A systematic review of the Diagnostic and Statistical Manual diagnostic criteria for nicotine dependence

Joseph R. DiFranza, University of Massachusetts Medical School
W. W. Sanouri A. Ursprung, University of Massachusetts Medical School
Beatrice Lauzon, Université de Montréal
Christina Bancej, Public Health Agency of Canada
Robert J. Wellman, University of Massachusetts Medical School
Douglas M. Ziedonis, University of Massachusetts Medical School
Sun S. Kim, University of Massachusetts Medical School
Andre Gervais, Direction de santé publique de Montréal, Quebec
Bruce Meltzer, University of Massachusetts Medical School
Colleen E. McKay, University of Massachusetts Medical School
Jennifer O'Loughlin, University of Montreal
Chizimuzo T. C. Okoli, University of British Columbia
Lisa R. Fortuna, University of Massachusetts Medical School
Michele Tremblay, Médecin Conseil Institut National de Santé Publique du Québec

Abstract

The Diagnostic and Statistical Manual diagnostic criteria for nicotine dependence (DSM-ND) are based on the proposition that dependence is a syndrome that can be diagnosed only when a minimum of 3 of the 7 proscribed features are present. The DSM-ND criteria are an accepted research measure, but the validity of these criteria has not been subjected to a systematic evaluation. To systematically review evidence of validity and reliability for the DSM-ND criteria, a literature search was conducted of 16 national and international databases. Each article with original data was independently reviewed by two or more reviewers. In total, 380 potentially relevant articles were examined and 169 were reviewed in depth. The DSM-ND criteria have seen wide use in research settings, but sensitivity and specificity are well below the accepted standards for clinical applications. Predictive validity is generally poor. The 7 DSM-ND criteria are regarded as having face validity, but no data support a 3-symptom ND diagnostic threshold, or a 4-symptom withdrawal syndrome threshold. The DSM incorrectly states that daily smoking is a prerequisite for withdrawal symptoms. The DSM shows poor to modest concurrence with all other measures of nicotine dependence, smoking behaviors and biological measures of tobacco use. The data support the DSM-ND criteria as a valid measure of nicotine dependence severity for research applications. However, the data do not support the central premise of a 3-symptom diagnostic threshold, and no data establish that the DSM-ND criteria provide an accurate diagnosis of nicotine dependence.