Alcohol and anesthesia: intoxication, chronic use and withdrawal

August 28, 2026 · Read 12 min

“Social.” That answer closes most alcohol histories taken at the preoperative visit, and what went unasked reappears in the operating room or after surgery. One in five surgical patients meets criteria for alcohol use disorder. Management turns on two things: the usual daily amount and the time of the last drink.

1. Focused preoperative assessment

The useful history at the clinic visit is short: how much is drunk on a usual day, when the last drink was, and whether the patient has been through withdrawal before. Quantity stratifies surgical risk. The interval since the last drink predicts when the first signs of withdrawal are likely to appear after surgery. A previous episode indicates prophylaxis from admission.

Above 60 g of ethanol a day — about four standard drinks — postoperative complications are two to four times more frequent. The risk is concentrated in the preceding weeks: a remote history does not carry the same probability as drinking that is still active at the time of surgery.

The history underestimates actual intake: examination and laboratory studies stratify risk independently of the answer given. How the question is put also changes the answer, and stigmatizing terms — “alcoholic”, “addict” — are associated with worse outcomes.

The surgical setting shifts the prior probability: in emergency trauma and in tumours of the head, neck and upper gastrointestinal tract, half of patients have unhealthy alcohol intake and a quarter are dependent.

Examination looks for stigmata of chronic liver disease such as ascites and asterixis, since these point to decompensated cirrhosis and change the plan. Blood pressure behaves differently by stage: excessive intake raises it, and advanced cirrhosis lowers it through a hyperdynamic circulation. The isolated preoperative reading does not predict the hemodynamic response to induction.

An isolated abnormality of liver function tests has little predictive value and is not enough to reclassify risk. Platelet count and fibrinogen define the transfusion plan: in major surgery with decompensated disease, the targets are a platelet count of 50,000/µL and a fibrinogen of 100 mg/dL.

Electrolytes are checked for hyponatremia, which appears even without cirrhosis; for hypokalemia and hypophosphatemia, which produce muscle weakness; and for hypomagnesemia, which aggravates hypokalemia and lowers the seizure threshold.

2. What deferral achieves

Reduction over the preceding 4 weeks is the only interval with a randomised trial in its favour; 2 weeks of abstinence improve platelet function and thereby lower bleeding risk, and 8 weeks or more improve wound healing.

The benefit has a documented limit: preoperative interventions reduce intake, without any demonstrated reduction in complications, length of stay or mortality. Deferral does allow withdrawal to be managed as a planned event.

Chronic unhealthy intake is not a contraindication to ambulatory surgery, and same-day discharge avoids the withdrawal that precipitates admission. The indication is decided case by case, on dilated cardiomyopathy, cirrhosis, the state of coagulation and home circumstances. Discharge is into the care of a responsible adult, with follow-up contact details given in writing.

3. Anesthetic requirement by clinical state

Chronic intake induces cross-tolerance with hypnotics and volatile agents: the compensated patient needs a higher dose for the same depth of anesthesia. Acute intoxication shifts requirement the other way, since it adds its own central nervous system depression. Hepatic or cardiac dysfunction reduces requirement by a different route: cardiorespiratory vulnerability and altered pharmacokinetics.

Acute intoxication and the delayed gastric emptying of autonomic neuropathy both raise the risk of pulmonary aspiration. The intoxicated patient receiving general anesthesia for emergency surgery undergoes rapid sequence induction. In the stable chronic drinker, consensus favours intubation over a supraglottic airway device, and reserves rapid sequence induction for symptomatic reflux. Extubation is planned on the same reasoning.

Hemodynamic instability is usually multifactorial: autonomic neuropathy, dilated cardiomyopathy, pulmonary hypertension, hypovolemia and vasoplegia coincide in the same patient without any one of them accounting for the picture. Preparation includes large-bore venous access before induction, a low threshold for an arterial line, and ensuring a vasopressor is available; with liver disease or malnutrition, hourly blood glucose. Marked preoperative anxiety may represent early withdrawal, and is managed with midazolam, which also reduces induction requirement.

Regional anesthesia offers one specific advantage in this population: an awake patient allows a seizure to be recognized. Neither peripheral blocks nor neuraxial techniques have evidence of worsening alcoholic neuropathy; documentation of the pre-existing deficit precedes the block. The limits are coagulopathy, thrombocytopenia, and cooperation limited by intoxication or neurological impairment; hypoproteinemia is considered separately, since it alters local anesthetic distribution and with it the maximum doses.

4. Analgesia with hepatic dysfunction

Chronic intake blunts the effect of opioids, and pain control usually needs higher and more frequent doses; there are no data showing greater susceptibility to respiratory depression in this group. Prolonged exposure is associated with hyperalgesia, which persists in sustained abstinence and becomes more marked during the syndrome. Liver disease works against that tolerance: it reduces first-pass metabolism and doubles the half-life of opioids with active metabolites. Titration with short-acting agents compensates for both effects.

Uncontrolled pain is a relapse factor in the dependent patient, so restricting opioids pre-emptively carries its own risk. A multimodal approach spares opioids without losing analgesia: local infiltration by the surgeon, analgesics from several classes, and regional techniques where not contraindicated. Of the infusion adjuncts, magnesium also corrects a frequent deficit, and ketamine has evidence of its own in withdrawal.

Lidocaine infusion is reserved for the patient without established liver disease. With hepatic damage the plasma concentration rises by two routes: the half-life is prolonged up to 3.5-fold, and the fall in alpha-1-acid glycoprotein increases the free fraction in proportion to the severity of cirrhosis. It is the same hypoproteinemia that reduces maximum local anesthetic doses for blocks.

The drugs prescribed for dependence are handled differently before surgery. Disulfiram is stopped one to two weeks before the procedure and acamprosate is continued unchanged.

Naltrexone needs the most precise planning. It blocks opioid analgesia and abolishes postoperative rescue: the oral form is stopped at least 72 hours before elective surgery, and the last depot intramuscular dose 25 days beforehand, with the oral form bridging the interval. Where opioid-free anesthesia and analgesia are reliable, naltrexone is continued. Restarting is deferred 7 to 10 days from the last opioid dose. Stopping it in itself raises the risk of relapse, so the decision is taken with the prescriber.

5. Perioperative withdrawal

The syndrome sets in between 6 and 48 hours from the last drink and usually settles by the end of the first week; seizures cluster between 12 and 48 hours. Its incidence in surgical patients is two to five times that of patients admitted for other reasons. Half of heavy long-term drinkers develop withdrawal, mild in nine out of ten cases; the remainder progress to a generalized tonic-clonic seizure or to delirium tremens, which appears in about 5%.

Severe withdrawal is predicted by a high blood alcohol concentration on admission, previous episodes, loss of control over drinking, and a falling platelet count and potassium with raised transaminases and gamma-glutamyl transferase. In the patient who has all of these, prophylactic medication is started before symptoms appear. If the patient arrives with a positive blood alcohol and deferral would compromise the outcome, surgery proceeds under benzodiazepine prophylaxis.

Benzodiazepines are the treatment, prescribed as a symptom-triggered regimen and given by nursing staff: diazepam 20 mg every 90 minutes for as long as symptoms persist. A long half-life benzodiazepine is preferred; with cirrhosis or in the older patient, only short half-life agents, because of the risk of encephalopathy and cumulative sedation. Phenytoin is not effective against these seizures.

In the patient who has just had surgery, severity is easily overestimated, because fever, residual sedation and difficulty communicating reproduce those signs. Deterioration of a concurrent condition is investigated separately before it is attributed to withdrawal. Autonomic symptoms are managed with clonidine, and persistent hypertension or tachycardia with beta-blockers; both are adjuncts that do not prevent seizures or delirium.

Thiamine is given as prophylaxis against Wernicke’s encephalopathy — ophthalmoplegia, ataxia and delirium — at 100 mg IV or IM daily for 3 to 5 days; the order relative to intravenous dextrose does not matter and the two may be given at the same time.