Proc. Natl. Acad. Sci. USA Vol. 82, pp. 2513-2517, April 1985 Medical Sciences Capillary number and percentage closed in human diabetic sural nerve (morphometry/diabetes/polyneuropathy/pathology) PETER JAMES DYCK, SEONA HANSEN, JEANNINE KARNES, PETER O'BRIEN, HITOSHI YASUDA, ANTHONY WINDEBANK, AND BRUCE ZIMMERMAN Peripheral Nerve Center, Mayo Clinic and Mayo Foundation, Rochester, MN 55905 Communicated by Ralph T. Holman, December 13, 1984 ABSTRACT The number of capillaries per mm2, mini- alteration interposed between metabolic derangement and mum intercapillary distance, number of endothelial nuclei per neuropathic dysfunction is inferred (24). capillary section, and percentage of capillaries closed were A functional and structural alteration of diabetic capillar- evaluated in transverse sections of fascicles of 45 control and ies should be considered for this intervening pathologic alter- 36 diabetic sural nerves. All controls and patients were pro- ation because (i) a generalized basement membrane abnor- spectively studied to ascertain their diabetic and neuropathic mality is characteristic of diabetes (25-27); (ii) such an ab- status. An index of pathology was introduced and it was found normality has also been described for diabetic nerve (28-31); to provide a sensitive and reliable measurement of the pres- (iii) increased permeability has been described for capillaries ence and severity of neuropathy. The number of capillaries of several tissues (32-34); (iv) the complications of retinopa- and minimum intercapillary distance of diabetic nerves were thy, nephropathy, and neuropathy are correlated in the same not significantly different from those of controls (P > 0.05). patient (5, 35) and, since capillary dysfunction is involved in Diabetic nerves exhibited a small but statistically significant retinopathy and nephropathy, it may also be involved in neu- increase in the number of endothelial nuclei per capillary that ropathy; (v) reduced nerve blood flow and oxygen tension was positively correlated with the severity of neuropathy. The have recently been described in streptozotocin diabetes (36); most striking abnormality was the statistically significant in- and (vi) the nerve conduction abnormality of streptozotocin crease in the percentage of capillaries closed in patients with diabetes can be partially prevented by oxygen supplementa- neuropathy as compared to those without neuropathy and con- tion (37). For nerve, however, capillaries have not been sys- trols. Among diabetics, this percentage increased with the se- tematically studied and severity of capillary derangement verity of neuropathy (P = 0.008). The two capillary abnormal- has not yet been critically related to neuropathic deficit. ities that have been demonstrated may play a role in the devel- opment of diabetic polyneuropathy. MATERIALS AND METHODS Patient Selection. An ankle-level fascicular sural nerve bi- Diabetic polyneuropathy, a common complication of diabe- opsy was performed on 45 healthy subjects, recruited by ad- tes mellitus, is typically expressed as sensory loss, pain, and vertisement from medical personnel and a recreational club autonomic dysfunction in the feet and legs (1). Although the for elderly people. Persons with a family history of neuropa- mechanisms underlying diabetic neuropathy remain un- thy or diseases known to be associated with predisposition known, chronic hyperglycemia may lead to metabolic de- to neuropathy were excluded. Controls underwent the same rangement that directly affects neurons (axons) or Schwann evaluation for neuropathy as did diabetics. cells or indirectly affects them by first altering another tis- A similar biopsy was performed on 36 patients with diabe- sue-e.g., vessels. Since a higher rate of atherosclerotic tes mellitus, as diagnosed by the National Diabetes Data heart disease and peripheral vascular disease occurs among Group criteria (38). Ten of these had insulin-dependent (ID) diabetics (2, 3) and since arteriosclerosis with vessel occlu- and 26 had non-insulin dependent (NID) diabetes by the cri- sion is reported for vasa nervorum of nerve (4-6), arterio- teria of the American Diabetes Association and by an algo- sclerosis has been postulated as a cause of diabetic neuropa- rithm developed by us. A neuropathy symptom score, a neu- thy. This hypothesis, however, may not explain the common rologic disability score (39), measured attributes of nerve occurrence of an abnormality of nerve conduction and devel- conduction of peroneal and sural (and other) nerves, and de- opment of a diffuse neuropathy among diabetic patients who tection thresholds of vibratory and cooling sensation on the do not manifest peripheral vascular disease (7-10). Atten- great toe and dorsal foot (40), respectively, were obtained tion, therefore, has been focused on metabolic derangements for each patient. that might affect neurons (or their peripheral axons) or The minimal criteria used for the diagnosis of neuropathy Schwann cells. Among the metabolic mechanisms that have was that two or more of the following evaluations gave ab- been considered are the following: (i) lipid alterations (11- normal results: (i) neuropathy symptom score - 1; (ii) neuro- 13); (ii) accumulation of sorbitol and fructose (14, 15); (iii) logic disability score . 6; (iii) computer-assisted sensory ex- decreased nerve myo-inositol (16); (iv) increased nonenzy- amination (CASE) of toe or foot; and (iv) nerve conduction matic glycosylation of protein (17, 18); (v) decrease of of limb nerves (two abnormal attributes in at least two Na+,K+-ATPase (19); (vi) increased intraaxonal sodium nerves and in at least two limbs). By these criteria, 32 of the (20); (vii) alterations in axonal flow (21) and axonal attenua- 36 diabetic patients had neuropathy (8 ID and 24 NID) and 4 tion (22); and (viii) tissue dehydration (23). Since the nerve did not ID 2 conduction abnormality is not readily reversed after near (2 and NID). Among neuropathy patients, the normalization of blood glucose for periods of up to 8 months, diagnosis was based on abnormalities in four evaluations in an irreversible nerve alteration or an intervening pathologic 14 patients, on three evaluations in 14 patients (but in 9 of Abbreviations: ID, insulin-dependent diabetes; NID, non-insulin- The publication costs of this article were defrayed in part by page charge dependent diabetes; lp, index of pathology; C/mm2, number of cap- payment. This article must therefore be hereby marked "advertisement" illaries per mm2; N/C, number of endothelial nuclei per capillary in accordance with 18 U.S.C. §1734 solely to indicate this fact. section; %CC, percentage of capillaries closed. 2513 Downloaded by guest on September 26, 2021 2514 Medical Sciences: Dyck et aL Proc. Nad Acad ScL USA 82 (1985) 120 [ 8 0 0 100 0 0 6 0 0 ~0 00 0 0 N 80 Z 4 0 -E 60 0 0 0 0 0 2 0f 0 0 40 0 0 20 0 "0 10 20 30 40 50 60 70 Age, yr 0 10 20 30 40 50 60 70 FIG. 2. Mean N/C from transverse sections of fascicles of ankle Age, yr sural nerves plotted against age for 45 healthy subjects. The regres- sion line, fitted by the method of least squares, does not show a FIG. 1. C/mm2 from transverse sections of fascicles of ankle significant slope (P > 0.05). A significantly positive slope was found The regres- sural nerves plotted against age for 45 healthy subjects. for males only (n = 27; 0.025 < P < 0.05). sion line, fitted by the method of least squares, has a negative slope (P = 0.008), indicating that, on average, there are fewer capillaries with increasing age. teased fibers with de- and remyelination; and percentage of teased fibers with axonal de- and regeneration) provided the these only 3 of those evaluations had been obtained), and on best predictor ofneuropathy using clinical, nerve conduction two evaluations in 3 patients (but in one of these only three and computer-assisted sensory examination criteria. The Ip evaluations were done). One patient, with only one abnor- provided the best measure, accurately predicting all cases mality ofthree evaluations performed, was added to the neu- with neuropathy followed in succession by number of my- ropathy group because of an unequivocally abnormal index elinated fibers per mm2, number of large myelinated fibers of pathology (Ip). Among the diabetics without neuropathy, per mm2, and then all teased fiber abnormalities. two had one abnormal evaluation but, in all cases, the Ip was The procedure of fascicular nerve biopsy, size of nerve normal. The polyneuropathy among diabetic patients was of sample, fixation, and histologic processing were the same mild or moderate severity. All of the patients were ambula- for diabetic patients and controls and as previously de- tory and most held jobs. None of these patients had prolifer- scribed (42). Fixation was in isotonic glutaraldehyde and ative retinopathy, renal failure, an elevated plasma creati- then the sample was immersed in isotonic osmium tetroxide nine value, or symptomatic peripheral vascular disease. Five at 10TC. The number of capillaries per mm2 (C/mm2), mini- patients had a history of coronary artery disease. By the mum intercapillary distance (MICD), number of endothelial World Health Organization criteria (41), definite hyperten- nuclei per capillary section (N/C), and the percentage of sion was found in 10 patients and borderline hypertension in capillaries closed (%CC) were evaluated by using our imag- 7 patients. ing system for nerve morphometry (ISNM) (43) in methylene Histologic Studies. To provide a measure of the severity of blue-stained, semithin (3/4-,um), transverse, epoxy sections of the pathologic abnormalities affecting nerve fibers, we used fascicles of sural nerve. All evaluations were performed on Ip. Ip is the product of the ratios of the observed-to-expected coded slides, from controls and diabetic patients with and number of myelinated fibers per mm2 and the observed-to- without neuropathy, randomly assigned and with identifica- expected percentage of graded normal teased fibers. The use tion marks covered. All vessels without muscle in their walls of Ip combines two components of pathologic abnormality were evaluated.