Contribution of non-neurologic disturbances in acute physiology to the prediction of intensive care outcome after head injury or non-traumatic intracranial haemorrhage.

OBJECTIVE: To study the additional contribution of non-neurologic disturbances in acute physiology and chronic health to the prediction of intensive care outcome in patients with head injury or non-traumatic intracranial haemorrhage. DESIGN: A nationwide study in Finland with prospectively collected data on all adult patients admitted to intensive care after head trauma or non-traumatic intracranial haemorrhage during a 14-month period. Two-thirds of the patients were randomly selected to derive predictive models, and the remaining one third constituted the validation sample. SETTING: A total of 25 medical and surgical ICUs in Finland (13 in tertiary referral centers). PATIENTS: 901 consecutive adult patients with head injury or non-traumatic intracranial haemorrhage. MEASUREMENTS AND RESULTS: Variables of the APACHE II including Glasgow Coma Score were collected at the time of ICU admission. Two predictive models were created to explain hospital mortality. The addition of variables describing acute physiology to a predictive model consisting of Glasgow Coma Score, age, diagnosis of head injury and the type of ICU admission did not increase its performance in discriminating between survivors and nonsurvivors, but the calibration accuracy of the predictive model especially at the high ranges of risk was improved. CONCLUSIONS: The non-neurologic disturbances in acute physiology have prognostic significance in the prediction of intensive care outcome in patients with head injury or non-traumatic intracerebral haemorrhage. The created predictive model may supplement clinical judgement of this patient group.

One-year outcome in early aneurysm surgery: prediction of outcome.

Predictors of one-year outcome were studied in patients treated for ruptured intracranial aneurysm. A total of 929 patients, who were treated conservatively or surgically, and 839 patients as a part of this population who were surgically treated, were randomly divided into two groups in order to create predictive models by logistic regression and to validate them. The models were derived from two-thirds of these two patient groups and the remaining one-thirds were used for validation. The pre-operative variables of both conservatively and surgically treated patients were studied by Model A. The pre-operative Grade (Hunt and Hess), age, and the presence of vasospasm on angiography were the three most important predictors of the one-year outcome (Glasgow Outcome Scale 1-2/3-5). Model B consisted of pre- and per-operative, and Model C pre-, per-, and post-operative variables collected from the surgically treated group. The pre-operative Grade, ligation of a major vessel and age were the three most powerful determinants of outcome in Model B. In Model C high Grade, post-operative CT-hypodensities and ligation of a major vessel were most closely associated with poor outcome. Model A, based on pre-operative data, most accurately predicted good outcome. All the 59 patients in the validation sample (n = 310) who were predicted to have a less than 5% probability of poor outcome had a favourable outcome (positive and negative predictive value 100%). Model C determined that 22 of 280 patients would have a more than 80% probability of poor outcome.(ABSTRACT TRUNCATED AT 250 WORDS)

Quantitative metabolite patterns of human brain tumors: detection by 1H NMR spectroscopy in vivo and in vitro.

OBJECTIVE: The aim of the present study was to investigate quantitative metabolite patterns in human brain tumors by 1H nuclear MR spectroscopy (1H MRS). MATERIALS AND METHODS: Single voxel 1H MRS was used in studying metabolites in 23 primary brain tumors in vivo. The T2 relaxation times and saturation factors were determined for N-acetylaspartate (NAA), total creatine (Cr), choline-containing compounds (Cho), and water, which was used as an internal standard in computations of metabolite concentrations in vivo. Metabolites in biopsy specimens from 75 tumors were quantified by means of 1H MRS in vitro. RESULTS: The NAA concentrations were lower in brain tumors than in normal tissue in vitro and in vivo, irrespective of the histological type. The NAA was some threefold higher in vivo than in vitro, which could be due to partial volume effect and/or contributions from other metabolites to the peak at 2.02 ppm in vivo. Ratios of Cho to Cr concentrations were elevated in tumors relative to normal brain in vivo. Absolute Cho was some 50% higher in both benign astrocytomas and oligodendrogliomas in vivo than in normal brain. However, Cho concentration in vitro was significantly higher only in pituitary adenomas than in healthy white matter. Total creatine was approximately 50% lower in malignant astrocytomas and meningiomas than in normal brain. CONCLUSION: It is concluded that absolute concentrations of metabolites in vivo yield a different picture of tumor metabolites than that derived from uncorrected metabolite ratios.