Competing mortality between arsenic induced cancer and cardiovascular diseases in different arsenic exposure concentrations

• Main Authors: Chi-Hen Lin, 林志恒
• Other Authors: Chien-Jen Chen, Professor
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/10197033028565579907

碩士 === 國立台北護理學院 === 醫護管理研究所 === 92 === It was well established that ingestion of inorganic arsenic can cause some cancer and cardiovascular diseases; and the estimation of mortality in cohort studies is important for public health by using survival analysis methods. However, subjects who die due to the event of non-interest during the observation period are treated as censored observations in traditional survival analytic techniques such as the Kaplan-Meier method. This method is inappropriate as it assumes that subjects die from the event of interest are still possible beyond the time at which the death of non-interested event (censoring) occurred, or a person who dies of cardiovascular diseases can die of cancer later. Treating such subjects as censors inflates the estimate of cumulative mortality. The age to be dead of cardiovascular diseases and the age to be dead of arsenic induced cancer were likely to be correlated following in similar arsenic exposure concentrations. Hence, an assessment of the effect of arsenic exposure concentrations on the mortality of arsenic induced cancer cannot be isolated from its effect on the mortality of cardiovascular diseases. Various authors have advocated the use of cumulative mortality function for a specific event which takes into consideration the presence of dying from other event disease within a competing risks framework. This method permits an unbiased analysis of competing mortality. Hence, this study uses cumulative mortality function which considers the ingestion of inorganic arsenic caused cancer and cardiovascular diseases. In this study both right censor and left truncation were used. The purpose of this study are to estimate the cumulative mortality of arsenic induced cancer and cardiovascular diseases by adjusting competing risks and compared this result with unadjusted competing risks. Database sources were conducted from the Graduate Institute of Epidemiology, College of Public Health, National Taiwan University. The study cohort consisted of two separate subcohorts, a subcohort of total 2,497 residents of arsenic diseases endemic area on the southwest coast of Taiwan from 1985 to 1990 and a subcohort of total 8,081 residents of Northeast Lanyang area of Taiwan from 1990 to 1995. We futhur linked the data to the cancer registed file and cause of death data file to 2002 from the Department of Health, a total of 10,585 residents whose last follow-up age at least 40 years. The final sample was 8,604 residents by excluding subjects of unknown arsenic concentrations for estimating competing mortality of arsenic induced cancers and cardiovascular diseases. The results show that unadjusted age cumulative mortality of cancer and cardiovascular diseases were greater than those respectively adjusted by competing risks. Age cumulative cancer mortality of unadjusting or adjusting was increased with increased arsenic exposure concentrations. Mostly, the greater arsenic exposure concentrations were, the greater age cumulative mortality of lung cancer and skin/ kidney/ bladder cancer adjusted by their respectively competing risks. However, since the age cumulative mortality of cardiovascular diseases were competing with that of arsenic caused cancer, it was smaller than in other arsenic exposure concentrations. For female, the age cumulative mortality of cerebrovascular diseases and ischemic heart diseases did not show statistical difference among different arsenic exposure concentrations. The results of this study were consistent with those of previous studies, furthermore, it was shown that competing causes of deaths existed and more integrated precise estimation of age cumulative mortality was in this study. It also provides suggestions for health policy making.