Physiological alcohol dependence as a "specifier" of risk for medical problems and relapse liability in DSM-IV.

OBJECTIVE: This study tested the ability of DSM-IV physiological alcohol dependence to predict multiple indices of medical problems and relapse behavior. It also tested the ability of three additional variables--DSM-IV nonphysiological dependence, an alternative dichotomous criterion for coding physiological dependence and a dimensional measure of physiological dependence--to predict medical problems and relapse behavior in alcoholism.
METHOD: A heterogeneous group of 365 patients was recruited from eight addictions treatment programs in the northeastern United States. A multidimensional assessment battery able to diagnose the presence of physiological dependence according to each of three systems--the criteria of DSM-IV, alternative dichotomous criteria and a dimensional scale-- was administered about 2 weeks after admission, and 241 subjects were reinterviewed 6 months later. The three systems were compared for their ability to predict a variety of external measures of medical complications and relapse liability.
RESULTS: Physiological alcohol dependence as diagnosed by DSM-IV bore no relationship to either risk for medical problems or relapse behavior. Further analyses showed that this failure was due to operational problems of physiological dependence in DSM-IV, rather than to a lack of conceptual merit for physiological dependence per se as a course specifier. Use of alternative criteria for coding physiological dependence which are difficult and less internally consistent, and use of a dimensional measure, found improved relationships with the external validators.
CONCLUSIONS: Contrary to early reports, physiological dependence can serve as a course specifier for alcohol problems, but must be more sensitively scaled than it was in DSM-IV. Tests of alternative options suggest that a multistage criterion to replace DSM-IV's dichotomous criterion is the best remedy.