The aims of this study were to examine long-term survival in a population-based sample of spinal cord injury (SCI) survivors in Great Britain, identify risk factors contributing to deaths and explore trends in cause of death over the decades following SCI. Current survival status was successfully identi®ed in 92.3% of the study sample. Standardised mortality ratios (SMRs) were calculated and compared with a similar USA study. Relative risk ratio analysis showed that higher mortality risk was associated with higher neurologic level and completeness of spinal cord injury, older age at injury and earlier year of injury. For the entire ®fty year time period, the leading cause of death was related to the respiratory system; urinary deaths ranked second followed by heart disease related deaths, but patterns in causes of death changed over time. In the early decades of injury, urinary deaths ranked ®rst, heart disease deaths second and respiratory deaths third. In the last two decades of injury, respiratory deaths ranked ®rst, heart related deaths were second, injury related deaths ranked third and urinary deaths fourth. This study also raises the question of examining alternative neurological groupings for future mortality risk analysis.

Study objectives: To describe the distribution of clinically apparent cardiovascular disease (CVD) in people with long-term spinal cord injury (SCI) according to neurologic level and severity of injury. Design: Historical prospective study. Setting: Two British Spinal Injuries Centers. Participants: Five hundred and forty-®ve individuals surviving at least 20 years with SCI were divided into three neurologic categories by level of injury and Frankel/ASIA grade as follows: Tetra ABC, Para ABC, and All D. Main outcome measures: Cardiovascular disease outcomes de®ned by ICD/9 codes 390 ± 448 and obtained through medical record review. Cardiovascular disease outcomes measured included All CVD, coronary heart disease (CHD), hypertension, cerebrovascular disease, valvular disease, and dysrhythmia. Results: After age-adjustment, the rates of All CVD were 35.2, 29.9, and 21.2 per 1000 SCI person-years in the Tetra ABC, Para ABC, and All D groups, respectively. Rates of All CVD increased with increasing age in all neurologic groups. Tetraplegic level of SCI conferred an excess 16% risk of All CVD (95% Con®dence interval [CI], 0.93 ± 1.46), a ®vefold risk of cerebrovascular disease (relative risk [RR] 5.06; 95% CI, 1.21 ± 21.15), and 70% less CHD (RR 0.30; 95% CI, 0.13 ± 0.70) when compared with paraplegics. More complete SCI was associated with an excess 44% All CVD risk (95% CI, 1.16 ± 1.77). Conclusions: Risk of All CVD increased with increasing age, rostral level of SCI, and severity of SCI. More rostral level of SCI was associated with cerebrovascular disease, dysrhythmia, and valvular disease. Conversely, there was an inverse relationship between level of SCI and CHD.

An extradural mass at the craniocervical junction causing progressive neurological disability in five elderly patients is described. The lesion, which might be confused with a meningioma or other tumour, is composed of amorphous degenerate fibrocartilaginous material and could be due to degeneration of the ligaments responsible for atlanto-axial stability. Recognition of the condition early is important as the patient's clinical condition will deteriorate without decompression. Anterior transoral removal is relatively simple, unlike surgery for tumours in the area, and will not destabilise the craniovertebral junction. It is likely that a proportion of these lesions are undetected, misdiagnosed or untreated to the detriment of the patient.

One hundred and fifty three patients who had sustained a spinal cord injury more than 20 years previously were assessed neurologically and by MRI scanning of their spinal cords. The spinal cord pathologies shown were, in order of prevalence, extended atrophy, malacia, syrinx, cyst, disruption and tethering. There was no relationship between the prevalence of any type of pathology and the degree of spinal canal compromise or angulation of the spine adjacent to the level of injury. Neurological changes after initial neurological stabilisation were seen only in patients with extended atrophy, malacia or a syrinx, not in those with only a cyst or cord disruption. Tethering is always associated with other lesion(s). Longer syrinxes were more likely to have associated neurological changes than shorter ones. The most common neurological change was pain.

Study Design: The functional outcome of the diaphragm after acute spinal cord injury was reviewed over a 16 year period for 107 patients who had required assisted ventilation in the acute phase. Objectives: To quantify the incidence of recovery of diaphragm function which occurred beyond the period of acute oedema; to produce a time-related pro®le of this as a guide to clinicians considering phrenic nerve pacing; and to assess the value of phrenic nerve testing in predicting recovery. Setting: The Southport Regional Spinal Injuries Centre, Southport, England. Methods: Bilateral phrenic nerve and diaphragm integrity was assessed clinically, by spirometry, and by¯uoroscopy without and with phrenic nerve stimulation. Results: Thirty-one per cent of all the ventilated patients (33 cases), with a level of injury between C1 and C4 (Scale A in ASIA Impairment Scale), had diaphragmatic paralysis at the time of respiratory failure. The subsequent diaphragm recovery which appeared in seven of these patients, between 40 and 393 days (mean 143), permitted weaning from ventilatory support at 93 to 430 days (mean 246) after the acute injury, with a vital capacity of over 15 ml kg 71 at that stage. The diaphragm recovery in a further ®ve patients, whose vital capacity remained below 10 ml kg 71 and who could not be fully weaned, occurred signi®cantly later, between 84 and 569 days (mean 290), P=0.053. Negative phrenic nerve tests were followed by weaning at a later interval in several cases. By contrast, one patient with an early positive phrenic stimulation test and subsequent diaphragm activity could not be weaned from the ventilator. Conclusion: Twenty-one per cent of the patients with initial diaphragm paralysis were ultimately able to breathe independently after 4 and 14 months, whilst a further 15% had some diaphragm recovery. Phrenic nerve testing should be repeated at 3 monthly intervals for the ®rst year after high tetraplegia.