Clinical Institute Withdrawal Assessment for Alcohol–Revised might be an unreliable tool in the management of alcohol withdrawal

Discussion

An important limitation of the CIWA-Ar is its heavily subjective nature. Only 3 of 10 components (tremor, paroxysmal sweats, agitation) can be rated by observation alone. The other 7 components require at least some discussion with the patient. Given that benzodiazepines are provided based on the CIWA-Ar score, there is risk of incorrect dosing when scores are unreliable, which harbours potential for patient harm. There are 2 primary reasons why the CIWA-Ar was unreliable in this case. First, there was a substantial language barrier preventing the discussions necessary for accurate scoring. This became clearer as the patient’s withdrawal improved and he was still unable to answer simple questions in English. Even with an interpreter available, CIWA-Ar might remain impractical, as it requires frequent reassessments and would necessitate 24-hour interpreter coverage. The second limitation of the CIWA-Ar was subtler; the patient was confused and disoriented, so even in the absence of a communication barrier, his responses might have been unreliable. In a hospitalized population this might be a common scenario; acute medical issues can contribute to delirium and complicate the clinical picture.

While this case illustrates 2 reasons to use an OAWS, other common reasons exist. These might include patients with a clouded sensorium from acute psychosis or severe dementia, those with mechanical communication problems including severe facial trauma limiting speech and vision, and those with intubation.

Alternative assessment tools and clinical pathways have been proposed for inpatient management of alcohol withdrawal, but like the CIWA-Ar, they often require a reliable history.^9,10 The approach presented here has proved reliable for treatment of complex alcohol withdrawal by a busy AMCT in a tertiary Canadian hospital. The OAWS in Box 1 is not intended as an alternative to the CIWA-Ar and therefore has not been validated as such. Rather, it is an approach to treatment that can be useful when other validated tools cannot be reliably applied. It is based on objective findings and can be modified to fit the clinical situation. For example, in a patient with poorly controlled hypertension, blood pressure could be excluded or a higher blood pressure cutoff chosen. Similarly, heart rate might be excluded for a patient with uncontrolled atrial fibrillation or sepsis, and tremor excluded for a patient with essential tremor or parkinsonism. Additionally, the OAWS can be modified by changing the cutoff for scores prompting doses of benzodiazepines. In this case, the patient was unwell and required high doses, so we opted for a liberal scale to minimize underdosing. In more moderate withdrawal or where there was concern for benzodiazepine toxicity, cutoffs of 3 or more, or 4 or more, would be more benzodiazepine sparing.