The effect of alcohol upon admission serum insulin levels was studied in 114 injured patients with mean blood alcohol levels of 130 +/- 71 milligrams per deciliter, ranging from 10 to 350. These data were compared with those of 133 similar patients with no measurable blood alcohol levels and with those of 47 healthy volunteers. Whereas marked hyperglycemia caused by trauma or shock was unaffected by alcohol ingestion, insulin levels were more depressed in patients with blood alcohol levels 29.5 +/- 15.3 microunits per milliliter versus 37.8 +/- 36.7 microunits per milliliter, p less than 0.025. In fatally injured patients with measurable alcohol concentrations, insulin and glucose levels were markedly different from those for the corresponding surviving patients, p less than 0.001 and less than 0.01, respectively, with particularly large differences in the insulin-glucose ratio, p less than 0.001. The spontaneous reappearance of the insulin response to hyperglycemia was dependent upon the elapsed time and the severity of injury. Regardless of the blood alcohol concentrations, in patients admitted within two hours, there was no correlation between the levels of insulin and glucose. In patients admitted within two to eight hours, a significant correlation was found only in those with no detectable alcohol blood levels, p less than 0.001. In addition, a remarkable similarity of mean different admission times. These findings suggest that, in alcohol intoxication, recovery of insulin secretions suppressed by acute trauma is delayed; metabolism of alcohol in acute trauma is probably reduced.