中文摘要：

目的：在横断面调查的基础上，采用生存分析方法（survival analysis）探讨乌鲁木齐市社区吸毒人群发生首次吸毒的情况及其影响因素。方法：于2005年4-6月，以应答驱动抽样为主要招募方式，在社区招募吸毒者参加调查；应用结构式问卷调查吸毒者的社会人口学特征、首次口吸和静脉注射毒品时间、方式，首次吸毒前的个人成瘾性行为（如吸烟、饮酒）及家庭成员、亲友吸毒情况等。结果：401名静脉注射毒品者进入本次横断面调查，首次吸毒方式为口吸和静脉注射毒品者分别为95.0％（381／401）和5.0％（20／401），首次使用的毒品均为海洛因，首次吸毒年龄19．8a±s5．0a，首次吸毒年龄最小值为6．4a，15a以前开始吸毒者占14.2％（57／401）。从出生到首次吸毒的发生率为5．0／100人年（95％CI：45-5．5），影响首次吸毒发生的变量为：维族（HR=2．161，95％CI：1．713-2．726）、15a以前开始吸烟（HR=1．273，95％CI：1．006-1．610）、16a以前开始饮酒（HR=1．774，95％CI：1．387-2．269）、有亲戚吸毒（HR=1．411，95％CI：1．032-1．928）。从出生到发生首次吸毒的时间分别为：维族17．9（95％CI：17．2-18．5）a，15a以前开始吸烟者为20．0（95％CI：20．0-20．8）a，16a以前开始饮酒者为17．0（95％CI：16.6-17.7）a，有亲戚吸毒者为19．2（95％CI：18．6-20．0）a。结论：提示吸毒流行严重地区，吸毒低龄化可能会加速艾滋病的流行。在青少年早期开展个人成瘾性行为（如吸烟、饮酒）的干预教育及艾滋病知识的宣传教育非常必要，维族青少年应该是重点干预人群。

英文摘要：

Objective：To identify the associated factors with initial drug use among drug users recruited from community in Urumqi, Xinjiang. Methods：From April to June 2005, 401 drug users were recruited from community using respondent - driven sampling, all of them were interviewed by trained health workers. Each participant completed an informed consent and a structured questionnaire was used to collect data of demographic characteristics, the time of initial drug use （ including oral drug use or injection）, and addiction behavioral, such as cigarette smoking and alcohol drinking, having relationship with drug users such as family members, friends, or relatives before initial drug use. Results ： Of the 401 participants, oral drug use and injection was 95.0% （381/401） and 5.0% （20/401）, respectively at initial drug use and the drug was heroin. The age at initial drug use was 19. 8 a±s 5.0 a, minimum was 6.4 a, the drug users started drug use at the age less than 15 years old was 14. 2% （57/401）. The average incidence rate was 5.0/100（95% CI： 4.5-5.5）person -years from birth to initial drug use. In Cox proportional hazard regression model, Uigur （ HR = 2. 161, 95% CI： 1.713 - 2.726） , starting smoking at the age less than 15 （HR = 1. 273,95% CI：1.006 - 1.610）, starting drinking at the age less than 16 （ HR = 1. 774, 95% CI ： 1. 387 - 2. 269） and having drug user relatives （ HR = 1.411, 95% CI ： 1. 032 - 1. 928） before initial drug use was independently associated with time from birth to initial drug use, respectively. Age on initial drug addiction was 17.9 （95% CI： 17.2 - 18. 5 ） a for Uigur, 20.0（95%CI：20. 0- 20. 8）a for drug users smoking at the age less than 15, 17.0（95% CI： 16.6 - 17.7 ） a for drug users alcoholic drinking at the age less than 16 and 19. 2 （ 95 % CI： 18.6 - 20. 0） a for drug users having drug addition relatives. Conclusion：Becoming drug users much younger can speed up AIDS epidemic in drug abuse high prevalence areas. Interve