Publications

Detailed Information

국민건강보험공단 표본데이터를 이용한 대장암 예후와 관련된 처방 약물 체계적 평가 : Systematic assessment of pharmaceutical prescriptions association with Colorectal Cancer prognosis in a nationwide population-based cohort study in South Korea

Cited 0 time in Web of Science Cited 0 time in Scopus
Authors

우형택

Advisor
신애선
Issue Date
2021-02
Publisher
서울대학교 대학원
Keywords
대장암처방 약물다중검정랜드마크 분석재현성선택적 보고Colorectal cancerprescription drugsmultiple testslandmark analysisreproducibilityselective reporting
Description
학위논문 (석사) -- 서울대학교 대학원 : 의과대학 의학과, 2021. 2. 신애선.
Abstract
Objectives
Previous epidemiologic studies examining the prognosis of colorectal cancer and treatment drugs for noncancerous chronic diseases commonly prescribed in colorectal cancer patients were hypothesis-driven studies. Since the relationship with the prognosis of colorectal cancer is studied by limiting to one or several drugs, biased results may result from selective reporting. Therefore, in this study, to prevent the problem of selective reporting, a Medication wide association study(MWAS) was conducted to systematically examine the relationship between all drugs prescribed in colorectal cancer patients and the prognosis of colorectal cancer.
Methods
Using the National Health Insurance Service Sample Cohort DB, followed from 2002 to 2013, all drugs prescribed in patients diagnosed with colorectal cancer were analyzed for drugs related to the patient's prognosis. A total of 2,178 subjects were divided into 1:1 through random sampling and used as a drug discovery set and a drug validation set. The increase in false positives according to the multiple test was controlled using the false detection rate, and the immortal time bias that occurred as drug use changed over time was controlled by increasing the landmark time point to 1, 3 years. Additional analysis was conducted to see if there was a difference in reproducibility of internal validation between the unselected drug and the selected drug.
Results
In the drug discovery set, a total of 773 drugs were analyzed, and the number of drugs found to be related to the prognosis of colorectal cancer according to the landmark time point was 117 at 1 year and 29 at 3 years. Regardless of the time point, 22 drugs showed significant relevance. When 124 drugs that were significant at least once in the drug discovery set were verified in the drug verification set, the significant drugs at 1 and 3 years were 95 and 61 respectively. A total of 17 drugs, including H2-Receptor Antagonists(ATC code: A02BA), Other Drugs For Peptic Ulcer And Gastro-Oesophageal Reflux Disease (GORD)(ATC code: A02BX), Mosapride(ATC code: A03FA09), Prednicarbate(ATC code: D07AC18), Methylprednisolone(ATC code: H02AB04), Fluconazole(ATC code: J02AC01), Eperisone(ATC code: M03BX09), Enzymes (Used In Musculo-Skeletal System Inflammatory Conditions)(ATC code: M09AB), Mometasone(ATC code: R01AD09), Pseudoephedrine(ATC code: R01BA02), Acetylcysteine(ATC code: R05CB01), Erdosteine(ATC code: R05CB15), Levodropropizine(ATC code: R05DB27), Chlorphenamine(ATC code: R06AB04), Levocetirizine(ATC code: R06AE09), Other Antihistamines For Systemic Use(ATC code: R06AX) were significant in both the drug discovery set and drug validation set. In further analysis of reproducibility through the internal validity evaluation conducted in the drug validation set, the overall reproducibility of drugs found through the MWAS methodology in the drug discovery set was higher.
Conclusions
In this study a total of 17 signals related to the prognosis of colorectal cancer were found and
it was confirmed that the MWAS methodology is not only useful for finding new hypotheses, but also can be used to increase the reproducibility of the study.
배경
대장암 환자에서 흔히 처방되는 당뇨, 고지질혈증 등 비암성(non-cancer) 만성질환의 치료 약물과 대장암 환자의 예후를 살펴본 연구들은 특정 가설 기반하에 한 가지 또는 몇 가지의 약물과 대장암 예후와의 관계를 본 가설주도(hypothesis-driven) 연구로 선택적 보고(selective reporting)가 발생할 경우 편향된 결과가 나올 수 있다. 이에 본 연구에서는 선택적 보고의 문제를 원천적으로 막기 위하여 대장암 환자에서 처방되는 모든 약물과 대장암 예후와의 관계를 체계적으로 살펴본 Medication wide association study (MWAS)를 수행하였다.

방법
2002년에서 2013년까지 추적관찰된 국민건강보험공단 표본코호트DB를 이용하여 대장암으로 진단된 환자에서 진단 이후 처방된 약물과 환자의 예후와의 관련성을 분석하였다. 총 2,178명의 대상자를 무작위 추출을 통해 1:1로 나눠 약물발견셋(drug discovery set) 및 약물검증셋(drug validation set)으로 이용하였다. 다중검정에 따른 위양성 증가는 위발견율(false discovery rate)을 이용하여 통제하였고 약물 사용이 시간에 따라 변함으로 인해 발생하는 불멸시간편향(immortal time bias)은 랜드마크 시점을 1,3년으로 늘려가며 분석을 진행하여 통제하였다. 약물발견셋에서 p값은 0.05 미만이었지만 선정되지 않은 약물과 약물발견셋에서 선정된 약물 간의 약물검증셋에서의 내적타당도 평가(internal validation)에서 재현성의 차이가 있는지 추가 분석을 진행하였다.

결과
약물발견셋에서 총 773개의 약물에 대해 분석을 진행하였고 대장암 예후와 관련이 있는 것으로 나타난 약물은 랜드마크 시점 1년에서 117개, 3년에서 29개였다. 랜드마크 시점과 관계없이 모두 관련성을 보인 약물은 22개였다. 약물발견셋에서 랜드마크 시점과 관계없이 한번이라도 유의했던 124개의 약물을 약물검증셋에서 검증하였을 때 랜드마크 시점 1년에서 유의한 약물은 95개, 3년에서 유의한 약물은 61개였고 약물발견셋과 약물검증셋의 모든 랜드마크 시점에서 유의했던 약물은 H2-Receptor Antagonists(ATC code: A02BA), Other Drugs For Peptic Ulcer And Gastro-Oesophageal Reflux Disease (GORD)(ATC code: A02BX), Mosapride(ATC code: A03FA09), Prednicarbate(ATC code: D07AC18), Methylprednisolone(ATC code: H02AB04), Fluconazole(ATC code: J02AC01), Eperisone(ATC code: M03BX09), Enzymes (Used In Musculo-Skeletal System Inflammatory Conditions)(ATC code: M09AB), Mometasone(ATC code: R01AD09), Pseudoephedrine(ATC code: R01BA02), Acetylcysteine(ATC code: R05CB01), Erdosteine(ATC code: R05CB15), Levodropropizine(ATC code: R05DB27), Chlorphenamine(ATC code: R06AB04), Levocetirizine(ATC code: R06AE09), Other Antihistamines For Systemic Use(ATC code: R06AX) 등 총 17개였다. 약물검증셋에서 진행한 내적타당도 평가를 통한 재현성 차이에 대한 추가 검증에서 약물발견셋에서 MWAS 방법론을 통해 발견된 약물이 그렇지 않은 약물에 비해 약물검증셋에서 전반적으로 더 높은 재현성을 보여주었다.

결론
본 연구를 통해 총 17개의 대장암 환자 예후와 관계된 실마리정보(signal)를 찾아냈으며 MWAS 방법론이 새로운 가설을 찾아내는데 유용할 뿐만 아니라 연구의 재현성을 높이는데 사용될 수 있음을 확인하였다.
Language
kor
URI
https://hdl.handle.net/10371/175686

https://dcollection.snu.ac.kr/common/orgView/000000164838
Files in This Item:
Appears in Collections:

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share