Wernicke's Encephalopathy Is Reversible Within Hours. Twenty-Six Studies Show Care Still Arrives Late.

The drug is a vitamin, off-patent, stocked in every hospital pharmacy in the country, and in high intravenous doses it can pull a patient back from a neurological emergency that otherwise ends in death or permanent amnesia. The obstacle is not price. According to a scoping review published Oct. 20 in Addiction Science & Clinical Practice, the obstacle is that the emergency is described four different ways, treated at doses that differ by a factor of three hundred, and frequently dosed as a tablet for patients who needed a drip.

Eight researchers led by André Akira Sueno Goldani at the Center for Alcohol and Drug Research at Hospital de Clínicas de Porto Alegre, working with colleagues at the Federal University of São Paulo, pooled 26 studies covering 44,240 adults with alcohol use disorder or clinically significant drinking. What they mapped is a care pathway with no standard: diagnostic definitions that vary between studies, prescribing audits that repeatedly catch low-dose oral thiamine in high-risk patients, and exactly two randomized trials of dosing, both of which they rated at high risk of bias.

The review's strongest conclusion is deliberately modest. It cannot name the right regimen. It can say where care reliably diverges, and it can say that the divergence is not a knowledge gap at the level of basic biology. Thiamine deficiency in chronic heavy drinking is well understood: reduced intestinal absorption, depleted hepatic stores, impaired conversion to the active pyrophosphate form, and simply not enough food. During withdrawal, a hyperadrenergic state and rising cerebral glucose use push demand higher still.

The Dose Range Spans Three Hundredfold

Across the included studies, reported daily thiamine doses ran from 5 mg to 1,500 mg, given orally, intramuscularly or intravenously. Three loose clusters emerged. Regimens at or below 50 mg a day appeared mainly as prophylaxis or maintenance. Doses of 100 to 200 mg a day turned up in preventive, equivocal and poorly specified contexts. High-dose intravenous regimens of 500 mg a day or more were the ones associated with suspected or established encephalopathy.

Frequency varied with the same looseness. Once-daily dosing was common across the literature; three-times-daily schedules showed up more often where the diagnosis was suspected or confirmed. A short intravenous course of one to three days followed by oral maintenance was a frequent pattern, but the review found that the boundary lines between prevention, treatment and maintenance were drawn inconsistently enough that the three categories often could not be compared.

That ambiguity has a cost measured in neurons. Thiamine is a cofactor for enzymes linking glycolysis to the tricarboxylic acid cycle and the pentose phosphate pathway, so acute depletion impairs oxidative ATP production and increases oxidative stress, leaving metabolically vulnerable regions — the thalamus and mammillary bodies among them — exposed to excitotoxic injury.

Two Trials, Both Rated High Risk of Bias

The evidence base for the actual question — what dose, by what route, for how long — is thinner than the volume of published recommendations suggests. Only two randomized controlled trials evaluated thiamine regimens, and the reviewers rated both at high risk of bias because of problems with outcome measurement, attrition or reporting.

One compared different intramuscular doses in patients with alcohol use disorder and reported modest improvements in working memory at the higher dose. The other compared three intravenous regimens in symptomatic and at-risk patients and found no clear dose-response difference in cognitive or neurological outcomes. Everything else in the review was observational, quasi-experimental, audit or implementation reporting: useful for describing routine practice, vulnerable to confounding, shifting case definitions and inconsistent outcome measures.

Set that against the guidance clinicians already have. The 2010 European Federation of Neurological Societies guideline recommends 200 mg of thiamine three times daily, preferably intravenously, given before any carbohydrate, a recommendation the authors graded at Level C — the weakest evidence tier the guideline uses for a treatment recommendation. Even the reference standard is a consensus built on thin trials.

Where the Banana Bag Fits

American hospital practice adds a wrinkle the review's authors do not need to name for the finding to land. The standard intravenous "banana bag" given to patients admitted with alcohol-related problems typically contains 100 mg of thiamine, a dose that functions as prophylaxis rather than treatment and that has been challenged in the clinical literature for years.

Read the audits next to that number and the pattern is not mysterious. A patient at high risk of encephalopathy who receives a banana bag receives a fifth of a single EFNS-recommended dose, once, in a bag that also contains folate, magnesium and multivitamins — a formulation that feels like comprehensive care and behaves, for this specific purpose, like a gesture. The review's auditors found low-dose oral thiamine given in acute or high-risk presentations often enough to call it a recurring feature rather than an outlier.

The Triad That Shows Up in a Minority of Cases

Part of why treatment arrives late is that the textbook picture is rare. Wernicke encephalopathy is classically taught as a triad: encephalopathy, gait ataxia and ocular motor dysfunction. Older neuropathological series found all three in only 16 to 20 percent of cases, and more recent meta-analytic evidence placed the complete triad in roughly one third of clinically identified cases.

The Caine criteria were built to compensate, requiring two of four features — dietary deficiency, oculomotor abnormalities, cerebellar dysfunction, and altered mental status or mild memory impairment. The review found that fewer than half of the included studies applied formal diagnostic definitions for either alcohol use disorder or Wernicke encephalopathy. Many relied on clinical impression or risk-based inference, or on alcohol-related complications and blood alcohol concentrations; relatively few used structured DSM or ICD criteria, or validated screens such as AUDIT-C or CAGE.

Imaging and labs did not resolve the ambiguity. Magnetic resonance imaging can support the diagnosis, but the reviewers did not treat it as a prerequisite for starting treatment, and biochemical markers of thiamine status were reported inconsistently and generally served as supportive rather than stand-alone criteria. Recognition is hardest exactly where these patients land: emergency departments, general medical wards, psychiatric units and medically assisted withdrawal services, where intoxication, withdrawal, delirium, malnutrition, trauma and infection all compete to explain the same confusion.

Twelve Percent at Autopsy

The consequence of missing it is not subtle. Postmortem studies have reported Wernicke-related lesions in as many as 12 percent of people with alcohol use disorder, including cases never diagnosed during life — a figure that points to underrecognition rather than to a dispute about treatment.

Untreated, the condition can progress to Korsakoff syndrome, with severe anterograde amnesia and the long-term care needs that follow. The review also flags a biological complication that cuts against simple screening logic: biochemical thiamine deficiency can occur without clinical encephalopathy, which means a low lab value does not equal disease — and, more dangerously for clinical decision-making, the absence of the classical triad does not exclude it. Susceptibility appears to vary with nutritional status, comorbid illness, genetic differences in thiamine transport and other vitamin deficiencies, so no clinician can predict which deficient patient will progress.

Protocols Change Prescribing. Nobody Has Shown They Change Brains.

The review's most actionable finding is also its least surprising. Studies describing more protocolized, guideline-concordant care more often reported earlier treatment initiation and greater use of parenteral thiamine in suspected or confirmed cases. Sites with electronic order sets, structured withdrawal protocols and explicit prescribing guidance tracked closer to recommended doses and routes.

The authors are careful about what that means. These are associations drawn from retrospective designs, with confounding and local differences in which formulations a pharmacy actually stocks, and the review explicitly declines to claim that any protocol prevented neurological harm. What it does claim is narrower and still useful: delayed recognition and inconsistent implementation are modifiable features of care, not fixed properties of the disease.

What This Means for Treatment

For families and for the detox and residential programs that admit heavy drinkers, the review turns a vague worry into a set of specific questions. What is the thiamine protocol here, and is it written down? What dose and route does a patient get if staff suspect encephalopathy, versus if they are simply at risk? Is the first dose given before or after a glucose-containing infusion? Is there a trigger that escalates from oral to parenteral, or does escalation depend on someone thinking of it at 3 a.m.?

Those questions matter most in alcohol use disorder treatment settings where withdrawal management happens on a schedule and the patient's nutritional history is often incomplete. The clinical literature has treated thiamine as a routine checkbox for decades. The audit data in this review suggest that in a meaningful share of high-risk encounters, it has functioned as exactly that.

What the Scoping Review Can't Tell You

The design sets hard limits. No included study enrolled an exclusively outpatient population. Long-term cognitive recovery and recurrence were rarely assessed. Every included study came from a high-income country, nine of them from the United States, and women, older adults and patients in low- and middle-income countries were underrepresented or absent. Because the goal was mapping rather than pooling, the reviewers performed no formal certainty assessment and offer no pooled treatment effect.

So the review cannot state an optimal dose, cannot establish a dose-response relationship, and cannot prove that any protocol reduces neurological harm. Its conclusion is descriptive, and it is uncomfortable: within a narrow and methodologically weak randomized base, diagnostic ambiguity, delayed treatment and inconsistent prescribing persist across every care setting where these patients are seen.

Where the Evidence Goes Next

The authors' prescription for the field is a sequencing argument. First separate four things the literature currently blurs — biochemical thiamine deficiency, suspected encephalopathy, confirmed encephalopathy and Korsakoff syndrome — then test not just doses but timing, route, duration and implementation in pragmatic trials that include outpatient settings.

That is a research agenda, not a fix for the patient in the emergency department tonight. The fix available now is cheaper and less elegant: a written protocol, a low threshold for the intravenous route in anyone high-risk, and a first dose that arrives before the carbohydrate does. The vitamin costs almost nothing. The window, the review suggests, is what gets spent.