Impact of statin therapy on plasma leptin concentrations: a systematic review and meta‐analysis of randomized placebo‐controlled trials

Abstract

Objectives

The effects of statins on insulin sensitivity, metabolic homeostasis and adipokines in humans are controversial. Several studies have investigated the impact of statin therapy on plasma leptin concentrations but the results have been inconsistent. The aim of the present study was to conduct a systematic review and meta‐analysis of available evidence to calculate the effect size of statin therapy in changing serum leptin concentrations.

Methods

A systematic search in PubMed‐Medline, SCOPUS, Web of Science and Google Scholar databases was performed to identify randomized placebo‐controlled trials investigating the effect of statins on plasma leptin concentrations. A random‐effects model and generic inverse variance method were used for meta‐analysis. Sensitivity analysis, risk‐of‐bias evaluation and publication bias assessment were carried out using standard methods. Random‐effects meta‐regression was used to evaluate the impact of treatment duration on the estimated effect size.

Results

Six trials, with a total of 425 subjects, met the eligibility criteria. Overall, statin therapy had no significant effect on leptin levels (weighted mean difference −0.32 ng ml^–1, 95% confidence nterval: −2.94, 2.30, P = 0.813). This effect was robust in the sensitivity analysis and in subgroup analyses of trials with <12 or ≥12 weeks' duration. There was no association between the duration of statin therapy and changes in plasma leptin levels. Furthermore, there was no differential effect of hydrophilic and lipophilic statins on plasma leptin concentrations.

Conclusion

Unless more consistent evidence becomes available in the future, the hypothesis of a relationship between statin use and serum leptin concentrations seem to be unfounded.

What is Already Known About this Subject

• Studies investigating the impact of statin therapy on plasma leptin concentration have found controversial results.

• Some studies suggest that the effect on leptin concentration could be limited to lipophilic statins due to their capacity of passing cell membrane.

What this Study Adds

• This study is the first metanalysis on randomized controlled trials investigating the impact of statin therapy on plasma leptin concentration.

• Six trials of 82 have been included in the meta-analysis, for a total of 425 subjects.

• The hypothesis of a relationship between statin use and serum leptin concentrations seem to be unfounded.

Tables of Links

TARGETS
Other protein targets 2
C reactive protein
Transporters 3
GLUT4

LIGANDS
TGF β
Cerivastatin
Pravastatin
Simvastatin
Atorvastatin
Lovastatin
Rosuvastatin
Leptin

These Tables list key protein targets and ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 1, and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 2, 3.

Introduction

Statins are the mainstay for treating dyslipidaemia, and their efficacy in reducing cardiovascular outcomes has been supported by several lines of robust evidence 4, 5. Additionally, statins possess several lipid‐independent (pleiotropic) properties, including anti‐inflammatory, renoprotective, immunoregulatory, hypouricaemic, antioxidant and anti‐thrombotic functions 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18.

Some experimental studies have demonstrated that lipophilic statins, especially at high doses, may cause unfavourable metabolic effects, such as reduction of insulin secretion and exacerbation of insulin resistance 19.

Some studies have shown that statins might improve insulin resistance 20 and insulin secretion 21, increase insulin‐mediated glucose uptake and have a protective effect on the development of diabetes 20.

Nonetheless, experimental studies, using in vitro and animal models, have demonstrated that statins can reduce glucose‐induced insulin secretion in rat islet β‐cells, block L‐type Ca^{2 +} channels 22, 23, produce insulin resistance in adipocytes, inhibit isoprenoid biosynthesis and protein farnesylation, and reduce the expression of glucose transporter 4 (GLUT4) at the cell surface 24.

These effects are particularly significant for lipophilic statins (simvastatin, atorvastatin, lovastatin) because of their high capacity to pass through the cell membrane and diffuse into extrahepatic tissue. Hydrophilic statins (pravastatin, rosuvastatin) do not show these interactions in in‐vitro studies 22 but in some studies even these statins have been associated with an increased risk of diabetes 25.

Leptin is an adipokine that plays an important role in metabolic homeostasis 26, 27, 28. It inhibits hunger by acting on specific receptors in the arcuate nucleus of the hypothalamus 29. In obese subjects, a decreased sensitivity to leptin and an increased leptin level have been reported, resulting in a reduced capacity to detect satiety 30. Moreover, leptin modulates several other activities, such as the immune response to atherosclerosis 31 and regulation of the inflammatory response 32. Some studies have also shown an association between leptin status and renal damage, which is exerted through both direct and indirect mechanisms 19.

Hitherto, several studies have investigated the impact of statin therapy on plasma leptin levels but the results have been equivocal, showing elevations, reductions and no change in plasma leptin levels following statin therapy.

We therefore performed a systematic review and meta‐analysis of the results of randomized placebo‐controlled trials in order to estimate the size of the effect of statin therapy on plasma leptin concentrations.

Methods

Search strategy

The present systematic review and meta‐analysis was designed according to the 2009 PRISMA statement 33. A literature search was performed in PubMed‐Medline, SCOPUS, Web of Science and Google Scholar databases. The following terms were searched in titles and abstracts: (atorvastatin OR simvastatin OR rosuvastatin OR fluvastatin OR pravastatin OR pitavastatin OR lovastatin OR cerivastatin OR ‘statin therapy’ OR statins) AND (leptin OR ‘ob gene product’ OR ‘ob protein’ OR ‘obese gene product’ OR ‘obese protein’) AND (placebo). The wild‐card character ‘*’ was used to increase the sensitivity of the search strategy.

Study selection

The inclusion criteria for original studies were:

• – Parallel or crossover placebo‐controlled clinical trials;
• – Evaluation of the impact of a statin on plasma/serum concentrations of leptin;
• – Statin tretament duration of ≥2 weeks; and
• – Availability of information on leptin levels at baseline and at the study end in both statin and placebo groups, and providing the net change in leptin levels in the study groups.

Observational studies or studies lacking a proper control group (e.g. nonplacebo controls), and studies recruiting patients on stable statin therapy were excluded.

Data extraction

Data regarding the authors, publication date, study design and location, number and demographic characteristics of participants, and baseline anthropometric and biochemical parameters [including total cholesterol, low‐density lipoprotein cholesterol (LDL‐C), high‐density lipoprotein cholesterol (HDL‐C), triglycerides, high‐sensitivity C‐reactive protein (hs‐CRP) and glucose] were collected from each study. In addition, baseline and follow‐up serum/plasma concentrations of leptin (or alternatively net change values if applicable) were retrieved from each included study.

Quality assessment

Risk of bias (selection bias, performance bias, detection bias, attrition bias, reporting bias and other sources of bias) in the included studies was evaluated according to the Cochrane criteria 34.

Quantitative data synthesis

All analyses were performed using Comprehensive Meta‐Analysis (CMA) V2 software (Biostat, NJ, USA) 35. Net changes in circulating leptin concentrations were calculated as previously reported 12, 36, 37. To convert values reported as median and interquartile range into mean and standard deviation (SD), we used the method described by Hozo et al. 35. Conversion of standard error of the mean (SEM) to SD was performed by multiplying SEM by the square root of the number of subjects.

Owing to the interstudy heterogeneity with respect to the demographic characteristics of study populations, design and protocol of treatment with statins, a random‐effects model (using the DerSimonian–Laird method) and the generic inverse variance method were used for the meta‐analysis 38. In studies with more than one statin treatment arm, the control (placebo) group was split into equal (where possible) parts to provide separate control arms and avoid unit‐of‐analysis error due to double counting of control subjects. Quantitative evaluation of interstudy heterogeneity was performed using the I² index. The weighted mean difference (WMD) and 95% confidence interval (CI) were used as the summary statistic of the estimated effect size. Leave‐one‐out sensitivity analysis was performed to assess the robustness of the effect size 39, 40.

Meta‐regression

In order to assess the association between estimated effect size and potential confounders, including changes in plasma LDL‐C concentrations (including potential confounders), meta‐regression analysis was conducted under a random‐effects model and unrestricted maximum likelihood method.

Publication bias

The presence of publication bias in the meta‐analysis was evaluated using visual inspection of Begg's funnel plot asymmetry, and Begg's rank correlation and Egger's weighted regression tests. Adjusted effect size for publication bias was calculated using the Duval and Tweedie ‘trim and fill’ method 41.

Results

The initial literature search yielded 82 articles. Of these, 13 were selected for full‐text assessment. After careful assessment, six articles met the inclusion criteria and were selected for meta‐analysis, with eight treatment arms globally. Reasons for rejecting the other seven articles were as follows: absence of the control group, not being a placebo‐controlled study, lack of measurement of leptin levels, and incomplete data. A summary of the study selection process is shown in Figure 1.

Figure 1.

Flow chart of the number of studies identified and included in the meta‐analysis

A total of 425 subjects were included in the six eligible studies, comprising 262 subjects in the statin group and 178 subjects in the control group (considering that in the Paniagua study 21 all the 15 subjects belong both to the study group and to the control group). Two studies divided the sample into two treatment arms and a single control group 42, 43. The largest study had a population size of 134 subjects 43, while the smallest study recruited 15 subjects 21. Included studies were published during the period 2002–2014, and were conducted in Lebanon 44, Korea 42, 43, Iran 45, Spain 21 and Poland 46. Three trials used simvastatin, at dosages of 20 mg 42, 43 or 40 mg 43, 46; two trials used pravastatin 42, 44, one trial used lovastatin 45 and another trial used cerivastatin 21. The duration of therapy ranged between 2 months and 3 months. Each trial was designed as a placebo‐controlled study. Two trials were double blind 21, 24, while the other four were single blind 42, 43, 45, 46. Selected studies were performed in subjects with obesity 21, diabetes 21, 45 and isolated hypercholesterolaemia 42, 43, 46. One trial was performed in healthy subjects 44. The demographic and baseline biochemical parameters of the included studies are shown in Table 1.

Table 1.

Demographic characteristics of the included studies

Study	Gannagé‐Yared et al. 2005 44	Gholamin et al. 2014 45	Koh et al. 2009 42		Koh et al. 2011 43		Krysiak et al. 2014 46	Paniagua et al. 2002 21
Location	Lebanon	Iran	Korea		Korea		Poland	Spain
Design	Double‐blind trial	Single‐blind trial	Single‐blind trial		Single‐blind trial		Single‐blind trial	Double‐blind trial
Duration of trial	12 weeks	3 months	2 months		2 months		12 weeks	3 months
Inclusion criteria	Healthy subjects	Type 2 diabetic patients	Hypercholesterolaemic patients		Hypercholesterolaemic patients		Hypercholesterolaemic patients	Obese, type II diabetes patients

			1st arm	2nd arm	1st arm	2nd arm
Statin used	Pravastatin	Lovastatin	Pravastatin	Simvastatin 20 mg	Simvastatin 20 mg	Simvastatin 40 mg	Simvastatin 40 mg	Cerivastatin
Participants	40	65	127		134		44	15
Study group	19	30	42	43	45	45	23
Placebo	21	35	42		44		21
Age (years)								57 ± 7
Study group	51.6 ± 13.0	54.5 ± 6.1	56 ± 2	58 ± 2	58 ± 2	61 ± 2	51.9 ± 2.7
Placebo	46.3 ± 9.7	50.5 ± 8.3	59 ± 1		57 ± 2		51.1 ± 2.6
Female (%)		N/A						47
Study group	47		62	60	56	58	39
Placebo	43		64		52		43
Smoker (%)		N/A	N/A		N/A			N/A
Study group	26						22
Placebo	48						24
BMI ( kg m – 2 )
Study group	26.8 ± 4.28	25.6	25.5 ± 0.5	25.3 ± 0.5	25.3 ± 0.5	25.4 ± 0.4	26.5 ± 2.6	32.1 ± 1.0
Placebo	26.1 ± 4.20	25.7	24.5 ± 0.5		24.9 ± 0.5		27.2 ± 2.6	31.8 ± 1.0
Glucose (mg dl –1 )								131 ± 5
Study group	92.8 ± 15.1	157.8	99 ± 3	97 ± 2	85 ± 2	91 ± 3	95 ± 4	123 ± 5
Placebo	89.5 ± 9.2	151.5	94 ± 3		94 ± 3		95 ± 5	121 ± 5
LDL‐C (mg dl –1 )								N/A
Study group	159.6 ± 43.9	116.2	170 ± 5	178 ± 5	176 ± 5	177 ± 5	186 ± 13
Placebo	149.7 ± 43.8	124.1	177 ± 6		177 ± 6		178 ± 12
HDL‐C (mg dl –1 )								N/A
Study group	44.1 ± 14.1	40.0	54 ± 2	51 ± 2	52 ± 2	56 ± 2	46 ± 4
Placebo	47.8 ± 13.5	39.8	55 ± 2		54 ± 1		47 ± 4
Total cholesterol (mg dl –1 )								217 ± 8
Study group	234 ± 51.7	199.0	254 ± 5	260 ± 5	258 ± 5	263 ± 4	258 ± 16	163 ± 5
Placebo	229.8 ± 48.9	194.3	267 ± 5		268 ± 5		246 ± 14	202 ± 9
Triglycerides (mg dl –1 )								164 ± 9
Study group	146.6 ± 56.9	175.3	162 ± 9	143 ± 13	142 ± 13	155 ± 14	122 ± 12	114 ± 6
Placebo	168.5 ± 100.7	172.5	147 ± 2		144 ± 12		120 ± 12	133 ± 6
SBP (mmHg)	N/A		N/A		N/A		N/A
Study group		139						149.0 ± 5.8
Placebo		137						146.0 ± 5.6
DBP (mmHg)	N/A		N/A		N/A		N/A
Study group		76						84.0 ± 3.7
Placebo		73						83.0 ± 1.9
CRP (mg dl –1 )	N/A		N/A		N/A		N/A	N/A
Study group		3.52
Placebo		2.75
hs‐CRP (mg l –1 )	N/A
Study group			1.45	1	1.1	0.8	3.4 ± 0.5	2.23 ± 0.4
Placebo			1		1		3.5 ± 0.6	3.57 ± 0.6
Leptin (ng ml –1 )
Study group	9.83 ± 3.93	10.78	7.43	6.73	9.5 ± 1.1	11.2 ± 1.1	24.5 ± 5.0	30.66 ± 5.5
Placebo	8.97 ± 4.56	9.95	9.51		10.3 ± 0.9		24.1 ± 4.9	28.49 ± 4.1

BMI, body mass index; CRP, C‐reactive protein; DBP, diastolic blood pressure; HDL‐C, high‐density lipoprotein cholesterol; hs, high‐sensitivity; LDL‐C, low‐density lipoprotein cholesterol; N/A, not available; SBP, systolic blood pressure

Some of these studies showed a significant reduction in leptin levels 45, 46 in patients treated with simvastatin or lovastatin. Other studies – using cerivastatin and pravastatin – showed no significant difference between the study and placebo groups in leptin levels after a 4‐week and 12‐week treatment period, respectively 21, 44. Finally, Koh et al. found increased leptin levels in patients treated with simvastatin, with no significant difference in those treated with pravastatin 42, 43.

Risk–of‐bias assessment

Most of the selected studies did not provide sufficient data about random sequence generation and allocation concealment. Nevertheless, other potential sources of bias were addressed adequately by most of the included trials. Financial support was not reported in two studies 42, 45. Details on the risk of bias among included trials are summarized in Table 2.

Table 2.

Risk of bias assessment in the studies considered for meta‐analysis

Study	Random sequence	Allocation concealment	Blinding	Incomplete outcome data	Selective reporting	Free of other bias
Gannagé‐Yared et al. 2005 44	U	U	L	L	L	L
Gholamin et al. 2014 45	U	U	H	L	L	L
Koh et al. 2009 42	U	U	H	L	L	L
Koh et al. 2011 43	U	U	H	L	L	L
Krysiak et al. 2014 46	H	H	H	L	L	H
Paniagua et al. 2002 21	U	L	L	L	L	L

Criteria defined for quality assessment are based on the Cochrane guidelines. H, high risk of bias; L, low risk of bias; U, unclear or unrevealed risk of bias

Effect of statin therapy on plasma leptin concentrations

Changes in plasma leptin concentrations following statin therapy were reported in eight treatment arms. No significant alteration in plasma leptin concentrations was observed following statin therapy (WMD: −0.32 ng ml^–1, 95% CI: −2.94, 2.30; P = 0.813). This effect was robust in the sensitivity analysis (Figure 2). In subgroup analysis, there was no significant change of plasma leptin concentrations either in the subset of trials of <12 weeks' (WMD: 0.49 ng ml^–1, 95% CI: −1.94, 2.92; P = 0.692) or ≥12 weeks' (WMD: −1.21 ng ml^–1, 95% CI: −7.70, 5.27; P = 0.713) duration (Figure 3). With respect to statin type, there was no differential effect of hydrophilic (WMD: −0.54 ng ml^–1, 95% CI: −3.54, 2.46; P = 0.722) and lipophilic (WMD: −0.25 ng ml^–1, 95% CI: −4.45, 3.95; P = 0.907) statins on plasma leptin concentrations (Figure 4). Consistently, the pooled effect size did not differ significantly between subgroups of studies in hypercholesterolaemic (WMD: −1.60 ng ml^–1, 95% CI: −5.22, 2.03; P = 0.388) and nonhypercholesterolaemic (including subjects with prediabetes, diabetes or obesity) (WMD: 2.33 ng ml^–1, 95% CI: −0.11, 4.77; P = 0.062) populations (Figure 5).

Figure 2.

Plot A is the forest plot displaying weighted mean difference and 95% confidence intervals for the impact of statin therapy on plasma leptin concentrations. Plot B shows the results of leave‐one‐out sensitivity analysis (removing one study each time and repeating the analysis)

Figure 3.

Forest plot displaying weighted mean difference and 95% confidence intervals (CIs) for the impact of statin therapy on plasma leptin concentrations in trials lasting <12 weeks (A) and ≥12 weeks (B)

Figure 4.

Forest plot displaying weighted mean difference and 95% confidence intervals (CIs) for the impact of lipophilic (A) and hydrophilic (B) statins on plasma leptin concentrations

Figure 5.

Forest plot displaying weighted mean difference and 95% confidence intervals (CIs) for the impact of statins on plasma leptin concentrations in hypercholesterolaemic (A) and non‐hypercholesterolaemic (including subjects with prediabetes, diabetes or obesity) (B) populations

Random‐effects meta‐regression was performed to evaluate if changes in plasma leptin concentrations are dependent on the magnitude of the LDL‐C reduction by statins. The results did not suggest any significant association between changes in plasma levels of leptin and LDL‐C concentrations (slope: 0.03; 95% CI: −0.18, 0.24; P = 0.763) (Figure 6).

Figure 6.

Meta‐regression plots of the association between mean changes in plasma leptin and low‐density lipoprotein cholesterol (LDL‐C) concentrations

Publication bias

The funnel plot of standard error by effect size (mean difference) was slightly asymmetric. This asymmetry was addressed by imputing two potentially missing studies on the left‐hand side of funnel plot using ‘trim and fill’ correction. The imputed effect size was −1.12 ng ml^–1 (95% CI: −3.45, 1.20). There was no sign of publication bias according to the results of Begg's rank correlation (Kendall's tau with continuity correction = −0.07, z = 0.25, two‐tailed P‐value = 0.805) and Egger's linear regression (intercept = 0.77, standard error = 2.55; 95% CI = −5.48, 7.01; t = 0.30, df = 6.00, two‐tailed P = 0.774) tests (Figure 7).

Figure 7.

Funnel plot detailing publication bias in the studies reporting the impact of statin therapy on plasma leptin concentrations. Open diamond represents observed effect size; closed diamond represents imputed effect size. SE, standard error

Discussion

The present meta‐analysis of randomized controlled trials did not support the hypothesis that statin treatment is associated with an alteration in serum leptin concentration. There are different reasons that may explain this finding. First, there are many different factors regulating leptin synthesis and secretion that may be affected by statin treatment, and these factors are also influenced by interindividual difference and predisposition. Furthermore, although it has been observed in an in vitro model that simvastatin treatment produces a suppression of leptin gene expression in adipocytes, without any role of protein farnesylation 47, no in vivo evidence is available at the moment.

Our findings indicated that both the reported pleiotropic effects of statins, such as improvement in endothelial function, increased bioavailability of nitric oxide, and antioxidant and anti‐inflammatory properties 48, and the reported negative metabolic effects of statins on insulin sensitivity and glucose homeostasis are probably independent of a reduction or increase in leptin concentrations.

It should be noted that leptin and statins act on different pathways. Statins inhibit the mevalonate pathway, interfering with the prenylation of GTPase, such as the Rho GTPase 49, involved in insulin signal transduction, and also leads to reduced GLUT4 expression. Leptin signalling occurs typically through the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway, where STAT dimers translocate into the nucleus and act as transcription factors by binding to specific response elements in the promoter of their target genes, such as suppressor of cytokine signalling (SOCS) and pro‐opiomelanocortin 50, and also reduce tyrosine phosphorylation of the insulin receptor 51.

In the same way, statins and leptin influence the inflammatory response via different pathways.

Statins have many different effects that diminish innate and adaptive immune responses, and thereby reduce inflammation. They can reduce inflammation, diminishing the innate and adaptative response, with inhibition of isoprenoid synthesis through the mevalonate pathway (however, this is not the only anti‐inflammatory effect of statins) 52, 53. They can also block the inflammatory responses of endothelial cells, increase regulatory T‐cell differentiation, reduce T‐cell cytotoxicity and reduce the antigen‐presenting function of dendritic cells 54. By contrast, leptins can activate the JAK/STAT pathway through an interaction with specific receptors on neutrophils, monocytes, macrophages and subpopulations of T and B cells 55.

According to our results, it should be noted that, even if some studies have found a correlation between statin therapy and leptin levels, no authors have proposed or described a specific biological pathway by which this therapy could influence leptin levels.

The results were consistent in subgroups with different treatment durations and types of statin (hydrophilic vs. lipophilic), and with or without hypercholesterolaemia. The latter result deserves consideration because lipophilic statins are expected also to act on fat tissue and muscle, whereas hydrophilic statins cannot diffuse into extrahepatic tissues 22. Accordingly, a greater increase in leptin concentrations would be expected in patients taking a lipophilic statin. The study showing the largest reduction in leptin levels, however, was the one by Krysiak et al. 46, which investigated the effect of simvastatin (a lipophilic statin), and this result was in contrast to that of other studies investigating the same molecule. Even though there appears to be no difference in patient characteristics in the studies with simvastatin 42, 43, 46, the divergent results obtained by Krysiak et al. may have been due to a higher level of hs‐CRP in treated patients.

In the interpretation of our results, there are certain limitations that need to be taken into account. First, all the studies had a small sample size, ranging from 15 to 134, and with a total of 425 participants. In addition, the two studies by Koh et al. 42, 43 included some of the same patients. Furthermore, recruited participants had different backgrounds. For example, four studies included hypercholesterolaemic patients with a mean body mass index ranging from 25 kg m⁻² to 27 kg m^–2, one study included obese patients with type 2 diabetes, and another study included patients with advanced‐stage type 2 diabetes complicated by nephropathy. The focus on diabetic patients is of relevance because some studies have shown that lipophilic statins are taken up by the pancreas, fat tissue and muscle, where they may cause a reduction in insulin secretion, exacerbation of insulin resistance and new‐onset type 2 diabetes mellitus 22, 42, 56. This is likely to be particularly important in patients with metabolic disturbances, such as reduced glucose tolerance or type 2 diabetes. In the present study, we tried to minimize the impact of interstudy heterogeneities by applying a random‐effects model for quantitative data synthesis. In addition, the results of sensitivity analysis revealed that the estimated effect size was not sensitive to any single study included in the meta‐analysis.

In conclusion, the present meta‐analysis did not suggest any significant change in plasma leptin concentrations following statin therapy. Unless more consistent evidence becomes available in the future, the hypothesis of a relationship between statin use and serum leptin concentrations seem to be unfounded.

Competing Interests

There are no competing interests to declare.

Contributors

All the authors of this paper contributed significantly to its drafting and revision.

Sahebkar, A. , Giua, R. , and Pedone, C. (2016) Impact of statin therapy on plasma leptin concentrations: a systematic review and meta‐analysis of randomized placebo‐controlled trials. Br J Clin Pharmacol, 82: 1674–1684. doi: 10.1111/bcp.13086.