Comments on “CYP3AP1*3 Allele Is Associated With Lipid‐Lowering Efficacy of Simvastatin and Atorvastatin in Chinese Women”

Li et al1 reported an interesting association between the CYP3AP1*3 polymorphism and lipid-lowering efficacy of simvastatin and atorvastatin in Chinese patients, which was only significant in the female subgroups. However, the mean reductions from baseline in low-density lipoprotein cholesterol (LDL-C) after 4 weeks of treatment with both simvastatin 20 mg and atorvastatin 20 mg were almost identical at 27.3%, which is at variance with results from multiple previous studies showing that atorvastatin is approximately twice as potent in lowering LDL-C as simvastatin. A recent meta-analysis of 37 studies, 2 using mainly intention-to-treat analyses, found that LDL-C was reduced by 33% and 41.4% with at least 4 weeks of treatment of simvastatin 20 mg and atorvastatin 20 mg, respectively. Another study in Asian patients with good compliance showed that treatment with atorvastatin 10 mg daily for 12 weeks reduced LDL-C by 40.2%,3 and our data in Hong Kong Han Chinese patients with almost 100% medication compliance showed a 41.0% reduction in LDL-C after 4 weeks treatment with simvastatin 20 mg.4 The reductions in LDL-C reported by Li et al are therefore much less than expected, which may suggest poor drug or dietary compliance or that the starting lipid values were not a true baseline. Li et al did not report drug or dietary compliance; hence, any changes in the lipid profile may not be fully ascribable to the drug therapy alone. The authors mentioned that subjects were not allowed to take other lipid-lowering medications for 7 days before the study, but it usually takes up to 4 weeks for the lipid-lowering effects of statins to wear off,5 and fibrates require a longer washout period because of their diverse pharmacodynamic effects on parameters like lipid and lipoprotein metabolism, plasma creatinine, and alterations in gene transcription mediated by the alpha form of peroxisome proliferator-activated receptors.6 The finding by Li et al that the reduction in total cholesterol (TC) with atorvastatin was significantly lower in female CYP3AP1*3/*3 carriers than in CYP3AP1*1 carriers (17.0% vs 18.6%) is also difficult to interpret. Reductions in TC with statin treatment represent a composite effect, largely driven by the primary reduction in LDL-C with small contributions from reduction in cholesterol carried in very low-density lipoprotein particles, reflected by the reduction triglycerides, and the change in high-density lipoprotein cholesterol (HDL-C), which is usually a small increase. There was no difference in the reduction in LDL-C between these 2 genotype groups, and the main component contributing to the greater reduction in TC in CYP3APq*1 carriers was a smaller increase in HDL-C. The most important determinant of the increase in HDL-C is the baseline level,7 and in these subgroups the mean baseline HDL-C levels were numerically lower in the CYP3AP1*3/*3 group, which might predict a greater percentage increase with statin treatment. Thus, although the differences in change in TC in these subgroups may be statistically significant, they have no biological meaning and are likely to be a chance finding. CYP3AP1 is a pseudogene that is usually nonfunctional.8 The CYP3AP1*3 polymorphism is almost completely linked with the CYP3A5*3 variant.9 Although CYP3A5*3 has been associated with increased systemic exposure to simvastatin in healthy volunteers10 and an early small study in European Caucasian patients treated with various statins metabolized through CYP3A reported that there were greater lipid-lowering responses in the homozygous CYP3A5*3 enzyme nonexpressers (n = 39) compared with CYP3A5*1 wild-type enzyme expressers (n = 7),11 some other large studies (n > 100) failed to replicate significant associations between the CYP3A5*3 polymorphism and lipid responses to simvastatin or atorvastatin in Caucasians patients,12–15 whereas others reported opposite associations.16 In the pharmacogenetic analyses of the response to atorvastatin in the Treating to New Targets (TNT) and PROVE IT-TIMI 22 studies involving large numbers of subjects (n = 5745 and n = 686, respectively), the CYP3A4 and CYP3A5 polymorphisms showed no associations with the lipid responses to atorvastatin.14,15 Furthermore, a recent in vitro study suggested that atorvastatin was preferentially metabolized by CYP3A4 rather than by CYP3A5, and the authors suggested that CYP3A5 polymorphisms were unlikely to affect the disposition and efficacy of atorvastatin.17 Thus, the effects of CYP3A polymorphisms on the responses to statins have been quite variable between studies,18 and further studies are necessary to corroborate the observations of Li et al. Financial disclosure: Brian Tomlinson is a Fellow of the American College of Clinical Pharmacology (FCP); has received research funding from AstraZeneca, Bayer, Boehringer Ingelheim, Daiichi Sankyo, Kowa, Merck, Merck Sharp and Dohme, Novartis, Otsuka, Pfizer, Roche, Sanofi-Aventis, and Servier to perform clinical studies; and has acted as a consultant or speaker on occasions for AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Kowa, Merck, Merck Sharp and Dohme, Novartis, Pfizer, Roche, Sanofi-Aventis, Schering-Plough, and Servier.