2.38 ± 0.90b	2.31 ± 0.92c	2.29 ± 0.84c	0.0039
HDL-C (mmol/l)	1.13 ± 0.28	1.16 ± 0.27	1.18 ± 0.29a	1.22 ± 0.29b	0.0853
cIMT (mm)	0.97 ± 0.18	0.89 ± 0.14c	0.82 ± 0.13c	0.78 ± 0.20c	<0.0001

Data are presented as mean ± SD

CV waist circumference, BMI body mass index, HbA1c glycated hemoglobin, TC total cholesterol, TG triglyceride, LDL low-density lipoprotein, HDL high-density lipoprotein, cIMT carotid-intima media thickness

Statistically significant p values are in italics

^ap < 0.05

^bp < 0.01

^cp < 0.001 (all vs baseline)

We performed Spearman correlation analysis in order to search for potential associations between changes in all the evaluated metabolic parameters after 18 months of liraglutide treatment (data not shown). We found that the reduction in HbA1c was significantly correlated with changes in body weight (r = 0.203; p = 0.025), waist circumference (r = 0.272; p = 0.003), BMI (r = 0.187; p = 0.040), fasting glycemia (r = 0.407; p < 0.0001), HDL-C (r = −0.225; p = 0.013) and TG (r = 0.192; p = 0.035). We also found that the reduction in glycemia was significantly associated with changes in TC (r = 0.233; p = 0.010), TG (r = 0.207; p = 0.022) and LDL-C (r = 0.242; p = 0.008). Changes in plasma lipids significantly correlated to each other, while the reduction in cIMT was significantly correlated with changes in TG (r = 0.362; p < 0.0001). Finally, the prevalence of the MetS significantly reduced during the study (p < 0.0001, Fig. 2) and the 26% of subjects no longer fulfilled the criteria for the MetS after 18 months of liraglutide treatment.

We also performed multivariate analysis (data not shown) in order to assess the independent effect of clinical and laboratory parameters on changes in cIMT, and we found that only TG had a predictive role (p = 0.0002).

There is a close association between T2DM and the incidence or development of the MetS. The presence of abnormal metabolic parameters such as central obesity, dyslipidemia, and hypertension further increase the risk of CVD in both T2DM and MetS [13], which remains elevated despite intensive pharmacological treatment [14]. In this 18-month prospective, real-world study we found that liraglutide improved several cardio-metabolic risk factors, including cIMT, in patients with T2DM and the MetS.

Several studies have shown that liraglutide has a positive effect on anthropometric parameters, in particular on body weight (as seen in the LEAD [Liraglutide Effect and Action in Diabetes] studies (reviewed in [3]) and waist circumference [15]. Liraglutide’s benefit was noted even in non-diabetic subjects [2, 8] and in those at high CV risk [16]. In the present study liraglutide progressively reduced waist circumference and BMI over time. Body weight significantly reduced after 6 months of therapy; however, no further change was found during the subsequent 18-month follow-up. A greater liraglutide effect on waist circumference, rather than BMI, may be linked to a potential more powerful effect of liraglutide on abdominal obesity. In this context, there is evidence that the addition of liraglutide to insulin is able to reduce waist circumference more than just increasing the insulin dosage in obese T2DM patients [17].

Moreover, several studies in the last few years have demonstrated a beneficial effect of liraglutide on body weight, independently of the effect on glucose metabolism; such studies typically used greater daily doses of liraglutide than those approved for the management of T2DM. Astrup et al. [2] reported a significant reduction in body weight after 20 weeks of liraglutide therapy in obese non-diabetic subjects with a dose of 3.0 mg/day—a higher amount than what was used in the present study (1.2 mg/day). However, in Japanese patients with T2DM, the reduction in body weight induced by liraglutide was maintained up to 2 years with much lower doses of the drug (ranging from 0.3 to 0.9 mg/day) [18].

In the present study we found a strong and significant reduction in glucose levels as well as a significant reduction in plasma lipids, consistent with established data [8, 19–21]. These latter findings are supported by the positive correlation between improvements in fasting glycemia and plasma lipids (namely TC, TG and LDL-C) as well as by the significant correlation between changes in HbA1c and TG. There was also an 8% increase in HDL-C, although it did not achieve statistical significance. The LEAD studies similarly showed a non-significant increase in HDL-C with liraglutide administration [21, 22]. More recently, the multicenter ROOTS study, which followed 245 diabetic subjects on liraglutide therapy (1.2 or 1.8 mg) for 12 months, showed a significant reduction in TC and LDL-C without changes in HDL-C [23].

We also evaluated the effect of liraglutide on cIMT as a surrogate marker of subclinical atherosclerosis. Some studies have shown that assessing the number of markers of MetS is more useful than the binary definition of MetS for predicting the atherosclerotic burden [24]. We found in the present study a significant association between cIMT changes and the disappearance of the MetS, highlighting the role of subclinical atherosclerosis in such high-risk patients (data not shown). Evidence from pre-clinical studies suggests that liraglutide may have beneficial effects on atherosclerotic plaque formation, progression and its stability [25], in addition to cardioprotective effects leading to improved outcomes and enhanced survival after acute myocardial infarction [26]. In the present study liraglutide significantly reduced cIMT in patients with the MetS during the 18-month follow-up, with the reduction gaining statistical significance after only 6 months of treatment.

The rapid regression of cIMT with the use of liraglutide, previously reported by our group [5] has generated great interest. Liraglutide has been shown to reduce cIMT in a stronger fashion than other incretin-based therapies, including sitagliptin, vildagliptin and exenatide (as reported in the meta-analysis by Song et al. [27]). The macrovascular benefits of liraglutide have been demonstrated with the publication of the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) study [28]. LEADER investigators have suggested that the regression in atherosclerosis induced by liraglutide may have contributed to the significant CV benefit. In the present study, liraglutide therapy was associated with normalization of cIMT values into a normal range; note that the joint European Society of Hypertension/European Society of Cardiology guidelines [29] consider cIMT <0.9 mm as normal. Since cIMT is a surrogate marker of early, subclinical atherosclerosis [30], our findings showing regression of cIMT with the use of liraglutide in MetS patients seem to be congruent with the emerging CV actions of liraglutide.

Interestingly, we also found a significant association between changes in cIMT and those in TG and, at multivariate analysis, TG was the only parameter with a predictive role for changes in cIMT. It is known that TG-metabolism plays a central role in CVD risk [31] and higher TG levels in T2DM subjects are associated with less favorable cardio-metabolic phenotypes and subclinical atherosclerosis [32], that may lead to increased CVD mortality [33]. Additionally, antioxidative markers may be inversely associated with TG levels in early preclinical atherosclerosis even with normal levels of plasma cholesterol [34], and our group has previously demonstrated that liraglutide is able to reduce markers of oxidative stress in patients with T2DM [35]. Kim SK et al. [36] have interestingly reported that nonalcoholic fatty liver disease (NAFLD) is very common in subjects with T2DM, but NAFLD not accompanied by insulin resistance is not associated with a carotid atherosclerotic burden. However, in this latter study [36], coexistence of NAFLD and insulin resistance seemed to be an independent predictor of increased cIMT. Therefore, a greater cIMT reduction in subjects with NAFLD and insulin resistance in our study remains a strong possibility.

As shown in Fig. 2, the prevalence of the MetS significantly decreased during the study. This finding is of considerable clinical value, and we observed a higher decrease in the prevalence of the MetS in comparison with the results reported by Astrup et al. [8], where a non-significant decrease of 12% was recorded. However, it should be highlighted that in the latter study all subjects were obese and only 21 (4%) had diabetes, while in our study 56% of subjects were obese and all of them had diabetes. In this context, a recent study showed that patients with the MetS were more likely to respond in a favorable manner to GLP-1 receptor agonist therapy (in that case, exenatide) than to a dipeptidyl peptidase 4 inhibitor [9].

Since the effect of liraglutide on cardio-metabolic parameters in patients with the MetS has not been evaluated, our data may have important clinical implications. The significant cardio-metabolic effect that we report with the use of liraglutide in the MetS is even higher than what was reported in most of the published studies with the use of this drug in T2DM patients [37]. This finding is of considerable interest, and it is consistent with that reported by Fadini et al. with the use of another GLP-1 analogue in an Italian cohort of patients. In the latter study, MetS patients had a better response to exenatide treatment than those without the MetS [9]. We hope that our findings will help pave the way for future research in the field.

Although liraglutide has been extensively studied in observational studies and clinical trials in cohorts of T2DM patients, to our knowledge ours is the first study directly investigating the effect of liraglutide on several cardio-metabolic risk factors in MetS subjects. It is likely that many subjects included in previous studies involving liraglutide fulfilled the criteria for the MetS; however, the uniqueness of the present study lies in the reporting of novel data that contribute to the understanding of the cardio-metabolic effects of liraglutide beyond glycemic control. Strengths of our study include the real-world setting, 18-month follow-up period, blinded measurements of cIMT, and high compliance rate with liraglutide therapy. Although the use of concomitant medications and CV agents was extensive, their doses were kept largely unchanged during the course of the study in order to minimize potential confounding effects. Further, multivariate analysis found no potential impact of concomitant therapy on changes in cIMT.

A limitation of our study stems from the lack of a control group with a metformin-only arm. However, previous studies have shown that metformin therapy has little or no impact on cIMT [38, 39], and at best, only a modest effect on waist circumference, weight or BMI [40, 41], and lipids [42, 43]. It is therefore unlikely that metformin was responsible for the significant modulation in anthropometric variables, plasma lipids and cIMT that we found in our study. A potential interpretation of the findings is that liraglutide’s effect on cIMT might have been mediated by significantly improved glycemic and metabolic profile; however, no significant correlations between changes in cIMT and those in metabolic parameters were evident. We cannot therefore exclude that the significant reduction in cIMT may be linked to favourable liraglutide effects beyond glycemic control, such as those on oxidative stress, inflammation and atherogenic lipoproteins, although not measured in the present study [3]. Finally, the statistical analysis revealed that the effect of liraglutide on several parameters, including cIMT, was very strong and highly significant; therefore, the probability of the occurrence of our findings by chance alone is extremely low.

This is the first report of a study prospectively investigating the effects of liraglutide on cardio-metabolic risk factors in subjects with the MetS. Liraglutide improved waist circumference, glucose parameters, plasma lipids and cIMT in an 18-month follow-up, although a major limitation is the lack of a control group. Our findings are of considerable clinical value, but their direct impact on CV outcome remains to be elucidated. The recently published liraglutide CV outcome study reported that liraglutide was able to reduce death from CV causes, nonfatal myocardial infarction and nonfatal stroke in patients with T2DM [28]. Whether this significant CV protection by liraglutide can be extended in patients with T2DM and the MetS, as those included in our study, is at present unknown.