Clinical Study Of Single Nucleotide Polymorphisms (SNPs) Of CYP2D6 Versus Personalized Analgesia

Background:Optimal postoperative analgesia is necessary both for reliefing patients'suffering and rehabilitation. Currently, it is concerned that patient controlled analgesia may be a good choice for pain relief postoperatively. But it fails to include genetic factors and poor analgesia often torments both doctors and patients. Drug-metabolizing enzyme CYP2D6 is one of the factors affecting analgesia during PCA intervention, especially with tramadol analgesia. Some SNPs of CYP2D6 play a key role in activity of its product. However, current studies were focused mainly on some exon polymorphisms of CYP2D6. Researches on other genes suggested that intron polymorphisms may affect the function of the products. We hypothesise that certain SNPs of introns of CYP2D6 may also play a role in analgesic efficacy besides those of the exons. This study was designed to detect SNPs and haplotypes of CYP2D6 in Chinese Han and investigate effects of SNPs/haplotypies on postoperative analgesia under certain environmental factors.Methods:A randomized sample survey was conducted from November 2009 to August 2010, and 245 patients,18-60 year, ASAâ… -â…¡, were recruited. The main inclusion criteria included elective upper abdominal procedures and Chinese Han.48 patients withdrew because of failing to collect blood, staying in ICU unexpectedly, mechanical errors of a PCA pump, and severe side effects including dizzy, nausea and vomiting. Pain scores and other possible effects were evaluated at the time of the surgery completing(Ohr),2hr,4hr,6hr,12hr, 24hr and 48hr postoperatively in sequence.Data analysis was performed using SPSS 15.0. The primary outcomes were pain scores and tramadol consumption and the secondary outcomes included sedation state, nausea and vomiting, itching and satisfaction.Results:The patients were classified into 2 groups using hierarchical clustering with tramadol consumption at the planned time points(n1=133 and n2=34, respectively). Age, surgical duration and smoking had significant differences between the groups(P<0.005), while sex, occupation, surgical type, blood loss, drinking(tea, coffee or alcohol), and history of surgery/diabetes mellitus(DM) all had no statistical differences(P>0.1). Additionally,15 sites of SNPs in the gene were detected with a minimal genotype frequency of 4%. LD analysis indicated that there was a haplotype block in intron 1 (SNP2, SNP5, SNP6 and SNP7). The frequency of haplotypes GTGA and CCCC was 79.3% and 17.7% respectively, and the two haplotypes had significant difference between the two groups(P<0.0216).On the other hand, there was a significant difference between the two groups when the variables, including age, surgical duration, smoking, SNP1, SNP2 and the interactions between SNP1 and the two environmental factors(age and smoking), were analyzed using binary logistic regression(P<0.022). In the established model, the Ward statistic of surgical duration had significant contribution to the formula(P=0.047). And Hosmer and Lemeshow test implied that the model was good(P=0.154). Press's Q test also suggested that there was a significant difference between the predictve clusters using the model and the clusters by randomized assignment(P<0.05), suggesting the model was effective for predicting. Compared with a hit ratio of 79.6% by experience, the one predicted by the model was about 91.6%.Furthermore,86.7% of cases in the test sample were classified correctly using the model, and Press's Q test showed that the model had acceptable discriminant power(P<0.05).Conclusions:Age, surgical duration, smoking, SNP1 and SNP2(SNP5, SNP6 and SNP7) are associated significantly with analgesic efficacy of postoperative PCA with tramadol. And the three main interactions, that is, SNP1 and smoking, SNP1 and age, SNP1 and SNP2, also show statistical association with pain scores. Though the established logistic regression model has good discriminant power, it should be cautiously explained with the variables for their small contributions to the model.In addition, other genetic factors, including polymorphisms of receptors and/or targets and drug transportors may also be involved in the development of postoperative pain. And the variables involved should be concerned systamatically. Obviously, further work is encouraging and challenging.