Levin F R, Lehman A F
Department of Psychiatry, University of Maryland School of Medicine, Baltimore 21201.
J Clin Psychopharmacol. 1991 Dec;11(6):374-8.
Many pharmacologic regimens have been recommended as supplemental treatment strategies for cocaine abuse. Using the technique of meta-analysis, we compared and combined the data from six studies to examine the efficacy of desipramine as an adjunctive treatment approach for cocaine addiction. These studies were randomized, placebo-controlled clinical trials of desipramine and involved a total of 200 cocaine-addicted patients. We found that desipramine is no better than placebo in retaining patients in treatment. However, while patients are in treatment, desipramine is helpful in promoting abstinence (p less than 0.001). There were several compromises made to use meta-analysis for these studies. These included comparing patient populations with different psychiatric and substance use diagnoses, comparing studies that used varying amounts of psychotherapy, and comparing studies that defined certain outcome variables in diverse ways. There are other limitations in the use of meta-analysis that were not specific to the studies we compared. These include such issues as statistical bias and publication bias. The purpose of this paper is to suggest the use of meta-analysis as one way to help the clinician evaluate the literature and to point out the areas of difficulty when assessing the literature.
许多药物治疗方案已被推荐作为可卡因滥用的辅助治疗策略。运用荟萃分析技术，我们比较并整合了六项研究的数据，以检验去甲丙咪嗪作为可卡因成瘾辅助治疗方法的疗效。这些研究是去甲丙咪嗪的随机、安慰剂对照临床试验，共涉及200名可卡因成瘾患者。我们发现，在使患者持续接受治疗方面，去甲丙咪嗪并不比安慰剂效果更好。然而，在患者接受治疗期间，去甲丙咪嗪有助于促进戒断（p值小于0.001）。对这些研究使用荟萃分析存在一些妥协之处。这些包括比较具有不同精神疾病和物质使用诊断的患者群体、比较使用不同心理治疗量的研究，以及比较以不同方式定义某些结果变量的研究。在使用荟萃分析方面还存在其他并非我们所比较研究特有的局限性。这些包括统计偏差和发表偏倚等问题。本文的目的是建议将荟萃分析作为帮助临床医生评估文献的一种方法，并指出评估文献时的困难领域。
