Dm
ORIGINAL ARTICLES
The Effect of y-linolenic Acid on Human Diabetic Peripheral Neuropathy: A Double-blind Placebocontrolled Trial G.A. Jamal”, H. Carmichaelb “Clasgow University Department of Neurology, lnstitute of Neurological Sciences, bVale of Leven Hospital, Alexandria, Clasgow, UK
Twenty-two patients with distal diabetic polyneuropathy confirmed both clinically and by objective nerve function studies, completed a double-blind, placebo-controlled study to assess the effect of dietary supplementation with y-linolenic acid on their neuropathy. Patients received either 360 mg y-linolenic acid (1 2 patients) or indistinguishable placebo capsules (10 patients) for 6 months. All patients were assessed at the beginning and end of the study period by neuropathy symptom and sign scoring, motor and sensory nerve conduction studies, and thermal threshold measurements. When compared with the placebo group, patients on y-linolenic acid showed statistically significant improvement in neuropathy symptom scores (p < O.OOl), median nerve motor conduction velocity ( p < 0.01) and compound muscle action potential amplitude ( p < 0.01), peroneal nerve motor conduction velocity (p < 0.05) and compound muscle action potential amplitude ( p < 0.05), median (p < 0.01) and sural ( p < 0.001) sensory nerve action potential amplitude and ankle heat threshold (p < 0.001) and cold threshold ( p < 0.01) values. yLinolenic acid therapy might have a useful role in the prevention and treatment of distal diabetic polyneuropathy. KEY WORDS
Essential fatty acids y-Linolenic acid Diabetic neuropathy Nerve conduction Thermal thresholds
Introduction There is growing evidence that in diabetes mellitus there are abnormalities of plasma, platelet, and red cell n-6 series essential fatty acids, both in experimental animals’,2 and in human^.^-^ In particular, the conversion of linoleic acid (1 8:2n-6) to y-linolenic acid (1 8:3n-6) by 6desaturation is impaired.’,2,5-7 Most of the biological effects of linoleic acid depend on its conversion to 6desaturated metabolites, which act as key components of membrane structure and also as precursors of the prostaglandins and other eicosanoids.8 The longer chained essential fatty acids produced by this pathway form key structural components of the neuronal cell membrane and of myelin.5,”2 The degree of disturbance of metabolism of linoleic acid has been correlated with glycosylated haemoglobin concentration.I3 In a controlled prospective s t ~ d y , it’ ~was demonstrated that very high doses of linoleic acid given to patients with Type 2 diabetes for 6 years were associated with a significant reduction of the risk of progression to microvascular complications. It is possible that very high levels of linoleic acid given overcame the partial block
Correspondence to: Dr G.A. Jamal, Clasgow University Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, Clasgow G51 4TF, UK.
0742-3071/90/0403 19-05$05.00 & Sons, Ltd.
0 1990 by John Wiley
of 6-desaturation. Another study suggested that the severity of microvascular disease in diabetes mellitus is positively correlated with an increasing level of linoleic acid and a decreasing level of arachidonic acid, consistent with reduced conversion of linoleic acid to arachidonic acid.“ The present study was designed to investigate whether established distal diabetic polyneuropathy could be reversed by direct supplementation of y-linolenic acid. y-Linolenic acid was given as the seed oil from a variety of the evening primrose (Oenothera spp.), selected and bred to produce oil of constant quality and composition, (Efamol evening primrose oil, Scotia Pharmaceuticals, Guildford, UK).
Patients and Methods Twenty-two patients, 10 with Type 1 and 1 2 with Type 2 diabetes, were admitted to the study. Age range was 21 to 74 years (Table I ) , and the minimum duration of symptoms of neuropathy was 6 months and of diabetes 3 years. All the patients had an abnormal HbA,. Each patient had one or more neuropathic symptom, accompanied by one or more objective physical sign of distal ’symmetrical’ polyneuropathy. All patients had one or more abnormality on nerve conduction studies, and also an abnormality of thermal threshold measurement. Accepted 8 January 1990 DIABETIC MEDICINE, 1990; 7: 319-323
31 9
Drn
ORIGINAL ARTICLES Table 1. Clinical characteristics of the diabetic patients studied Treatment group y-Linolenic acid n
Type of diabetes (112) Age (years) Sex (MIF)
Duration of diabetes (years) Duration of neuropathy (years) Number, or mean &
*
10 416 55 ? 15 (23-74) 515 15.2 t 4.2 3.1 0.9
*
SD (range)
Patients with unrecordable values of the measured nerve functions were excluded from the study. Repeated neurological assessment was performed in all patients and included semi-quantitative symptom scoring for pain,' 'i parasthesia, numbness, weakness, and abnormal sensation for heat and cold. These symptoms were scored as follows: absent 1, present 2, present and severe 3, giving an overall symptom score ranging from 5 to 15. Neurological examination included reflexes and semi-quantitative sensory examination to pinprick, touch, and vibration. Dichotomous scoring was chosen, to enhance reliability. Supinator and ankle reflexes were assessed and scored on the dominant side, with a scoring of 1 for normal, 2 for presence on reinforcement only, and 3 for absent. Sensory examination results for each of the modalities tested were scored as: 1 , no change; 2, a change below wrist and/or below ankle; 3, a change above wrist and/or above ankle. Overall score for examination therefore ranged from 4 to 12. Standardized nerve conduction studies under wellcontrolled conditions'" were performed in every patient. These included motor conduction studies for the right median and common peroneal nerves, and sensory nerve action potential (SNAP) measurements for the right median (orthodromic) and sural (antidromic) nerves at a limb temperature of 34 "C. Warm and cold threshold measurements were performed o n all patients at the ankle and wrist using the technique of Jamal et a/.,l7,I8 which has good reproducibility and a narrow normal range.",'9 In all nerve function studies, the average of two recordings within 1 week was taken as baseline in the week immediately prior to the beginning of the trial, and again in the week immediately after the end of the trial. None of the patients abused drugs or alcohol, and none received any medication other than their usual treatment for diabetes. All other causes of neuropathy were carefully excluded. None of the patients had their diabetic management regimen changed substantially in the 6 months prior to entry into the study. All the patients gave informed consent and the study was approved by the local ethical committee.
320
12 517 53 19 (21-74) 517 14.8 ? 3.6 3 . 2 ? 0.7
Placebo
Pria I Design The trial was a randomized, double-blind, parallel, placebo-controlled study of 6 months duration. During this period, patients received either 8 x 500 m g capsules of evening primrose oil or indistinguishable placebo capsules, 4 twice daily, providing a daily dose of 360 mg of y-linolenic acid. Randomization assigned 12 patients to the active and 10 patients to the placebo group. There was no significant difference between the placebo and the active group in age, sex, or duration of either diabetes or neuropathy (Table 1). All routine biochemical studies, glycosylated haemoglobin, and estimation of the various essential fatty acid profiles were performed at the beginning and the end of the trial period for all patients. H b A , concentration was measured using the Gelman method (normal reference range 4.0-6.0 %). Plasma phospholipid fatty acid composition was determined as described previously.2" Briefly, plasma samples were extracted with chloroformimethanol ( 2 : l ) . The lipid fractions were separated by thin layer chromatography. Total phospholipids were methylated using boron trifluorideimethanol. The fatty acid methyl esters were analysed on Hewlett Packard 5880-A gas chromatograph. Normal values for these fatty acids for 12 age-matched normal subjects are provided in Table 3. Compliance was assessed by counting the number of tablets left at the end of the study. Statistical analyses were done by using McNemars Test and Spearman's rank correlation. A level of p < 0.05 was taken as significant.
Resu I ts Table 2 summarizes the results of the clinical and neurophysiological assessments of the patients at the beginning and end of the trial. A comparison at baseline between the active and placebo groups showed no significant difference. Statistical assessment at the end of the trial for each of the variables showed a significant improvement in the active compared with the placebo group in 9 of the 12 variables (Table 2). C A. JAMAL, H CARMICHAEL
Dm
ORIGINAL ARTICLES
Table 2. Effects of 6-months y-linolenic acid therapy on clinical and neurophysiological measures in diabetic neuropathy Baseline Active Symptom scores Sign scores Median MCV
8.5 t 0.5 8.4 0.7 47.2 t 1.5
(FV) Sural SNAP (JLV) Ankle HT ("C) Ankle CT ("C) Wrist HT ("C) Wrist CT ("C)
Placebo
Difference
p
Active
NS
6.4 0.3 8.2 t 0.7 48.8 1.5
* *
9.0 t 0.4 0.4 8.1 45.3 t 2.6
< 0.001 -2.08
NS NS
P
Placebo
Active
P
Placebo
*
0.48 0.90 ? 0.61 < 0.01 -0.25 t 0.12 0.20 ? 0.19 NS 1.38*0.78-1.42? 1.10
ORIGINAL ARTICLES
The Effect of y-linolenic Acid on Human Diabetic Peripheral Neuropathy: A Double-blind Placebocontrolled Trial G.A. Jamal”, H. Carmichaelb “Clasgow University Department of Neurology, lnstitute of Neurological Sciences, bVale of Leven Hospital, Alexandria, Clasgow, UK
Twenty-two patients with distal diabetic polyneuropathy confirmed both clinically and by objective nerve function studies, completed a double-blind, placebo-controlled study to assess the effect of dietary supplementation with y-linolenic acid on their neuropathy. Patients received either 360 mg y-linolenic acid (1 2 patients) or indistinguishable placebo capsules (10 patients) for 6 months. All patients were assessed at the beginning and end of the study period by neuropathy symptom and sign scoring, motor and sensory nerve conduction studies, and thermal threshold measurements. When compared with the placebo group, patients on y-linolenic acid showed statistically significant improvement in neuropathy symptom scores (p < O.OOl), median nerve motor conduction velocity ( p < 0.01) and compound muscle action potential amplitude ( p < 0.01), peroneal nerve motor conduction velocity (p < 0.05) and compound muscle action potential amplitude ( p < 0.05), median (p < 0.01) and sural ( p < 0.001) sensory nerve action potential amplitude and ankle heat threshold (p < 0.001) and cold threshold ( p < 0.01) values. yLinolenic acid therapy might have a useful role in the prevention and treatment of distal diabetic polyneuropathy. KEY WORDS
Essential fatty acids y-Linolenic acid Diabetic neuropathy Nerve conduction Thermal thresholds
Introduction There is growing evidence that in diabetes mellitus there are abnormalities of plasma, platelet, and red cell n-6 series essential fatty acids, both in experimental animals’,2 and in human^.^-^ In particular, the conversion of linoleic acid (1 8:2n-6) to y-linolenic acid (1 8:3n-6) by 6desaturation is impaired.’,2,5-7 Most of the biological effects of linoleic acid depend on its conversion to 6desaturated metabolites, which act as key components of membrane structure and also as precursors of the prostaglandins and other eicosanoids.8 The longer chained essential fatty acids produced by this pathway form key structural components of the neuronal cell membrane and of myelin.5,”2 The degree of disturbance of metabolism of linoleic acid has been correlated with glycosylated haemoglobin concentration.I3 In a controlled prospective s t ~ d y , it’ ~was demonstrated that very high doses of linoleic acid given to patients with Type 2 diabetes for 6 years were associated with a significant reduction of the risk of progression to microvascular complications. It is possible that very high levels of linoleic acid given overcame the partial block
Correspondence to: Dr G.A. Jamal, Clasgow University Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, Clasgow G51 4TF, UK.
0742-3071/90/0403 19-05$05.00 & Sons, Ltd.
0 1990 by John Wiley
of 6-desaturation. Another study suggested that the severity of microvascular disease in diabetes mellitus is positively correlated with an increasing level of linoleic acid and a decreasing level of arachidonic acid, consistent with reduced conversion of linoleic acid to arachidonic acid.“ The present study was designed to investigate whether established distal diabetic polyneuropathy could be reversed by direct supplementation of y-linolenic acid. y-Linolenic acid was given as the seed oil from a variety of the evening primrose (Oenothera spp.), selected and bred to produce oil of constant quality and composition, (Efamol evening primrose oil, Scotia Pharmaceuticals, Guildford, UK).
Patients and Methods Twenty-two patients, 10 with Type 1 and 1 2 with Type 2 diabetes, were admitted to the study. Age range was 21 to 74 years (Table I ) , and the minimum duration of symptoms of neuropathy was 6 months and of diabetes 3 years. All the patients had an abnormal HbA,. Each patient had one or more neuropathic symptom, accompanied by one or more objective physical sign of distal ’symmetrical’ polyneuropathy. All patients had one or more abnormality on nerve conduction studies, and also an abnormality of thermal threshold measurement. Accepted 8 January 1990 DIABETIC MEDICINE, 1990; 7: 319-323
31 9
Drn
ORIGINAL ARTICLES Table 1. Clinical characteristics of the diabetic patients studied Treatment group y-Linolenic acid n
Type of diabetes (112) Age (years) Sex (MIF)
Duration of diabetes (years) Duration of neuropathy (years) Number, or mean &
*
10 416 55 ? 15 (23-74) 515 15.2 t 4.2 3.1 0.9
*
SD (range)
Patients with unrecordable values of the measured nerve functions were excluded from the study. Repeated neurological assessment was performed in all patients and included semi-quantitative symptom scoring for pain,' 'i parasthesia, numbness, weakness, and abnormal sensation for heat and cold. These symptoms were scored as follows: absent 1, present 2, present and severe 3, giving an overall symptom score ranging from 5 to 15. Neurological examination included reflexes and semi-quantitative sensory examination to pinprick, touch, and vibration. Dichotomous scoring was chosen, to enhance reliability. Supinator and ankle reflexes were assessed and scored on the dominant side, with a scoring of 1 for normal, 2 for presence on reinforcement only, and 3 for absent. Sensory examination results for each of the modalities tested were scored as: 1 , no change; 2, a change below wrist and/or below ankle; 3, a change above wrist and/or above ankle. Overall score for examination therefore ranged from 4 to 12. Standardized nerve conduction studies under wellcontrolled conditions'" were performed in every patient. These included motor conduction studies for the right median and common peroneal nerves, and sensory nerve action potential (SNAP) measurements for the right median (orthodromic) and sural (antidromic) nerves at a limb temperature of 34 "C. Warm and cold threshold measurements were performed o n all patients at the ankle and wrist using the technique of Jamal et a/.,l7,I8 which has good reproducibility and a narrow normal range.",'9 In all nerve function studies, the average of two recordings within 1 week was taken as baseline in the week immediately prior to the beginning of the trial, and again in the week immediately after the end of the trial. None of the patients abused drugs or alcohol, and none received any medication other than their usual treatment for diabetes. All other causes of neuropathy were carefully excluded. None of the patients had their diabetic management regimen changed substantially in the 6 months prior to entry into the study. All the patients gave informed consent and the study was approved by the local ethical committee.
320
12 517 53 19 (21-74) 517 14.8 ? 3.6 3 . 2 ? 0.7
Placebo
Pria I Design The trial was a randomized, double-blind, parallel, placebo-controlled study of 6 months duration. During this period, patients received either 8 x 500 m g capsules of evening primrose oil or indistinguishable placebo capsules, 4 twice daily, providing a daily dose of 360 mg of y-linolenic acid. Randomization assigned 12 patients to the active and 10 patients to the placebo group. There was no significant difference between the placebo and the active group in age, sex, or duration of either diabetes or neuropathy (Table 1). All routine biochemical studies, glycosylated haemoglobin, and estimation of the various essential fatty acid profiles were performed at the beginning and the end of the trial period for all patients. H b A , concentration was measured using the Gelman method (normal reference range 4.0-6.0 %). Plasma phospholipid fatty acid composition was determined as described previously.2" Briefly, plasma samples were extracted with chloroformimethanol ( 2 : l ) . The lipid fractions were separated by thin layer chromatography. Total phospholipids were methylated using boron trifluorideimethanol. The fatty acid methyl esters were analysed on Hewlett Packard 5880-A gas chromatograph. Normal values for these fatty acids for 12 age-matched normal subjects are provided in Table 3. Compliance was assessed by counting the number of tablets left at the end of the study. Statistical analyses were done by using McNemars Test and Spearman's rank correlation. A level of p < 0.05 was taken as significant.
Resu I ts Table 2 summarizes the results of the clinical and neurophysiological assessments of the patients at the beginning and end of the trial. A comparison at baseline between the active and placebo groups showed no significant difference. Statistical assessment at the end of the trial for each of the variables showed a significant improvement in the active compared with the placebo group in 9 of the 12 variables (Table 2). C A. JAMAL, H CARMICHAEL
Dm
ORIGINAL ARTICLES
Table 2. Effects of 6-months y-linolenic acid therapy on clinical and neurophysiological measures in diabetic neuropathy Baseline Active Symptom scores Sign scores Median MCV
8.5 t 0.5 8.4 0.7 47.2 t 1.5
(FV) Sural SNAP (JLV) Ankle HT ("C) Ankle CT ("C) Wrist HT ("C) Wrist CT ("C)
Placebo
Difference
p
Active
NS
6.4 0.3 8.2 t 0.7 48.8 1.5
* *
9.0 t 0.4 0.4 8.1 45.3 t 2.6
< 0.001 -2.08
NS NS
P
Placebo
Active
P
Placebo
*
0.48 0.90 ? 0.61 < 0.01 -0.25 t 0.12 0.20 ? 0.19 NS 1.38*0.78-1.42? 1.10
