Impact of fibrates on circulating cystatin C levels: a systematic review and meta-analysis of clinical trials

Abstract

Aims: To assess the effect of fibrates on circulating cystatin C levels.

Material and methods: Clinical studies evaluating the effect of a fibrate on circulating cystatin C levels were searched in PubMed-Medline, SCOPUS, Web of Science, and Google Scholar databases. A random-effect model and generic inverse variance method were used for quantitative data synthesis, sensitivity analysis conducted using the leave-one-out method, and weighted random-effects meta-regression performed to evaluate potential confounders on cystatin C levels.

Results: This meta-analysis of data from nine published studies (16 treatment arms) involved a total of 2195 subjects. In a single-arm analysis of clinical trials (without control group; eight studies comprising 14 treatment arms), fibrate therapy increased circulating cystatin C concentrations (WMD: 0.07 mg/dL, 95% CI: 0.04, 0.10, p < .001; I² = 82.66%). When the analysis was restricted to randomized controlled trials (four studies comprising six treatment arms), again elevation of circulating cystatin C levels was observed (WMD: 0.06 mg/L, 95% CI: 0.03, 0.09, p < .001; I² = 42.98%). Elevated cystatin C levels were only seen with fenofibrate and not with other fibrates.

Conclusions: The results suggest that fenofibrate treatment adversely affects cystatin C levels and might partially explain the limited efficacy of fenofibrate in reducing cardiovascular events.

Key message

• Fenofibrate treatment adversely affects cystatin C levels and might partially explain the limited efficacy of fenofibrate in reducing cardiovascular events.

Keywords:

• Fibrates
• cystatin C
• fenofibrate

Introduction

The reduction of low-density lipoprotein cholesterol (LDL-C) levels with the use of statins is a central strategy in primary and secondary prevention of cardiovascular diseases (CVD) [1]. Despite statin therapy, however, the residual risk for CVD remains high [1]. Hypertriglyceridemia, the elevation of triglyceride (TG)-rich lipoproteins, represents an independent risk factor for cardiovascular (CV) events, and is therefore another key therapeutic target [2]. Fibrates, which activate peroxisome proliferator-activated receptor-α (PPARα), are the most effective agents for lowering triglyceride levels; they show some efficacy in the treatment of hypercholesterolemia and combined dyslipidemias [1,3–7], and possess lipid-independent pleiotropic activities [8–10]. Combination therapy, such as fenofibrate and simvastatin as used in the Action to Reduce Cardiovascular Risk in Diabetes (ACCORD) trial, showed an improvement in TG and HDL-C levels to a greater degree than with simvastatin alone [11]. While generally well-tolerated, side effects associated with fibrate treatment have been well-documented. These adverse drug reactions may include elevations in liver enzymes, serum creatinine, and homocysteine [12].

Cystatin C is a strong inhibitor of cysteine proteases [13] that is involved in several physiological processes, including protein turnover, precursor protein activation, innate immune cell regulation, antigen presentation, and apoptosis [14]. Notably, cystatin C has been proposed as a more reliable marker of glomerular filtration and kidney function than creatinine [15]. In addition, cystatin C appeared to be a better unfavourable predictor of CVD and mortality than creatinine [15,16]. Additional studies have shown that levels of cystatin C correlate with those of homocysteine [17], the latter marker increasing during fibrate treatment and exerting adverse effects upon endothelium, platelets, coagulation factors [18] as well as upon CVD risk independent of traditional risk factors [19].

The purpose of this systematic review and meta-analysis was to look at the impact of fibrates on the concentration of cystatin C in the circulation.

Methods

Search strategy

This study was designed in accordance with the instructions of the 2009 preferred reporting items for systematic reviews and meta-analysis (PRISMA) statement. SCOPUS, PubMed-Medline, ISI Web of Science, and Google Scholar databases were searched using the following search terms in titles and abstracts: (fenofibrate OR pemafibrate OR bezafibrate OR clofibrate OR ciprofibrate OR gemfibrozil OR “fibric acid” OR “clofibric acid” OR procetofen) AND (“cystatin C” OR cystatinC). The wild-card term “*” was used to increase the sensitivity of the search strategy. The search was limited to studies in humans. The literature was searched from inception to 2 April 2018.

Study selection

Clinical trials evaluating the effect of a fibrate on circulating (serum or plasma) cystatin C levels were included in this meta-analysis. Non-interventional studies and studies not providing sufficient information on circulating cystatin C concentrations were excluded from the meta-analysis. Before excluding a study for the latter reason, the author(s) were contacted and asked to provide the necessary data.

Quality assessment

Risk of bias in the studies considered in this meta-analysis was evaluated according to the Cochrane instructions [20]. Selection bias, performance bias, attrition bias, detection bias, reporting bias, and other sources of bias were judged to be high, low, or unclear in each of the included studies.

Data extraction

Studies meeting the inclusion criteria were reviewed and data regarding authors, study location, publication date, number of studied population, trial design, dose and duration of intervention, baseline characteristics of studied population [including age, gender, systolic and diastolic blood pressure, serum creatinine, body mass index (BMI), and plasma lipid concentrations], and changes in circulating concentrations of cystatin C. When the values were only presented as graph, GetData Graph Digitizer 2.24 software (http://getdata-graph-digitizer.com/) was used to digitize and extract the data.

Quantitative data synthesis

Comprehensive meta-analysis (CMA) V2 software (Biostat, NJ) software was used for statistical procedures. All reported cystatin C concentrations were unified in mg/L. Inverse-variance weighted mean difference (WMD) and 95% confidence limits were used as the summary statistic, considering a correlation coefficient (R) of 0.5. Conversion of median and inter-quartile range to mean and standard deviation was performed as suggested by Wan et al. [21]. Meta-analysis was performed using a random-effects model (using DerSimonian–Laird method) and the generic inverse variance weighting method. Heterogeneity was quantitatively assessed using I² index. Sensitivity analysis was performed using leave-one-out method [22–24]. A subgroup analysis was conducted to explore the impact of different types of fibrates on circulating cystatin C levels. Also, a separate analysis was performed in randomized controlled trials.

Publication bias

Presence of publication bias in the meta-analysis was investigated using assessment of Begg’s funnel plot, and statistical tests as previously described [24,25]. The “trim and fill” method was used to adjust the effect size for potential publication bias [26].

Results

Flow of included studies

Briefly, after multiple database searches, 39 published studies were identified and the abstracts reviewed. Twenty-five articles did not meet the inclusion criteria and were excluded. Next, 14 full-text articles were carefully assessed and reviewed, of which five studies were excluded for not measuring cystatin C concentrations. Finally, nine studies with 16 treatment arms were found to be eligible and included in the systematic review and meta-analysis. The study selection process is shown in Figure 1.

Figure 1. Flow chart of the number of studies identified and included into the meta-analysis.

Characteristics of the studies included

Data were pooled from nine clinical trials comprising a total of 2195 subjects, including 1261 and 934 participants in the treatment and control arms (individuals of the cross-over trials were considered in treatment and control groups), respectively. Included studies were published between 1999 and 2018. The clinical trials used different types and/or doses of fibrates [27–35]. The range of intervention periods was from 6 weeks [27–29,35] to 5 years [30]. Study designs of included trials were parallel [28,30–33] and cross-over group [27,29,34,35]. Selected studies enrolled subjects with dyslipidemia [28,29,31,32,34,35], type 2 diabetes [30,33], and healthy subjects [27]. Characteristics of the included clinical trials are shown in Table 1.

Table 1. Demographic characteristics of the included studies.

Author	Study design	Target population	Treatment duration	n	Study groups	Age, years	Female (n, %)	BMI (kg/m^2)	Total cholesterol (mg/dl)	LDL cholesterol (mg/dl)	HDL cholesterol (mg/dl)	Triglycerides (mg/dl)	Cystatin C (mg/L)
Ansquer et al. [27]	Randomized, double-blind, placebo-controlled, cross-over	Healthy people	6 weeks	24	Overall	ND	7 (30)	24.0*
				24	Placebo				162 ± 31	98 ± 24	44 ± 11	88 ± 74	0.85 ± 0.28
				24	Fenofibrate 160 mg/day				166 ± 28	105 ± 24	46 ± 11	97 ± 108	0.75 ± 0.22
Dierkes et al. [28]	Open-label, controlled	Men with hypertriglyceridemia	6 weeks	10	Fenofibrate 200 mg/day	ND	0 (0)	ND	ND	ND	ND	363 (203–850)^†	0.75 (0.61–0.82)^†
				10	Bezafibrate 400 mg/day	ND	0 (0)	ND	ND	ND	ND	310 (203–726)^†	0.71 (0.53–0.94)^†
Dierkes et al. [29]	Randomized, double-blind, placebo-controlled, cross-over	Men with hyperlipidemia	6 weeks	29	Overall	46 ± 8	0 (0)	28.8 ± 2
				25	Fenofibrate 200 mg/day			0.8	286 (185–417)^†	59 (10–364)^†	35 (21–63)^†	496 (168–1691)^†	0.76 (0.57–0.85)^†
				25	Fenofibrate 200 mg/day + vitamins				324 (193–437)^†	59 (13–220)^†	33 (23–58)^†	558 (141–1594)^†	0.76 (0.52–0.89)^†
Forsblom et al. [30]	Randomized, double-blind, placebo-controlled	Type 2 diabetes	5 years	83	Placebo	61.3 ± 6.7	21 (25)	29.7 (26.8–33.0)^†	189 ± 27	116 ± 23	42 (38–50)^†	150 (115–186)^†	0.85 ± 0.13
				87	Fenofibrate 200 mg/day	62.5 ± 6.3	24 (27)	29.1 (26.3–32.5)^†	201 ± 27	127 ± 23	42 (38–54)^†	132 (97–194)^†	0.92 ± 0.17
Foucher et al. [31]	Randomized, double-blind, active-controlled	Mixed dyslipidemia	12 weeks	109	Fenofibrate 145 mg/day + Simvastatin 20 mg/day	59.6 ± 10.0	38 (35)	32.0 ± 5.2	176.7 ± 23.2	101 ± 18	43 ± 10	226 (117–601)^‡	0.96 ± 0.17
				114	Simvastatin 20 mg/day	60.5 ± 9.3	40 (35)	31.1 ± 4.8	175.2 ± 22.0	100 ± 16	46 ± 12	194 (90–592)^‡	0.95 ± 0.18
				110	Fenofibrate 145 mg/day + Simvastatin 40 mg/day	60.4 ± 8.1	43 (39)	31.4 ± 4.6	179.8 ± 25.5	101 ± 17	46 ± 11	208 (110–1103)^‡	0.92 ± 0.17
				112	Simvastatin 40 mg/day	59.3 ± 9.1	39 (35)	31.3 ± 5.3	181.0 ± 23.2	104 ± 17	46 ± 12	216 (122–576)^‡	0.93 ± 0.14
				111	Fenofibrate 145 mg/day	60.4 ± 9.0	38 (34)	31.6 ± 5.1	184.8 ± 25.9	105 ± 18	46 ± 10	236 (126–843)^‡	0.95 ± 0.16
Ishibashi et al. [32]	Randomized, double-blind, active-controlled	Dyslipidemia	24 weeks		Pemafibrate 0.2 mg/day	53.1 ± 10.8	15 (20)	26.1 ± 3.5	232 ± 30	143 ± 30	42 ± 4	239 ± 53	0.8 ± 0.1
				74	Pemafibrate 0.4 mg/day	53.4 ± 10.3	13 (17)	26.9 ± 4.2	228 ± 30	143 ± 27	42 ± 4	230 ± 62	0.8 ± 0.1
				76	Fenofibrate 106.6 mg/day	52.9 ± 11.6	14 (18)	26.0 ± 3.7	228 ± 27	139 ± 27	42 ± 4	239 ± 70	0.8 ± 0.1
Mychaleckyj et al. [33]	Randomized, placebo-controlled	Type 2 diabetes	12 weeks	321	Fenofibrate 160 mg/day (cases)	67.2 ± 6.2	97 (30)	ND	ND	ND	41.5 ± 10.3	136 ± 74	1.03 ± 0.35
				175	Fenofibrate 160 mg/day (controls)	66.9 ± 6.5	70 (40)	ND	ND	ND	42.6 ± 11.1	160 ± 227	0.95 ± 0.35
				564	Placebo	67.3 ± 7.1	197 (35)	ND	ND	ND	41.5 ± 9.1	163 ± 93	0.95 ± 0.17
Ncube et al. [34]	Open-label	Hyperlipidemia	12 weeks	12	Fenofibrate 267 mg/day	43.5 (33–70)^†	2 (16)	ND	232*	ND	45*	264*	0.79 (0.72–0.92)^†
Westphal et al. [35]	Randomized, active-controlled, cross-over	Men with hypertriglyceridemia	6 weeks	22	Overall	36–66^ǂ	0 (0)	25–37^ǂ
				22	Gemfibrozil 900 mg/day				ND	ND	32 (17–73)^†	682 (221–4694)^†	0.73 (0.57–1.48)^†
				22	Fenofibrate 200 mg/day				ND	ND	34 (15–65)^†	575 (203–5376)^†	0.68 (0.54–1.56)^†

Values are expressed as mean ± standard deviation (SD).

* Mean.

† Median (IQR).

‡ Median (min-max).

ǂ Range.

ND: no data; BMI: body mass index; IQR: interquartile range.

Cystatin C assay methods

Different assay methods were used to determine plasma cystatin C levels. In this regard, three trials quantified cystatin C concentrations using particle-enhanced immuno-nephelometry assay [27,28,35]. One study measured cystatin C levels in plasma by an immuno-precipitation method [30], while another trial determined cystatin C concentrations using a latex-enhanced immuno-turbidimetric assay [34]. Finally, four studies did not provide the method used to determine plasma cystatin C levels [29,31–33].

Quality assessment

Regarding sequence generation, two studies showed high risk of bias for this parameter [28,34]. Three trials had a high risk of bias about allocation concealment [28,34,35]. With respect to blinding of participants, personnel and outcome assessors, three studies exhibited high risk of bias [28,34,35]. Nevertheless, all evaluated trials were characterized by low risk of bias for incomplete outcome data and selective outcome reporting. Details of the risk of bias assessment are shown in Table 2.

Table 2. Quality of bias assessment of the included studies according to the Cochrane guidelines.

Study	Sequence generation	Allocation concealment	Blinding of participants, personnel and outcome assessors	Incomplete outcome data	Selective outcome reporting	Other sources of bias
Ansquer et al. [27]	U	U	U	L	L	U
Dierkes et al. [28]	H	H	H	L	L	U
Dierkes et al. [29]	U	U	U	L	L	U
Forsblom et al. [30]	U	U	U	L	L	U
Foucher et al. [31]	L	U	L	L	L	U
Ishibashi et al. [32]	L	L	L	L	L	L
Ncube et al. [34]	H	H	H	L	L	U
Westphal et al. [35]	U	H	H	L	L	U
Mychaleckyj et al. [33]	L	U	L	L	L	U

L: low risk of bias; H: high risk of bias; U: unclear risk of bias.

Quantitative data synthesis

In a single-arm analysis of clinical trials (without control group; eight studies comprising 14 treatment arms), fibrate therapy was found to increase circulating cystatin C concentrations (WMD: 0.07 mg/dL, 95% CI: 0.04, 0.10, p < .001; I² = 82.66%) (Figure 2). This effect size was robust in the leave-one-out sensitivity analysis and not significantly altered by removing any single-study arm from the meta-analysis (Figure 2). When the analysis was stratified to individual fibrates, only fenofibrate was associated with cystatin C increase (WMD: 0.09 mg/L, 95% CI: 0.07, 0.11, p < .001; I² = 38.38%), while pemafibrate, bezafibrate, and gemfibrozil did not alter cystatin C levels (Figure 3).

Figure 2. Forest plot detailing weighted mean difference and 95% confidence intervals for the impact of fibrates on plasma cystatin C concentrations in single-arm uncontrolled trials. Lower plot shows the results of leave-one-out sensitivity analysis.

Figure 3. Forest plot detailing weighted mean difference and 95% confidence intervals for the impact of individual fibrates on plasma cystatin C concentrations in single-arm uncontrolled trials. Lower plot shows the results of leave-one-out sensitivity analysis.

When the analysis was restricted to randomized clinical trials (RCTs) (four studies comprising six treatment arms), again a significant elevation of circulating cystatin C levels was observed (WMD: 0.06 mg/L, 95% CI: 0.03, 0.09, p < .001; I² = 42.98%) (Figure 4). All of the RCT arms included fenofibrate as the fibrate intervention, and the meta-analysis was found to be robust in the leave-one-out sensitivity analysis (Figure 4).

Figure 4. Forest plot detailing weighted mean difference and 95% confidence intervals for the impact of fibrates on plasma cystatin C concentrations in randomized controlled trials. Lower plot shows the results of leave-one-out sensitivity analysis.

A subgroup analysis was performed to assess the impact of fibrates on serum creatinine concentrations. In the single-arm analysis of fibrate treatment arms (n = 13), a significant elevation of serum creatinine levels was observed (WMD: 0.09 mg/dL, 95% CI: 0.002, 0.01, p = .033; I² = 98.07%). A trend toward increased creatinine levels of was also observed when the analysis was limited to RCTs (four studies comprising six treatment arms) (WMD: 0.12 mg/dL, 95% CI: −0.01, 0.25, p = .068; I² = 97.16%).

Publication bias

The funnel plot of the study standard error by effect size (mean difference) was asymmetric and suggested potential publication bias in the meta-analysis of fenofibrate’s effects on circulating cystatin C levels. However, the results of Begg’s rank correlation (Kendall’s τ with continuity correction =0.18, z = 0.72, two-tailed p value =0.474) and Egger’s linear regression (intercept =0.83, standard error =0.71; 95% CI = −0.82, 2.47, t = 1.16, df =8, two-tailed p = .279) did not suggest potential publication bias. An attempt was made to address publication bias using trim-and-fill correction. Two potentially missing studies on the left side of funnel plot were imputed leading to a corrected effect size that was still significant (WMD: 0.08 mg/L; 95% CI: 0.06, 0.11). The “fail safe N” method indicated that 294 theoretically missing studies would be required to make the overall estimated effect size non-significant. Funnel plot of the impact of fenofibrate on circulating cystatin C levels is illustrated in Figure 5.

Figure 5. Funnel plot detailing publication bias in the studies reporting the impact of fibrate fibrates on plasma cystatin C concentrations in single-arm uncontrolled trials. Open circles represent observed published studies; closed circles represent imputed unpublished studies.

Discussion

Fibrates are among the most effective agents for lowering triglyceride levels, as well as having efficacy in the treatment of hypercholesterolemia and combined dyslipidemias [1,3–6]. These agents are often used in combination with statins, especially in patients with atherogenic dyslipidemia [11]. However, some adverse drug effects, including elevations in liver enzymes, serum creatinine, and homocysteine [12,36] might limit their use. As cystatin C levels are inversely associated with kidney function and GFR [15], as well as directly with plasma homocysteine and CVD risk [15], testing the effects of fibrates on cystatin C levels might provide useful information on both kidney function and CVD risk.

In this systematic review and meta-analysis, the impact of fibrates on the concentration of circulating cystatin C was considered. Our analysis of the published data showed that, overall, fibrate monotherapy did cause a small but significant elevation in circulating levels of cystatin C. Interestingly, however, when the data was stratified for individual fibrates, only fenofibrate was associated with elevated cystatin C levels, while pemafibrate, bezafibrate, and gemfibrozil had no effect on circulating cystatin C levels. Interestingly, in one study by Dierkes et al. [28], bezafibrate, as well as fenofibrate, increased serum homocysteine levels, but only fenofibrate increased cystatin C levels. Further, as it is known that the plasma homocysteine concentration is dependent on renal function, the impairment of renal function by fenofibrate may explain the elevation in plasma homocysteine levels [35]. Likewise, cystatin C is considered to be an excellent endogenous marker of glomerular filtration rate (GFR) [34].

This meta-analysis is in accord with that of Jun et al. [37] in that we also found an increase in serum creatinine with fibrate therapy. However, Jun et al. [37] found that there was an improvement in lipid profile and a reduction in cardiovascular events in subjects with chronic kidney disease, but the implications of the raised creatinine on chronic kidney disease outcomes remains unclear.

Limitations of this systematic review and meta-analysis include the small number of studies that were found to be eligible, and that the clinical trials used different fibrates and/or differing doses of fibrates for differing periods of time. Study design also differed, some using parallel and other using cross-over design, as well as differing methods used to assay cystatin C. A further limitation is that the majority of the studies included in this meta-analysis involved only fenofibrate, with only three studies designed to look at additional fibrates [32,35,37]. Moreover, apart from one study that assessed the impact of pemafibrate [32], there is scant evidence as to the effect of selective PPAR modulators [38,39] on plasma cystatin C levels. Finally, taking into account the heterogeneity of the included studies, the obtained results should be obviously treated with caution. It is especially important to explore if the small increase of cystatin C (<0.1 mg/dL) is of clinical relevance.

In conclusion, the results of this meta-analysis suggest that fenofibrate treatment adversely affects cystatin C levels, what might influence its clinical effectiveness in reducing the residual cardiovascular risk. We might obviously hypothesize that it might be one of the reasons on why fenofibrate has only limited clinically efficacy while added to statins, but we need further well-designed studies in order to confirm whether this observation might have any influence on the reduction of residual risk.

Disclosure statement

No potential conflict of interest was reported by the authors.