Phenobarbital vs. Benzodiazepine-Based Alcohol Withdrawal Protocols: A Refractory Case
A patient in alcohol withdrawal has already crossed the threshold most definitions use for benzodiazepine-refractory disease. The question is whether to keep giving more of the same drug or add one that works by a different mechanism.
Four hours and more than 40 mg of lorazepam-equivalent dosing after arrival, D.H. is not improving — and that fact, more than how he got here, is what the team is actually contending with right now. He was found shaking against a bus shelter badly enough that a passerby called 911 rather than assuming it was the cold; he is forty-seven, has no fixed address currently, and has been staying between a shelter and a cousin's couch since losing his apartment eight months ago, when the construction job he'd held for over a decade ended in a company downsizing. He told EMS he'd been drinking roughly a fifth of liquor daily for “years,” with his last drink sometime the previous morning, though he isn't entirely certain of the timeline himself — he thinks he ran out of money for more sometime after that. This is his third emergency department presentation for withdrawal in the past year, and the first this severe; the two prior visits resolved on the hospital's standard symptom-triggered lorazepam protocol without needing anything further, the same protocol he was started on this time. It didn't hold. Over the following four hours his required dosing crossed the threshold most clinicians use to define benzodiazepine-refractory withdrawal, without his tremor, diaphoresis, or agitation meaningfully improving. He has no other known chronic illness anyone can confirm from prior records and no documented psychiatric history — this is, as far as the chart shows, an otherwise uncomplicated withdrawal that simply isn't answering the standard approach. He remains tachycardic at 128, hypertensive at 168/102, and is now visibly more agitated and less redirectable than he was less than an hour ago, pulling at his IV line and no longer tracking questions the way he did at triage.
Escalate the same drug again, or change mechanism
I'd add a single dose of IV phenobarbital now rather than continue escalating lorazepam alone. Rosenson's randomized, placebo-controlled trial found that a single 10 mg/kg phenobarbital dose added to a standard lorazepam protocol cut ICU admissions from 25% to 8%, with no increase in adverse events. Benzodiazepines increase how often the GABA-A channel opens; barbiturates increase how long it stays open once it does — that's a real mechanistic reason the combination can succeed where more of the same drug alone is stalling.
I take the mechanism seriously, but I want to name that his airway is still protected right now, and benzodiazepine monotherapy remains the guideline-standard approach for a reason — it's the most familiar tool to every nurse and physician on this unit tonight. Four hours of escalation without response is concerning, but it isn't yet the same as proven treatment failure.
And I'd push back on the channel-kinetics argument specifically: open-frequency versus open-duration is a real mechanistic difference, but it's mechanism, not outcome. Stacking a barbiturate on top of forty milligrams of lorazepam-equivalent is also the combination that ends with an unplanned intubation, and his airway is the one thing I'm not willing to trade for a faster response.
Both of you are working from real data, but I want to be precise about how strong that data actually is. Rosenson's result is a genuine randomized trial, but it's one moderately sized study; Hendey and colleagues ran a separate ED trial comparing phenobarbital directly to lorazepam and found no significant difference in length of stay, admission rate, or symptom score at 48 hours. Phenobarbital also has a narrow therapeutic index and a long half-life, so this isn't a risk-free substitution.
But "the most familiar tool" isn't the same as "still working." He is already past the 40 mg-in-four-hours threshold most refractory-withdrawal definitions use, and his trend is worsening, not plateauing — an intubation risk you're weighing against a benefit you're assuming continued escalation will eventually deliver.
Given where he is right now — already refractory, still protecting his airway, trending worse — I'd support a single weight-based phenobarbital dose now, in a monitored setting with someone at bedside watching for oversedation, rather than either open-ended benzodiazepine escalation or a full phenobarbital-only pathway.
Agreed: a single weight-based IV phenobarbital dose given now under continuous monitoring, lorazepam continued and reassessed rather than further escalated, with a low threshold to involve the ICU team if his mental status or respiratory status changes.
Not agreed: whether tonight's decision should shift the unit's default protocol toward earlier phenobarbital use in future refractory cases, or whether this remains a case-by-case judgment call. The critical care physician sees a real, replicable pattern worth protocolizing; the clinical pharmacologist points to Hendey's null result as a reason to keep it individualized rather than standard. The emergency medicine physician, having watched tonight's response, said only that he'd want to see it work here before deciding what it means for the next patient.