Study On Adverse Reactions Of Statins Based On Data Mining

Background:According to the latest epidemiological reports,in general,the prevalence and mortality of cardiovascular disease(CVD)were still in the rising stage,and were receiving more and more attention.Studies had shown that controlling dyslipidemia was the most effective measure to prevent CVD.Among the many lipid-lowering drugs,statins were the drugs most used in the prevention and treatment of CVD.Statins were hydroxymethylglutaryl coenzyme A reductase inhibitors,which competitively inhibit endogenous cholesterol synthesis rate-limiting enzymes,blocking intracellular hydroxyvalerate metabolism pathways,Thereby reducing the synthesis and secretion of triglyceride-rich apolipoprotein and lipoprotein.With the widespread use of statins in the clinic,their effectiveness had been fully demonstrated,but their safety has also received increasing attention.Since the clinical application of cerivastatin ceased in 2001,the safety of statins has attracted great attention from the medical community.In response to the adverse reaction(ADR)situation of statins,since 2012,various national food and drug administrations had issued revised announcements on statin instructions,requiring attention to the safety of statins.However,the current research on statin lipid-lowering drug ADR was still relatively simple,failing to reflect the relationship between the various factors involved from a multi-dimensional perspective,so it is still urgent to explore an effective ADR evaluation method.With the advent of the era of big data,data mining has been widely used in the medical field,but there have been few reports in the study of statins ADR.In this study,Modeler data mining was used to analyze the statins ADR,in order to revealing the occurrence of ADR and promoting clinical safe medication.Objective:1.Retrospective summary and analysis of statins ADR were conducted to explore the occurrence characteristics of statins ADR.2.Used data mining technology to reveal the occurrence of statin ADR and provide a scientific basis for safe medication.Methods:1.Retrospectively analyzed statin ADR characteristicsCollected a report of 2070 cases of ADR occurred in a university system hospital from January 2009 to December 2017 using statins.Used Excel software and manual screening methods to screen and clean the data,and finally obtain effective data of1490 cases of statin ADR,thenclassified and counted the effective data,and used SPSS16.0 for analysis of variance,chi-square test or rank sum test for statistical analysis.The test level wasα= 0.05,and p<0.05 was considered statistically significant.According to the patient’s general information(gender,age,disease basis,etc.),medication and ADR occurrence(drug name,time of occurrence,clinical manifestations,combined medication,etc.)and ADR effects(severity,outcome,impact on the original disease,etc.)were retrospectively analyzed.2.Association analysis of statin ADR based on data mining technologyBased on the effective data of 1490 statins ADR in method 1,SPSS Modeler 18.0was used to conduct network graph analysis,Bayesian network model analysis and association rule analysis on the correlation and importance between statins ADR and its related factors(such as gender,age,drug type,dosage,etc.).Results:1.Among the 1490 patients in this study,833 were male(55.91%)and 657 were female(44.09%),mostly concentrated in the 41-80 age group(84.10%),with an average age of 61.15 ±14.75 years.There was a significant difference in the sex ratio of different statins(P<0.05).The primary diseases of these patients mainly focus on hyperlipidemia,cerebrovascular disease,atherosclerosis or coronary atherosclerotic heart disease,hypertension,diabetes,and kidney disease.5.7% of these patients had a previous history of ADR,and 20.13% of patients had combined medication.The top three statins were atorvastatin(50.07%),simvastatin(23.76%),and rosuvastatin(14.16%).In terms of the time of ADR,24.23% of patients had ADR on the day after medication,61.14% of patients had ADR within 7 days after medication,and 79.60% of patients had ADR within 15 days after medication,and the average occurrence time was33.40±139.23 after treatment.There was statistically significant difference in the distribution of the occurrence time of ADR among various statins(P<0.05).The results showed that ADR of different statins occurred at different times..The first four ADR reported cases were abnormal liver function(38.11%),musculoskeletal damage(21.81%),skin symptoms(16.31%)and gastrointestinal reactions(11.88%).The organ/ system involvement of ADR caused by various statins was statistically significant(P<0.05).The severity of ADR in 77.58% of cases was evaluated as average.The ADR of 87.45% of patients had no obvious effect on the original disease.The outcome of ADR in 85.44% of patients was cured or improved,and the outcome of ADR of statins was different(P<0.05).2.Through Bayesian network model,we found that the most important influencing factors of ADR involved organs / systems were the occurrence time of ADR and drug varieties,while the most important influencing factors of ADR outcome were ADR involved organs / systems,drug varieties and ADR occurrence time.In the analysis of network diagrams and association rules,there were obvious differences in adverse reactions in drugs,gender,time of occurrence,and age.Middle-aged and elderly male patients taking atorvastatin and rosuvastatin were likely to have abnormal liver function within 15 days,middle-aged and elderly male patients take simva Statin and rosuvastatin were more likely to have musculoskeletal damage after one week,skin symptoms and gastrointestinal reactions were more common in elderly patients within 1 day of medication.Severe ADR mainly included two categories.Severe liver function abnormalities were more common in middle-aged and older male patients taking atorvastatin within 1 month.Severe musculoskeletal damage was common in elderly male patients taking simvastatin and rosuvastatin for more than 2 months.There were obvious gender differences in adverse reactions.Abnormal liver function,musculoskeletal damage and skin system damage were more common in male patients.Related factors for better ADR outcome were simvastatin,rosuvastatin,atorvastatin,gastrointestinal reactions,skin system,liver function abnormalities,musculoskeletal system and medication time within 1 week.Conclusions:Based on the data mining technology,we can find the correlation and rules among gender,age,drug type,ADR occurrence time,prognosis,ADR involving organs/systems and other relevant factors in the ADR of statins,which can provide theoretical basis for guiding the safe use of drugs in clinical and monitoring of ADR.