Heterogeneity And Meta Analysis Of The Essential Hypertension Risk Assessments

Objective:Exploring the application of single and multiple Meta regression analysis, to find the heterogeneity sources and their affects to the results of essential hypertension risk assessment studies . At the same time, Through the sub-group and mixed model Meta analysis, merge into various risk factors odds ratio (OR). The Purpose is to have lessons learned and identify problems.finally ,more reliable systematic analysis data was achieved and we can Provide references for the prevention and control of hypertension.Methods: Search domestic and foreign hypertension risk factors literatures from 2000.1 to 2009.1 and Screen according to standard procedures, to achieve required literature.After making quality assessment to the literatures, we extracted the required data,verified and Recorded. Using stata10 statistics software made the heterogeneity test to all of the literatures.Then we made single and multiple Meta regression analysis from"study time, male-female ratio, average age, sample size, case-control ratio, longitude, latitude"to selecte Heterogeneity sources for"BMI, gender, family history, salt intake, smoking, drinking"risk factors with domestic literatures. At last ,Meta-analysis and subgroup analysis, forest mapping, and publication bias tests were made according to heterogeneity sources. At the same time, Meta-analysis,forest mapping and publication bias test were made for the factos that foreign literatures have sufficient analysis number .Results:"Sample size"was the suspicious factor that caused the heterogeneity of"BMI"OR in the hypertension risk factors assessments (P = 0.055);"male-female ratio"caused the heterogeneity of"gender,smoking"OR in the literatures(P=0.024,P=0.000);The heterogeneity of"family history,salt intake"OR were caused by"study time"(P=0.000,P=0.000);"latitude"was the factor that caused the heterogeneity of"BMI,salt intake"OR(P=0.001,P=0.006);At last,we found that"average age"caused the heterogeneity of"BMI,drinking"OR in the hypertension risk factors assessments(P=0.001,P=0.000).The results of sub-group Meta analysis were:The sub-group Meta analysis of"BMI" OR by"latitude"showed that the greater latitude related to the greater OR(OR= 1.21,1.79,2.33);Analysis by"average age"showed that The older sub-group got the greater OR;Analysis by"Sample size"acheived the result was larger sample size study got the moderate OR usually. The sub-group Meta analysis of"gender"OR by"male-female ratio"showed that study with more men than women was the larger OR obtained ,Conversely, the smaller(OR=1.46,OR=1.17). Analysis by"average age"showed that <45 years sub-group got the greater OR(OR=1.51),and≥45 years sub-group got the smaller OR(OR=1.10). The sub-group Meta analysis of"family history"OR by"study time"showed that the more recent study got the greater the odds ratio (OR = 1.27,1.74,2.50). Analysis of"salt intake"OR by"latitude"showed that the south sub-group obtained (OR = 1.31) greater OR than the northern (OR = 1.07); Analysis by"study time"showed that the more recent study got the greater the odds ratio(OR=1.31,1.08). Analysis of"smoking"OR by"male-female ratio"showed that the subgroup of Male- female ratio is less than1 obtained the smaller OR (OR = 1.09) than the subgroup of Male- female ratio is more than1 (OR = 2.05). Analysis of"drinking"OR by"average age"showed that <45 years sub-group got the smaller OR(OR=1.03),and≥45 years sub-group got the greater OR(OR=1.38)".Physical exercise"combined odds ratio and 95% CI was 0.93 (0.80-1.08). "Noise" combined odds ratio and 95% CI was 1.27 (1.15-1.41) that was a risk factor.Conclusion:The main heterogeneity of risk factors OR in these essential hypertension risk assessment studies were:"Sample size,study time,male-female ratio,average age, latitude ";This study showed "BMI,gender,family history,salt intake,noise" is very determined risk factors of essential hypertension.But "Drinking, smoking, physical exercise"were uncertain factors。We need explore the complex interference factors and specify measurement standards to obtained the results. There are population differences in"BMI, salt intake"risk factors."Family history, salt intake"OR would change with years. Interference factors that related to pre-analysis factors were tent to causer heterogeneity.