The acute effects of glycemic control on nerve conduction in human diabetics

Clinical Neurophysiology 116 (2005) 270–274 www.elsevier.com/locate/clinph

The acute effects of glycemic control on nerve conduction in human diabetics Yuriko Kikkawaa,b, Satoshi Kuwabaraa,*, Sonoko Misawaa, Noriko Tamuraa, Yukiko Kitanoa, Kazue Ogawaraa, Takamichi Hattoria a

Department of Neurology, Chiba University School of Medicine, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan b Department of Neurology, Narita Red Cross Hospital, 90-1, Iida-cho, Narita 286-8523, Japan Accepted 27 August 2004 Available online 29 September 2004

Abstract Objective: To investigate acute changes in nerve conduction associated with glycemic control. In diabetes, nerve dysfunction can result from reversible metabolic factors associated with hyperglycemia, as well as structural changes. Methods: Multiple nerve conduction parameters including F-wave latencies were measured in 47 diabetic patients with prominent hyperglycemia before and after intensive insulin treatment. Results: Four weeks after the start of treatment, there was a significant improvement in minimal F-wave latencies of the median (P! 0.001) and tibial (P!0.001) nerves, and in distal latencies (PZ0.01) and sensory nerve conduction velocities (P!0.001) of the median nerves. Amplitudes of motor and sensory responses did not change significantly. These findings were similar for patients with type 1 (nZ8) and those with type 2 (nZ39) diabetes. Patients with poorer glycemic control or milder neuropathy tended to show greater changes after treatment. Conclusions: Glycemic control quickly alters the speed of nerve conduction. F-wave latencies and conduction times across the carpal tunnel are very sensitive parameters. Significance: Serial nerve conduction studies can detect reversible slowing of nerve conduction presumably caused by metabolic factors, such as decreased NaC/KC-ATPase activity, the altered polyol pathway, and tissue acidosis. q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Diabetic neuropathy; Nerve conduction study; F-wave; Glycemic control

1. Introduction Peripheral neuropathy is a very frequent complication in patients with diabetes mellitus (Dyck et al., 1993; Watkins and Thomas, 1998). The role of hyperglycemia in the pathogenesis of nerve dysfunction is not well understood, but metabolic factors that mediate acutely reversible functional impairment are suggested to contribute to axonal dysfunction in diabetic neuropathy (Greene et al., 1984, 1988; Low, 1987). The metabolic factors include decreased NaC/KC-ATPase activity, increased anaerobic glycolysis, accumulation of polyols, increased protein glycation, * Corresponding author. Tel.: C81-43-222-7171x5414; fax: C81-43226-2160. E-mail address: [email protected] (S. Kuwabara).

decreased myoinositol content, and nerve ischemia due to microangiopathy (Dyck et al., 1986; Greene et al., 1988; Quasthoff, 1998). Nerve conduction studies are the most sensitive, reliable, objective and non-invasive means of investigating diabetic polyneuropathy (Claus et al., 1993), and has been widely used for the assessment of diabetic polyneuropathy not only to evaluate the degree of abnormality but also to document serial changes in the clinical course in general and drug effect in particular (Kohara et al., 2000). Among nerve conduction parameters, F-wave latency is the most sensitive and reproducible (Andersen et al., 1997; Kohara et al., 2000). Moreover, the amplitude ratio between sural and radial sensory responses has been suggested to be a very sensitive parameter to detect neuropathy in diabetic patients (Pastore et al., 1999).

1388-2457/$30.00 q 2004 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2004.08.011

Y. Kikkawa et al. / Clinical Neurophysiology 116 (2005) 270–274

There are a number of reports investigating short-term changes in nerve conduction associated with glycemic control in diabetic patients (Fraser et al., 1977; Graf et al., 1981; Pietri et al., 1980; Ward et al., 1971), and some studies demonstrate a significant improvement in nerve conduction. However, these studies measured only conduction velocity, employing neither F-wave latency nor amplitude, and the numbers of patients were small (10–18 patients). To elucidate the acute changes in nerve conduction associated with glycemic control, and investigate which parameters are sensitive to detect the sequential changes, we performed extensive nerve conduction studies including analyses of F-waves and amplitude in diabetic patients with prominent hyperglycemia.

2. Methods

271

For sensory nerve conduction studies, median, superficial radial, and sural nerves were tested. Sensory nerve action potentials (SNAPs) were recorded using antidromic technique. Median SNAPs were recorded from the index fingers after wrist stimulation. For radial nerve studies, SNAPs were recorded at the anatomical snuff-box after stimulation at 12 cm proximal to the recording electrodes. For sural nerve studies, SNAPs were recorded at the lateral malleolus, and stimulation was delivered 14 cm proximally. Skin temperatures were monitored, and maintained above 32 8C on the forearm and 31 8C on the mid-leg using a heat lamp, if necessary. The meanGSD temperature was 33.4G1.1 8C on the arm and 33.1G1.3 8C on the leg before treatment, and 33.3G1.0 8C on the arm and 33.1G2.0 8C on the leg after treatment. Most of the nerve conduction studies were performed by two (Y.K. and S.K.) of the authors, who were not blinded to clinical and laboratory information. 2.3. Statistical analyses

2.1. Patients Forty-seven consecutive patients (17 women, 30 men) with diabetes mellitus, hospitalized for control of hyperglycemia during the study period (from January to December, 2003), were studied. Their mean age was 55 years (range 22–78 years), with the mean duration of diabetes of 8.6 years (range 1–30 years). Thirty-nine patients had type 2 diabetes, and the remaining 8 had type 1 diabetes. Of the 47 patients, 42 had mild to moderate polyneuropathy (decreased vibratory sensations at the ankle, absent Achilles tendon reflexes, or paresthesias in the distal extremities), and the remaining 5 had no evidence of neuropathy (normal neurological examination and nerve conduction study results). Hemoglobin A1c (HbA1c) levels ranged from 9.1 to 14.0% (mean 11.1%) before treatment. Based on patients’ history, and physical and laboratory examinations, causes of neuropathy other than diabetes were excluded. HbA1c measurements and electrodiagnostic studies were performed before and 4 weeks after the start of intensive insulin treatment (4 times a day, subcutaneous injection). 2.2. Nerve conduction studies Nerve conduction studies were performed using the conventional procedures and a standard electromyography machine (Viking 4, Nicolet Biomedical Japan, Tokyo, Japan). For motor nerve conduction studies, median and tibial nerves were tested. F-waves were elicited after stimuli at the wrist or ankle, and consecutive 16 responses were recorded. F-wave was defined as a potential with an amplitude exceeding 20 mV and a reasonable range for the investigated nerve, excluding spurious voluntary activity. The latency to onset of the first deflection from baseline was measured for each trace, and the shortest latency was determined as minimal F-wave latency (Andersen et al., 1997). F-wave latencies were corrected to height of 165 cm.

Almost all the values of nerve conduction parameters showed the Gaussian distribution except amplitudes of motor and sensory responses. However, because the purpose of this study is to investigate which parameters are sensitive to detect changes associated with glycemic control, all the parameters should be analyzed by the same statistical method. Therefore, differences in mean values of all nerve conduction parameters before and after treatment were tested by the Wilcoxon test. The non-parametric Spearman’s correlation coefficient rank test was used for correlation between variables. All statistical analyses were performed using the Statcel software for Windows 2001 (OMS Co. Ltd, Tokyo, Japan). 3. Results 3.1. All diabetic patients All 47 patients had marked hyperglycemia on admission (fasting blood glucose, meanGSD, 236G93 mg/dl; HbA1c, 11.1G1.3%). Four weeks after the start of intensive insulin treatment, fasting blood glucose decreased to 136G 42 mg/dl (meanGSD; P!0.001), and the level of HbA1c was 8.1G1.3% (P!0.001). There was an improvement in general symptoms such as general malaise, polyuria, and thirst in some patients, but neurological symptoms/signs did not change obviously, except two patients who experienced slight lessening of limb paresthesias after treatment. A summary of nerve conduction study results is given in Table 1. In motor nerve studies, there was a significant improvement in distal latencies and F-wave latencies of the median nerve, and in nerve conduction velocities and F-wave latencies of the tibial nerve. Amplitudes of compound muscle action potentials did not change significantly. In sensory nerve studies, median and radial conduction velocities became faster after treatment, but

272

Y. Kikkawa et al. / Clinical Neurophysiology 116 (2005) 270–274

Table 1 Changes in nerve conduction parameters before and after insulin treatment Before treatment Motor nerve Median Distal latency (ms) Conduction velocity (m/s) CMAP amplitude (mV) F-wave latency (ms) Tibial Distal latency (ms) Conduction velocity (m/s) CMAP amplitude (mV) F-wave latency (ms) Sensory nerve Median Conduction velocity (m/s) SNAP amplitude (mV) Radial Conduction velocity (m/s) SNAP amplitude (mV) Sural Conduction velocity (m/s) SNAP amplitude (mV) Sural/radial SNAP anplitude ratio (%)

After treatment

Table 2 Changes in nerve conduction parameters in patients with type 1 diabetes

P value

4.5 (0.9) 50 (6)

4.3 (1.0) 51 (6)

0.01 NS

7.2 (2.3)

7.5 (2.7)

NS

30.2 (3.8)

29.0 (4.1)

!0.001

4.7 (1.1) 40 (6)

4.6 (1.1) 43 (7)

NS !0.001

9.1 (4.8)

9.0 (4.9)

NS

52.5 (3.7)

50.5 (4.4)

!0.001

45 (8)

49 (10)

16.2 (13.3)

17.3 (10.3)

53 (6)

57 (6)

23.6 (11.0)

24.8 (13.9)

NS

46 (9)

48 (11)

NS

8.5 (7.3)

8.3 (6.1)

NS

38 (33)

44 (29)

NS

!0.001 NS

0.004

Motor nerve Median Distal latency (ms) Conduction velocity (m/s) CMAP amplitude (mV) F-wave latency (ms) Tibial Distal latency (ms) Conduction velocity (m/s) CMAP amplitude (mV) F-wave latency (ms) Sensory nerve Median Conduction velocity (m/s) SNAP amplitude (mV) Radial Conduction velocity (m/s) SNAP amplitude (mV) Sural Conduction velocity (m/s) SNAP amplitude (mV) Sural/radial SNAP anplitude ratio (%)

Before treatment

After treatment

P value

4.2 (0.5) 54 (7)

3.6 (0.4) 57 (6)

0.02 NS

7.8 (2.4)

8.9 (2.5)

NS

27.7 (2.5)

25.7 (2.3)

0.02

4.8 (0.9) 42 (6)

4.2 (0.7) 48 (7)

NS 0.02

13.4 (6.0)

14.5 (6.3)

NS

50.6 (5.0)

47.5 (5.0)

0.02

49 (8)

55 (9)

0.03

28.4 (20.7)

27.7 (10.6)

NS

51 (6)

57 (5)

0.03

25.3 (9.3)

35.0 (18.1)

0.05

49 (4)

53 (7)

0.02

15.1 (10.3)

12.6 (4.0)

NS

67 (55)

41 (20)

NS

CMAP, compound muscle action potential; SNAP, sensory nerve action potential; NS, not significant; data are given as mean (SD).

CMAP, compound muscle action potential; SNAP, sensory nerve action potential; NS, not significant; data are given as mean (SD).

other parameters including SNAP amplitudes and sural/ radial amplitude ratios did not change significantly. In summary, F-wave latencies and conduction times across the carpal tunnel (median distal latencies and sensory nerve conduction velocities) were sensitive parameters.

3.2.2. Effects of hemoglobin A1c levels The correlation of the extent of changes in electrophysiological parameters with HbA1c levels before treatment was analyzed. Whereas patients with a higher HbA1c level tended to show greater improvement in a number of parameters, only tibial F-wave latencies showed a significant correlation with HbA1c levels (PZ0.02). Next, the relationship of changes in electrophysiological parameters with changes in HbA1c levels was analyzed. A greater change in HbA1c levels was associated with greater improvement in median (PZ0.04) and tibial (PZ0.02) F-wave latencies.

3.2. Subgroup analyses 3.2.1. Patients with type 1 diabetes Because there is a possibility that type 1 and 2 diabetes may respond differently to glycemic control, data of patients with type 1 diabetes (nZ8) were analyzed separately. Table 2 shows changes in nerve conduction parameters in these patients. Results were similar to those of all diabetic patients (Table 1). F-wave latencies of the median and tibial nerves, and conduction times across the carpal tunnel (distal latency and sensory nerve conduction velocity in the median nerve) were, again, sensitive to detect changes associated with glycemic control.

3.2.3. Effects of baseline structural changes To look for the influence of the baseline structural changes on improvement in nerve conduction, patients were classified into the two subgroups according to the amplitude of sural SNAPs (Table 2). The cut-off value was 5 mV, which is the lower limit of normal in our laboratory.

Y. Kikkawa et al. / Clinical Neurophysiology 116 (2005) 270–274 Table 3 Changes in nerve conduction parameters in the two subgroups of diabetic patients Before treatment Sural SNAP!5 mV (nZ16) Median Distal latency (ms) 4.6 (2.8) F-wave latency (ms) 32.2 (1.2) Sensory conduction 44 (4) velocity (m/s) Tibial Motor conduction 40 (6) velocity (m/s) F-wave latency (ms) 55.1 (3.2) Sural SNAPO5 mV (nZ31) Median Distal latency (ms) 4.5 (0.8) F-wave latency (ms) 29.3 (2.2) Sensory conduction 46 (4) velocity (m/s) Tibial Motor conduction 40 (6) velocity (m/s) F-wave latency (ms) 51.5 (2.4)

After treatment

P value

4.5 (0.8) 31.1 (1.4) 45 (6)

NS NS NS

41 (5)

NS

53.2 (4.7)

0.03

4.2 (0.8) 27.9 (2.1) 50 (7)

0.003 0.001 0.004

43 (6)

0.02

49.4 (2.8)

!0.001

SNAP, sensory nerve action potential; NS, not significant; data are given as mean (SD).

Parameters, which showed a significant difference in Table 1, are compared in Table 3. In patients with a decreased sural SNAP amplitude, the extent of improvement was smaller, and a difference before and after treatment was significant only in tibial F-wave latency. In contrast, patients with a normal sural SNAP showed more obvious improvement after treatment.

4. Discussion Our results confirm that in diabetic patients, slowing of nerve conduction can quickly respond to glycemic control, and showed that F-wave latencies and conduction times across the carpal tunnel are sensitive parameters to detect the acute changes in nerve conduction associated with glycemic control. The extent of improvement in each parameter was small, but highly significant (Table 1). These findings suggest that nerve function is fluctuating according to the state of glycemic control in diabetic patients. The changes in nerve conduction associated with glycemic control were characterized by an increase in the speed of nerve conduction, but no obvious effects on amplitudes of CMAP or SNAP. The San Antonio Conference on diabetic neuropathy (Kahn et al., 1988) recommended the inclusion of F-wave analysis in the battery of electrodiagnosis. Recent reports showed the high reliability and sensitivity of F-wave analysis, although determination of minimal F-wave latencies is occasionally difficult, and could include a subjective interpretation

273

(Andersen et al., 1997; Kohara et al., 2000). Our findings indicate that F-wave test is useful to document short-term changes in nerve conduction in diabetic patients. The Fwave is a late response, and results from the backfiring of antidromically activated anterior horn cells (Kimura, 1989). The F-wave latency is a measure of the conduction time over a longer segment of the nerve. Thus compared with distal latency or nerve conduction velocity over a shorter segment, the same measurement error in latency contributes less in percentage. In contrast, amplitudes of CMAPs and SNAPs did not change significantly. This is presumably because glycemic control does not affect the number of conducting fibers. The sural/radial amplitude ratio therefore did not change significantly. This ratio is sensitive to detect a dying-back process in diabetic polyneuropathy (Pastore et al., 1999), but our results showed that it is not useful to demonstrate changes associated with glycemic control in diabetic patients. Another feature of our findings included significant improvement in motor distal latency and sensory nerve conduction velocity in the median nerve after treatment. These parameters reflect conduction times across the carpal tunnel. Median mononeuropathy is frequent in diabetic patients, and the incidence was reported to be 23% (Albers et al., 1996). The precise mechanisms for preferential nerve involvement at the carpal tunnel in diabetes are unclear, but at the common entrapment sites, intra-axonal NaC accumulation, mediated by increased sorbitol and paralysis of NaC-KC pump in diabetes, would further increase the intra-axonal pressure, and these regions may be particularly vulnerable to metabolic abnormalities in diabetes. This study confirms reversible, and probably, nonstructural factors in human diabetic neuropathy. Multiple metabolic factors such as intra-axonal NaC accumulation, tissue acidosis, and high serum osmolarity, could contribute to the reversible nerve dysfunction (Greene et al., 1984, 1988; Kuwabara et al., 2002). Further studies are needed to elucidate the mechanisms of functional nerve impairment in diabetic patients. Excitability testing using threshold tracking, which can provide insights into NaC and KC conductances in human axons non-invasively (Bostock et al., 1998; Burke et al., 2001), is potentially helpful to investigate the ionic mechanisms responsible for reversible conduction abnormalities in diabetic nerves.

References Albers JW, Brown MB, Sima AA, Greene DA. Frequency of median mononeuropathy in patients with mild diabetic neuropathy in the early diabetes intervention trial (EDIT). Muscle Nerve 1996;19:140–6. Andersen H, Sta˚lberg E, Flack B. F-wave latency, the most sensitive nerve conduction parameter in patients with diabetes mellitus. Muscle Nerve 1997;20:1296–302. Bostock H, Cikurel K, Burke D. Threshold tracking techniques in the study of human peripheral nerve. Muscle Nerve 1998;21:137–58.

274

Y. Kikkawa et al. / Clinical Neurophysiology 116 (2005) 270–274

Burke D, Kiernan MC, Bostock H. Excitability of human axons. Clin Neurophysiol 2001;112:1575–85. Claus D, Mustafa C, Vogel W, Herz H, Neundorfer D. Assessment of diabetic neuropathy: definition of norm and discrimination of abnormal nerve function. Muscle Nerve 1993;19:757–68. Dyck PJ, Karnes JL, O’Brien P, Okazaki H, Lais A, Engelstad J. The spatial distribution of fiber loss in diabetic polyneuropathy. Ann Neurol 1986; 19:440–9. Dyck PJ, Kratz KM, Karnes JL, Lichy WJ, Pach JM, Wilson DM, O’Brien PC, Melyon LJ, Service FJ. The prevalence by staged severity of various type of diabetic neuropathy, retinopathy, and nephropathy in a population-based cohort: the Rochester Diabetic Neuropathy Study. Neurology 1993;43:817–24. Fraser DM, Campbell IW, Ewing DJ, Murray A, Nielson JMM, Clarke BF. Peripheral and autonomic nerve function in newly diagnosed diabetes mellitus. Diabetes 1977;26:546–60. Graf RD, Halter JB, Pfeifer MA, Halar E, Brozovich F, Porte D. Glycemic control and nerve conduction abnormalities in non-insulin-dependent diabetic subjects. Ann Intern Med 1981;94:307–11. Greene DA, Yagihashi S, Lattimer SA, Sima AAF. Nerve NaC/KCATPase, conduction, and myo-inositol in the insulin deficient BB rat. Am J Physiol 1984;247:E534–E539. Greene DA, Lattimer SA, Sima AAF. Are disturbance of sorbitol, phosphoinositide, and NaC/KC-ATPase regulation involved in pathogenesis of diabetic neuropathy? Diabetes 1988;37:688–93. Kahn R, Asbury AK, Porte DJ, Genuth SM, Griffin JW, Hatler JB, Kimura J, Kuller LH, McLeod JG, Ochoa JL, Ward JD, Albers JW,

Arezzo JC, Benneett PH, Daube JR, Dyck PJ, Ewing DJ, Franklin GM, Gilliat R. Report and recommendation of the San Antonio Conference on diabetic neuropathy. Diabetes 1988;37:1000–4. Kimura J. Electrodiagnosis in diseases of nerves and muscles: principles and practice. Philadelphia, PA: Davis; 1989 p. 332–353. Kohara N, Kimura J, Kaji R, Goto Y, Ishii J, Takiguchi M, Nakai M. Fwave latency serves as the most reproducible measure in nerve conduction studies of diabetic neuropathy: multiple centre analysis in healthy subjects and patients with diabetic neuropathy. Diabetologia 2000;43:915–21. Kuwabara S, Ogawara K, Hattori T, Suzuki Y, Hashimoto N. The acute effects of glycemic contrtol on axonal excitability in human diabetics. Intern Med 2002;41:360–5. Low PA. Recent advances in the pathogenesis of diabetic neuropathy. Muscle Nerve 1987;10:121–8. Pastore C, Izura V, Emilio G-B, Dominguez J-R. A comparison of electrophysiological tests for the early diagnosis of diabetic neuropathy. Muscle Nerve 1999;22:1667–73. Pietri A, Ehle AL, Raskin P. Changes in nerve conduction velocity after six weeks of glucoregulation with portable insulin infusion pumps. Diabetes 1980;29:668–71. Quasthoff S. The role of axonal ion conductances in diabetic neuropathy: a review. Muscle Nerve 1998;21:1246–55. Ward JD, Barnes CG, Fisher DJ, Jessop JD, Baker RWR. Improvement in nerve conduction following treatment in newly diagnosed diabetics. Lancet 1971;7696:428–30. Watkins PJ, Thomas PK. Diabetes mellitus and the nervous system. J Neurol Neurosurg Psychiatry 1998;65:620–32.