건강보험심사평가원 청구자료를 이용한 약물과민반응의 역학 분석
Epidemiology of Drug Hypersensitivity Reactions Using National Health Insurance Claim Data
- 주제(키워드) Adverse drug reactions , Drug hypersensitivity , Health insurance database
- 발행기관 아주대학교
- 지도교수 이숙향
- 발행년도 2019
- 학위수여년월 2019. 2
- 학위명 박사
- 학과 및 전공 일반대학원 약학과
- 실제URI http://www.dcollection.net/handler/ajou/000000028388
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
Background and Objective Drug hypersensitivity reactions (DHRs) constitute a large portion of adverse drug reactions (ADRs), but studies for DHR incidence based on national data are scarce. This study aimed to estimate the incidence and patterns of DHR in a Korean population and the associated utilization of medical resources using the national claims data. Stevens-Johnson syndrome (SJS) and Toxic epidermal necrolysis (TEN) are representative one of the severe ADRs. To determine and dechallenge culprit drug is essential process for SJS and TEN patients for safety. Causative drugs have been proved mainly through case report or spontaneous report data in Korea. Analysis using nationwide database would be a new aspect finding out culprit drugs. Methods The retrospective cohort study performed using the national insurance claim database of the Health Insurance Review and Assessment (HIRA) in Korea. The International Classification of Disease 10th revision code was used to identify DHR including SJS and TEN. The claim data from January 2009 through June 2015 were analyzed. The incidence rates of DHR were calculated to 2009-2014. For the detailed analysis of SJS and TEN, the annual incidence and the 6.5-year incidence rates were calculated again in PART II analysis using newly defined real event dataset. Incidence rate coefficients were analyzed by sex, age, and year. The risk of ED or ICU visits with SJS and TEN were analyzed. The algorithm for assessment of drug causality in SJS and TEN (ALDEN) score tool was applied to analyze the causative drug of the SJS and TEN events. The relative incidence of two risk periods (risk 1 period, 1-28 day; risk 2 period, 29-56 day) and other baseline periods were compared for causative drugs which ALDEN evaluated ‘probable’ or ‘very probable’. Meta-analysis was performed overall IRR for study drugs. Results A total of 535,049 patients with 1,083,507 claims were assessed in the HIRA database for 6 years. DHR incidence was high in the elderly. DHRs related to blood were high in the young age. The number of SJS and TEN events that occurred during January 2009 to June 2015 were 1,297 (1,280 patients, SJS) and 300 (294 patients, TEN). All claims prescribed these patients were 13,223 (SJS) and 3,935 (TEN). IV steroid was most general treatment for SJS and TEN. Recurrent rate was 1.7%. Three patients experienced multiple events. Total person-year of SJS and TEN was 322,468,037 and 322,471,229. Incidence crude rates of SJS and TEN in the 6.5-year were calculated 4.0 million person-year (MPY) and 0.9 MPY, respectively. Age was risk factor as increasing incidence for SJS and TEN. Incidence rates coefficient by 5 year age band were 1.15 (1.13-1.18, p <0.0001, SJS) and 1.20 (1.16-1.24, p <0.0001, TEN). ICU admission risk was increased by age, TEN event and several comorbidities. Cirrhosis was the highest OR as risk factor of ICU admission (6.84, 2.60-18.01, p=0.0001). For analysis of determination culprit drugs in Korea, 8,943 cases / 1,280 patients (SJS) and 2,564 cases / 294 patients (TEN) were included for ALDEN assessment and self-controlled case series analysis, finally. The ‘probable’ causality by ALDEN evaluated to 117 cases (SJS 86, TEN 31). The number of culprit drug was 25 types. Sulfamethoxazole+trimethoprim (SMX+TMP) and allopurinol were most assessed as ‘probable’. Antibiotics (36%), antiepileptics (AEDs) (25%) occupied most of the culprit drugs. Highest incidence rate ratio (IRR) drug was azithromycin (1-28 day (risk 1): 318.33 [95% confidence interval 89.64-1130.48]). Ceftriaxone (1-28 day (risk 1): 99.18 [34.21-287.54], 29-56 day (risk 2): 49.59 [13.92-176.60]) and ciprofloxacin (1-28 day (risk 1): 88.64 [26.79-293.25], 29-56 day (risk 2): 44.32 [10.5-186.94]) was evaluated also high IRR drugs. Overall IRR of total culprit drugs were 36.68 (18.67-72.06, risk 1) and 12.78 (4.20-38.91, risk 2), respectively. Lamotrigine was highest relative risk as culprit drugs (10.00 [1.56-64.20], p=0.0001). As culprit drug latency on SJS and TEN, risk 1 period was more predominant than risk 2. (Odds ratio 4.25 [2.19-8.26], p <0.0001). Conclusion Incidence of DHRs in the real-world clinical practice was higher in the elderly and female. Incidence of SJS and TEN were also higher in the elderly, but not significant by sex. Clinical consequence was more severe in the elderly. SMX+TMP, allopurinol, lamotrigine, ceftriaxone, azithromycin and ciprofloxacin were major of culprit drug caused SJS and TEN in Korea. SMX+TMP and allopurinol was assessed high frequent as culprit dugs with SJS and TEN by ALDEN. Azithromycin, ceftriaxone and ciprofloxacin were shown high IRR compared other drugs. Even if the frequency was low, attention to the AED should continue. It is because the observed AED exposure events in this study are likely to be the actual causative agent, highly. Drugs 1-28 days before taking have a most significant effect on SJS and TEN. Efforts to prevent recurrence are needed through systematic supplement such as drug utilization review (DUR) system.
more초록/요약
연구배경 및 연구목적 약물과민반응은 대표적인 약물유해 반응의 종류이지만, 낮은 발생률로 인해 대규모 코호트 분석 연구가 부족하다. 스티븐스-존슨 증후군 (Stevens-Johnson Syndrome, 이하 SJS)과 독성 표피 괴사 융해증 (Toxic epidermal necrolysis, 이하TEN)은 심각한 약물과민반응의 한 종류이다. 한국에서의 원인약물 분석은 단일기관연구나 자발적 보고자료를 이용한 분석이 대부분이다. 원인약물을 밝혀내고 재투여 하지 않는 것은 SJS와 TEN 환자의 안전을 위하여 중요한 과정이다. 국민 대부분이 포함되어 있는 건강보험심사평가원 (이하 ‘심평원’) 자료를 이용하여 한국에서의 발생률과 발생양상, 원인약물에 대한 분석은 상기 약물유해반응의 현황을 파악하는 데에 중요한 근거자료가 될 것이다. 연구방법 2009년 1월부터 2015년 6월까지의 심평원 자료를 활용하여 해당 연구기간 동안의 발생률과 성별, 나이, 연도에 따른 발생률 계수를 산출하였다. 약물과민반응 종류 별 발생률의 산출은 2009년 1월부터 2014년 12월까지의 6년간 자료를 활용하였다. 발생률의 정확한 분석을 위하여 실제 유해반응 발생과 관련 된 명세서를 재 분류 한 후, SJS와 TEN의 발생률을 다시 산출하였다. 응급실과 집중치료실 방문의 위험도를 분석하였다. SJS와 TEN의 원인약물을 파악하는 도구인 ‘ALDEN’을 이용하여 한국에서 다빈도로 평가 된 원인약물의 종류와 빈도를 파악하였다. ALDEN으로 ‘상당히 관련 있음 (probable)’ 이상으로 평가 된 원인약물이 해당 유해반응에 영향을 미치는 위험기간을 1-28일, 29-56일 전의 두 구간으로 나누어 각각 구간의 기준구간에 대한 발생률 비를 분석하였다. 연구결과 본 연구에 활용 된 심평원 자료는 1,083,507 건의 명세서 (환자수 535,049명)를 포함하고 있었다. 약물과민반응의 발생률을 노령인구에서 높게 나타났다. 단, 혈액과 관련 된 약물과민반응의 경우는 어린 연령층에서 높게 나타났다. SJS와 TEN의 6.5년간 발생률은 각각 4.0 million person-year(MPY)와 0.9 MPY이었다. 5년 연령을 그룹화하여 발생률을 분석하였을 때 연령 그룹이 한 계단 상승할 때 위험도는 SJS에서 1.15 (95% 신뢰구간, 1.13-1.18, p<0.0001), TEN에서 1.20 (1.16-1.24, p<0.0001) 상승하는 것으로 분석되었다. 집중치료실 방문 정도는 연령과 병존질환 여부가 영향을 주는 것으로 분석되었다. 간경화증의 유무가 가장 높은 오즈비를 가짐을 확인하였다 (6.84, 2.60-18.01, p=0.0001). Sulfamethoxazole+trimethoprim (SMX+TMP)과 allopurinol은 다빈도 의심약물로 선정되었다. 항생제 (36%)와 항전간제 (25%)는 ALDEN 분석을 통하여 의심약물로 평가 된 약물들의 주요 효능군이었다. Azithromycin, ceftriaxone과 ciprofloxacin은 두 위험구간에서 높은 발생비를 보였다. 두 위험구간 중 1-28일 전 구간에 약물 노출이 많았다(오즈비 4.25 [2.19-8.26], p<0.0001). 결론 약물과민반응의 발생률은 고령인구와 여성에게 높게 나타났다. SJS와 TEN의 발생은 성별의 차이는 없었으나 고령인구에서 높게 나타났으며, 심각성 또한 해당인구에서 높은 것으로 분석되었다. SMX+TMP, allopurinol, lamotrigine, ceftriaxone, azithromycin 과 ciprofloxacin은 한국에서 발생 된 SJS와 TEN의 주요 원인약물로 선정되었다. 증상 발생 1-28일 전 복용한 약물이 유효한 영향을 미쳤다.
more목차
PART I. Epidemiology of Drug Hypersensitivity Reactions Using 6-year National Health Insurance Claim Data 1
I. Introduction 2
I-1. Adverse drug reactions (ADRs): definition and classification 2
I-2. Purpose of this study 3
II. Methods 4
II-1. Data source 4
II-2. Study design 5
II-3. Measure, outcome and statistical analysis 7
II-4. Ethics approval 8
III. Results 9
IV. Discussion 16
IV-1. Comparison with previous published research 16
IV-2. Strength and limitation of study 18
V. Conclusion 19
PART II. Severe DHR, Stevens-Johnson Syndrome (SJS) and Toxic Epidermal Necrolysis (TEN): Analysis of Epidemiology and Culprit Drugs 20
I. Introduction 21
I-1. Background of the study 21
I-2. Drugs are most common causes with SJS and TEN. 24
I-3. Review of previous studies 28
I-4. Purpose of this study 32
II. Methods 33
II-1. Data source 33
II-2. Study design 34
II-3. Measure and outcome 38
II-4. Statistical analysis 46
II-5. Ethics approval 47
III. Results 48
III-1. Epidemiology: The incidence generally increased with age. 50
III-2. High frequent culprit drugs of SJS and TEN in Korea: Antibiotics are noteworthy as causative agents. 69
III-3. Determination of high risk as culprit drugs: Allopurinol, ceftriaxone and azithromycin had high quality as culprit drugs. 76
IV. Discussion 94
IV-1. Epidemiologic features of SJS and TEN in Korea 94
IV-2. Age-related risk for SJS and TEN 97
IV-3. Drugs that should be noted as the cause of SJS and TEN in Korea 100
IV-4. The drug utilization review (DUR) system can be used to prevent SJS and TEN. 105
IV-5. Strength of study 107
IV-6. Limitation of study 108
IV-7. Further study direction 109
V. Conclusion 111
Reference 112
Abstract (in Korean) 130
