Primary traumatic brain damage may be compounded by secondary pathophysiological insults that can occur soon after trauma, during transfer to hospital or subsequent treatment of the head-injured patient. The aim of this prospective study was to quantify the burden of a wide range of secondary insults occurring after head injury and to relate these to 12-month outcome. In 124 adult head-injured patients studied during intensive care using a computerized data collection system, < or = 14 clinically indicated physiological variables were measured minute-by-minute. Verified values falling outside threshold limits for > or = 5 min, as defined by the Edinburgh University Secondary Insult Grading scheme, were analysed by insult grade and duration. A greater incidence of secondary insults was detected than previous studies have indicated. Insults were found in 91% of patients and occurred in all severities of head trauma, at all ages, and at every level of Injury Severity Score (ISS). The cumulative durations were much greater than previously recorded although 85% of the total time was at the least severe grade. Short duration insults were common. In 71 patients, in whom 8 insults could be assessed (intracranial pressure, arterial hypo- and hypertension, cerebral perfusion pressure, hypoxemia, pyrexia, brady- and tachycardia), outcome at 12 months was analysed using logistic regression to determine the relative influence of age, admission Glasgow Coma Sumscore, ISS, pupil response on admission, and insult duration on both mortality and morbidity. The most significant predictors of mortality in this patient set were durations of hypotensive (p = .0064), pyrexic (p = .0137), and hypoxemic (p = .0244) insults. When good versus poor outcome was considered, hypotensive insults (p = .0118) and pupil response on admission (p = .0226) were significant.

A group of 74 patients with head injury (54 severe, 17 moderate and 3 minor) had continuous monitoring of both arterial and intracranial pressure with computer-based registration of these pressures, cerebral perfusion pressure and other vairables. In 60 patients cerebral perfusion pressure CPP fell below 60 mm Hg for periods of 5 minutes or longer. The distribution over time of these reductions in CPP during up to 12 days of monitoring was studied, and each episode of reduced CPP was attributed to a fall in arterial pressure, an increase in intracranial pressure, or both. Two clusters of reduced CPP were found, one during the first 24 hours of monitoring, when reduced CPP was mainly caused by a reduction in arterial pressure, and the other at 5 or 6 days after injury, when reduced CPP was due mainly to an increase in intracranial pressure.

The Glasgow Coma Scale (GCS) and the Swedish Reaction Level Scale (RLS85), two level-of-consciousness scales used in the assessment of patients with head injury, were compared in a prospective study of 239 patients admitted to a regional head injury unit over a 4-month period. Assessments were made by nine staff members ranging from house officer to registrar, after briefing about the two scales. Data were also collected on age, nature of injuries, surgical treatment, and condition at discharge or transfer using the Glasgow Outcome Scale. Both the GCS and the RLS85 reliably identified comatose patients and those with minor head injury, but were much less effective in defining the response level in patients considered to have a moderate head injury. Only 41% of the patients allocated to a moderate-head-injury category by the GCS and the RLS85 were common to both groups. Where a mismatch occurred, neither scale allocated patients to a 'better' or 'worse' category more frequently than the other. Assessment of patients' conscious levels using the GCS was difficult in only two cases. One patient had facial injuries, and the other was intubated. The RLS85 was reported by all users to be simpler to use than the GCS, but the latter is much more widespread in use. Both scales function well in cases of severe and minor head injury, but have weaknesses when defining moderate head injury. Level-of-consciousness scales are only an aid to assessment and the final choice between the two scales must remain a matter of personal or departmental preference.

Three 1-year surveys of head injury management spanning a 9-year period in a single regional centre are presented, There was a reduction in total numbers of head injury admissions after guidelines for admission and referral were implemented. More liberal use of computed tomography resulted in detection of a greater number of intracranial haematomas with the majority detected in non-comatose patients. The early mortality rate in severe head injury fell from 45 per cent to 34 per cent despite referral of large numbers of patients with multiple injuries and a substantial proportion (12per cent) of patients aged more than 70 years in whom outcome did not improve. Total occupied bednights and bednights occupied per surviving patient with severe head injury fell over the period of study. Care for patients with significant head injury should be based on regional neurosurgical units associated with trauma services.

One hundred and eighty-one patients with single aneurysms involving the anterior circulation were treated either by wrapping (60 cases) or clipping (121 cases), and 96.1% of them were followed for 10 years. The rate of early rebleeding (less than 6 months) from wrapped aneurysms was 8.6% (confidence interval 1.4% to 15.8%) and the late rebleeding (6 months to 10 years) rate was 1.5% per annum (confidence interval 0.3% to 2.5%/yr). Wrapping an aneurysm offers some protection from rebleeding, particularly during the first 6 months when the risk of rebleeding is high.

The water content of samples of normal and oedematous brain in lobectomy specimens from 16 patients with cerebral tumours has been measured by gravimetry and by wet and dry weighing. Uncorrected gravimetry underestimated the water content of oedematous peritumoural cortex by a mean of 1.17%, and of oedematous peritumoural white matter by a mean of 2.52%. Gravimetric correction equations calculated theoretically and from an animal model of serum infusion white matter oedema overestimate peritumoural white matter oedema in man, and empirical gravimetric error correction factors for oedematous peritumoural human white matter and cortex have therefore been derived. These enable gravimetry to be used to accurately determine peritumoural oedema in man.

Brain water content was measured in tissue samples taken at operation from 19 patients with intrinsic cerebral tumours imaged preoperatively by magnetic resonance. A high correlation (r = 0.94, p less than 0.0001) between white matter water content and the longitudinal relaxation time (T1) enabled water content to be estimated from T1 to within 4%. 11 patients received dexamethasone and improved clinically but their T1, and thus brain water content, was unchanged an average of 6 days later. Intravenous infusion of 20% mannitol in 11 patients significantly reduced T1 in oedematous white matter and tumour within 15 min of administration, and by 30 min the T1 of oedematous white matter had fallen to a mean of 32.4 (SEM 7.1) ms, corresponding to a reduction in water content of 1.4 (0.3)%.

A relationship has been demonstrated between nuclear magnetic resonance (NMR) longitudinal relaxation times (T1 values) obtained in vivo in both normal and oedematous (peritumoral) brain tissue, and measurements of brain water obtained by gravimetric analysis of operative samples. Significant correlations were found in seven patients in both cortex (r = 0.97, P less than 0.001) and white matter (r = 0.92, P less than 0.001). These findings suggest that NMR may prove a useful technique for monitoring brain oedema.