Multiple elements modulate drug use, including sleep, which is increasingly being considered as an important contributor to substance use and abuse. The present study aimed to evaluate the association between sleep, psychiatric and socioeconomic/demographic factors and substance use in a large-scale representative sample from the city of São Paulo, Brazil. Data from the 2007 São Paulo Epidemiological Sleep Study (EPISONO) database were used. In the EPISONO study, volunteers underwent a polysomnographic exam and completed a series of questionnaires to assess objective and subjective sleep quality and associated comorbidities. Drug use was assessed using the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST). Linear (univariate and multivariate) and logistic regressions were performed to identify factors associated with the use of the 4 most commonly used substances in the sample (tobacco, alcohol, cannabis and cocaine/crack). Structural equation models were used to establish theoretical networks to explain the relationship between sleep, psychiatric and socioeconomic factors and use of these substances. The logistic regression results showed that psychiatric symptoms, lower income, and poorer subjective sleep were the main factors associated with tobacco consumption; gender and occupational status with alcohol intake; age and occupation with cannabis use; and education with cocaine/crack use. The structural equation models partially supported these findings and identified significant effects of psychiatric symptoms on tobacco consumption, both directly and mediated by sleep. Our results reinforce previous findings concerning factors associated with generally misused substances and suggest that sleep should be considered as an important element in future substance use disorder studies.

多种因素调节药物使用，包括睡眠，这越来越被认为是药物使用和滥用的重要因素。本研究旨在评估来自巴西圣保罗市的大规模代表性样本中睡眠、精神和社会经济/人口因素与物质使用之间的关联。使用了来自 2007 年圣保罗流行病学睡眠研究 (EPISONO) 数据库的数据。在 EPISONO 研究中，志愿者接受了多导睡眠图检查并完成了一系列问卷调查，以评估客观和主观的睡眠质量和相关的合并症。使用酒精、吸烟和物质参与筛查测试 (ASSIST) 评估药物使用情况。进行线性（单变量和多变量）和逻辑回归以确定与样本中最常用的 4 种物质（烟草、酒精、大麻和可卡因/快克）的使用相关的因素。结构方程模型被用来建立理论网络来解释睡眠、精神和社会经济因素之间的关系以及这些物质的使用。Logistic回归结果显示，精神症状、收入较低、主观睡眠较差是与烟草消费相关的主要因素；饮酒的性别和职业状况；使用大麻的年龄和职业；和可卡因/快克使用教育。结构方程模型部分支持了这些发现，并确定了精神症状对烟草消费的显着影响，直接和睡眠介导。我们的研究结果强化了先前关于与普遍滥用物质相关因素的研究结果，并建议在未来的物质使用障碍研究中，睡眠应被视为一个重要因素。