Association between SLCO1B1 521 T C and 388 A G Polymorphisms and Statins Effectiveness: A Meta-Analysis

Transcription

1 796 Original Article Association between SLCO1B1 51 T C and 388 A G Polymorphisms and Statins Effectiveness: A Meta-Analysis Rong Dai 1, Jing Feng 1, Yang Wang 1, Yuan Yang, Changkai Deng 3, Xiaojun Tang 1, Yong Zhao 1, Hao Zhou 1 and Fan Zhang 1 Rong Dai and Jing Feng contributed equally to this work. 1 School of Public Health and Management, Research Center for Medicine and Social Development, The Innovation Center for Social Risk Government in Health, Chongqing, China Department of Cardiovascular Medicine, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China 3 Chengdu Women s and Children s Central Hospital, Chengdu, China Aim: Previous studies on the association between the SLCO1B1 51 T C and 388 A G polymorphisms and statin effectiveness have been inconsistent. We performed this meta-analysis to provide a more comprehensive estimation of this issue. Methods: Multiple electronic literatures databases were searched on March 5th 014. A quality assessment was performed using the Methodological Index for Non-Randomized Studies (MINORS) criteria. A meta-analysis, sub-group analysis, sensitivity analysis (RevMan 5.), publication bias measuring and meta-regression analysis were conducted utilizing the Stata software program (version 1.0). Results: A total of 13 studies were included in the final meta-analysis, which included 7,079 participants. Overall, there was no statistically significant association in the four genetic models of hypolipidemic effect. For the 51 T C polymorphism, significant associations were found for the longterm effectiveness of lowering the low-density lipoprotein cholesterol (LDL-C) and in non-asian populations in the dominant model [(CC TC vs. TT: mean difference (MD) 1.44, 95% CI: , p 0.0) and (CC TC vs. TT: MD 1.38, 95% CI: , p 0.01)], the recessive model [(CC vs. TT TC: MD 3.31, 95% CI: , p 0.04) and (CC vs. TT TC: MD.83, 95% CI: , p 0.03)], and the homozygote comparison [(CC vs. TT: MD 3.68, 95% CI: , p 0.03) and (CC vs. TT: MD 3.33, 95% CI: , p 0.01)], respectively. There were no significant differences for the other analyses of the 51 T C polymorphism or all the analyses of the 388 A G polymorphism. Conclusions: The overall results suggest that the SLCO1B1 51 T C and 388 A G polymorphisms do not affect the lipid-lowering effectiveness of statins. However, allele C of the SLCO1B1 51 T C polymorphism leads to an attenuated effect on lowering the LDL-C in non-asian populations and the long-term effectiveness of statin treatment. J Atheroscler Thromb, 015; : Key words: SLCO1B1, Statin, Meta-analysis Address for correspondence: Fan Zhang, Department of Epidemiology, School of Public Health and Management, Chongqing Medical University, No. 1 Yixueyuan Road, Yuzhong District, Chongqing , China ava11@16.com Received: June 30, 014 Accepted for publication: January 4, 015 Introduction Cardiovascular diseases account for 9.7% of the worldwide burden of all causes of death and disability 1). To reduce the cardiovascular risk, various regimens of adjusted lifestyle have been recommended (such as smoking cessation, limiting alcohol and dietary

2 797 therapy). The disorder of the lipid metabolism is one of the main determinants of cardiovascular risk. Drugs are given to decrease the patient s cholesterol levels, among which 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) are currently the most widely used for cholesterol-lowering therapy. The effectiveness of statin treatment in reducing the risk of coronary heart disease (CHD) was indicated by previous studies -1). Statins reduce cholesterol synthesis by competitively inhibiting HMG-CoA reductase. The inhibition of HMG-CoA reductase enhances the removal of low-density lipoprotein (LDL) particles from the blood and reduces the total and low-density lipoprotein concentrations in the plasma. The organic anion transporting polypeptide 1B1 (OATP1B1), also known as OATP-C, is encoded by the solute carrier organic anion transporter1b1 (SLCO1B1). The OATP1B1 transporters are present in the basolateral membrane of hepatocytes and play a key role in the hepatic uptake of a number of endogenous and exogenous substances, such as pravastatin, rosuvastatin, pitavastatin and atorvastatin 13-15). SLCO1B1 displays a number of single nucleotide polymorphisms (SNPs). Some SNPs have been identified and some of the non-synonymous SNPs have been found to alter its transport activities 16-18). Among them, the SLCO1B1 51 T C and 388 A G polymorphisms have been the most studied; however, the findings have been controversial. The SLCO1B1 388 A G polymorphism is present in exon 4 of chromosome 1 and the 51 T C polymorphisms is present in exon 5. An in vivo clinical pharmacokinetic trial determined that the two polymorphisms were associated with a significant reduction in the oral clearance of a single dose of pravastatin 16). However, the association between the SLCO1B1 gene polymorphisms and statin effectiveness has been difficult to determine. For 51 T C, some trials identified that the polymorphism modulated the lipidlowering effectiveness of HMG-CoA reductase inhibitors 19-1). Conversely, some trials did not find a significant effectiveness -7). In addition, Hedman et al. found that decreases in the total and low-density lipoprotein cholesterol by pravastatin were significantly smaller, and the increase in the high-density lipoprotein cholesterol (HDL-C) was greater in transplant recipients with the 51TC genotype compared with patients with the 51TT reference genotype 8). For 388 A G, Rodrigues et al. and Sortica et al. found that the polymorphism caused a significant increase in the response to statins and may be an important marker for predicting the effectiveness of lipid-lower- ing therapy 9, 30). Specially, Shabana et al. maintained that the polymorphism showed gender-related effects on TG change following the atorvastatin treatment 31). Although the benefits of statins in the prevention of coronary heart disease has been demonstrated by primary and secondary prevention trials -1), there is inter-individual variability among the statin therapeutic response and some studies 3-37) recognized that genetic polymorphisms may influence the drug disposition, side effects and statin effectiveness. Due to the inconsistency in the existing literature and in order to overcome the limitation of individual studies, we performed this meta-analysis to provide a more precise and comprehensive estimation of the association between the SLCO1B1 51 T C and 388 A G polymorphisms and statin effectiveness. Methods Data Source Four databases were electronically searched to retrieve studies on the associations between the SLCO1B1 51 T C and 388 A G variants and the effectiveness of statin-induced therapy on March 014 (PubMed, Web of Science, Embase and Cochrance Library). The search was based on the following keywords: HMG-CoA reductase inhibitor, statin, simvastatin, lovastatin, fluvastatin, atorvastatin, pravastatin, rosuvastatin, cerivastatin, or mevastatin combined with SLCO1B1. Furthermore, we checked the reference lists of retrieved articles to identify more pertinent studies. Inclusion Criteria Studies were considered eligible if they met all of the following criteria: 1) the study explored the relationship between SLCO1B1 gene polymorphisms and the effectiveness of statins; ) the SLCO1B1 51 T C or 388 A G polymorphisms were tested in the study; 3) the papers identified the sample size, distribution of alleles, genotypes, and the mean percentage change level of low-density lipoprotein cholesterol (LDL-C), total cholesterol (TC), HDL-C, or triglyceride (TG); 4) the results were expressed as the mean standard deviation (M SD), or other information that was useful in inferring the results; and 5) the manuscript was written in English. Exclusion Criteria Studies that met any of the following exclusion criteria were excluded: 1) studies based on animals or cells rather than human populations; ) reviews, editorials, meeting abstracts, and commentaries; and 3)

3 798 articles with no target data or relevant outcomes. When multiple reports concerning the same cohort were identified, we only included the report which had the longest treatment duration in our meta-analysis Data Extraction Two reviewers (Rong Dai and Jing Feng) independently searched and selected the literature and then extracted relevant data according to a pre-formed data extraction form. Disparities were solved by discussion, and a third party (Fan Zhang) was involved when necessary. The extracted information including: first author, year of publication, country of origin, ethnicity of the study population, mean age of the cases, sample size, disease of the included subjects, genotyping methods, dosage and frequency of medication, outcomes and any other information our researchers thought to be pertinent. Quality Assessment A quality assessment of the included articles was performed using the Methodological Index for Non- Randomized Studies (MINORS) criteria 38). It is a validated, 1-item scoring tool for non-randomized studies. The items in the MINORS scale are scored 0 (not reported), 1 (reported but inadequate), or (reported and adequate). The global ideal score is 16 for noncomparative studies and 4 for comparative studies. Statistical Analyses The LDL-C, TC, HDL-C and TG changes were shown as percentage changes in different genotypes and were calculated with the formula p ((after treatment baseline)/ baseline) 100% in the original studies. Two studies 5, 9) calculated the percentage change with p ((baseline after treatment)/ baseline) 100%, so we adjusted their minus signs. In addition, the mean SD for all of the results were obtained. Two studies 7, 30) demonstrated their results as mean standard error (M SE). We used the formula SD SE n to get SD. For three of the studies 0, 1, 30), we calculated the SD according to a 95% confidence interval (CI) with the formula previously used in a study by Chowbay 39). One article 5) reported the effectiveness of two types of statins in two groups of participants, and we treated this study as two studies. Another cross-over study reported two results; however, we did not include one of the reults 6) due to the unclear washout period. Ultimately, 13 articles with 14 studies were included for the final analysis. The mean difference (MD) of the mean percentage change of LDL-C, TC, HDL-C and TG levels were used to evaluate the strength of the association between the SLCO1B1 51 T C and 388 A G polymorphisms and the effectiveness of statins. Heterogeneity among the included studies was assessed by the chi-square-based Q test and I test. If the heterogeneity between studies was not significant (p 0.10, I 50%), the inverse variance method with a fixed-effects model was used. Otherwise, the inverse variance method with a random effects model was chosen. Four comparison genetic models were used to assess the association: the dominant model (TC CC vs. TT and GG AG vs. AA), the recessive model (CC vs. TT TC and GG vs. AA AG), the heterozygote comparison (TC vs. TT and AG vs. AA), and the homozygote comparison (CC vs.tt and GG vs. AA). All statistical tests were conducted with the Cochrane Collaboration meta-analysis software, Review Manager 5.. Furthermore, if included studies did not provide relevant information on the Hardy-Weinberg Equilibrium (HWE), we calculated the HWE using an online HWE calculation tool 40). A p value of 0.05 for any test or model was considered to be statistically significant unless otherwise specified. Subgroup Analyses The treatment duration was not consistent among included studies. Subgroup analyses were conducted according to the treatment duration to explore the possible explanations for heterogeneity. We defined a treatment duration 6 months as long-term effectiveness; any treatment duration less than 6 months was defined as short-term effectiveness. We also conducted a subgroup analysis on studies specifically targeting non-asian populations to explore the influence of ethnicity. Sensitivity Analyses Sensitivity analyses were performed by omitting an individual study, one-by-one, to investigate the influence of any single study on the overall effect. Publication Bias Potential publication bias was assessed by the Begg rank correlation test and the Egger linear regression test using Stata 1.0 software. Meta-regression Analyses Meta-regression analyses were performed using the Stata 1.0 software program to identify study-level factors contributing to the heterogeneity between studies.

4 799 Potential relevant citations (n=676) Pubmed: 170 Embase: 47 Web of science: 4 Cochrane: 17 Excluded (n=657) Duplicate studies: 334 Reviews, case reports, editorials, mee ng abstracts: 135 Cell culture or animal experimental studies: 36 polymorphisms and the e cacy of sta n-induced: 15 Articles requiring full text review (n=19) Additional study met the inclusion criteria (n=1) Excluded (n=7) No target data: 5 No relevant outcome: 1 Review: 1 Articles meeting criteria for meta analysis (n=13) Fig.1. Schematic of studies included in the meta-analysis. Results Study Selection We initially identified 676 potentially eligible studies. Many of the studies were excluded after the screening of titles and abstracts, mainly due to duplication, reviews, non-population-based studies or irrelevancy to the SLCO1B1 51 T C or 388 A G polymorphisms and the effectiveness of statins. After assessing the full-text of 19 potentially relevant articles, we identified 1 eligible articles. The main reasons for exclusion were as follow: five studies lacked target data 41-45) (frequency of genotypes, mean percentage change of LDL-C, TG, TC or HDL-C levels and standard deviation in various genotype groups or any information was pertinent to infer them); one study focused on the pharmacokinetics of statins 46) ; and one study was a review 47). Another relevant publication was found through reference screening (Fig. 1). In total, 13 articles were included in the meta-analysis. Characteristics of Included Studies The basic characteristics of included studies are presented in Table 1. All studies were published between 005 and 013 and included a total of 7,079 subjects. Four studies were conducted in China, two in Brazil, one in Britain, one in France, one in Egypt, one in Japan, and one in Finland; two of the studies

5 800 Table 1. Characteristics of the thirteen studies included in the meta-analysis First author Year Country Ethnicity Mean age (M SD or range) Sample size (male/female) Disease of included subjects Genotyping methods Statin types study type of RCT basleline details Length of follow-up Dosage/ frequency (mg/day) or (mg/kg) SLCO1B1 Polymorphism MINORS scale 51T C, Akao 01 USA NA (NA) Vascular disease TaqMan pravastatin Y Y 1m A G 51T C, Couvert 008 France Caucasian (98/3) hyperlipidemia TaqMan fluvastatin Y Y 8w A G 51T C, Rodrigues 011 Brazil Brazilian NA 136 (44/9) hyperlipidemia TaqMan atorvastatin N Y 4w A G Shabana 013 Egypt Egyptian (3/7) hyperlipidemia PCR-RFLP atorvastatin N Y 4w A G 8 51T C, Yang 010 China Chinese (41/44) hyperlipidemia PCR-RFLP pitavastatin N Y 8w A G Tachibana- Iimori 004 Japan Japanese (17/49) hyperlipidemia TaqMan pitavastatin N Y NA NA 51T C 6 AS-PCR; Fu 013 China Chinese (NA) hyperlipidemia PCR-RFLP atorvastatin; N Y 4w 0 simvastatin 51T C, 7 388A G rosuvastatin Hu 01 China Chinese (100/147) hyperlipidemia TaqMan N Y 6w simvastatin T C, 7 388A G Martin 01 UK NA (NA) NA PCR pravastatin N Y 1m 40 51T C Hedman 006 Finland NA (7/13) FH; PCDTR TaqMan pravastatin N N m 1 (3/9) T C 8 Zhang 007 China Chinese (8/17) CHD ARMS-PCR pravastatin N Y 1m 0 51T C 9 51T C, Sortica 01 Brazil Brazilian (66/150) hyperlipidemia TaqMan simvastatin N Y 6m A G pravastatin thompson 005 USA Caucasian NA 190 (NA) NA TaqMan N N 6m 0 51T C 8 fluvastatin Abbreviations: SD, standard deviation; NA, not available; CHD, coronary heart disease; FH, familial hypercholesterolemia; PCDTR, pediatric cardiac transplant recipients; m, month; w, week; mg/kg; Y, the study provided the baseline details and the study type is RCT; N, the study didn t provide the baseline details and the study type is not RCT. There is control group in the study.

6 801 Table. Comparison results of the 51 T C polymorphism Genetic model MD (95% CI) Z p value I % Phet LDL-C [19,, 3, 5-7, 30] CC: TT [19-3, 5-8, 30] TC: TT [19,, 3, 5-7, 30] CC: TC TT [19, -7, 9, 30] CC TC: TT TC [, 3, 5, 30] CC: TT [0-3, 5, 8, 30] TC: TT [, 3, 5, 30] CC: TC TT [-5, 9, 30] CC TC: TT HDL-C [3, 5, 7, 30] CC: TT [0, 1, 3, 5, 7, 8, 30] TC: TT [3, 5, 7, 30] CC: TC TT [3-5, 7, 30] CC TC: TT TG [5, 30] CC: TT [0, 1, 5, 8, 30] TC: TT [5, 30] CC: TC TT [4, 5, 30] CC TC: TT 1. ( 0.33,.78) 0.80 ( 1.1,.71) 1.03 ( 0.48,.54) 0.66 ( 0.14, 1.46) 0.39 ( 1.99, 1.) 1.61 ( 0.35, 3.57) 0.39 ( 1.97, 1.0) 0.16 ( 0.95, 0.63) 0.53 ( 4.3, 5.8) 0.7 (.3, 1.78) 0.0 ( 4.51, 4.55) 0.03 (.07,.01) 0.90 ( 5.94, 4.18) 0.04 (.5,.18) 0.74 ( 5.79, 4.30) 0.40 (.5, 1.7) Begg s test Egger s test z p value t p vaule Abbreviations: MD, mean difference; CI, confidence interval; Phet, p value of heterogeneity; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride. p 0.05 stands for significance. R, random-effect model; others use fixed-effect model Quantitative Data Synthesis SLCO1B1 51 T C Polymorphism and Statin Effectiveness No statistically significant association was observed for all of the four genetic models of lowering the LDL- C, TC, and TG, and increasing HDL-C (Table ). In the subgroup analyses stratified by the treatment duration, the study 0) which did not provide the length of follow-up did not influence the results when put it in either of the subgroups. We therefore included this study in the short-term effectiveness group. A metaanalysis showed that a significant association was found for the long-term effectiveness of lowering LDL-C in the dominant model (CC TC vs. TT: MD 1.44, 95% CI: , p 0.0), the recessive model (CC vs. TT TC: MD 3.31, 95% CI: , p 0.04) and the homozygote comparison (CC vs. TT: MD 3.68, 95% CI: , p 0.03). No significant relationship was found for the heterozygote comparison (TC vs. TT: MD.00, 95% CI: , p 0.73) (Fig. -5, respectively). After stratifying by the treatment duration, there were no statistically sigdid not mention where the study was conducted. We assumed they enrolled the local people who were Americans. There were 6 studies involving Asian participants, two studies involving Brazilian participants, and two studies involving Caucasian participants; the ethnicities of three of the studies were not readily ascertained from reading the manuscripts, however, we believed they were not Asians. There were 8 studies which included subjects with essential hyperlipidemia, and two studies which included subjects with coronary heart disease (CHD); in two studies, this information was unavailable. All of the studies enrolled participants aged 45 years old except one study 8) which concentrated on children aged years old. The sample sizes ranged from 3 to,891. Pravastatin, fluvastatin, atorvastatin, simvastatin and rosuvastatin were used in the eligible studies, and the dosage of the drugs ranged from mg/day to 80 mg/day. The follow-up duration ranged from 1 month to 1 year and not all studies reported both SLCO1B1 51 T C or 388 A G polymorphisms.

7 80 Fig.. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the dominant model (CC TC VS. TT) as stratified by the treatment duration. Fig.3. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the recessive model (CC VS. TT TC) as stratified by the treatment duration.

8 803 Fig.4. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the homozygote comparision (CC VS. TT) as stratified by the treatment duration. Fig.5. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the heterozygote comparison (TC VS. TT) as stratified by the treatment duration.

9 804 nificant differences for any of the four genetic models with regard to the response to TC, a decrease in TG levels (Supplemental Table 1) or an increase in HDL- C. Upon exclusion of the Asian populations, a significant association between the 51 T C polymorphism and statin effectiveness of lowering the LDL-C was found in the dominant model (CC TC vs. TT: MD 1.38, 95% CI: , p 0.01), the recessive model (CC vs. TT TC: MD.83, 95% CI: , p 0.03) and the homozygote comparison (CC vs. TT: MD 3.33, 95% CI: , p 0.01). Conversely, no significant relationship was found for the heterozygote comparison (TC vs. TT: MD 1.5, 95% CI: , p 0.57) (Fig. 6-9, respectively). In addition, the outcomes for HDL-C, TG and TC indicated no statistical significance. SLCO1B1 388 A G Polymorphism and Statin Effectiveness There were no statistical differences in the four genetic models regarding the association between the SLCO1B1 388 A G polymorphism and the effectiveness of lowering LDL-C, TC, TG and increasing HDL-C (Table 3). Moreover, no associations were found in the subgroup analyses between the polymorphism and the treatment duration or ethnicity (Supplemental Table 1 and ). Sensitivity Analysis For the 51 T C polymorphism, the results were not significantly altered except one study 5) was excluded in the dominant model of lowering LDL-C (CC TC vs. TT: MD 1.5, 95% CI: , p 6). The value of I (assessment of heterogeneity) changed from 36.0% to 0.0% after omitting the study. In addition, we performed sensitivity analyses for the subgroup analyses that stratified by the treatment duration and ethnicity. When one study was omitted (Akao, 01), any result of the subgroup analyses that was significant became statistically non-significant (Fig , respectively). For the 388 A G polymorphism, none of the results were significantly altered. Quality Assessment The scores ranged from 6 to 1. Most of the included studies obtained scores for item 1 (a clearly stated aim) and item 7 (a loss to follow-up less than 5%) (Table 1). Publication Bias The Begg rank correlation test and the Egger linear regression test indicated that there was no evidence of publication bias except for two of the genetic models (Table and Table 3). Meta-regression Analyses Meta-regression analyses suggested that the ethnicity, the mean age, the sample size, the genotyping method, the type of statin and the length of follow-up did not contribute to the heterogeneity between studies. Discussion Statins are highly efficacious not only in reducing blood lipid level, but also in reducing cardiovascular mortality. However, inter-individual variability exists in the response to statin treatment. Thus, in this metaanalysis, our goal was to determine whether 51 T C and 388 A G variations in the SLCO1B1 gene affects statin effectiveness. Thirteen studies with 7,079 participants were included for the final analyses. The results indicate that there were no significant associations between the 51 T C and 388 A G polymorphisms of the SLCO1B1 gene and the effectiveness of statins. However, in the subgroup analyses stratified by the duration of treatment, the results suggest that the 51 T C polymorphism has a significant relationship with the statinreduced LDL-C level for long-term effectiveness. Individuals with the homozygote TT alleles demonstrate a higher reduction compared with homozygote CC and allele C carriers. In addition, the non-asian populations showed a significant association between the 51 T C polymorphism and statin effectiveness of lowering the LDL-C. Our study has some important strengths. While the individual studies had inadequate statistical power, our meta-analysis of 13 studies involving 7,079 participants had an increased power and was able to detect potential associations thereby provides more reliable estimates. Furthermore, heterogeneity was not present in most of the genetic model comparisons, although we found that the conclusions of the original studies were inconsistent. To support our findings, some data indicate that the presence of the 51 T C non-synonymous SNPs in the SLCO1B1 gene has a weaker statin-induced reduction of LDL-C 19-1). Conversely, other data suggest that there is no statistically significant difference among patients with wild type and 51 T C variants of the SLCO1B1 gene in the statin-induced lipid-lowering effectiveness -7). Additionally, some confounding factors exist, such as differences in study quality, demographic characteristics of the subjects (age, gender, race, etc.), inclusion and

10 805 Fig.6. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the dominant model (CC TC VS. TT) as stratified by Asian vs. non-asian populations. Fig.7. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the recessive model (CC VS. TT TC) as stratified by Asian vs. non-asian populations.

11 806 Fig.8. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the homozygote comparision (CC VS. TT) as stratified by Asian vs. non-asian populations. Fig.9. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the heterozygote comparison (TC VS. TT) as stratified by Asian vs. non-asian populations.

12 807 Fig.10. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the homozygote comparision (CC VS. TT) as stratified by the treatment duration (omitting the study by Akao, 01). Fig.11. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the recessive model (CC VS. TT TC) as stratified by the treatment duration (omitting the study by Akao, 01).

13 808 Fig.1. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the dominant model (CC TC VS. TT) as stratified by the treatment duration (omitting the study by Akao, 01). Fig.13. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the homozygote comparision (CC VS. TT) as stratified by Asian vs. non-asian populations (omitting the study by Akao, 01).

14 809 Fig.14. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the recessive model (CC VS. TT TC) as stratified by Asian vs. non-asian populations (omitting the study by Akao, 01). Fig.15. Forest plots for the association between the 51 T C polymorphism of the SLCO1B1 gene and statin effectiveness on lowering the LDL-C in the dominant model (CC TC VS. TT) as stratified by Asian vs. non-asian populations (omitting the study by Akao, 01).

15 810 Table 3. Comparison results of 388 A G polymorphism Genetic model MD (95% CI) Z p value I % Phet LDL-C [19,, 4, 6, 7, 30, 31] GG: AA [19,, 4, 6, 7, 30, 31] AG: AA [19,, 4, 6, 7, 9-31] GG: AA AG [19,, 4, 6, 7, 30, 31] GG AG: AA TC [, 4, 30, 31] GG: AA [, 4, 30, 31] AG: AA [, 4, 9-31] GG: AA AG [, 4, 30, 31] GG AG: AA HDL-C [4, 30, 31] GG: AA [4, 30, 31] AG: AA [4, 30, 31] GG: AA AG [4, 30, 31] GG AG: AA TG [4, 30, 31] GG: AA [4, 30, 31] AG: AA [4, 30, 31] GG: AA AG [4, 30, 31] GG AG: AA 0.77 ( 3.63,.09) 0.41 ( 0.60, 1.41) 0.11 ( 1.33, 1.1) 0.5 ( 0.4, 1.46).84 ( 6.37, 0.68) 1.61 ( 4.17, 0.95).04 ( 4.30, 0.) 1.90 ( 4.34, 0.53).78 ( 11.59, 6.03) 6.01 ( 15.6, 3.60).6 ( 4.80, 10.04) 4.5 ( 1.81, 3.76) 1.00 ( 31.7, 7.73) ( 36.04, 4.34) 5.38 ( 66.79, 16.03) ( 31.38, 4.44) Begg s test Egger s test z p value t p value Abbreviations: MD, mean difference ; CI, confidence interval; Phet, p value of heterogeneity; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride; p 0.05 stands for significance. R, random-effect model; others use fixed-effect model. the subjects with the 51C allele showed an attenuated LDL-C-lowering effect compared with those homozygous for the 51T allele. These findings suggest that the treatment duration and ethnicity may influence the statin effectiveness of lowering the LDL-C in patients with the SLCO1B1 51 T C polymorphism. In addition, when one study was omitted from the analyses (Akao, 01), all of the subgroup analyses results changed, suggesting that our results may be unstable and influenced by this study. Thus, the results should be interpreted with caution. Further studies regarding the treatment duration and ethnicity are necessary. Although previous studies have reported that pravastatin therapy for 4 weeks in hyperlipidemia patients was sufficient to decrease the total cholesterol level by 17-4% and LDL-C levels by 3-35% 48), the length of statin treatment duration in the original studies was varied. Consequently, we defined treatment duration 6 months as long-term effectiveness (a duration of approximately 6 months to 1 year), otherwise the duration was classified as short-term effectiveexclusion criteria for the cases, the baseline levels of LDL-C, HDL-C, TG and TC, duration of treatment, type of statin treatment, dosage and frequency of statins and the time of follow-up. Thus, in an attempt to explore the potential causes, both subgroup analyses and sensitivity analyses were performed. The risk of publication bias was also assessed with pre-defined criteria in order to obtain higher powered results. We performed leave-one-out sensitivity analyses. The exclusion of one particular study considerably altered the results of lowering LDL-C in the dominant model for the 51 T C SNP 5). In the study, the TT homozygotes (n 13) were found at a 0 times higher frequency than the CC homozygotes (n 5), which may have affected the sample power and led to an adverse result. To explore the effect of ethnicity on the effectiveness of statins with different genotypes, we excluded the studies involving Asian populations and a statistically significant association between the 51 T C polymorphism and statin effectiveness of lowering LDL-C was found in the dominant model, the recessive model and the homozygote comparison. Thus,

16 811 ness (a duration ranging from 1 to months). Our results suggest that allele C of the SLCO1B1 51 T C gene polymorphism has an attenuated statin effect on the long-term effectiveness of lowering LDL-C. In the clinical practice, most patients require long-term drug use. Therefore, the results of our study indicate that we must pay attention to the inter-individual variability in the statin therapeutic response in order to achieve a maximal effectiveness and minimal adverse effect. Limitations There are some limitations associated with this meta-analysis which may influence the findings. First, we searched for relevant studies in only four databases. Other relevant articles may exist in other databases. Second, we performed our analyses based on data adjusted for gender, body mass index, age, alcohol, smoking, diabetes, and location of the study. There may be potential confounding factors. Third, we focused on individual SNPs. Polymorphisms occur at low frequencies in many ethnic groups and are unlikely to explain the variations in response when solely considered. Furthermore, our meta-analysis did not take into consideration the possibility of SNP- SNP or gene-gene interactions or the possibility of linkage disequilibrium between the 51 T C and 388 A G SNPs. Instead, this study separately discussed those SNPs of the SLCO1B1 gene that may have a critical role in statin effectiveness of lowering cholesterol. Finally, the ethnicity of the participants in three studies (Akao, Martin, and Hedman) was not available and the treatment durations were inconsistent. Thus, with different treatment durations, the same polymorphism may play lead to different results different genetic backgrounds. Conclusion In conclusion, our study suggests that the SLCO1B1 51 T C and 388 A G polymorphisms do not affect the lipid-lowering effectiveness of statins. However, allele C of the SLCO1B1 51 T C gene polymorphism has an attenuated statin effectiveness on lowering LDL-C in non-asian populations and a long-term effectiveness. Thus, well-designed original studies involving more ethnicities and larger sample sizes are needed to confirm our findings. Acknowledgements and Notice of Grant Support We thank Hang Zhu (Center of Disease Control of Chengdu) and Jia Lu (Chongqing Medical University) for their assistance in modifying the article and consulting. The authors declare that there is no conflict of interest associated with this manuscript. This work was supported by scientific research project for young backbone teachers in School of Public Health and Management of Chongqing Medical University: Epidemiology research base building on chronic disease of university and college staff (Project No.: Gwzk01303) References 1) Murray CJ, Lopez AD: Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet, 1997; 349: ) Shepherd J, Cobbe SM, Ford I, Isles CG, Lorimer AR, MacFarlane PW, McKillop JH, Packard CJ: Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia Atheroscler Suppl, 004; 5: ) Jukema JW, Bruschke AV, van Boven AJ, Reiber JH, Bal ET, Zwinderman AH, Jansen H, Boerma GJ, van Rappard FM, Lie KI, et al.: Effects of lipid lowering by pravastatin on progression and regression of coronary artery disease in symptomatic men with normal to moderately elevated serum cholesterol levels. The Regression Growth Evaluation Statin Study (REGRESS). Circulation, 1995; 91: ) Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG, Brown L, Warnica JW, Arnold JM, Wun CC, Davis BR, Braunwald E: The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med, 1996; 335: ) Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. N Engl J Med, 1998; 339: ) Pravastatin use and risk of coronary events and cerebral infarction in japanese men with moderate hypercholesterolemia: the Kyushu Lipid Intervention Study. J Atheroscler Thromb, 000; 7: ) Pedersen TR, Kjekshus J, Berg K, Haghfelt T, Faergeman O, Faergeman G, Pyorala K, Miettinen T, Wilhelmsen L, Olsson AG, Wedel H: Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S) Atheroscler Suppl, 004; 5: ) Matsuzaki M, Kita T, Mabuchi H, Matsuzawa Y, Nakaya N, Oikawa S, Saito Y, Sasaki J, Shimamoto KItakura H: Large scale cohort study of the relationship between serum cholesterol concentration and coronary events with low-dose simvastatin therapy in Japanese patients with hypercholesterolemia. Circ J, 00; 66:

17 81 9) Mabuchi H, Kita T, Matsuzaki M, Matsuzawa Y, Nakaya N, Oikawa S, Saito Y, Sasaki J, Shimamoto K, Itakura H: Large scale cohort study of the relationship between serum cholesterol concentration and coronary events with low-dose simvastatin therapy in Japanese patients with hypercholesterolemia and coronary heart disease: secondary prevention cohort study of the Japan Lipid Intervention Trial (J-LIT). Circ J, 00; 66: ) Matsuzawa Y, Kita T, Mabuchi H, Matsuzaki M, Nakaya N, Oikawa S, Saito Y, Sasaki J, Shimamoto K, Itakura H: Sustained reduction of serum cholesterol in low-dose 6-year simvastatin treatment with minimum side effects in 51,31 Japanese hypercholesterolemic patients. Circ J, 003; 67: ) Pitt B, Waters D, Brown WV, van Boven AJ, Schwartz L, Title LM, Eisenberg D, Shurzinske L, McCormick LS: Aggressive lipid-lowering therapy compared with angioplasty in stable coronary artery disease. Atorvastatin versus Revascularization Treatment Investigators. N Engl J Med, 1999; 341: ) Sever PS, Dahlof B, Poulter NR, Wedel H, Beevers G, Caulfield M, Collins R, Kjeldsen SE, McInnes GT, Mehlsen J, Nieminen M, O Brien EOstergren J: Rationale, design, methods and baseline demography of participants of the Anglo-Scandinavian Cardiac Outcomes Trial. ASCOT investigators. J Hypertens, 001; 19: ) Kim RB: 3-Hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors (statins) and genetic variability (single nucleotide polymorphisms) in a hepatic drug uptake transporter: what s it all about? Clin Pharmacol Ther, 004; 75: ) Chung JY, Cho JY, Yu KS, Kim JR, Oh DS, Jung HR, Lim KS, Moon KH, Shin SG, Jang IJ: Effect of OATP1B1 (SLCO1B1) variant alleles on the pharmacokinetics of pitavastatin in healthy volunteers. Clin Pharmacol Ther, 005; 78: ) Lau YY, Okochi H, Huang Y, Benet LZ: Multiple transporters affect the disposition of atorvastatin and its two active hydroxy metabolites: application of in vitro and ex situ systems. J Pharmacol Exp Ther, 006; 316: ) Tirona RG, Leake BF, Merino G, Kim RB: Polymorphisms in OATP-C: identification of multiple allelic variants associated with altered transport activity among European- and African-Americans. J Biol Chem, 001; 76: ) Nishizato Y, Ieiri I, Suzuki H, Kimura M, Kawabata K, Hirota T, Takane H, Irie S, Kusuhara H, Urasaki Y, Urae A, Higuchi S, Otsubo K, Sugiyama Y: Polymorphisms of OATP-C (SLC1A6) and OAT3 (SLCA8) genes: consequences for pravastatin pharmacokinetics. Clin Pharmacol Ther, 003; 73: ) Niemi M, Schaeffeler E, Lang T, Fromm MF, Neuvonen M, Kyrklund C, Backman JT, Kerb R, Schwab M, Neuvonen PJ, Eichelbaum M, Kivisto KT: High plasma pravastatin concentrations are associated with single nucleotide polymorphisms and haplotypes of organic anion transporting polypeptide-c (OATP-C, SLCO1B1). Pharmacogenetics, 004; 14: ) Akao H, Polisecki E, Kajinami K, Trompet S, Robertson M, Ford I, Jukema JW, de Craen AJ, Westendorp RG, Shepherd J, Packard C, Buckley BM, Schaefer EJ: Genetic variation at the SLCO1B1 gene locus and low density lipoprotein cholesterol lowering response to pravastatin in the elderly. Atherosclerosis, 01; 0: ) Tachibana-Iimori R, Tabara Y, Kusuhara H, Kohara K, Kawamoto R, Nakura J, Tokunaga K, Kondo I, Sugiyama Y, Miki T: Effect of genetic polymorphism of OATP-C (SLCO1B1) on lipid-lowering response to HMG-CoA reductase inhibitors. Drug Metab Pharmacokinet, 004; 19: ) Zhang W, Chen BL, Ozdemir V, He YJ, Zhou G, Peng DD, Deng S, Xie QY, Xie W, Xu LY, Wang LC, Fan L, Wang A, Zhou HH: SLCO1B1 51T-->C functional genetic polymorphism and lipid-lowering efficacy of multiple-dose pravastatin in Chinese coronary heart disease patients. Br J Clin Pharmacol, 007; 64: ) Couvert P, Giral P, Dejager S, Gu J, Huby T, Chapman MJ, Bruckert E, Carrie A: Association between a frequent allele of the gene encoding OATP1B1 and enhanced LDLlowering response to fluvastatin therapy. Pharmacogenomics, 008; 9: ) Martin NG, Li KW, Murray H, Putt W, Packard CJ, Humphries SE: The effects of a single nucleotide polymorphism in SLCO1B1 on the pharmacodynamics of pravastatin. Br J Clin Pharmacol, 01; 73: ) Yang GP, Yuan H, Tang B, Zhang W, Wang LS, Huang ZJ, Ou-Yang DS, Zhang GX, Zhou HH: Lack of effect of genetic polymorphisms of SLCO1B1 on the lipid-lowering response to pitavastatin in Chinese patients. Acta Pharmacol Sin, 010; 31: ) Fu Q, Li YP, Gao Y, Yang SH, Lu PQ, Jia M, Zhang LR: Lack of association between SLCO1B1 polymorphism and the lipid-lowering effects of atorvastatin and simvastatin in Chinese individuals. Eur J Clin Pharmacol, 013; 69: ) Hu M, Mak VW, Tomlinson B: Intronic variants in SLCO1B1 related to statin-induced myopathy are associated with the low-density lipoprotein cholesterol response to statins in Chinese patients with hyperlipidaemia. Pharmacogenet Genomics, 01; : ) Thompson JF, Man M, Johnson KJ, Wood LS, Lira ME, Lloyd DB, Banerjee P, Milos PM, Myrand SP, Paulauskis J, Milad MA, Sasiela WJ: An association study of 43 SNPs in 16 candidate genes with atorvastatin response. Pharmacogenomics J, 005; 5: ) Hedman M, Antikainen M, Holmberg C, Neuvonen M, Eichelbaum M, Kivisto KT, Neuvonen PJ, Niemi M: Pharmacokinetics and response to pravastatin in paediatric patients with familial hypercholesterolaemia and in paediatric cardiac transplant recipients in relation to polymorphisms of the SLCO1B1 and ABCB1 genes. Br J Clin Pharmacol, 006; 61: ) Rodrigues AC, Perin PM, Purim SG, Silbiger VN, Genvigir FD, Willrich MA, Arazi SS, Luchessi AD, Hirata MH, Bernik MM, Dorea EL, Santos C, Faludi AA, Bertolami MC, Salas A, Freire A, Lareu MV, Phillips C, Porras- Hurtado L, Fondevila M, Carracedo A, Hirata RD: Pharmacogenetics of OATP Transporters Reveals That SLCO1B1 c.388a>g Variant Is Determinant of Increased Atorvastatin Response. Int J Mol Sci, 011; 1:

18 813 30) Sortica VA, Fiegenbaum M, Lima LO, Van der Sand CR, Van der Sand LC, Ferreira ME, Pires RC, Hutz MH: SLCO1B1 gene variability influences lipid-lowering efficacy on simvastatin therapy in Southern Brazilians. Clin Chem Lab Med, 01; 50: ) Shabana MF, Mishriki AA, Issac MS, Bakhoum SW: Do MDR1 and SLCO1B1 polymorphisms influence the therapeutic response to atorvastatin? A study on a cohort of Egyptian patients with hypercholesterolemia. Mol Diagn Ther, 013; 17: ) Ghatak A, Faheem O, Thompson PD: The genetics of statin-induced myopathy. Atherosclerosis, 010; 10: ) Link E, Parish S, Armitage J, Bowman L, Heath S, Matsuda F, Collins R: SLCO1B1 variants and statin-induced myopathy--a genomewide study. N Engl J Med, 008; 359: ) Linde R, Peng L, Desai M, Feldman D: The role of vitamin D and SLCO1B1* 5 gene polymorphism in statinassociated myalgias. Dermatoendocrinol, 010; : 77 35) Carr DF, O Meara H, Jorgensen AL, Campbell J, Hobbs M, McCann G, Pirmohamed M: SLCO1B1 genetic variant associated with statin-induced myopathy: a proof-ofconcept study using the clinical practice research datalink. Clin Pharmacol Ther, 013; 94: ) Voora D, Shah SH, Spasojevic I, Ali S, Reed CR, Salisbury BA, Ginsburg GS: The SLCO1B1* 5 genetic variant is associated with statin-induced side effects. J Am Coll Cardiol, 009; 54: ) Ho RH, Choi L, Lee W, Mayo G, Schwarz UI, Tirona RG, Kim RB: Effect of drug transporter genotypes on pravastatin disposition in European-and African-American participants. Pharmacogenet Genomics, 007; 17: ) Slim K, Nini E, Forestier D, Kwiatkowski F, Panis Y, Chipponi J: Methodological index for non-randomized studies (minors): development and validation of a new instrument. ANZ J Surg, 003; 73: ) Chowbay B, Li H, David M, Cheung YB, Lee EJ: Metaanalysis of the influence of MDR1 C3435T polymorphism on digoxin pharmacokinetics and MDR1 gene expression. Br J Clin Pharmacol, 005; 60: ) Rodriguez S, Gaunt TR, Day IN. Hardy-Weinberg equilibrium testing of biological ascertainment for Mendelian randomization studies. Am J Epidemiol, 009; 169: ) Hubacek JA, Dlouha D, Adamkova V, Lanska V, Ceska R, Vrablik M: Possible gene-gender interaction between the SLCO1B1 polymorphism and statin treatment efficacy. Neuro Endocrinol Lett, 01; 33 Suppl : -5 4) Igel M, Arnold KA, Niemi M, Hofmann U, Schwab M, Lutjohann D, von Bergmann K, Eichelbaum M, Kivisto KT: Impact of the SLCO1B1 polymorphism on the pharmacokinetics and lipid-lowering efficacy of multiple-dose pravastatin. Clin Pharmacol Ther, 006; 79: ) Hopewell JC, Parish S, Offer A, Link E, Clarke R, Lathrop M, Armitage J, Collins R: Impact of common genetic variation on response to simvastatin therapy among participants in the Heart Protection Study. Eur Heart J, 013; 34: ) Chasman DI, Giulianini F, MacFadyen J, Barratt BJ, Nyberg F, Ridker PM: Genetic determinants of statininduced low-density lipoprotein cholesterol reduction: the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) trial. Circ Cardiovasc Genet, 01; 5: ) Singer JB, Holdaas H, Jardine AG, Fellstrom B, Os I, Bermann G, Meyer JM: Genetic analysis of fluvastatin response and dyslipidemia in renal transplant recipients. J Lipid Res, 007; 48: ) Pasanen MK, Miettinen TA, Gylling H, Neuvonen PJ, Niemi M: Polymorphism of the hepatic influx transporter organic anion transporting polypeptide 1B1 is associated with increased cholesterol synthesis rate. Pharmacogenet Genomics, 008; 18: ) Patel J, Abd T, Blumenthal RS, Nasir K, Superko HR: Genetics and personalized medicine--a role in statin therapy? Curr Atheroscler Rep, 014; 16: ) Pan HY, DeVault AR, Swites BJ, Whigan D, Ivashkiv E, Willard DA, Brescia D: Pharmacokinetics and pharmacodynamics of pravastatin alone and with cholestyramine in hypercholesterolemia. Clin Pharmacol Ther, 1990; 48: 01-07

19 814 Supplemental Table 1. Subgroup analyses of some indicators stratified by the length of follow-up Gene Indicator Genotype Length of follow-up N MD p value 51T C 388A G TG TG TC HDL-C CC: TT TC: TT CC: TC TT CC TC: TT GG: AA AG: AA GG: AA AG GG AG: AA GG: AA AG: AA GG: AA AG GG AG: AA GG: AA AG: AA GG: AA AG GG AG: AA short-term [5] 1 1 short-term [0, 1, 5, 8] 1 4 short-term [5] 1 1 short-term [4, 5] 1 short-term [4, 31] 1 short-term [4, 31] 1 short-term [4, 31] 1 short-term [4, 31] 1 short-term [, 4, 31] 1 3 short-term [, 4, 31] 1 3 short-term [, 4, 9, 31] 1 4 short-term [, 4, 31] 1 3 short-term [4, 31] 1 short-term [4, 31] 1 short-term [4, 31] 1 short-term [4, 31] [ 47.70, 1.30] 0.40 [ 5.56, 4.75] [ 40.01, 19.1] 0.0 [.0,.4] [ 43.46, 13.86] 0.30 [ 5.4, 4.83] [ 39.07, 17.07] 0.34 [.46, 1.79] 0.30 [ 36.48, 37.08] [ 40.33, 6.41] 0.30 [ 36.90, 37.50].59 [ 46.63, 1.44] 0.10 [ 31.38, 31.58] [ 94.63, 19.63] 0.30 [ 3.89, 33.49] [ 40.40,.15] 7.10 [ 1.33, 7.13].57 [ 6.1, 1.08] 5.40 [ 19.86, 9.06] 1.48 [ 4.08, 1.1] 3.50 [ 15.67, 8.67] 1.99 [ 4.8, 0.31] 5.90 [ 18.80, 7.00] 1.76 [ 4.4, 0.7].43 [ 16.59, 11.73] 3.00 [ 14.6, 8.6] 3.43 [ 17.19, 10.33] 8.46 [ 1.89, 4.96] 0.0 [ 11.9, 11.5] 4.51 [ 5.08, 14.10] 3.13 [ 15.64, 9.38] 5.61 [ 16.66, 5.44] Abbreviations: MD, mean difference; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride; N, Number of included studies in each subgroup

20 815 Supplemental Table. Subgroup analyses of some indicators stratified by ethnicity Gene Indicator Genotype Ethnicity N MD p value 51T C 388A G TG TG TC HDL-C CC: TT TC: TT CC: TC TT CC TC: TT GG: AA AG: AA GG: AA AG GG AG: AA GG: AA AG: AA GG: AA AG GG AG: AA GG: AA AG: AA GG: AA AG GG AG: AA Asian [5] non-asian [30] 1 1 [0, 1, 5] Asian 3 non-asian [8, 30] Asian [5] non-asian [30] 1 1 [4, 5] Asian non-asian [30] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [, 30, 31] 1 3 Asian [4] non-asian [, 30, 31] 1 3 Asian [4] non-asian [, 9, 30, 31] 1 4 Asian [4] non-asian [, 30, 31] 1 3 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] 1 Asian [4] non-asian [30, 31] [ 5.56, 4.75] [ 47.70, 1.30] 0.10 [.13,.33] 11. [ 31.18, 8.75] 0.30 [ 5.4, 4.83] [ 43.46, 13.86] 0.34 [.46, 1.79] [ 39.07, 17.07] [ 47.1, 13.1] 8.4 [ 34.34, 17.87] [ 73.09, 0.91] 6.1 [ 30.57, 18.15] [ 94.14, 39.46] 5.00 [ 5.08, 15.08] 4.40 [ 53.3, 4.5] 6.66 [ 9.48, 16.16] 3.00 [ 1.48, 15.48].84 [ 6.43, 0.76] 3.00 [.1, 16.1] 1.58 [ 4.16, 1.00] 0.60 [ 8.98, 7.78].15 [ 4.49, 0.19] 3.00 [ 1.5, 15.5] 1.88 [ 4.34, 0.57] 4.00 [ 17.78, 9.78] 1.94 [ 13.40, 9.5] [ 3.98, 4.98] 3.5 [ 14.40, 7.89] 7.30 [ 5.73, 0.33] 0.37 [ 8.67, 9.40] 7.70 [ 1.81, 6.41].85 [ 13.08, 7.37] Abbreviations: MD, mean difference; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride; N, Number of included studies in each subgroup