The main finding of the present study was to observe significant differences between the allelic and genotypic frequencies of SLCO1B1 rs4149056 polymorphism associated with statin-induced myopathy, according to ethnicity in the Brazilian general population. To our knowledge this is the first study indicating that Brazilian Amerindians present higher frequencies of rs4149056 C variant allele (28.3%) and probably higher risk for statin-induced myopathy. These findings could be useful in the strategic planning of the implementation of pharmacogenetic testing according to ethnicity.
The SLCO1B1 rs4149056 C variant allele frequency found in our study was higher (p < 0.001) in Amerindian subjects (28.3%) compared with other ethnic groups and with Turkish  (12.2%); English  (13.0%); European American  (14.3%); and German  (15.0%) subjects. However, the Caucasian descent (14.8%) and Mulatto (14.9%) ethnic groups of our study presented similar (p > 0.050) rs4149056 C variant allele frequency compared with Caucasian descent subjects from Europe and USA. In addition, our African descent (5.7%) subject group had similar (p > 0.050) frequency compared with African American  (2.3%), Uganda African  (3.9%), and Tanzania  (6.0%) subjects. Interestingly, the frequency observed for the Mulatto subjects indicates that the Amerindian ancestry may be present in this multi-ethnic group, besides the presence of known African ancestry.
For the SLCO1B1 rs4363657 polymorphism, C allele frequency found in Caucasian descent in our study (15.4%) was similar (p > 0.050) to that reported in English subjects (13.0%) who participated as controls in a genome-wide association study (GWAS) . The non-coding rs4363657 polymorphism was in nearly complete linkage disequilibrium (LD = 97) with the non-synonymous rs4149056 polymorphism, in the population sample in which the cited GWAS was performed . In our population sample plus Amerindians, it is possible to observe that the SLCO1B1 rs4149056 and rs4363657 are in different linkage disequilibrium patterns depending on the ethnic origin. The African descent (LD = 89) and Mulatto (LD = 81) ethnic groups present higher degree of linkage disequilibrium. This highlights two main points: first, even with the different linkage disequilibrium among ethnic groups, the haplotype analysis are not need because only the rs4149056 polymorphism have been described as functional [11, 24]. Second, the different frequencies among ethnic groups for the polymorphisms used as markers in the GWAS platforms are limitations/peculiarities of the GWAS methodology , which may be acting as interferents in replication studies, mainly in countries with admixed population.
The association between SLCO1B1 rs4149056 with myopathy (odds ratio: 4.5 - 95% CI 2.6 to 7.7, per copy of the C allele; and odds ratio: 16.9 - 95% CI 4.7 to 61.1, among CC homozygotes as compared with TT genotypes) was observed in 85 subjects with myopathy taking 80 mg of simvastatin daily as part of a trial involving 12,000 participants; plus replication in a trial of 40 mg of simvastatin daily involving 20,000 participants . The authors of this GWAS plus replication study concluded that common variants in the SLCO1B1 gene are strongly associated with an increased risk of statin-induced myopathy and genotyping may help to obtain benefits of statin therapy more safely and effectively .
Other studies were performed to replicate this association. Brunham et al observed that rs4149056 genotype was significantly associated (odds ratio 2.3 per C allele) with myopathy in patients receiving simvastatin (mean ~30 mg daily), but not in patients who received atorvastatin . In this way, genotype CC individuals compared with wild-type TT individuals, the area under the plasma concentration time curve for simvastatin acid was increased 221% and for atorvastatin was increased 145% [12, 13]. Voora et al confirmed, in 99 subjects with statin-related adverse events, the association previously identified, but also expanded this finding to the most commom statin-induced side effects (e.g., myalgia or muscle ache without significant creatine kinase elevations) . Puccetti et al reported the same association, but in a group of 46 subjects affected by familial hypercholesterolemia treated with atorvastatin and developing myopathy .
There are potential limitations in our study. First, the self-referred ethnicity could be creating wrong separation. However, this form of classification is the one most probable to be encountered in real-life situations in the clinical practice. In addition, even self-referred ethnicity was able to clearly differentiate groups of individuals with rather different allele and genotype frequencies. Second, in order to implement a genotyping program in this context, some points could still be evaluated, for example: other genetic markers, primary disease, use of concomitant medications, types of statins, age and gender .