2015. 5. 7-8. 대한당뇨병학회 Cardiovascular risk models for diabetic patients in Korea The Korean Heart Study 지선하, KHS 연구팀 연세대학교보건대학원
Background 2008 년전세계 mortality 의 30% 인 1,730 만명이심뇌혈관질환 (ASCVD) 으로사망 (WHO) 2030 년까지 2,330 만명으로증가예측 (WHO) 흡연, 고혈압, 당뇨병, 고지혈증개선으로 ASCVD 발생을 80% 이상예방가능 (AHA) 단일위험요인보다는여러가지위험요인을통합한예측모형을통해고위험군을사전에예측하고예방, 관리하는것이중요함 (D Agostino RB 등. JAMA, 2001)
Risk Prediction Model 1976년 Kannel 등에의해심혈관질환예측을위한로지스틱모형발표 ( 연령, 흡연, 혈압, 당뇨병, 총콜레스테롤, HDL콜레스테롤, 좌심실비대등 ) 2001년 D'Agostino RB 등에의해 CHD 예측모형발표 (JAMA 2001;286(2):180-187) 2008년 D'Agostino RB 등에의해 primary care 를위한심뇌혈관예측모형발표 (Circulation 2008;117:743-753) 2013년 11월 ACC/AHA에서 Combined risk 모형발표 각나라에서자국민의유용성평가가절실히요구됨. 각질환군 ( 예, 당뇨병 ) 에서 CVD 예측모형의필요성있는지?
Framingham risk score 성별에따라연령, 흡연, 수축기혈압, 총 콜레스테롤, HDL- 콜레스테롤을점수화 하여 10 년간심장병발생위험예측 장점 심장병발생고위험대상발견 치료에대한동기부여 실제임상에적용간편 제한점 유럽, 중국등에서실제위험도보다과대평가 (Hense et al. EHJ 2003, Liu et at. JAMA 2004)
Framingham CHD risk 유럽인에서과대평가 Predicted By Framingham Predicted By Framingham Observed Observed Predicted By Framingham Predicted By Framingham Observed Observed
Framingham CHD risk 중국인에서과대평가 5.3 배과대추정 4.7 배과대추정
ACC/AHA new guideline ARIC (Atherosclerosis Risk in Communities) study, Cardiovascular Health Study, the CARDIA (Coronary Artery Risk Development in Young Adults) study, Framingham Original and Offspring Study cohorts. ACC/AHA 모형변수소개 Ln Age, Ln Age squared, Ln Total cholesterol, Ln Total cholesterol*ln Age, Ln HDL, Ln HDL*Ln Age, Log Treated SBP, Log Treated SBP*Ln Age, Log Untreated SBP, Log Untreated SBP*Ln Age, Current Smoker, Current Smoker*Log age, Diabetes American College of Cardiology/American Heart Association
10y ASCVD risk prediction
Conclusion 일반인구집단에서개발한 ASCVD, CHD 모형에대한평가 : FRS, SCORE 등모형은 ASCVD 를과도추계 한국인에서도비슷한결과를보임
In 2001, the UKPDS risk engine for CHD Duration of diabetes, glycaemic control In US, ADA and NCEP-ATPIII Diabet. Med. 22, 554 562 (2005)
Comparison of the proportions of actual and predicted CHD events using the Framingham model and the UKPDS risk engine 70 70 60 60 50 50 40 40 30 20 10 Actual Predicted 30 20 10 Actual Predicted 0 0 Framingham model UKPDS risk engine
ROC: 0.704 aroc: 0.737
The Korean Heart Study (KHS) KHS 의공식자문학회 : 대한심장학회
Study population 연구대상 한국인심장연구 (Korean Heart Study) 전국 18 개종합검진센터에서 1996-2004 년에검진을받은 430,920 명으로구축된코호트연구 연령 30-74 세 세브란스병원 IRB 승인번호 (4-2007-0065) 최종연구대상 10 년추적기간을확보하고자 1996-2001 년에검진을받은대상으로국한
Age distribution 70-74 65-69 60-64 55-59 50-54 45-49 여자 남자 40-44 35-39 30-34 0 5 10 15 20 25 % The Korean Heart Study (KHS)
Level of Education in Men 45 40 35 30 25 20 15 10 5 0 Mean education year KNHANES: 11.8 y KHS: 13.7 y Difference: 1.9 y KNHANES KHS KNHANES: Korean National Health And Nutrition Examination Survey, 2001 KHS: Korean Heart Study, 1996-2004
Level of Education in Women 45 40 35 30 25 20 15 10 5 0 Mean education year KNHANES: 9.7 y KHS: 11.7 y Difference: 2 y KNHANES KHS KNHANES: Korean National Health And Nutrition Examination Survey, 2001 KHS: Korean Heart Study, 1996-2004
Outcome (ASCVD) definition (Circulation, 2013) Methods Myocardial Infarction (ICD-10, I21-22) IHD death (ICD-10, I20-25) Stroke (ICD-10, I60-69) Follow up: 1997.1-2012. 12 국민건강보험공단과공동연구를통해 F/U Outcome validation First study: 1997-2007 event: accuracy 73% (Kimm 등, 2012) Second study: 2008-2011 event: accuracy 93% (Kimm 등, 2013)
The Korean Society of Cardiology 추계학술대회 Validity of the diagnosis of Acute Myocardial Infarction in Korean National Medical Health Insurance Claims Data 2008-2011 and Its Trend in 10 Years: The Korean Heart Study[2] Heejin Kimm, MD 1, Ha-yeoun Kim, MPH 1, Sungha Park, MD 2, Yangsoo Jang, MD 2, Sun Ha Jee, PhD 1 1 Institute for Health Promotion & Department of Epidemiology and Health Promotion, Graduate School of Public Health, Yonsei University, Seoul, Republic of Korea 2 Cardiology Division, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea Background and Objectives: The validity of Medical insurance claims (MIC) data was analyzed in the Korean Heart Study from 1994 to 2007. We aimed to evaluate the change in the accuracy and validity of ICD codes of acute myocardial infarction (AMI) from Korean National MIC data of 2008-2011 in the same cohort with its trend of the whole 18 years. Subjects and Methods: From 2004 to 2009, baseline information of the cohort was collected in 7 health examination centers of Korea. AMI incidence outcome was from hospital admission discharge records from 2008 to 2011. Among the 164 participants detected to have the ICD code of AMI incidence (I21-) in their hospital admission discharge records, data of 90 cases (81 men, 36-83 years of age) was obtained through structured questionnaires. The ICD code accuracy, trend of accuracy, and reliability (kappa) for AMI were analyzed. Results: Hospitals in Seoul area reported 51(56.7%) cases. The accuracy of AMI was 93.3% according to World Health Organization (WHO) criteria and 90.0% according to European Society of Cardiology/American College of Cardiology (ESC/ACC) criteria. Concordance rate between the two criteria was 96.4%(Kappa 0.78). The annual accuracy of both criteria has been improved from 1999(n=7, 42.9%, 57.1%) to 2011 (n=30, 96.7%, 90.0%). Conclusion: In the diagnosis of AMI by ICD-10 codes in Korean MIC data, accuracy was improved gradually and is within 85-100% since 2007. It coincides between two criteria well and the reliability was fair to good The Korean Heart Study (KHS)
10 year CHD risk Hard CHD events, comprising acute myocardial infarction, sudden death, and other coronary deaths The Korean Heart Study (KHS)
10 year CHD risk (P) F(x.M) = ß 1 (x 1 -M 1 )+ +ß k (x k -M k ) P = 1 S (t) exp (f[x,m]) x 1 x k represent an individual risk factors M 1 M k are the mean values of the risk factors in the cohort study S(t) is the survival rate at the mean values of the risk factors at 10 years The Korean Heart Study (KHS)
A coronary heart disease prediction model: the Korean Heart Study The Korean Heart Study (KHS)
The Korean Heart Study (KHS)
Objectives 한국인의당뇨병인구의심장병예측모형을개발하고, 평가한다.
A coronary heart disease prediction model: the Korean Heart Study Diabetic patients : 20,048 (7.4%) The Korean Heart Study (KHS)
General characteristics of diabetic patients Mean (SD) Age, y 52.0 (9.5) Total cholesterol, mg/dl 204.5 (39.4) HDL cholesterol, mg/dl 48.1 (11.6) Systolic blood pressure 129.7 (19.9) Fasting serum glucose, mg/dl 146.2 (53.7) % Sex (male) 68.6 Diabetes 100.0 Current smoking 35.0
ASCVD survival curve among Korean men S(t) 1 0.98 0.96 0.94 0.92 0.9 0.88 0.86 General population Diabetic patients 0.84 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Follow-up year
ASCVD survival curve among Korean women S(t) 1 0.98 0.96 0.94 0.92 0.9 0.88 0.86 General population Diabetic patients 0.84 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Follow-up year
55 M, TC 213, SBP 120, HDL 50, nonsmoker, DM 일반인구모형 (8% 저추계 ) 당뇨병환자모형
55 F, TC 213, SBP 120, HDL 50, nonsmoker, DM 일반인구모형 (13.6% 저추계 ) 당뇨병환자모형
Ten-year Prediction of ASCVD events using KHS G equations among Diabetic Patients in Men Prediction of ASCVD Event 0.24 0.21 0.18 0.15 0.12 0.09 0.06 0.03 0 Predicted Actual AUROC: 0.710 (0.710-0.723) 1 2 3 4 5 6 7 8 9 10 Decile of Predicted Risk Based on KHS Functions
Ten-year Prediction of ASCVD events using KHS G equations among Diabetic Patients in Women Prediction of ASCVD Event 0.24 0.21 0.18 0.15 0.12 0.09 0.06 0.03 0 Predicted Actual AUROC: 0.705 (0.685-0.726) 1 2 3 4 5 6 7 8 9 10 Decile of Predicted Risk Based on KHS Functions
Ten-year Prediction of ASCVD events using KHS D equations among Diabetic Patients in Men Prediction of ASCVD Event 0.24 0.21 0.18 0.15 0.12 0.09 0.06 0.03 0 Predicted Actual AUROC: 0.717 (0.703-0.732) 1 2 3 4 5 6 7 8 9 10 Decile of Predicted Risk Based on KHS Functions
Ten-year Prediction of ASCVD events using KHS D equations among Diabetic Patients in Women Prediction of ASCVD Event 0.24 0.21 0.18 0.15 0.12 0.09 0.06 0.03 0 Predicted Actual AUROC: 0.714 (0.693-0.734) 1 2 3 4 5 6 7 8 9 10 Decile of Predicted Risk Based on KHS Functions
Conclusion FRS, SCORE 등모형은 UKPDS 모형에비해당뇨병환자에서 ASCVD 를저추계한다. 한국인심장연구 (Korean Heart Study) 에서개발한 ASCVD 10 년위험예측모형은남녀모두에서한국인심뇌혈관질환을잘예측하였음. 앞으로좀더심층연구를통해검증작업이필요함.
Acknowledgement 대한심장학회자문위원 박성욱, 승기배, 오동주, 오병희, 이상훈교수 ( 가나다순 ) 연구자료기여기관의연구자 고상백, 김문찬, 김응수, 김현숙, 김홍규, 김성희, 박샛별, 박종규, 박주성, 성지동, 오장균, 유병연, 유철인, 이규장, 이덕철, 이상현, 이순영, 이용진, 이태용, 조비룡교수 ( 가나다순 ) 심장병 event validation 위원 김병욱, 김영대, 김철호, 박종선, 박창규, 성인환, 안영근, 안태훈, 양주영, 윤정한, 이상학, 주승제, 조명찬, 채제건, 탁승제, 한기훈, 홍택종교수 ( 가나다순 ) 자료관리및국가 DB 연계 국민건강보험공단보험정책연구원윤영덕연구원
Thank You for Your attention