In able-bodied individuals, insulin action is related to the level of physical activity (22). Correspondingly, individuals with spinal cord injuries (SCI), in whom major muscle groups are paralyzed, show reduced glucose tolerance (2,11) and exaggerated hyperinsulinemia during glucose loading (2). Insulin resistance and increased insulin response to an oral glucose challenge are associated with development of type 2 diabetes mellitus, atherosclerosis, and ischemic heart disease (2,4,11,13,29). In line with this, the prevalence of these conditions is found to be higher in SCI than in able-bodied individuals (2,24,25,31). It follows that, in the population with SCI, morbidity and mortality may be prevented or diminished by reducing insulin resistance.
In able-bodied subjects, physical training with large muscle groups increases the effect of insulin on glucose uptake in skeletal muscle (8). In addition, the GLUT 4 glucose transporter protein, the membrane protein responsible for the insulin mediated transport of glucose into skeletal muscle cells, is increased in response to training in able-bodied individuals (6,9). In contrast to able-bodied persons, the individuals with SCI are unable to perform voluntary physical training with large muscle groups. However, it is now possible to use computer controlled sequential electrical stimulation of the leg muscles to induce cycle movements, i.e., functional electrical stimulation (FES).
Thus, the aim of this study was to examine whether regular electrically induced cycle training increases insulin sensitivity, oral glucose tolerance, and the GLUT 4 content in the stimulated muscles in individuals with SCI. To evaluate the importance of the necessary amount of training, three FES training bouts were performed per week in 1 yr, which thereafter was reduced to one training bout per week.
MATERIALS AND METHODS
The study was approved by the Regional Ethics Committee of Copenhagen and Frederiksberg Municipalities. Ten subjects with SCI were enrolled after oral and written information, and they all signed an informed consent form. Before the study, none of the subjects had any experience with regular training using electrical stimulation. They were all suffering from traumatic SCI, seven due to traffic accidents, two due to accidents in sporting activities, and one due to shallow water diving. Eight men and two women were included, and the average age was 35 (range 27–45) yr, weight 73 ± 5 kg (mean ± SE). Six subjects were tetraplegic (level C6) and four were paraplegic (level Th4). Time since injury was 12 yr (range 3–23 yr), and all subjects were neurologically stable. None of the subjects suffered from diseases or disabilities other than the SCI.
For the training of the subjects with SCI, a computer-controlled FES exercise ergometer (REGYS I Clinical Rehabilitation System, Therapeutic Technology Inc., Tampa, FL) was used. This system has been described elsewhere (1,18,27). Briefly, the hamstring, quadriceps, and gluteal muscles of both legs were sequentially stimulated via surface electrodes to perform pedalling on a modified cycle ergometer. The stimulation sequence was computer controlled based on the position of the crank of the ergometer. The target pedalling speed was 50 revolutions per minute. The current used was rectangular, monophasic AC, frequency 30 Hz, pulse duration 350 μs, and maximal intensity was preset to 120 mA.
Subjects trained 30 min three times per week for 1 yr. During the following 6 months, the training was reduced to 30 min once a week.
Insulin-stimulated glucose uptake.
Insulin-stimulated glucose uptake was measured by the euglycemic, hyperinsulinemic clamp technique before regular training was initiated, after 1 yr of high-intensity training and after 6 months of reduced training. The clamp procedure was performed 48 h after the last exercise bout. The subjects were fasted from 10.00 p.m. the day before the experiment and arrived at 8.00 a.m. They emptied their bladders and were weighed and placed supine in a bed. Precordial electrodes were placed for continuous monitoring of heart rate and ECG. A catheter for glucose and insulin infusion (Secalon Cathy R, 650 × 1.8 mm, Viggo, Helsingborg, Sweden) was inserted in a medial antecubital vein and advanced to the level of the brachiocephalic vein. For blood sampling, another catheter (Venflon 0.8, Viggo) was inserted in retrograde direction as distally as possible in a dorsal hand vein. The hand was placed between two electric heating pads and heated to 40°C to arterialize the hand vein blood. Patency was maintained by flushing with a Na-Citrate solution (3.8%, Pharmacy, Rigshospitalet, Copenhagen, Denmark).
Insulin infusates consisted of 2.5 mL of 20% human serum albumin (Statens Seruminstitut, Copenhagen, Denmark), Actrapid insulin (100 IU·mL−1, NOVO, Copenhagen, Denmark) and isotonic saline to 50 mL. The amount of insulin added to the infusate was calculated on basis of individual surface area to result in infusion rates of 50 mU·min−1·m−2 and 480 mU·min−1·m−2, respectively, during the two 2-h sequential clamps performed in each clamp procedure. These infusion rates ensured plasma insulin levels above 75 μU·mL−1 to minimize hepatic glucose production (22). The glucose infusion rates are thereby equivalent to whole-body glucose uptake rates from plasma in both steps of the clamp. In each of the two steps of the clamp procedure, a bolus of 2 mL of the insulin infusate was given before the insulin infusion. Plasma glucose concentrations were always clamped at the individual fasting concentration, measured before training began. Arterialized blood (0.3 mL) was drawn from the heated hand vein for glucose measurements at least every 5th min. The blood was centrifuged for 10 s in a microfuge, and 25 μL of plasma was analyzed for glucose concentrations using a YSI 2300 STAT (Yellow Springs Instruments, Yellow Springs, OH). The glucose infusion rate was controlled by a computer. During the last 30 min of each clamp step, glucose infusion had reached steady state, and the average of the infusion rates in this period was calculated. This value represented the steady state glucose infusion rate (SSGIR).
Blood samples for determination of plasma insulin, C-peptide, free fatty acids, urea, β-hydroxybutyrate, glycerol, epinephrine, and norepinephrine were in the basal state as well as at the end of each clamp steps and kept frozen at −80°C (free fatty acids and catecholamines) and −20°C (other samples) until analyzed as described previously (7,14). The production of urine and urinary glucose excretion were measured in each clamp step.
Oral glucose tolerance test (OGTT).
OGTTs were performed before and at the end of 12 and 18 months of training. The OGTTs were carried out 48 h after exercise and after an overnight fast. The OGTTs and the clamps were separated by at least 1 week. After weighing the subjects, a cannula (Venflon 0.8, Viggo) was inserted in a dorsal hand vein for blood sampling. The first blood sample (time 0) was drawn after 20 min of rest, and the subject then ingested a 20% glucose solution containing 75 g of glucose, followed by 100 mL of water. Blood samples were drawn every 5 min and immediately analyzed for plasma glucose on an automatic glucose analyzer (YSI 2300 STAT, Yellow Springs Instruments). For insulin analysis, blood was sampled in ice-chilled tubes containing 500 KIU Trasylol (aprotinin) and 10 IU heparin per mL blood and immediately centrifuged at 4°C. The samples were then stored at −20°C until analyzed by a previously described radioimmunoassay (14). Plasma glucose and insulin responses during the OGTTs were compared with previously published data (7) from a group (N = 8) of young (24 (range 20–28) yr) healthy, untrained (V̇O2max: 48 (range 41–52) mL·min−1·kg−1) men, in whom the OGTT was carried out in a similar fashion.
Quantitation of GLUT 4 glucose transporter protein.
Needle biopsies were taken from the middle portion of the vastus lateralis muscle before training and after 3, 6, and 12 months of intensive training and finally after the subsequent 6 months of reduced training. Local anesthesia (2-mL lidocaine, 1%) of the subcutis was administered before incision of the skin to avoid reactions that could generate spasms during the procedure. Biopsies were immediately frozen in liquid nitrogen and stored at −80°C until analyzed. Approximately 50 mg of tissue was freeze dried and dissected free of connective tissue, fat, and blood, resulting in a mean of 13 mg of muscle powder. The isolation of the crude membranes and determination of protein content in this muscle powder were performed as described by Dela et al. (6). Samples containing 15-μg protein (in four preparations, 6–9 μg of protein were applied due to low protein content in the crude membrane preparation) of each crude membrane preparation were together with reference samples analyzed in duplicate by SDS-PAGE and Western blotting as described by Handberg et al. (15). Samples from the same subject before and during training were analyzed on the same gel. Quantitation of GLUT 4 protein immunoreactivity was done by densitometric scanning (LKB - Vitroscan XL, Bromma, Sweden).
The nonparametric paired Wilcoxon signed rank test was used to test whether the measured values were significantly different before and after training. Values of P < 0.05 was considered significant in two tailed testing. All data are presented as mean ± SE.
Insulin-stimulated glucose uptake.
Plasma insulin concentrations in the two steps were in the physiological and pharmacological range, respectively (Table 1). The concentration of insulin in plasma was similar during all three clamps (Table 1). Plasma glucose concentrations did not differ between the three clamp studies or between the clamp steps. SSGIRs before training were 4.9 ± 0.5 and 9.0 ± 0.8 mg·min−1·kg−1 in the first and second step of the clamp, respectively (Fig. 1). In response to 1 yr of training, SSGIR in step 1 of the clamp increased significantly to 6.2 ± 0.6 mg·min−1·kg−1 (28% increase, P < 0.05), whereas the increase seen in step 2 (to 10.6 ± 0.8) did not attain statistically significance (P = 0.129). After the following 6 months with reduced training, the SSGIR in step 1 decreased from 6.2 ± 0.6 mg·min−1·kg−1 (the value at 12 months) to 5.2 ± 0.5 mg·min−1·kg−1 (P < 0.05) and was no longer different from before training (P > 0.05). SSGIR in step 2 did not change significantly (10.6 ± 0.8 and 9.8 ± 0.8 mg·min−1·kg−1 between 12 and 18 months, respectively) over the period of training (P > 0.05).
Plasma concentrations of C-peptide, free fatty acids, glycerol, and β-hydroxybutyrate decreased during insulin infusion and did not change over the course of training (Table 1) and can therefore not be responsible for the changes in insulin action over the training period.
Norepinephrine concentrations were at step one higher after 12 months of training compared with before training (Table 1). The urinary excretion of glucose during the clamps was less than 1% of SSGIR in all subjects and therefore not taken into account.
Oral glucose tolerance test.
Plasma glucose concentrations during the OGTTs did not differ significantly in response to training (Fig. 2 A). Furthermore, the glucose concentrations in the subjects with SCI were not different from the glucose concentration in healthy, untrained men (Fig. 2 A) The plasma insulin response to the oral glucose load did not change with training (Fig. 2 B). However, compared with the healthy control group, insulin concentrations were always higher in subjects with SCI (P < 0.05) (Fig. 2 B).
The content of GLUT 4 transporter protein increased 105% (from 20 ± 6 units before training to 41 ± 9 units after 1 yr of training) (Fig. 3, P < 0.05). A significant increase was seen already after 3 months of training (P < 0.05). After 6 months of reduced training the GLUT 4 content was no longer significantly different from before training (Fig. 3).
The present study has shown that 1 yr of regular FES training of paralyzed leg muscles result in an increased whole-body insulin sensitivity. Reduction of the training regimen from three to only one session per week is not sufficient to maintain the increase in insulin sensitivity. Neither glucose tolerance nor the insulin response to an oral glucose load showed an improvement after this kind of training. The increase in insulin sensitivity may be seen as a result of an increase in leg muscle mass, the large increase in GLUT 4 in the stimulated muscle, and possibly as a result of fiber type transformation in the muscles during the period of stimulation.
In able-bodied individuals, the insulin-stimulated glucose uptake rate is known to increase after a period of regular physical activity (22). It has been demonstrated that muscle accounts for approximately 85% of the glucose metabolism during euglycemic clamp studies (5). We have previously reported that the individuals included in the present study had an increase in muscle cross sectional area of the thigh of 12% in average measured by magnetic resonance imaging (MRI) (23). Accordingly, the increase in glucose infusion rates found in the present study after training could be associated with this increase in muscle volume. If, in these individuals with SCI, the increase in thigh muscle area of 12% found by MRI reflects an increase in volume of 12% in all electrically stimulated muscle, the training induced increase in muscle volume can be estimated to be 6–8% of whole-body muscle mass (23). Because the increase in SSGIR was 28% per kg overall body weight, the increase in muscle volume as determined by MRI could per se only partially account for the increase in glucose infusion rate. However, the amount of interfibrillar tissue in the muscles was not determined by MRI, and the magnitude of muscle volume increase could be underestimated due to the fact that interfibrillar tissue in stimulated muscle decreased as a result of training (23).
Other investigators have used the euglycemic-hyperinsulinemic clamp technique to study insulin sensitivity in individuals with SCI. Insulin sensitivity in seven paraplegics (17) has been studied with insulin in plasma at the same level as in the first step of our clamp, and the glucose infusion rates (4.37 mg·min−1·kg−1) were equivalent to the rates we found before training. In sedentary, young, able-bodied individuals studied during euglycemic clamping at the same plasma insulin level as in step 1 in the present study (90 μU·mL−1), we have previously found a steady state glucose infusion rate of 9 mg·min−1·kg−1(21). Despite an increase in glucose uptake of 28% after training, this value was not reached by the individuals with SCI in the present study, showing that the training program could not completely normalize insulin-mediated glucose metabolism. During training, V̇O2max in the subjects with SCI increased 23% to an average of 1.48 L·min−1(23). The fact that the V̇O2max also remained subnormal after training is in agreement with a previously demonstrated association between insulin-mediated glucose metabolism and maximal oxidative capacity in healthy subjects (28).
In the individuals included in the present study, fiber type composition changed from mixed type IIB/IIA fibers to pure type IIA fibers in response to FES training (1,23). A relationship between muscle fiber type and insulin-stimulated glucose uptake has been clearly demonstrated in rat muscle (20). In humans, indices of insulin action are positively correlated with the percentage of type I fibers and negatively correlated with the percentage of type IIB fibers (19). These correlations and the evidence from animal studies indicate that the fiber type transformation from dominance of type IIB fibers toward dominance of type IIA fibers observed in our subjects as a result of training may contribute to the increase in insulin-stimulated glucose uptake.
Neither the plasma glucose nor the plasma insulin during the OGTT did change throughout training periods, indicating no training effect on this parameter in subjects with SCI (Fig. 2, A and B). The 2-h plasma glucose value exceeded 7.2 mM, indicating glucose intolerance both before and after training. The fact that no training-induced change was found in the OGTT despite an increase in insulin-stimulated glucose uptake during clamp procedures may reflect 2that much more training than performed in this study is needed to increase the glucose tolerance or that glucose tolerance is influenced by other factors than insulin sensitivity, e.g., glucose uptake in the gut and pancreatic β-cell secretion.
An increase in GLUT 4 glucose transporter protein as a result of electrical stimulated training has before only been demonstrated in five individuals with SCI (3). In this recent study, 8 wk of training using the same protocol as in first part of the present study showed an increase of 72% in the amount of GLUT 4. This corresponds to the relative increase in GLUT 4 seen in our study (75%) after the first 3 months of training with similar intensity.
Because the electrical stimulated contraction does not involve central mechanisms it indicates that the stimuli for synthesis of GLUT 4 is a local contraction-dependent phenomenon. The observed relative increase in GLUT 4 glucose transporter protein in quadriceps muscle was five times the relative increase seen in insulin-stimulated glucose uptake in the whole body over the 1-yr training period. However, GLUT 4 content in leg muscle has been shown not to vary in parallel with whole-body insulin sensitivity (6). In line with the present findings (Fig. 3), electrical stimulation of animal muscle (predominantly fast-twitch plantaris muscle in rats) has been shown to result in a 82% increase in GLUT 4 transporter protein content after 10–20 d of stimulation (10 Hz, 8 h·d−1), and a prolongation of the stimulation period did not result in further enhancement of GLUT 4 transporter protein content (12).
A relationship between insulin-stimulated glucose uptake and fiber type distribution in humans has been mentioned above. To our knowledge, no studies in humans exist that were designed to investigate a possible relationship between fiber type composition and GLUT 4 transporter protein, but studies in rats have revealed a positive correlation between proportion of oxidative fibers and GLUT 4 transporter protein in hindlimb muscles (16,20). This indicates that the fiber type transformation found over the training period in the present study (1,23) could, in part, be responsible for the increase in GLUT 4 transporter protein content. However, indicating that other mechanisms also were involved in man (6) as well as in rat (26), increases in muscle GLUT 4 transporter protein concentration have been observed in response to training regimens that did not alter the fiber type composition.
Dela et al. (6) have described changes in GLUT 4 transporter protein concentration in skeletal muscle in healthy subjects after a 10-wk period of one leg exercise for 30 min·d−1, 6 d·wk−1 at 70% of V̇O2max. The concentration of GLUT 4 transporter protein increased 26% in the trained leg only, also indicating that the increase is due to a local, contraction-dependent adaptation, and not to systemic factors. The absolute amount of exercise administered per day during training in the study by Dela et al. (6) was much higher than in the present study. The fact that the increase in concentration of GLUT 4 transporter protein was, nevertheless, smaller (26% vs 105%) probably reflects that initial values most likely (a direct comparison is unfortunately not possible) were higher in able-bodied compared with subjects with SCI.
In conclusion, this study has demonstrated that intense FES training improves insulin sensitivity in subjects with spinal cord injuries. However, a reduced training regimen of only one training session per week is not sufficient to maintain the increased insulin sensitivity. The increase in insulin sensitivity may be ascribed to increased muscle mass, conversion to more oxidative fiber types, and increased GLUT 4 content in the trained muscle. Because individuals with SCI have a high incidence of cardiovascular diseases (2,10) and because insulin resistance is known to be a risk indicator in cardiovascular disease (4,13,29,30), the results from the present study encourages the use of FES training in the population with spinal cord injuries.
The authors thank the subjects included in this study for their participation, physiotherapists Per Tornøe, Susse Broberg, and Karin Christophersen for carrying out the training, and technicians Gerda Hau, Lisbeth Kall, Regitse Kraunsøe, and Vibeke Staffeldt for their skilled technical assistance. Financial support was received from the Danish Medical Research Council (J. no. SSF 12-9360), the National Society of Polio and Accident Victims, the Danish National Research Foundation (504-14), The Hamilton Foundation, Danish Hospital Foundation for Medical Research (Region of Copenhagen, The Faroe Islands, and Greenland), Idrættens Forskningsråd, the Danish Heart Foundation, the Foundation of J. and O. Madsen, the Danish Medical Association Research Found, and Team Denmark.
Address for correspondence: Thomas Mohr, M.D., Ph.D., Department of Anaesthesiology, Glostrup University Hospital, Ndr. Ringvej, DK 2600 Glostrup, Denmark; E-mail: [email protected]
1. Andersen, J. L., T. Mohr, F. Biering-Sorensen, H. Galbo, and M. Kjaer. Myosin heavy chain isoform transformation in single fibres from m. vastus lateralis in spinal cord injured individuals: effects of long-term functional electrical stimulation (FES). Pflugers Arch. Eur. J. Physiol. 431: 513–518, 1996.
2. Bauman, W. A., A. M. Spungen, M. Raza, et al. Coronary artery disease: metabolic risk factors and latent disease in individuals with paraplegia. Mt. Sinai J. Med. 59: 163–168, 1992.
3. Chilibeck, P. D., G. Bell, J. Jeon, et al. Functional electrical stimulation exercise increases GLUT-1 and GLUT-4 in paralyzed skeletal muscle. Metabolism 48: 1409–1413, 1999.
4. Defronzo, R. A., and E, Ferrannini. Insulin resistance: a multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14: 173–194, 1991.
5. Defronzo, R. A., E. Jacot, E. Jequier, E. Maeder, J. Wahren, and J. P. Felber. The effect of insulin on the disposal of intravenous glucose: results from indirect calorimetry and hepatic and femoral venous catheterization. Diabetes 30: 1000–1007, 1981.
6. Dela, F, A. Handberg, K. J. Mikines, J. Vinten, and H. Galbo. GLUT 4 and insulin receptor binding and kinase activity in trained human muscle. J. Physiol. (Lond.) 469: 615–624, 1993.
7. Dela, F., K. J. Mikines, M. Von Linstow, and H. Galbo. Effect of training on response to a glucose load adjusted for daily carbohydrate intake. Am. J. Physiol. 260: E14–E20, 1991.
8. Dela, F., K. J. Mikines, M. Von Linstow, N. H. Secher, and H. Galbo. Effect of training on insulin-mediated glucose uptake in human muscle. Am. J. Physiol. 263: E1134–E1143, 1992.
9. Dela, F., T. Ploug, A. Handberg, et al. Physical training increases muscle GLUT4 protein and mRNA in patients with NIDDM. Diabetes 43: 862–865, 1994.
10. Devivo, M.J., K. J. Black, and S. L. Stover. Causes of death during the first 12 years after spinal cord injury. Arch. Phys. Med. Rehabil. 74: 248–254, 1993.
11. Duckworth, W. C., P. Jallepalli, and S. S. Solomon. Glucose intolerance in spinal cord injury. Arch. Phys. Med. Rehabil. 64: 107–110, 1983.
12. Etgen, G. J., Jr., R. P. Farrar, and J. L. Ivy. Effect of chronic electrical stimulation on GLUT-4 protein content in fast-twitch muscle. Am. J. Physiol. 264(4 Pt 2): R816–R819, 1993.
13. Ferrannini, E., and A. Natali. Essential hypertension, metabolic disorders, and insulin resistance. Am. Heart J. 121(4 Pt 2): 1274–1282, 1991.
14. Galbo, H., J. J. Holst, and N. J. Christensen. The effect of different diets and of insulin on the hormonal response to prolonged exercise. Acta Physiol. Scand. 107: 19–32, 1979.
15. Handberg, A., A. Vaag, P. Damsbo, H. Beck-Nielsen, and J. Vinten. Expression of insulin regulatable glucose transporters in skeletal muscle from type 2 (non-insulin-dependent) diabetic patients. Diabetologia 33: 625–627, 1990.
16. Henriksen, E. J., R. E. Bourey, K. J. Rodnick, L. Koranyi, M. A. Permutt, and J. O. Holloszy. Glucose transporter protein content and glucose transport capacity in rat skeletal muscles. Am. J. Physiol. 259(4:Pt 1): E593–E598, 1990.
17. Karlsson, A. K., S. Attvall, P. A. Jansson, L. Sullivan, and P. Lonnroth. Influence of the sympathetic nervous system on insulin sensitivity and adipose tissue metabolism: a study in spinal cord-injured subjects. Metab. Clin. Exp. 44: 52–58, 1995.
18. Kjaer, M., G. Perko, N. H. Secher, et al. Cardiovascular and ventilatory responses to electrically induced cycling with complete epidural anaesthesia in humans. Acta Physiol. Scand. 151: 199–207, 1994.
19. Lillioja, S., A. A. Young, C. L. Culter, et al. Skeletal muscle capillary density and fiber type are possible determinants of in vivo insulin resistance in man. J. Clin. Invest. 80: 415–424, 1987.
20. Megeney, L. A., P. D. Neufer, G. L. Dohm, et al. Effects of muscle activity and fiber composition on glucose transport and GLUT-4. Am. J. Physiol. 264: E583–E593, 1993.
21. Mikines, K. J., B. Sonne, P. A. Farrell, B. Tronier, and H. Galbo. Effect of physical exercise on sensitivity and responsiveness to insulin in humans. Am. J. Physiol. 254: E248–E259, 1988.
22. Mikines, K. J., B. Sonne, P. A. Farrell, B. Tronier, and H. Galbo. Effect of training on the dose-response relationship for insulin action in men. J. Appl. Physiol. 66: 695–703, 1989.
23. Mohr, T., J. L. Andersen, F. Biering-Sorensen, et al. Long-term adaptation to electrically induced cycle training in severe spinal cord injured individuals [published erratum appears in Spinal Cord 35:262, 1997]. Spinal Cord 35: 1–16, 1997.
24. Nakajima, A., and S. Honda. Physical and social condition of rehabilitated spinal cord injury patients in Japan: a long-term review. Paraplegia 26: 165–176, 1988.
25. Nakajima, A., S. Honda, S. Yoshimura, Y. Ono, J. Kawamura, and N. Moriai. The disease pattern and causes of death of spinal cord injured patients in Japan. Paraplegia 27: 163–171, 1989.
26. Ploug T, Stallknecht BM, Pedersen O, Kahn BB, Ohkuwa T, Vinten J, Galbo H. Effect of endurance training on glucose transport capacity and glucose transporter expression in rat skeletal muscle. Am. J. Physiol. 259(6:Pt 1): E778–E786, 1990.
27. Ragnarsson, K. T., S. F. Pollack, W. O’Daniel, Jr, R. Edgar, J. S. Petrofsky, and M. S. Nash. Clinical evaluation of computerized functional electrical stimulation after spinal cord injury: a multicenter pilot study. Arch. Phys. Med. Rehabil. 69: 672–677, 1988.
28. Rosenthal, M., W. L. Haskell, R. Solomon, A. Widstrom, and G. M. Reaven. Demonstration of a relationship between level of physical training and insulin-stimulated glucose utilization in normal humans. Diabetes 32: 408–411, 1983.
29. Stout, R. W. Insulin and atheroma: 20-yr perspective. Diabetes Care 13: 631–54, 1990.
30. Stout, R. W. The impact of insulin upon atherosclerosis. Hormone Metab. Res. 26: 125–128, 1994.
31. Yekutiel, M., M. E. Brooks, A. Ohry, J. Yarom, and R. Carel. The prevalence of hypertension, ischaemic heart disease and diabetes in traumatic spinal cord injured patients and amputees. Paraplegia 27: 58–62, 1989.