Effects of sorbinil therapy in diabetic patients with painful peripheral neuropathy and autonomic neuropathy

Effects of Sorbinil Therapy in Diabetic Patients with Painful Peripheral Neuropathy and Autonomic Neuropathy

JONATHAN B. JASPAN, M.B.B.Ch., M.R.C.P. KEVAN HEROLD, M.D. CYNTHIA BARTKUS, R.N. Chicago,

Illinois

From the Department of Medicine, University of Chicago, and the Diabetes Research and Training Center of the University of Chicago, Chicago, Illinois. This work was supported by grant AM-20595 from the National Institutes of Health. This article includes a summary of material that has previously been published in the Lancet (1983; II: 758-765). Requests for reprints should be addressed tq Dr. Jonathan B. Jaspan, University of Chicago, Department of Medicine, Box 435, 5841 South Maryland Avenue, Chicago, Illinois 60637.

24

November

15,1985

The American

Journal

Clinical investigations with the aldose reductase inhibitor sorbinil in patients with peripheral neuropathy due to diabetes are described. After an improvement in motor and sensory nerve conduction velocities was demonstrated in asymptomatic diabetic patients taking sorbinil (compared with velocities during a placebo period), 11 patients with painful diabetic neuropathy were treated with sorbinil for three weeks without alterations in diabetic management or control. Therapy was placebo-controlled in a single-blind fashion in eight patients. Pajn (assessed by or on a zero to 20 rating scale), which had been constant for many months before entry into the study and unresponsive to numerous medications, improved from a mean score of 16 to 6 and returned when the drug was discontinued. Objective improvement in sensation and strength were observed in some cases. Improvements in nerve conduction velocity and cardiac autonomic function were also documented. Cardiac autonomic neuropathy was studied in 36 patients in a double-blind, placebocontrolled, randomized, noncrossover trial. Patients received one 250-mg sorbinil tablet or one placebo tablet daily for six weeks, after a one-week baseline period. Glycemic control did not change during the study period, as indicated by unaltered glycohemoglobin levels. Response was assessed by expiration-inspiration ratios, obtained on electrocardiography during six cycles per minute respiration, and by resting minima! heart rate, both measures of vagal function. In the sorbinil-treated group, expiration-inspiration ratios improved from 4 674 f 0.012 to 1.096 2 0.020 (p ~0.03). There was a slight decrease in the ratios in the placebo-treated group, from 1.112 + 0.023 to 1.105 + 0.023 (not significant). The difference between the Week 0 to Week 6 changes in each group was significant (p ~0.01). Resting minimal heart rate decreased in the sorbinil-treated group from 76.4 + 2.3 to 66.6 f 2.6 + 2.4 beats per minute (p
of Medicine

Volume

79 (suppl

5A)

SYMPOSIUM

PERONEAL MOTOR NERVE Washout-Sorbinil

6

o-l=l

-

REDUCTASE

INHIBITION-JASPAN

ET AL

With the observation of increased polyol pathway activity in the lens and peripheral nerve of animals and patients with diabetes and the suggestion that intracellular sorbitol accumulation may play a role in the pathogenesis of diabetic cataract and neuropathy, it was realized that the polyol pathway could be pharmacologically inhibited, thereby opening possibilities for a clinical application. Some of the earliest clinical trials with aldose reductase inhibitory agents supported this notion. improvements in sensory nerve function, assessed subjectively and objectively by clinical sensory scoring and sensory threshold determinations, and in motor nerve conduction velocities were reported in 30 patients with diabetes in a 12-week, double-blind, placebo-controlled trial with alrestatin (Ayerst 22, 284) [l]. Subsequently, in a randomized, double-blind, crossover trial in 39 patients, sorbinil, administered for nine weeks, produced a small but significant improvement in motor and sensory nerve conduction velocities when compared with the conduction velocities during an equivalent placebo period (Figures 1 and 2) [2]. These studies implicated the polyol pathway in human diabetic neuropathy. However, the overall contribution of sorbitol pathway activity to the development of diabetic neuropathy remained uncertain. The maximal improvement .in nerve conduction velocity occurred after three weeks of sorbinil therapy (Figure 2), which was considered to be compatible with a maximal effect, rapidly achieved, on some sorbitol-related reversible component of the neuropathy that overall made a small contribution to

which was sometimes accompanied by fever. These effects disappeared rapidly after discontinuation of the drug. The results of this study are promising, suggesting the possibility of a major role for aldose reductase inhibitors in the treatment of, and perhaps prophylaxis against, both somatic and autonomic components of diabetic peripheral neuropathy.

Sorbinil-Placebo

ON ALDOSE

T.'

. . .

lPlacebo 1 I

0 Figure 1. Effect of sorbinil on peroneal motor nerve conducf tion velocity in 39 patients with diabetes. The left-hand column shows the difference between the median values of each of the three measurements of peroneal nerve conduction velocity at three weekly intervals during the sorbinil period and of the three measurements during the placebo period for each patient (mean difference equals +0.70 + 0.24, p 4008). The right-hand, column shows the difference (mean difference equals -0.80 + 0.28, p ~0.008) between the nerve conduction velocity three weeks after discontinuing sorbinil therapy (drug washout period) and nine weeks after starting sorbinil in each patient. Reproduced with permission from [2].

November

15, 1985

3

1 f

6

9 Time

. F‘igure

_

Sorbinil I I

12

15

Wash1 out I

18

21

(weeks)

-.

2. Time course of the effect of sorbinil on peroneal nerve conduction velocity in 20 patients with diabetes who received placebo followed by sorbinil. Peroneal nerve conduction velocityhad increased at three weeks after the initiation of sorbinil treatment, remained stable during the sorbinil period, and then decreased within three weeks after the end of the treatment (drug washout period). The standard errors for each point were calculated from the residuals .of a repeated-measure analysis of variance and represent only within-subject variability. Reproduced with permission from

m.

The

American

Journal

of Medicine

Volume

79

(suppl 5A)

25

SYMPOSIUM

TABLE

l l l l

l

l

ON ALDOSE

I

Patient

REDUCTASE

INHIBITION-JASPAN

Characteristics

(Study

ET AL

I)

sion in the study included absence of major systemic diseases on physical examination, normal serum chemistry profile and blood cell counts, and absence of other possible causes of neuropathy, including alcoholism, nutritional deficiency, neurotoxic drug use, renal failure, and significant peripheral ischemic vascular disease. Patient characteristics are shown in Table I. Of note, four patients (all men) had signs and symptoms characteristic of severe diabetic amyotrophy, with marked weight loss and proximal leg muscle weakness in addition to severe pain. Involvement was symmetric in two patients and asymmetric in the other two. All 11 patients had distal, predominantly sensory, polyneuropathy, and two had clinical manifestations of severe autonomic neuropathy, which included refractory gastroparesis in one. Chronic polyneuropathy was present in all but one patient, who had acute-onset neuropathy that developed soon after a diagnosis of diabetes was made. Informed written consent was obtained from all patients. The severity of pain made it impossible to have a placebocontrolled period for every patient or to rigidly define the length of the placebo period. Eight patients received placebo in a single-blinded fashion for periods varying from four days to three weeks before going on to sorbinil, at a dose of

Six men, five women Seven type I diabetics, four type II diabetics; all receiving insulin Mean age 42 years (range 28 to 74) Mean duration of diabetes 16 years (range 6 months to 36 years) Distal, predominantly sensory, polyneuropathy associated with severe, lower-extremity neuropathic pain, which was constant for at least three months; pain unresponsive to numerous medications Four men with severe amyotrophy

the neuropathy (or exerted a small influence on the neuropathic process). However, an alternative interpretation was that nine weeks represents a very brief period in the pathogenesis of diabetic neuropathy and that administration of sorbinil for longer periods might increase the magnitude of the changes. In any event, these studies provided a framework for subsequent clinical studies with sorbinil in patients with diabetic neuropathy. Disabling pain is perhaps the most frequent symptom of diabetic neuropathy and is often the major or sole symptomatic component of peripheral symmetric sensorimotor polypeuropathy, amyotrophy, and abdominal radiculopathy in patients with diabetes. In the first study to be described, the clinical benefits of aldose reductase inhibition in symptomatic somatic neuropathy were examined. PATIENTS

AND METHODS

(STUDY

250

I)

Eleven patients with diabetes and severe lower-extremity pains associated with neuropathy were included in a clinical trial of sorbinil. In all cases, pain had been constant for at least three months prior to the study and was responsible for major sleep interference. All patients had consulted several physicians and received nymerous medications, alone or in combination, for a period of at least four months and, in most cases, for more than six months. These included most of the drugs commonly used in the treatment of painful neuropathy, ranging from simple and narcotic analgesics, vitamins, hypnotics, and sedatives to anticonvulsants, phenothiazines, and tricyclic antidepressants. In all cases, minimal, if any, pain relief had been obtained with use of these agents for periods ranging from three to four weeks or lonier. Criteria for inclu-

No

-1 __ 2 __ 3 __ 4 __ 5 _ --.---------6 7

Pain

Figure

26

3.

s

7

1

i

h

t

m

Rating scale for pain. Reproduced

November

15, 1985

The American

Journal

mg once

per day,

8

9

10 -- 11 .- 12 .- 13 -- 14'

0

d

e

with permission

of Medicine

for three

weeks

in sjx patients

and four

weeks in two. Placebo was readministered in a single-blind fashion to six of these eight patients for three weeks after discontinuation of sorbinil therapy. Placebo tablets were visually indistinguishable from sorbinil tablets. The three patients who received no placebo participated in an open trial of the drug for periods of five weeks, three weeks, and ten days, respectively. Sorbinil, a noncompetitive inhibitor of the enzyme aldose reductase, is rapidly and almost totally absorbed, has a plasma half-life of 34 to 52 hours, and reaches plateau levels six to 10 days after commencement of a singledaily-dosage regimen [3]. With such dosing, almost total inhibition of aldose reductase over a 24-hour period has been demonstrated [4]. Because of the association between improvement in glycemic control and improvement in nerve function [5,6], only patients in whom borderline or better glycemic control was stable were selected, and no attempts to improve control were made during the study. Patients selected for this investigation agreed not to take any analgesics, sedatives, hypnotics, or other medications during the study period. A physical examination, including neurologic assessment, was performed at the start of the trial, after the baseline or placebo period, upon completion of treatment, and three weeks later. Pain was assessed by means of a pain rating scale (Figure 3), as suggested by Scott and Huskisson [7].

F

a

t-

e

from [l 11.

Volume

79

(suppl 5A)

15-. 16 _----.--17 s

e

v

18

19

20

Worst

e

r

c

Pain

SYMPOSIUM

Patient

Weeks

Patient

1

REDUCTASE

INHIBITION-JASPAN

ET AL

Patient 5

Patient 4

Patient 3

2

ON ALDOSE

-

0 % 0

2

4

6-10

2

4

6-3

0

2

4

6-2

0

2

4

6

8-2

0

2

4

6 -2

0

1.5

4 5

igure 4. Individual pain scores in 11 patients with diabetes. Dotted areas represent placebo periods, white areas represent sorbinil therapy, and cross-hatched areas represent periods when neither placebo or sorbinil was administered. Reproduced with permission from [I 11.

Patients were familiarized with the scoring system upon entry into the study, at which time they recorded the score of their dominant symptom or symptoms. Every morning during the trial they recorded an average pain score for the previous 24 hours on specially constructed charts. An average daily score was calculated for each week of the trial, and this was recorded during each of the four outpatient clinic visits. Blood glucose and total glycosylated hemoglobin levels were also determined during these visits. Conduction velocities and cardiac autonomic function tests were each performed in seven patients at the end of the placebo (or baseline) period, the end of the treatment period, and three weeks later. Cardiac autonomic nerve function was assessed by measurement of the R-R interval on an electrocardiogram during slow respiration, at six cycles per minute, while in a supine position, after a 30-minute period of rest. The ratio of the longest R-R interval after expiration to the shortest after inspiration was obtained for each of eight respiratory cycles, and a mean expiration-inspiration ratio was obtained at each point in the study. Respiratory variation in heart rate as assessed by expiration-inspiration ratios (i.e., heart rate differences between inspiration and expiration) is predominantly a measure of parasympathetic (vagal) nerve function [8]. Nerve conduction velocity was measured in the peroneal and median motor nerves and the median sensory nerve. The conduction velocities of the peroneal and median motor nerves were measured with the use of the superimposed wave technique after double stimulation, which has approximately one third of the variability of standard methods [9]. The conduction velocity of the median sensory nerve was determined antidromically by stimulating the nerve at the elbow and recording impulses from the second and third fingers [lo].

November

15, 1985

RESULTS Pain Scores. Symptomatic responses in each of the 11 patients during the course of the study, assessed by pain scores, are illustrated in Figure 4. Baseline pain scores (range: 14 to 20) were constant in all cases, as they had been for at least the previous three months and for the previous six months or more in the majority of patients. Eight of the 11 patients reported a substantial diminution of pain, which generally began on the third or fourth day of treatment with sorbinil; five of the eight did not know that they were receiving sorbinil. One patient (Patient 8) had no relief of pain, and two others (Patients 5 and 7) had equivocal relief of pain. After discontinuation of sorbinil, five of the eight responders reported a recurrence of pain of about pretreatment severity; two reported a recurrence of pain that was perceptibly less severe than before therapy, and one patient (Patient 9) continued to show improvement. In one of the equivocal responders, pain appeared to worsen slightly after discontinuation of sorbinil. In six of the seven responders whose pain became more severe after discontinuation of sorbinil, there was approximately a one-week delay before pain relapsed. This observation and the continued improvement in one patient suggested a possible carryover effect of sorbinil. In all patients reporting a major reduction in pain, sleep patterns were noted by the patients, spouses, or roommates to be substantially improved and, in some cases, normalized. Moreover, the mood of seven of these .patients, which had been unchanged or worse during the placebo period, was noted by family and friends to be improved, The American

Journal

of Medicine

Volume

79

(suppl 5A)

27

SYMPOSIUM

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

+4

-12

-16

Baseline

End of Sorbinil

3 Weeks Later

rgure 3. uverail pam responses arter baseline pain scores were standardized to zero. Negative scores represent improvement (end-of-sorbinil score represents average daily pain score during last week of sorbinil therapy). Reproduced wiih permission from [I 11.

along with morale, during sorbinil therapy. Psychologic improvement in these seven patients was also clinically obvious while fhey were receiving sorbinil. Since pain sensitivity varies among individuals, baseline pain scores were standardized to zero to facilitate the assessment of change across the group. The overall and relative response in each patient is shown in Figtire 5. The average daily pain scores during the last week of sorbinil therapy were taken as the “end of sorbinil” scores. The mean -t standard error (SE) of improvement in pain scores in the group as a whole was 7.7 ? 1.3. The corresponding value obtained using the average pain score for the entire treatment period was 6 + 1. This difference reflects the tendency for there to be a delay in the onset of drug action. In the three weeks after discontinuation of drug treatment, the pain scores increased by 6.9 2 1.2 for the treatment group as a whole, excluding the two patients in whom pain did not recur after discontinuing sorbinil and

28

November

15, 1985

The American

Journal

of Medicine

the one patient in whom there was no response to therapy. It was interesting that, as a group, the patients with amyotrophy showed the most dramatic response, with improvements in pain scores of IO + 1.4. It was noteworthy that in the patient who had no change in pain response over the three-week period of sorbinil use (Patient 8) and in one of the equivocal responders (Patient 7), decreases in pain were noted with prolonged administration of sorbinil after the conclusion of this trial (beginning at four to five weeks for Patient 8 and at eight weeks for Patient 7). After four months, their scores of 16 and 13 declined to three and one, respectively, and persisted at these low levels during the ensuing one year of therapy. Three patients had objective evidence of improvement in sensory perception; of these, two also reported subjective improvement in sensation, as did one other jpatient. Improvement in strength was noted by all four patients with the amyotrophy syndrome, and this was subdtantiated by neurologic examination, which revealed an Increase in the power of hip flexion and knee extension from two to three out of five to three to four out of five. Three weeks after discontinuation of drug therapy, strength had deteriorated in three of these four patients and was maintained in the fourth. Sensory testing showed deterioration in one patient. No improvements in ankle reflexes, which were universally absent, were noted during the study. Other patients with severely painful diabetic neuropathy who have received sorbinil on a compassionate basis have reported a reduction in pain. In a number of these patients, there have been periods of interruption of medication. In all cases, pain has reportedly worsened after discontinuation, with the first noticeable deterioration occurring rather consistently at two to three weeks and a return to pre-sorbinil levels at four to six weeks. Pain remained at these levels until sorbinil was readministered. After restarting sorbinil, pain levels again decreased, usually beginning at the end of the first week and steadily returning to their previous nadir at about three to four weeks. Peripheral and Autonomic Nerve Function. lmprovements in nerve conduction velocities (at two or three sites) were documented in a majority of these patients; these improvements were reversed after discontinuation of therapy. It was noteworthy that improvement in nerve conduction velocity was most striking in three of the four patients with the most dramatic clinical response to sorbinil (Patients 2, 3, and 9), which was of the order of IO m per second jn some cases. In each of these patients, clinical deterioration after discontinuation of sorbinil was paralleled by deterioration in nerve conduction velocities. A carryover effect was apparent, however, as some of the improvement in nerve conduction velocities persisted three weeks after discontinuation of therapy [11,12]. There was, in addition, evidence of improvement in car-

Volume

79

(suppl5A)

SYMPOSIUM

disc autonomic nerve function in these patients during treatment with sorbinil. Thus, expiration-inspiration R-R interval ratios, an index of vagal function, which were markedly subnormal, improved in all but one patient, and deteriorated upon discontinuation of therapy. These changes were significant for the group as a whole. Similarly, resting heart rate, another measure of vagal function, declined significantly with sorbinil therapy and increased again after cessation of treatment. These observations suggested sorbinil-related improvements in autonomic nerve function. Patient Preference. Upon completion of the study, patients were asked the questions, “With which agent did you feel better?” and “Would you want to continue taking this medication for the relief you experienced?” Eight of the 11 patients said they were clearly better during treatment with sorbinil, and three reported no noticeable differences. Seven of the eight who claimed a benefit from sorbinil expressed a desire to continue taking it. No sorbinil toxicity was observed in these 11 patients, although a pruritic, macular, erythematous rash, a wellrecognized toxic effect of the drug [2,11,13], developed in a 12th patient, a 35year-old woman, on the sixth day of therapy, and she was withdrawn from the study. No biochemical or hematologic abnormalities were detected, and the rash disappeared uneventfully after four days. It was, however, of interest that this patient claimed to have noted pain relief and was reluctant to stop the medication despite the presence of a florid rash over much of her body surface. Thus, although this study was not placebo-controlled in three cases, in the other eight ‘cases, patients did not know whether they were receiving the active drug. Furthermore, all patients had previously received many other medications for weeks or months without relief of pain. This investigation suggested that a potent aldose reductase inhibitor drug is capable of producing considerable relief of neuropathic pain in patients with diabetes. AUTONOMIC

NEUROPATHY

Autonomic neuropathy represents a major source of morbidity in patients with diabetes [14,17] and is also associated with increased mortality [18]. Although the pathogenesis of autonomic neuropathy remains uncertain, it tends to develop parallel to and sometimes precedes somatic neuropathy, is frequently generalized, and is widely believed to arise from a generalized metabolic disturbance [14,19]. Activity of the polyol pathway in autonomic nerves has also been demonstrated. In one of the first clinical trials of an aldose reductase inhibitor, the possibility of autonomic benefits was suggested [I]. During the study of painful neuropathy described earlier, sorbinil-related improvements in cardiac autonomic nerve function were suggested by virtue of increased cardiac beat-to-beat variability and slowing of resting heart rate [ll]. More re-

November

15, 1985

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

cently, the influence of sorbinil on autonomic nerve function was investigated in a double-blind, placebo-controlled trial. The results of this study, which indicated a beneficial effect of aldose reductase inhibition on cardiac autonomic neuropathy, are described next. PATIENTS

AND METHODS

(STUDY

II)

Thirty-six patients with diabetes and objective clinical evidence of moderate to severe distal, symmetric sensorimotor neuropathy were selected for the study. Patients were selected from the diabetic population at the University of Chicago Diabetes Research and Training Center or from those referred from outside the institution for evaluation of neuropathy. inclusion criteria required absence of uremia, malnutrition, severe peripheral ischemia, alcoholism, the use of neurotoxic drugs, other possible causes of neuropathy, and any major systemic illness. Numerous reports have described the coexistence of autonomic neuropathy, as evidenced by subnormal cardiac beat-to-beat variation with respiration, and severe peripheral neuropathy in patients with diabetes [15,19]. This was confirmed in the current study. Informed written consent was obtained from patients in the study, which was approved by the Institutional Clinical Investigation Committee. For the duration of the study, patients agreed not to consume alcoholic beverages with a total daily content of more than 30 ml of ethyl alcohol and to abstain from taking any medication known to influence nerve function. Normal renal and hepatic biochemistry test results and hematologic indices were demonstrated in screening evaluations. Women of childbearing age or with childbearing potential were excluded from the study. The 36 patients consisted of 24 men and 12 women, with a mean age of 52.5 t 2.3 years, mean duration of known diabetes of 17 + 1.9 years, and mean body weight of 79.9 + 2.2 kg. Antihyperglycemic therapy consisted of insulin in 29, an oral antidiabetic agent plus diet in five, and diet alone in two. No attempt to improve glycemic control was made during the study period. Six patients were receiving beta-adrenergic blocking therapy, of whom four were in the treatment group and two were in the placebo group. The study was conducted from November 1983 to June 1984. Patient characteristics are provided in Table II. There were no significant differences between the groups in terms of age, gender, body weight, diabetes type, duration, treatment, or associated complications. Patients were randomly assigned, by means of computergenerated random numbers, to receive sorbinil or placebo. After a baseline period of one week, during which time all patients ingested identically appearing placebo tablets, Group I patients received one 250-mg tablet of sorbinil and Group II patients one placebo tablet every morning for six weeks. During the study period, patients were requested not to alter their dietary habits or antihyperglycemic therapy and not to start or discontinue any other medications. Patients were evaluated one week before the study began, at the beginning of the study (Week 0), and at weeks one, two, three, and six. At each visit, blood cell counts and serum chemistry tests were done. Random, rather than fasting, plasma glucose levels were measured because of scheduling considerations that made it impossible to have patients come

The American

Journal

of Medicine

Volume

79

(suppl 5A)

29

SYMPOSIUM

TABLE

II

Group

ON ALDOSE

Clinical

REDUCTASE

INHIBITION-JASPAN

Characteristics

Gender

of Patients

ET AL

(Study

II)

Aoe(Yea@

Weight (kg)

Duration of Diabetes (years)

Associated Complications Diabetes Type

Diabetes Treatment

Sorbinil (n = 19)

13 men 6 women

52.7 t 3.7 (17-74)

79.7 t 3.1 (55-96.8)

15.2 rfr 1.8 (2.5-30)

7 type 12 type

1 II

16 insulin 3 oral agent plus diet

Placebo (n = 17)

11 men 6 women

52.2 f 2.8 (33-69)

80.1 r+ 3.3 (59.5-l 19)

19.0 + 3.4 (l-52)

8 type 9 type

I II

13 insulin 2 oral agent plus diet 2 diet alone

Range is given in parentheses *Less than 1 g per 24 hours.

below

mean

November 15,1985

8 unaffected 9 background retinopathy 2 proliferative retinopathy 8 unaffected 7 background retinopathy 2 proliferative retinopathy

Proteinuria

diabetic

14 unaffected 5 mild*

diabetic

14 unaffected 3 mild*

value.

in at different times of the day in a fasting state. Glycosylated hemoglobin, red blood cell sorbitol, and plasma sorbinil levels were assessed at Weeks 0 and 6. Red blood cell sorbitol levels were measured by a modification of methods previously described [20], and plasma sorbinil levels by high-pressure liquid chromatography after extraction of serum samples at a neutral pH in diethyl ether 131. Total glycosylated hemoglobin levels were measured by the thiobarbituric acid method [21] and plasma glucose levels by the glucose-oxidase method. At Weeks 0 and 6, cardiac autonomic nerve function was assessed by means of electrocardiographic tracings taken during steady constant breathing at six breaths per minute over two minutes, as previously described [22-241. In each case, tracings were taken after patients had been in a supine position for 30 minutes in temperature-controlled rooms. After familiarizing patients with the procedure, the two-minute recording period was preceded by a practice period of one minute, during which the patients breathed at the timed rate. Thereafter, the recording period was initiated without the patients’ knowledge. An expiration-inspiration ratio was calculated for each of the twelve breath cycles, which was derived from the ratio of the longest R-R interval after expiration to the shortest R-R interval after the corresponding inspiration. Therefore, a mean (and standard error) expiration-inspiration ratio for the 12 respiratory cycles was obtained for each patient. In addition, since maximal heart rate slowing is under parasympathetic (vagal) control, a minimal heart rate was obtained for each patient by obtaining the mean of the slowest heart rate during the expiratory phase of each of the 12 breath cycles. Normal values for these cardiac parameters have previously been described [22,23,27-301. Nevertheless, to facilitate the evaluation of the data in our patients and to validate our methods, we performed similar tests on a small number of normal volunteers, represented by physicians, nurses, technicians, and secretaries at our institution (six men, six women). Although these controls were somewhat younger (43.6 2 4.5 years) and of lower weight (76.5 ? 2.8 kg), these differences were not statistically significant. All electrocardiograms were interpreted and the corre-

30

Retinopathy

The American Journal of Medicine

sponding calculations performed by individuals who were uninformed as to the patients’ identities and group assignments. Statistical Analysis. For all the variables measured, the difference between the Week 0 and Week 6 data was calculated for each patient. A mean and standard error of the mean were obtained for the treatment group and the placebo-treated group; the differences between the two groups were assessed by the non-paired Student t test. Within-group changes were evaluated by paired t tests, using the Week 0 and Week 6 data. In both cases, t tests were done as two-tailed tests, significance being set at a p value of less than 0.05. Relationships between cardiac autonomic response and levels of red blood cell sorbitol, serum sorbinil, glycosylated hemoglobin, and plasma glucose were assessed by regression analyses. RESULTS Because two patients dropped out and replacements were randomly assigned to groups, it transpired that 19 patients received sorbinil and 17 received placebo. Electrocardiograms in two patients, one from each group, could not be analyzed because of frequent atrial or ventricular extrasystole activity. Another patient from the placebo-treated group was also omitted from the analysis because electrocardiography was inadvertently not performed at Week 0. Thus, the treatment group comprised 18 patients and the placebo group 15 patients. Data are provided on three cardiac autonomic parameters. First, respiratory variability in heat-l rate was expressed as expiration-inspiration ratios, as described in the Patients and Methods section. Second, this variability was also expressed as the difference between the maximal R-R interval in milliseconds (occurring during expiration) and the minimal R-R interval (inspiration). Third, a minimal heat-l rate was obtained from the slowest rate during expiration, as described earlier. The baseline expiration-inspiration ratio was 1.091 *

Volume

79 (suppl

5A)

SYMPOSIUM

TABLE

III

Changes in Cardiac with Diabetes

Autonomic

Nerve Function

Parameters

ON ALDOSE

in Sorbinil-

Sorbinil (n = 18) Week 0 Expiration/inspiration R-R interval ratio Mean maximal minus minimal R-R interval (milliseconds) Minimal heart rate (beats/minute)

1.074 54t

* 0.012 10

76.4 t 2.3

NS = not significant. *Paired t test of within-group changes. tTwo-sample Student t test of Week 0 minus

p Value’

t 0.020

<0.03

17

CO.01

82?

66.8 + 2.4

Week

Week 0

and Placebo-Treated

ET AL

Patients

between

0.013 in the patients and 1.295 -t 0.046 in the control subjects (p
15, 1985

Week 6

p Value

f 0.023

1 ,105 t 0.023

NS


16

73 r 14

NS

CO.003

77.9 k 3.9

77.5 t 3.3

NS

10.001

1.112 782


6 differences

November

INHIBITION-JASPAN

Placebo (n = 15)

Week 8 1.096

REDUCTASE

p Value+

groups.

treated patients (of a lesser magnitude than the majority of sorbinil-related improvements), virtually no change in six, and deterioration in seven. Although resting heart rate and heart rate variability during quiet deep breathing are predominantly under parasympathetic control, beta-adrenergic blockade, even if partial, might influence the integrated reflex autonomic control mechanisms for intrinsic cardiac rate variability. In consideration of this, we analyzed the data excluding the six patients receiving beta blocking drugs (27 patients: 14 in the sorbinil-treated group, 13 in the placebo-treated group). In the sorbinil-treated group, the expiration-inspiration ratio increased from 1.084 r 0.015 to 1 .114 IL 0.024 (p ~0.02) and in the placebo-treated group, it decreased from 1,118 t 0.026 to 1.111 + 0.026 (not significant). The difference between the within-group changes was significant (p ~0.01). The measure of R-R variability during baseline, at 65 -t 9 milliseconds, was markedly subnormal in the patients with diabetes compared with the value of 233 ? 36 milliseconds in the control subjects (p
The American

Journal

of Medicine

Volume

79 (suppl

5A)

31

SYMPOSIUM

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

A

I

SORBINIL

Heart Rate/min

A E/I Ratio

70

- Week

6

B 1 PLACEBO

1

B

11

I

+0.140

PLACEBO

I

+ 0.120 +0.100

-I -I

Week0

tt-

u

November

70

70

50

50

+0.060 AEII Ratio

Week

15, 1985

90

+ 0.080

The American

Heart Rateimin

6

Figure 6. Expiration-inspiration ratios derived from electrocardiographic tracings (see text) at Week 0 and Week 6 in sorbinil- (top) and placebo-treated (bottom) patients with diabetes. For graphic representation, Week 0 values have been transformed to zero and data plotted as the change in the expiration-inspiration ratio.

32

90

Journal

of Medicine

Week

0

Week

6

Flgure 7. Resting minimal heart rate at Week 0 and Week 6 in sorbinil-treated (top) and placebo-treated (bottom) patients with diabetes.

Volume

79 (suppl

5A)

SYMPOSIUM

TABLE

milliseconds, not significant), and a significant difference between the within-group changes (p cO.003). Baseline minimal heart rate, which was used as an index of maximal vagal tone (see Patients and Methods section), was 77.2 t 2.2 beats per minute in the patients with diabetes, compared with 59.6 + 3.3 beats per minute in the control subjects (p
TABLE

V

Erythrocyte

Sorbitol

Levels

REDUCTASE

INHIBITION-JASPAN

Mean Total Glycosylated Levels (percent) Week 6

Week OWeek 6 Values

11.5 i 0.6 NS’ 10.6 k 0.6 NY

NS = not significant. *Compared with Week

ET AL

Hemoglobin

Week 6

Sorbinil Placebo

:;:;

; ;:j)

NS

0 levels.

ated hemoglobin values from baseline to Week 6 in each group indicated essentially stable glycemic control during the two months prior to entry into the study and during the six-week period of study. Plasma glucose changes within each group and Week 0 to Week 6 glucose differences between the two groups were also not significantly different. Linear regression analysis showed no relationship between changes in glycosylated hemoglobin or plasma glucose levels and changes in red blood cell sorbitol levels or cardiac autonomic responses. Red blood cell sorbitol levels are shown in Table V. Complete data were available for 13 sorbinil- and 13 placebo-treated patients. It is evident that there was a significant reduction in red blood cell sorbitol levels in the sorbinil-treated group (p
(n&l/g hemoglobin)

Week 0 Sorbinil Placebo

IV

ON ALDOSE

Week OWeek 6 Values

Week 6

(n = 13) (n = 13)

17.91 r 3.18 p
NS = not significant.

November

15, 1985

The

American

Journal

of Medicine

Volume

79

(suppl5A)

33

SYMPOSIUM

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

of a macular, erythematous rash and fever on the tenth day of treatment. The latter patient had no biochemical or hematologic abnormalities and recovered uneventfully after discontinuation of sorbinil. Each of these patients was replaced by other patients who received the next randomized drug assignment number. In all other patients, the medication was well tolerated, and there were no untoward clinical or biochemical effects. COMMENTS

Widespread autonomic nerve dysfunction has repeatedly been demonstrated in patients with diabetes [14,16,17,19]. In recent years, extracardiac manifestations of diabetic autonomic neuropathy have received some attention [14,19], although the most well-documented and best-studied aspect of autonomic dysfunction in patients with diabetes had been that related to the cardiac autonomic nerves [8,18,22,25-271. Impairment of vagal control of the heart or’ a decreased vagal tone is commonly encountered in patients with diabetes [8,26,31]. Although usually asymptomatic, it may manifest as resting tachycardia, pa/nless myocardial infarction, or cardiac arrhythmia and sometimes spontaneous cardiac arrest. During the 197Os, a number of relatively simple bedside tests of cardiac autonomic nerve function were developed and refined [8,18,22,25-27,321. Of the various tests currently in use, measurement of heart-rate variability during deep breathing, generally at a rate of six breaths per minute, has been shown to be the most practical, reproducible, and sensitive method [22,26,27]. Coefficients of variation for repeated tests on the same subject on the same or different days are less than 10 percent in normal subjects [22,32] and approximately 5 percent or less in patients with diabetes and autonomic neuropathy [22,23,25-271. It has been pointed out that this particular method sensitively and accurately reflects parasympathetic or vagal nerve function [8,25-271. However, the neurogenic control of sinus arrhythmia involves a complex reflex process that includes afferent pathways, central processing, and efferent pathways. The sympathetic system is likely to have some influence on this overall mechanism, and it should not be considered to be purely under parasympathetic control. Malone and associates [20] recently demonstrated that erythrocyte sorbitol levels reflect those in the nerve and other polyol pathway tissues of rats with streptozotocininduced diabetes. These investigators suggested that reduction of red blood cell sorbitol levels in diabetic animals and diabetic patients receiving sorbinil is likely to reflect comparable effects of the drug in less accessible tissues affected by long-term diabetic complications. Red blood cell sorbitol levels in the current investigation were similarly elevated in both groups of patients prior to initiation of treatment and were comparable to those in the pla-

34

November

i&1995

The American

Journal

of Medicine

cebo-treated diabetic patients in the study by Malone et al [20]. In the current study, erythrocyte sorbitol levels in the sorbinil-treated group significantly decreased from 47 to 18 nM/gm hemoglobin, levels very similar to those reported by Malone and associates [20] in diabetic patients receiving the same dose of sorbinil for two weeks. The similarity between red blood cell sorbitol levels after two weeks of sorbinil in the study by Malone et al [20] and after six weeks in the current study suggests steady-state suppression of tissue sorbitol levels once plateau levels of ttie drug have been obtained after six to 10 days [3]. In accordance with the suggestions of Malone et al 1201,the red blood cell sorbitol data in the current study provide evidence for the desired metabolic blockage in sorbitol pathway activity in the autonomic nerves, specifically the vagus, of the sorbinil-treated patients. In addition, this data, combined with the data pertaining to plasma sorbinil levels, confirmed that appropriate medication was ingested by patients in each group. It is well recognized that an improvement in glycemic control is associated with improved peripheral nerve function [5,6] and presumably, although not directly studied, autonomic nerve’function as well. In this regard, it is noteworthy that there were no differences in glycemic control between the two groups of patients at entry, as assessed by total glycosylated hemoglobin levels, which indicates that there was no randomization bias. Although a less reliable indicator of overall glycemic control, random plasma glucose levels were also similar between the two groups at entry. Differential alterations in glycemic control as an explanation of the differences observed between the two groups is implausible as there were no significant changes in glycosylated hemoglobin or random plasma glucose levels in either group during the six-week study period. In addition, for both parameters, the difference between Week 0 and Week 6 were insignificant. In this investigation, significant improvements in vagal nerve function, as assessed by expiration/inspiration ratios and mean maximal minus minimal R-R variations during deep breathing, and a decrease in basal minimal heart rate were demonstrated in diabetic patients receiving aldose reductase inhibitor therapy during a period of six weeks. Previous investigations have demonstrated sorbinil-related improvements in nerve conduction velocity [2] and symptomatic neuropathy [I 1,131 in patients with diabetes. An increased basal heart rate in patients with diabetes compared with that of age- and gender-matched control subjects is widely recognized. The observations in this study were similar to those of Ewing et,al [29], who demonstrated that resting heart rate in normal subjects between the ages of 25 and 54 years was 73 beats per minute in men and 77 beats per minute in women, with rates in age- and gender-matched patients with diabetes of 87

Volume

79

(suppl 5A)

SYMPOSIUM

and 89 beats per minute, respectively (p ~0.001 in each case). Although the patients in this study were slightly older than those studied by Ewing et al [29], these rates, when expressed as a mean rather than as a minimal resting heart rate, are comparable to those of the patients described by these investigators. Indeed, other researchers have made similar observations regarding adult patients with diabetes [22,23,30,31]. Recently, in a pediatric population of diabetic patients (mean age 14 years), who, although it was not stated, obviously had diabetes of relatively short duration, smaller but nevertheless significant increases in resting heart rate were reported when compared with the rate in appropriate nondiabetic controls (74.5 versus 62.5 beats per minute, p ~0.05) [33]. This well-recognized abnormality in patients with diabetes is widely believed to be due to parasympathetic or combined parasympathetic and sympathetic damage [31]. Loss of parasympathetic control is more commonly encountered than loss of sympathetic function [8]. A number of studies, including some using computer-assisted simulation models, suggest that if nerve fibers are damaged randomly in a milieu of generalized metabolic disturbance, longer fibers are likely to have been damaged earlier [34]. Progression appears to be from vagal to sympathetic fibers, since the former are generally longer fibers. However, it is also possible that predominant or earlier vagal nerve disease is explicable on an entirely different basis, which could be related to neurotransmitter sensitivity rather than susceptibility of longer nerve fibers. Futthermore, if the proposed sequence is correct, the progression would be expected to be from normal to fast resting heart rates of 90 to 100 beats per minute or more, progressing to lower but still supranormal rates with advancing sympathetic nerve damage [31]. Indeed, similar heart rate changes can be induced in normal subjects by pharmacologic blockade with atropine, which results in considerable speeding of the heart rate that is reduced, but not normalized, by the addition of beta-adrenergic blockade [25,33]. Minimal heart rate was slowed by 10 beats per minute in the sorbinil-treated patients, compared with no change in the placebo-treated group. Reanalysis of data from a previous study [l 11, in which sorbinil was administered for three to four weeks to 11 patients with diabetes, revealed similar results with a smaller, but nevertheless significant, decrease in heart rate of 7.0 ? 1.3 beats per minute (p
November

15. 1995

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

feet of sorbinil on the sinoatrial node has not been excluded, this appears to be a somewhat remote possibility when considered with the other findings of this study and the well-known decrease in vagal tone in patients with diabetes. Simultaneous administration of beta-adrenergic blocking agents in six of these patients clearly did not influence the outcome of the study, since analysis with and without these patients yielded essentially similar results. On theoretic grounds, this outcome was to be expected since respiratory variation in heart rate is predominantly under parasympathetic control with the beta-adrenergic system having an apparently small influence. The improvement in expiration-inspiration ratios cannot be attributed to the slowing effect because if only heart rate slowed and there was no inherent change in the fractional increment in heart rate with inspiration, expirationinspiration ratios would remain unchanged. Furthermore, the absence of a correlation between expiration-inspiration ratio changes and heart rate changes suggests that, despite the fact they are both controlled by the vagus, they are independently regulated parameters. This conclusion was also suggested by Smith [23] and by Hilsted and Jensen [25], who reported no significant correlation between expiration-inspiration ratios and resting heart rate in a large group of subjects, which included patients with diabetes and reduced beat-to-beat variation. A significant linear regression relationship existed between the basal expiration-inspiration ratio and the expiration-inspiration ratio response (p <0.05), suggesting that during the six-week period, responses tended to be poorest in cases where cardiac autonomic nerve function was most severely affected. Although the improvements in vagal nerve integrity were small and subclinical, these tests are considered to be sensitive measures of existing autonomic dysfunction and, indeed, are frequently abnormal relatively early in the course of diabetes-well before objective clinical evidence of neuropathy, either somatic or autonomic, ensues. Thus, neuropathy, like many of the other target-organ complications of diabetes, is subclinical during a substantial period of its clinical course. In any event, these changes should be evaluated in the context of six weeks of therapy, clearly a very short period in the evolution of neuropathy in patients with diabetes. To date, there is no clinical information with respect to the possible use of aldose reductase inhibitors in symptomatic diabetic autonomic neuropathy. There is, however, nothing specific about vagal nerve dysfunction as compared with that in other autonomic nerves. In fact, with the exception of some diabetic men in whom impotence is the only manifestation of autonomic neuropathy, in diabetic patients with evidence of cardiac autonomic neuropathy, there are usually associated clinical features of widespread autonomic neuropathy [8,18]. Therefore, it is conceivable that other components of autonomic nerve dys-

The

American

Journal

of Medicine

Volume

79 (suppl

5A)

35

SYMPOSIUM

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

function in diabetes may be amenable to aldose reductase inhibitor therapy. In this context, the clinical response of one patient with painful diabetic neuropathy and symptoms and signs of severe generalized autonomic neuropathy, who received sorbinil, was of particular interest to us. In addition to a major decrease in his pain symptoms, he reported marked improvement in long-standing symptoms of gastroparesis, as well as a decrease in crampy abdominal pains and diarrhea. Improvement continued, and after 18 weeks of treatment, there was a complete absence of reported upper or lower intestinal disturbances. The patient also noted marked improvement in his dark-light pupillary adaptation and a return of previously absent sweating of his legs and feet. His resting pulse, initially between 100 and 104 beats per minute on three separate occasions, had declined to 80 beats per minute three weeks after starting sorbinil therapy. It has remained at this level over a period of more than one year of continuous sorbinil therapy. The patient reported that erectile function, previously reported to be absent and documented as absent by a nocturnal penile tumescence study, was nearly normal after 16 weeks of therapy. This has persisted over the subsequent nine months and has been shown by nocturnal penile tumescence testing and by nocturnal application of a snap gauge, a measure of penile rigidity, to be within the normal range. This patient’s clinical history and response to sorbinil are reported in more detail elsewhere [12]. Although somewhat anecdoctal, these data are not unique. Two other patients receiving sorbinil for painful neuropathy have reported improvement in erectile function, which had been absent for three or more years prior to treatment. Furthermore, in one of these patients, discontinuation of sorbinil on two separate occasions has resulted in a decline in erectile capacity, with return of function on readministration. Three other patients have also reported relief of severe diarrhea and symptoms of gastroparesis and two patients reported a reduction in orthostatic dizziness during treatment with sorbinil. Collectively, these isolated observations suggest that, in addition to the indications of a beneficial effect of sorbinil on symptomatic somatic neuropathy, symptomatic autonomic neuropathy may also be amenable to sorbinil. SOMATOSENSORY-EVOKED POTENTIALS Median nerve somatosensory-evoked potentials were studied before and after six weeks of sorbinil therapy in the 36 patients whose cardiac autonomic responses were described earlier. Significant improvements in the sorbiniltreated group for wrist-Erbs-point response and amplitude of the cortical response were observed, with no changes in the placebo-treated group [12]. These data are consistent with a sorbinil-related improvement in peripheral nerve integrity, with no significant change in central conduction.

36

November

15, 1965

The

American

Journal

of Medicine

TOXICITY OF SORBINIL The major side effect of sorbinil is an immune-complexmediated macular, erythematous rash with or without fever. The reaction appears to be restricted to the first four weeks of treatment with sorbinil and is rapidly reversible upon discontinuation of the agent. Over the past six years, we have been involved in the treatment of 106 diabetic patients with sorbinil. Reactions occurred in 12 of 106 or 11.3 percent (7 of 73 men or 9.6 percent compared with 5 of 33 women or 15.2 percent). In 11 of the 12 cases, the reactions manifested as a macular, erythematous rash over extensive areas of the body surface, which was associated with fever in eight. In all 11 patients, as well as in the twelfth patient who had a more severe reaction that included pancytopenia, the characteristic rash developed within the first 14 days of sorbinil. In each patient, recovery was rapid and complete after discontinuation of the medication. The toxicity of sorbinil is discussed in detail elsewhere [12]. In summary, although isolated instances of serious mucocutaneous toxicity have been observed, overall toxicity does not appear to be a major problem. Although caution is necessary, it does not appear that the use of sorbinil will be markedly restricted by its toxicity. In conclusion, our clinical experience with the aldose reductase enzyme inhibitor agent sorbinil (Pfizer-CP45; 634) has been described. Sorbinil-mediated benefits on somatic nerve function have been documented, including improvements in motor and sensory conduction velocities and in somatosensory-evoked potentials. Symptomatic improvement in neuropathic pain has been observed in many patients and, in a number of cases, has been notable. Pain relief is, nevertheless, not experienced in all cases and, when obtained, is often only modest. This is, however, not surprising, in view of the clinical and pathologic heterogeneity of diabetic neuropathy. There are as yet no data concerning an analgesic effect of sorbinil in nondiabetic neuropathies or non-neuropathic pain. Such studies may be necessary, although conceptually, there is no basis for anticipating such action. Objectively, sorbinilrelated improvements in motor and sensory deficits have been observed. With regard to the autonomic nervous system, the results of cardiac autonomic function tests-those predominantly related to vagal nerve function-have been improved with the use of sorbinil. Reduction in resting heart rate has perhaps been the most striking of these. However, these improvements have been at the preclinical stage of autonomic neuropathy. At the clinical level, sorbinil-related improvements in autonomic symptoms and signs of a more widespread nature have been observed in isolated cases. Improvements in clinical and test parameters related to both somatic and autonomic nerve function have been reversed upon discontinuation of sorbinil and,

Volume

79 (suppl

5A)

SYMPOSIUM

ON ALDOSE

REDUCTASE

INHIBITION-JASPAN

ET AL

Finally, based upon the suggested pathophysiology of the diabetic neuropathies and the presumed action of aldose reductase inhibitors in preventing the metabolic cascade that is believed to contribute to the neuropathic process, the ultimate application of such therapy may be as prophylaxis against diabetic neuropathy through the prevention or interruption at an earlier and more amenable stage of the pathogenetic process.

in most cases, beneficial responses were again observed with readministration. Sorbinil toxicity has been infrequent, mild, and readily reversible. The seemingly low toxicity of sorbinil and the accumulating evidence of its ability to relieve neuropathic pain and to improve, subjectively and objectively, peripheral somatic and autonomic nerve function strongly suggest that it may be important in treating diabetic neuropathy.

REFERENCES 1.

5. 6.

7. 8.

9.

10.

11.

12.

13.

14. 15.

16.

Fagius J, Jameson S: Effects of aldose reductase inhibitor treatment in diabetic polyneuropathy-a clinical and neurophysiological study. J Neurol Neurosurg Psychiatry 1981; 44: 9911001. Judzewitsch RG, Jaspan JB, Polonsky KS, et al: Aldose reductase inhibition improves nerve conduction velocity in diabetic patients. N Engl J Med 1983; 308: 119.-125. Foulds G, O’Brien MM, Bianchine JR, et al: Kinetics of an orally absorbed aldose reductase inhibitor, sorbinil. Clin Pharmacol Ther 1981; 30: 693-709. Peterson MJ, Sarges R: Aldinger CE, et al: CP-45,634: a novel aldose reductase inhibitor that inhibits polyol pathway activity in diabetic or galactosemic rats. Metabolism 197; 28 (suppl 1): 456-461. Porte D Jr, Graf RJ, Halter HB, et al: Diabetic neuropathy and plasma glucose control. Am J Med 1981; 94: 195-200. Graf RJ, Halter JB, Pfeifer MA, et al: Glycemic control and nerve conduction abnormalities in non-insulin-dependent diabetic subjects. Ann Intern Med 1981; 70: 307-311. Scott J, Huskisson EC: Graphic representation of pain. Pain 1976; 2: 175-184. Ewing DJ, Campbell IW, Clarke BF: Assessment of cardiovascular effects in diabetic autonomic neuropathy and prognostic implications. Ann Intern Med 1980; 92: 308-311. Halar EM, Venkatesh ME: Nerve conduction velocity measurements: improved accuracy using superimposed response waves. Arch Phys Med’Rehabil 1976; 57: 451-457. Melvin JL, Harris DH, Johnson EW: Sensory and motor conduction velocities in the ulnar and median nerves. Arch Phys Med Rehabil 1966; 47: 511-519. Jaspan JB, Maselli R, Herold K, et al: Treatment of severely painful diabetic neuropathy with an aldose reductase inhibitor: relief of pain and improved somatic and autonomic nerve function. Lancet 1983; II: 758-762. Jaspan J, Towle VL, Maselli R, et al: Clinical studies with an aldose reductase inhibitor in the autonomic and somatic neuropalhies of diabetes. Metabolism (in press). Young RJ, Ewing DJ, Clarke BF: A controlled trial of sorbinil, an aldose reductase inhibitor, in chronic painful diabetic neuropathy. Diabetes !983; 32: 938-942. Clarke BF, Ewing DJ, Campbell I: Diabetic autonomic neuropathy. Diabetologia 1979; 17: 195-212. Sundkvist G: Autonomic nervous function in asymptomatic diabetic patients with signs of peripheral neuropathy. Diabetes Care 1981; 4: 529-534. Keen H: Autonomic neuropathy in diabetes mellitus. Postgrad Med J 1959; 35: 272-280.

November

15, 1985

17. 18. 19. 20.

21.

22. 23.

24.

25. 26.

27.

28. 29.

30.

31. 32.

33.

34.

The

Rundles RW: Diabetic neuropathy. Medicine 194.5; 24: 1 lo160. Ewing DJ, Campbell JW, Clarke BF: The natural history of diabetic autonomic neuropathy. Q J Med 1980; 49: 95-108. Hosking DJ, Bennett T, Hampton JR: Diabetic autonomic neuropathy. Diabetes 1978; 27: 1043-1057. Malone JI, Leavengood H, Peterson MJ, et al: Red blood cell sorbitol as an indicator of polyol pathway activity. Inhibition by sorbinil in insulin-dependent diabetic subjects. Diabetes 1984; 33: 45-49. Gabbay KH, Sosenko JM, Banuchi GA, et al: Glycosylated hemoglobins: increased glycosylation of hemoglobin A in diabetic patients. Diabetes 1979; 28: 337-340. Sundkvist G, Almer L-O, Lilja B: Respiratory influence on heart rate. Br Med J 1979; 1: 924-925. Smith S: Reduced sinus arrhythmia in diabetic autonomic neuropathy: diagnostic value of an age-related normal range. Br Med J 1982; 285: 1599-1601. Bernstein R: R-R interval studies: a simple office protocol for evaluation of cardiac autonomic neuropathy. Diabetes Care 1984; 7: 510-513. Hilsted J, Jensen SB: A simple test for autonomic neuropathy in juvenile diabetes. Acta Med Stand 1979; 205: 385-387. Bennett T, Fentem PH, Fitton D, et al: Assessment of vagal control of the heart in diabetes. Measures of R-R interval variation under different conditions. Br Heart J 1977; 39: 25-28. Ewing DJ, Borsey DQ, Bellavere F, et al: Cardiac autonomic neuropathy in diabetes; comparison of measures of R-R interval variations. Diabetologia 1981; 21: 18-24. Murray A, Ewing D, Campbell J, et al: R-R interval variations in young male diabetics. Br Heart J 1975; 37: 882-885. Ewing DJ, Irving JB, Kerr F, et al: Cardiovascular responses to sustained handgrip in normal subjects and in patients with diabetes mellitus. Clin Sci 1974; 46: 295-306. Campbell IW, McGarry S, Smith DN, et al: Continuous electrocardiographic recording during exercise in young male diabetes. A computer study. Br Heart J 1975; 37: 277-282. Ewing DJ, Campbell IW, Clarke BF: Heart rate changes in diabetes mellitus. Lancet 1981; I: 183-186. Bennett T, Farquhar I, Hosking D, et al: Assessment of methods for estimating autonomic nervous control of the heart in patients with diabetes mellitus. Diabetes 1978; 27: 1167-l 174. Sosenko JM, Pasin RA, Kubrusky DB, et al: Heart rate, autonomic function, and glycemia in young type I diabetic patients. Diabetes 1984; 33: 361. Waxman SG: Pathophysiology of nerve conduction: relation to diabetic neuropathy. Ann Intern Med 1980; 92: 304-307.

American

Journal

of Medicine

Volume

79 (suppl

5A)

37