Inappropriate use of symptom-triggered therapy for alcohol withdrawal in the general hospital

To determine if hospitalized medical and surgical patients were placed inappropriately on symptom-triggered therapy (STT) for alcohol withdrawal syndrome (AWS) and if certain conditions were more likely to be associated with inappropriate STT use or adverse events.

We randomly selected 124 (25%) of the 495 Mayo Clinic inpatients who received STT according to the Revised Clinical Institute for Withdrawal Assessment for Alcohol (CIWA-Ar) protocol in 2003 and assessed them for STT appropriateness, defined as having both intact verbal communication and recent alcohol use. Adverse events, including delirium tremens, seizures, or death, were correlated with CIWA-Ar appropriateness.

Of the 124 randomly selected patients, only 60 (48%) met both inclusion criteria. Of the remaining 64 patients, 9 (14%) were drinkers but could not communicate, and 35 (55%) could communicate but had not been drinking. Twenty (31%) met neither criterion. Univariate analysis identified a significant association between inappropriate initiation and chronic heart failure, postoperative status (POS), liver disease (LD), nonmetastatic cancer, and chemical dependency consultation. On multivariate analysis, only LD (P equals .02) and POS (P equals .01) retained significance, with LD more and POS less likely to predict appropriateness. Seven of 11 patients who experienced adverse events had received STT according to the CIWA-Ar protocol (P equals .05). Univariate analysis identified a significant association between adverse events and a history of alcohol dependence or AWS. Multivariate analysis showed significance only for a history of alcohol dependence (P equals .049).

Fewer than half of the randomly selected patients met both of the inclusion criteria for the CIWA-Ar instrument, leading us to conclude that more stringent evaluation is needed. Particularly postoperatively, alternative explanations for putative AWS should be sought. Health care professionals should more aggressively seek information on recent alcohol use from medical records, family members, and patients themselves.