Abstract: Statins inhibit cholesterol synthesis by blocking 3-hydroxy-3-methylglutaryl coenzyme A reductase in the liver, thereby ameliorating hypercholesterolemia. Thus, to determine statins with the best efficacy, a meta-analysis was performed to compare the effects of statins against hypercholesterolemia. Comprehensive literature searches were established, from Cochrane library, Pubmed, Embase. The studies were performed to randomize controlled trials (RCTs), cohort studies or case-control studies about efficacy of different statin drugs and dose against hypercholesterolemia published between 1997 and 20 February, 2017. Study qualities were assessed according to Cochrane collaboration recommendations. The non-programming software Aggregate Data Drug Information System (ADDIS) (version 1.16.5) was used to perform Bayesian network meta-analysis and compare treatments using the Markov Chain Monte Carlo (MCMC) method. Overall, 28 RCTs studies, including 12855 patients, met the inclusion criteria. Total cholesterol (TC) levels significantly reduced (p<0.05) using 2 mg Pitavastatin (Pit) than those using 20 mg Pravastatin (Pra), 10 mg Simvastatin (Sim) or 10 mg Atorvastatin (Ato). Similarly, triglyceride (TG) levels reduced using 2 mg Pit than those using 20 mg Pra (p<0.05), 10 mg Sim (p<0.05) or 20 mg Sim (p<0.05) and reduced apolipoprotein B (Apo B) levels were observed than those using 10 mg Ato or 20 mg Pra (p<0.05). Rosuvastatin (Ros) significantly reduced TC and TG levels (p<0.05) when administered at 20 and 10 mg Ros treatments ameliorated percentage changes in low-density lipoprotein cholesterol more than the other drugs (p<0.05) and increased high-density lipoprotein cholesterol levels more effectively than 10 mg Ato (p<0.05), 20 mg Pra (p<0.05) or 10 mg Sim (p<0.05). Increases in Apo A1 levels did not differ between treatments (p>0.05). Among the present statin drug regimens, 2 mg Pit and 10 or 20 mg Ros had the highest efficacy against hypercholesterolemia.
INTRODUCTION
Hypercholesterolemia is a lipid metabolism disorder that is characterized by very high cholesterol levels in the blood and increased risks of coronary heart disease (CHD)1. Approximately 1 in 300-500 people in most countries carry inherited familial hypercholesterolemia, which can result in extremely high cholesterol levels (above 300 mg dL–1)2,3. Multiple studies show decreased the risks of CHD in patients with hypercholesterolemic receiving treatments with statins4,5, which are 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) inhibitors6.
Because HMG-CoA reductase is necessary for the production of cholesterol7, the statins can block HMG-CoA rosuvastatin (Ros), atorvastatin (Ato), simvastatin (Sim), pravastatin (Pra), lovastatin (Lov) and pitavastatin (Pit) inhibit cholesterol synthesis and increase low-density lipoprotein (LDL) uptake in the liver8,9. At doses of 5, 10 and 20 mg, Ros significantly reduces LDL and improves flow-mediated dilation as well as reduces adiponectin levels and insulin sensitivity in patients with hypercholesterolemia after 2 months10,11. In contrast, Ato inhibits calcification of the aortic valves by inhibiting the LRP5 receptor (murine LDL receptor homologue) pathway in hypercholesterolemia mice12. Sim is an effective lipid-lowering drug that can decrease LDL levels by up to 50%13,14. Pit decreases the levels of serum lectin-like oxidized LDL receptor 1 ligand and membrane-type 1 matrix metalloproteinase expression in CD14+mononuclear cells in patients with hypercholesterolemia15. In addition, in a previous dose specific meta-analysis, statins acted as anti-hypercholesterolemia agents by reducing the total cholesterol (TC) and LDL cholesterol (LDL-C)16,17. However, the relative efficacy of these statins remains unclear.
In this study, a comprehensive literature search was conducted from the Cochrane Library, Pubmed, Embase databases upto 20 February, 2017. These studies were included in the present study, such as randomized controlled trials (RCTs), cohort studies or case-control studies on statin drug and dose effects in patients with hypercholesterolemia. Subsequently, the quality of studies was assessed and aggregate data drug information system software was used to compare treatments in a Bayesian meta-analysis. This Cochrane systematic review provides high-level comparisons of the effects of various statins and provides essential guidance for the treatment of hypercholesterolemia.
MATERIALS AND METHODS
Search strategy: The Cochrane Library, Pubmed, Embase databases were comprehensively searched for literature published before 20 February, 2017 using the key search terms "statin", "rosuvastatin", "atorvastatin" or "atorvastatine", "simvastatin" or "simvastatine", "pravastatin" or "pravachol", "lovastatin", "pitavastatin", "hypercholesteremia" or "hypercholesterolemia" or "hypercholesteroli" and "randomized controlled trial (RCT)". Only studies published in English were retrieved.
Selection criteria: Included studies met the following inclusion and exclusion criteria for the meta-analysis: (1) They reported the efficacy of statins as hypercholesterolemia treatments and were published in English language journals and (2) Their design included RCTs, cohort studies or case-control studies on the efficacy of varying doses of statin drugs and determination of TC, LDL-C, high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), apolipoprotein B (Apo B) and Apo A1 (% change). Reviews, reports, letters and comments were excluded.
Study characteristics: In the initial network meta-analysis, screened 2384 relevant studies from the database searches. 972 duplicated articles and 1249 articles that did not completely meet the inclusion criteria. These were excluded without retrieving full papers (Fig. 1). Subsequently, from the remaining 135 articles, review articles (30), letters/editors/comments (29), case series/reports (19), articles in the same area (11) and data from a total of 12855 patients with hypercholesterolemia from different countries in 28 RCT studies. Among these patients, 3732 were treated with Ato, 1246 were treated with Pra, 6207 patients treated with Ros, 386 were treated with Pit, 1284 were treated with Sim. As shown in Table 1, these studies were published between 1997 and 2015, included patients groups that did not differ in ages, sex or BMI, were followed for 6-12 weeks and were treated with drugs at dose of 1-40 mg. TC, LDL-C, HDL-C, TG, Apo B and Apo A1 (change %) levels were reported in all included studies (Table 2).
Data extraction and quality evaluation: To reduce bias, three investigators independently reviewed and extracted the information from all enrolled studies and reached a consensus on all items during discussion with an additional investigator. The following data were extracted from each eligible study:
| Fig. 1: | Diagram of the study screening protocol |
| Table 1: | Characteristics of included studies from online literature databases (1997-2017) |
m(se)/m±sd: Mean (standard error)/mean±standard deviation, Ato: Atorvastatin, Ros: Rosuvastatin, Pit: Pitavastatin, Pra: Pravastatin, Sim: Simvastatine, NA: Not available | |
name of the first author, year of publication, case number, age, sex and body mass index (BMI), follow-up time, methods of intervention, kinds of statin, dose, TC, LDL-C, HDL-C, TG, Apo B and Apo A1 (change %). Subsequently, the risk of bias was assessed according to the recommendations of the Cochrane collaboration.
Statistical analysis: The non-programming software Aggregate Data Drug Information System (ADDIS) (version 1.16.5), was used to perform Bayesian network meta-analysis and compare treatments using the Markov Chain Monte Carlo (MCMC) method18,19. All data are presented as Means±Standard Deviation (M±SD) or 95% confidence interval (CIs). A random-effects model was used in all test and node-splitting analyses were used to assess inconsistencies in network meta-analysis20. The model was considered consistent when p>0.05 and convergence in the model was identified according to Potential Scale Reduction Factor (PSRF) using Brooks-Gelman-Rubin method. In general, PSRF that are close to 1 indicate good convergence21. Finally, it is evaluated ranking probabilities of drugs to identify the best therapy for hypercholesterolemia.
| Table 2: | Indices of hypercholesterolemia are listed with drug names and doses |
1: Lipid-lowering therapy, 2: Switched, Change (%): Percent change from base line, Values obtained by Mean±Standard deviation (M±SD) |
|
RESULTS
Quality evaluation displayed that all 28 studies were of relatively high quality (Fig. 2a). However, blinding of participants and personnel (performance bias) and blinding of outcome assessments (detection bias) were not mentioned in many studies, potentially introducing a source of bias (Fig. 2b).
Comparative effects of statins on TC levels: Node-splitting analysis indicated sufficient consistency (p>0.05) and PSRF ranged from 1.00-1.02. Thus, a consistency model was used for meta-analysis. As shown in Fig. 3a, regiments of Pit at 2 mg doses had the highest probability of reducing TC levels followed by those of Ros at 20 mg doses. The effects of 2 mg Pit regimens differed significantly from those of 20 mg Pra (p = 0.003), 10 mg Sim (p = 0.020), 10 mg Ato (p = 0.006). 20 mg Ros were significantly different from those of 20 mg Praregimens (p = 0.040).
Comparative effect of statins on TG levels: Treatments with Pit at 2 mg doses had the highest probability of reducing TG levels (Fig. 3b), followed by those with Pit at 1 mg and then Ros at 20 mg. The effects of 2 mg Pit regimens were significantly superior to those of 20 mg Pra (p = 0.006), 10 mg Pra (p = 0.048), 40 mg Pra (p = 0.002), 10 mg Sim (p = 0.008) and Sim 20 mg (P = 0.049). No significant differences were identified between the effects of 1 mg Pit and those of other drugs. However, 20 mg Ros regimens had significantly greater efficacy than40 mg Pra (p = 0.001), Ros 5 mg (p = 0.040), Sim 10 mg (p = 0.002) and Sim 20 mg (p = 0.030).
Comparative effects of statins on percent changes in LDL-C levels: In assessments of consistency, node-splitting analysis was used. p>0.05 and PSRF ranged from 1.00-1.01, warranting use of the consistency model. As shown in Fig. 3c, 10 mg Ros regimens had the highest probability of reducing percent changes in LDL-C levels and were significantly superior to the drug regimens (p<0.05).
Comparative effects of statins on percent changes in HDL-C levels: As shown in Fig. 3d, 1 mg Pit had the highest probability of increasing changes in HDL-Clevels followed by 10 mg Ros regimens. In addition, the efficacy of 1 mg Pit significantly differed from that of 20 mg Ato (p = 0.020) and that of 10 mg Ros doses significantly differed from that of 10 mg Ato (p<0.001), 20 mg Ato (p = 0.001), 20 mg Pra (p = 0.004), 40 mg Pra (p = 0.001), 10 mg Sim (p = 0.002) and 20 mg Sim regimens (p = 0.006).
| Fig. 2(a-b): | (a) Quality evaluation of included studies was performed according to the recommendations of the Cochrane collaboration (b) Blinding of participants and personnel (performance bias) and blinding of outcome assessments (detection bias), introducing a source of bias |
| Fig. 3(a-f): | Statins were ranked according to the probability of efficacy as treatments for hypercholesterolemia on (a) TC, (b) TG, (c) LDL-C, (d) HDL-C, (e) Apo B and (f) Apo A1 |
Comparative effects of statins on Apo B levels: Among the present statins, regimens of 2 mg Pit had the highest probability of reducing changes in Apo B levels followed by regimens of 20 mg Ros (Fig. 3e). Moreover, the effects of 2 mg Pit were significantly different f rom those of 10 mg Ato (p = 0.030) and Pra 20 mg (p = 0.010). However, no significant differences were identified between the effects of 20 mg Ros and other drugs (p>0.05).
Comparative effects of statins on Apo A1 levels: Although the highest probability of increasing changes in Apo A1 levels were achieved with 10 mg Ros followed by 10 mg Sim (Fig. 3f).
DISCUSSION
In the present systematic review, 28 studies from global databases met the inclusion criteria and contained data from a total of 12855 patients with hypercholesterolemia. Subsequent network meta-analysis showed that TC levels are significantly reduced in patients with hypercholesterolemia receiving 2 and 20 mg doses of Pit and Ros, (p<0.05), respectively. However, although 2 mg Pit, 1 mg Pit and 20 mg Ros regimens reduced TG levels. 10 mg Ros regimens had the highest probability of reducing the changes in LDL-C levels. Moreover, the probability of increasing changes in HDL-C levels was greater with 1 mg Pit and 10 mg Ros regimens, whereas 2 mg Pit and 20 mg Ros had the highest probability of reducing the changes of Apo B. However, although 10 mg Ros and 10 mg Sim regimens were more likely to increase Apo A1 levels, no significant differences in the levels with other drugs were identified.
Multiple studies show that statins decreased the levels of TC, TG, LDL-C and Apo B levels and significantly increase HDL-C and Apo A1 levels in patients with hypercholesterolemia22-49. In a study by Avis50, the efficacy and safety of statin therapies were assessed in children with heterozygous familial hypercholesterolemia but the reported effects were inconsistent with the findings of the present meta-analysis. For example, at doses of 10-20 mg Ato reduced the levels of LDL-C by 39%, TC by 30% and Apo B by 34%, whereas, at doses of 40 mg, Pra only increased HDL-C levels by 9% and Apo A1 levels by 5%50. In contrast, significantly reduced TC, TG and Apo B levels following treatments with 2 mg Pit or 20 mg Ros and 10 mg Ros treatments had the highest probability of reducing changes in LDL-C levels and increasing HDL-C and Apo A1 levels. These discrepancies likely reflect differing patients groups in the present studies compared with those in the study by Avis50, which only included children with heterozygous familial hypercholesterolemia. Additionally, Yokote et al.51 found that 2 mg Pit and 10 mg Ato regimens reduced TC, TG and LDL-C in Japanese patients with hypercholesterolemia. In particular, 2 mg Pit decreased Apo B levels with an efficacy equal to that of 10 mg Ato in patients with primary hypercholesterolemia52,53. Treatments with 10 mg Ros significantly decreased LDL-C levels and increased HDL-C levels comparison with those with 20 mg Sim 20 and 40 mg Pra54. Moreover, treatments with 10 mg Ros reduced LDL-C levels more than those with 20 mg Ato, whereas, HDL-C levels were similarly increased by both treatments55. These data were similar to the findings of the present meta-analysis, which indicated that treatments with10 mg Ros show the highest chances of reducing the percent changes of LDL-C levels and increasing HDL-C levels. Previously, treatments with Ros at 10 or 20 mg were effective and safe in patients at high risks of CHD and acted by reducing TC, LDL-C and Apo B levels56. Ros also reduced TC with significantly greater efficacy than the other statins and reduced TG levels significantly more than 10-80 mg treatments with Sim and Pra57. The present meta-analysis provides strong evidence of the relative efficacies of statins types and doses and could be used to inform future selections were helpful for patients statin drugs patients with hypercholesterolemia. However, (1) The subgroup analyses were not performed due to the absence of relevant data lacked in some included studies, (2) Numbers eligible studies were low and other indexes were insufficiently reported to form a closed cycle network meta-analysis, (3) ADDIS is a non-programming software with limited facilities and (4) ‘No’ safety data for statins were investigated.
CONCLUSION
Based on the present data from 28 studies, 2 mg Pit or 10 and 20 mg Ros ameliorated hypercholesterolemia with greater efficacy than those with other statin drugs. The present evidence will inform further investigations and clinical selections of statins for the treatment of patients with hypercholesterolemia.
SIGNIFICANCE STATEMENTS
The results of meta-analysis show that 2 mg Pit and 10 and 20 mg Ros produce better outcomes than other statin drugs in patients with hypercholesterolemia. These findings provide important information for the selection of statin drugs for the treatment of the patients with hypercholesterolemia.
ACKNOWLEDGMENT
This study was supported by Beijing Lisheng Cardiovascular Health Foundation Linghang Funds (Program No. LHJJ20159227.