A comparison of diagnostic criteria for neuroleptic malignant syndrome.

BACKGROUND

A variety of diagnostic criteria for neuroleptic malignant syndrome (NMS) have been used in clinical studies of this disorder, but it is not known if different criteria consistently identify NMS. This study examines agreement between three frequently used sets of diagnostic criteria in a series of possible NMS episodes.

METHOD

All clinically suspected NMS episodes occurring at a large tertiary psychiatric facility during a 6-year period were evaluated by three different sets of diagnostic criteria. Agreement among these criteria was quantified statistically by means of the kappa and intraclass correlation coefficients.

RESULTS

The NMS diagnostic criteria examined generally demonstrated only fair agreement with one another in the diagnosis of NMS. Agreement was best among these criteria when the "probable" category was employed. A complex interaction involving both definition and structure of individual diagnostic criteria and designation of criteria as major or minor appears to contribute to these findings.

CONCLUSION

The published diagnostic criteria used in this study do not consistently identify NMS episodes and demonstrate different thresholds for assigning this diagnosis. These differences are not due solely to different definitions of individual criteria (e.g., fever). Possible implications of these findings for clinical practice and research are discussed.