Clinical, electrophysiological and ultrastructural morphometric observations were made in 5 diabetic non-neuropathic patients, 5 diabetic patients with mild neuropathy and 11 diabetic patients with severe neuropathy. Capillary abnormalities were assessed in simultaneous nerve, muscle and skin biopsies and compared with results from 6 age-matched, non-diabetic control subjects. Nerve capillaries demonstrated markedly greater pathology than skin and muscle capillaries. Endoneurial capillary density was significantly reduced in severely neuropathic diabetic patients (p less than 0.01) when compared with control subjects. Capillary basement membrane (p less than 0.002), endothelial cell (p less than 0.003) and total diffusion barrier (endothelial cell, pericyte, basement membrane) (p less than 0.001) thickness were significantly increased, and oxygen diffusing capacity was significantly reduced (p less than 0.001) in the nerves of patients with severe diabetic neuropathy when compared to control subjects. Endothelial cell profile number and luminal perimeter were significantly increased in asymptomatic (p less than 0.01), (p less than 0.05) and severely neuropathic (p less than 0.001), (p less than 0.05) diabetic patients respectively. However, endothelial cell outer perimeter, a measure of capillary size, showed no significant increase in diabetic patients when compared with control subjects. An association was observed between neurophysiological and neuropathological measures of neuropathic severity. There was no significant correlation between the duration of diabetes and HbA1 levels with capillary pathology or with neuropathic severity. Very few abnormalities of muscle and skin correlated with neuropathic severity. However, all measures of nerve capillary pathology correlated significantly with neurophysiological and neuropathological measures of neuropathic severity.

Peripheral nerve oxygen tensions were assessed in vivo by using microelectrodes to measure endoneurial oxygen tension in exposed sural nerve. In 11 diabetic patients with chronic sensorimotor neuropathy the mean endoneurial oxygen tension was 39 7

SUMMARY Angiographic studies on 150 pituitary adenomas and 33 craniopharyngiomas presenting for surgical treatment are reviewed. Eleven incidental silent aneurysms (four arising from the intracavernous and four from the supraclinoid carotid artery, and three from the anterior cerebral artery complex) are shown. Intracavernous aneurysms were also present in two acromegalic patients who had been treated previously with yttrium implantation. Although encasement of vessels by these tumours is unusual, the relevance of vascular abnormalities to surgical treatment is sufficient to justify routine magnification angiography.

Insulin neuritis, or painful neuropathy following rapid improvement in glycaemic control, is well recognised but its aetiology is unclear. An understanding of the processes involved in the genesis of acute painful neuropathy of rapid glycaemic control may give an insight into the early pathogenetic factors leading to diabetic nerve damage in general. We have identified five subjects with insulin neuritis including one who developed severe autonomic neuropathy following treatment with insulin. Subjects underwent: 1) assessment of neuropathic symptom and deficit scores; 2) quantitative sensory and electrophysiological studies and 3) sural nerve epineurial vessel photography and fluorescein angiography in vivo. The sural nerve photographs were independently graded by an ophthalmologist. All subjects with insulin neuritis presented with severe sensory symptoms but clinical examination and electrophysiological tests were normal except in the subject with the severe autonomic neuropathy in whom all the tests were abnormal. On nerve photography, there was an abundance of epineurial nutrient vessels although these showed severe abnormalities including arteriolar attenuation, tortuosity and arterio-venous shunting in all subjects. Proliferating neural 'new vessels' which bear striking similarities to those found in the retina and that were more leaky to fluorescein than normal vessels, were observed in three subjects. Venous distension and/or tortuosity was also observed in three subjects and this was most marked in the subject with severe autonomic neuropathy. This study shows that epineurial nutrient vessel anatomy is abnormal in subjects with acute painful neuropathy of rapid glycaemic control, a condition previously thought to be purely metabolic in origin. The presence of epineurial arterio-venous shunting and a fine network of vessels resembling the new vessels of the retina, may lead to a 'steal' effect rendering the endoneurium ischaemic. This process may be important in the genesis of neuropathic pain, and further supports the importance of vascular factors in the pathogenesis of diabetic neuropathy.

Summary. New techniques of sural nerve photography and fluorescein angiography which are able to provide an index of nerve blood flow have been developed. Under local anaesthetic, 3 cm of sural nerve was exposed at the ankle using an operating microscope. Without disturbing the epineurium, vessels were identified and photographed at a standard magnification ( x 30). These were independently graded by an ophthalmologist not otherwise involved with the study. Fluorescein angiography was then carried out on the exposed nerve. The fluorescein appearance time and intensity of fluorescence were quantified, using computer analysis of digitised images. Thirteen subjects with chronic sensory motor neuropathy, five non-neuropathic diabetic and nine normal control subjects were studied. The mean epineurial vessel pathology score of the neuropathic group was significantly higher than the combined normal control and non-neuropathic diabetic groups (p < 0.01). Direct epineurial arteriovenous shunting was observed in six neuropathic and one non-neuropathic diabetic patients and not in any of the normal control subjects. The nerve fluorescein appearance time was significantly delayed in subjects with chronic sensory motor neuropathy (51.5 + 12 s) compared to both normal (34.7 _+ 9 s, p < 0.01) and non-neuropathic diabetic subjects (33.4 + 11 s,p < 0.025). The mean intensity of fluorescence at 96, 252 and 576 s, was significantly lower in subjects with chronic sensory motor neuropathy compared with both of the other groups (p < 0.05). The epineurial vessel pathology score was significantly related to reduced sural (p < 0.01) and peroneal (p < 0.001) nerve conduction velocities, elevated vibration (p < 0.01) and thermal (p < 0.001) perception and the severity of retinopathy (p < 0.002). The fluorescein appearance time was significantly related to reduced sural sensory (p <0.02) conduction velocity, elevated vibration (p < 0.01) perception and epineurial vessel (p < 0.002) pathology score, but it failed to relate to peroneal motor (p = 0.06) conduction velocity, thermal (p = 0.1) perception and the severity ofretinopathy (p = 0.3). Intensity of fluorescence was significantly related to fluorescein appearance time (at 96 s,p < 0.001; at 576 s,p < 0.05) but did not relate to measures of neuropathic severity. These techniques have enabled us to observe that epineurial vessel anatomy is abnormal and that nerve blood flow is impaired in subjects with chronic sensory motor neuropathy. In addition epineurial arterio-venous shunting may be a feature of diabetic neuropathy. These techniques may further be applied to study nerve blood flow in early diabetic neuropathy.