Dennis Haack

Relationship between admission hyperglycemia and neurologic outcome of severely brain-injured patients.

Severe head injury is associated with a stress response that includes hyperglycemia, which has been shown to worsen outcome before or during cerebral ischemia. To better define the relationship between human head injury and hyperglycemia, glucose levels were followed in 59 consecutive brain-injured patients from hospital admission up to 18 days after injury. The patients who had the highest peak admission 24-hour serum glucose levels had the worse 18-day neurologic outcome (p = 0.01). Patients with peak 24-hour admission glucose levels greater than 200 mg/dL had a two-unit increase in Glasgow Coma Scale score while patients with admission peak 24-hour serum glucose levels less than or equal to 200 mg/dL had a four-unit increase in Glasgow Coma Scale score during the 18-day study period (p = 0.04). There was a significant relationship between 3-month and 1-year outcome and peak admission 24-hour serum glucose level (p = 0.02 and p = 0.02, respectively). Those patients with admission peak 24-hour serum glucose levels less than or equal to 200 mg/dL had a greater percentage of favorable outcome at 18 days, 3 months, and 1 year than those with admission peak 24-hour glucose levels greater than 200 mg/dL (p = 0.0007, p = 0.03, and p = 0.005, respectively). A significant relationship between admission peak 24-hour Glasgow Coma Scale score and 18-day, 3-month, and 1-year outcomes was found (p = 0.0001, p = 0.0002, and p = 0.0002, respectively). Patients with mean admission peak 24-hour Glasgow Coma Scale scores of 3.5, 6, and 10 had mean admission 24-hour peak serum glucose levels of 252 +/- 23.5, 219.1 +/- 19, and 185.8 +/- 21, respectively (p = 0.05). These relationships were not significantly altered when confounding variables such as the amount of glucose given over the initial 24-hour postinjury period, the presence of diabetes or multiple injuries, and whether patients were given steroids, dilantin, or insulin were statistically incorporated. These data suggest that admission hyperglycemia is a frequent component of the stress response to head injury, a significant indicator of severity of injury, and a significant predictor of outcome from head injury.

Failure of Prophylactically Administered Phenytoin to Prevent Post-Traumatic Seizures in Children

We report the results of a randomized, double-blind, placebo-controlled study to determine whether phenytoin administered soon after a head injury lessens the incidence of late post-traumatic epilepsy in children. 41 patients were randomized into either a phenytoin or placebo group and followed for 18 months. The patients were administered phenytoin or placebo intravenously or intramuscularly within 24 h of hospital admission. The patients were parenterally administered phenytoin or placebo until oral doses could be tolerated. There was no significant difference in the percentage of children having seizures in the treated and placebo groups (p = 0.25).

A decision theoretic approach to a distribution–free compound classification problem

A compound decision problem with component decision problem being the classification of a random sample as having come from one of the finite number of univariate populations is investigated. The Bayesian approach is discussed. A distribution–free decision rule is presented which has asymptotic risk equal to zero. The asymptotic efficiencies of these rules are discussed. The results of a compter simulation are presented which compares the Bayes rule to the distribution–free rule under the assumption of normality. It is found that the distribution–free rule can be recommended in situations where certain key lo cation parameters are not known precisely and/or when certain distributional assumptions are not satisfied.

The Effect of Hyperglycemia on the Progression of Vasogenic Edema Following Cold Injury

Increased glucose availability can enhance cerebral ischemic damage (Pulsinelli et al. 1979; Rehncrona et al. 1981) and edema development (Myers 1979; Pulsinelli et al. 1979; Warner et al. 1987). Hyperglycemia is associated with poor prognosis following head injury (Merguerian et al. 1981) and the level of consciousness in head-injured patients has been correlated with fasting blood glucose levels (Pentelenyi et al. 1979). Traumatized brain may be susceptible to alterations in glucose availability as is ischemic brain. This study examined whether or not regional vasogenic edema development was altered over 24 hours in cold-injured rats of different glycemic states.

Phenytoin-induced hypotension in animals receiving sympathomimetic pressor support

The observation that intravenous phenytoin infusion in patients receiving dopamine hydrochloride causes pronounced hypotension stimulated a series of animal studies to investigate the mechanism of phenytoin-dopamine interaction and to determine if phenytoin can be used safely with other pressors. Intravenous phenytoin was administered at 25 or 50 mg/min to normotensive dogs and dogs rendered hypotensive by hemorrhage, coronary artery ligation, or endotoxin. Phenytoin infusion was superimposed on infusions of dopamine, norepinephrine, phenylephrine, or isoproterenol. Phenytoin infusion in hypotensive dogs on pressor support caused a significant decrease in systemic vascular resistance (P < 0.0001) and a substantial decrease in mean arterial pressure. Dogs supported by phenylephrine had the smallest observed change in mean arterial pressure. The decrease in systemic vascular resistance appears to be secondary to phenytoin induced peripheral vasodilation. It is concluded that when phenytoin infusion is required in the critically ill patient, it should be administered very slowly and with continuous monitoring. When pressor support is required concomitantly with phenytoin infusion, a pure a agonist may minimize the risk of an adverse reaction.

An adaptive approach to compound multiple decisions

We will investigate a multiple decision problem (m.d.p.). In particular, situations in which a number of related m.d.p. ’s are to be analyzed simultaneously, called the compound multiple decision problem (c.m.d.p.), are considered. Within the Bayesian framework the usual assumption that the random parameter vector is a random sample from some distribution is modified. An adaptive analysis of the resulting c.m.d.p. is discussed. This discussion includes a simulation of a compound test of normal means where an adaptive (empirical Bayes) rule is analyzed as well as a simulation of a compound selection problem. An illustration considering yields of 7 corn varieties at each of 6 test sites is given.

Stringency in the emprical bayes problem

The empirical Bayes approach to statistical decision problems is preasentad where the unknown prior distribution is estimated up to an equivalence −k. This approximation of the prior is than used in the decision process where an asymptotically stringent solution with respact to − k is sought. The equivalence relations − k are such that k becomes larger − k approachas equality and our decision rule becomes asymptotically stringent with respect to the unknown prior distribution.