Antiarrhythmic Drugs and Polyneuropathy J Neurol Neurosurg Psychiatry: First Published As 10.1136/Jnnp.57.3.340 on 1 March 1994

34030ournal ofNeurology, Neurosurgery, and Psychiatry 1994;57:340-343

Antiarrhythmic drugs and polyneuropathy J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.3.340 on 1 March 1994. Downloaded from

The Collaborative Group for the Study of Polyneuropathy

Istituto di Ricerche Farmacologiche factors of polyneuro- "Mario Negri", Abstract prevalence and risk Milano, Italy A total of 151 patients on chronic treat- pathy, using standard clinical and electro- E Beghi ment with amiodarone and other anti- physiological techniques. M L Monticelli arrhythmic drugs were subjected to Ospedale "San standard clinical and electrophysiologi- Gerardo", Monza, Italy cal investigation to assess the prevalence Materials and methods 15 to and seen consec- Clinica Neurologica and specificity of polyneuropathy. Subjects aged 70 years A Bono Twenty two untreated patients with car- utively in four cardiological outpatient E Beghi diac disorders and 246 normal subjects services in Italy (Monza, Turin, Rome, and G Bogliun as controls. Abnormal electro- San Giovanni Rotondo) were considered for N Curt6 served L Marzorati physiological findings supporting the inclusion if they had been first treated for no L Frattola diagnosis of polyneuropathy were pre- less than six months with the following drugs, Divisione Cardiologica sent in 38 subjects (25%) given antiar- singly or in combination: amiodarone, F Achilli rhythmic drugs, with even distribution propafenone, atenolol, flecainide, mexiletine, A Bozzano A small A Cadel among drugs, and four untreated propanolol, verapamil, quinidine. A Cazzaniga patients (18%). Concurrent clinical sample of patients who were considered B Pria abnormalities were present in five candidates for antiarrhythmic treatment, but A Vincenti treated patients (one each with amio- had not yet been treated, was included as a Ospedale "Molinette", darone, propafenone, and flecainide, and reference group. Torino, Italy two with multiple drugs). Therefore, A preliminary screening investigation was DEA Neurologia G Chianale electrophysiological abnormalities are a made by the consultant cardiologist during M De Mattei common, although non-specific, feature routine ambulatory practice. The cardiologist P Zaina in patients taking antiarrhythmic drugs. retained the patients fulfilling the inclusion Istituto di Cardiologia Amiodarone users do not seem at higher criteria and excluded those with other con- F Colivicchi risk of polyneuropathy than subjects current factors or disorders causing poly- R Fanelli F Gaita treated with other drugs or even un- neuropathy (indicated on a check list). The C Giosetto treated patients with cardiac disorders. reasons for exclusion were: diabetes, liver dis- Universita Cattolica, ease, porphyria, amyloidosis, uraemia, thyroid Roma, Italy (7 Neurol Neurosurg Psychiatry 1994;57:340-343) disorders, paraproteinaemia, collagen disease, Clinica Neurologica neoplasms (checked by history, reports of S Magi abnormal biochemical assays, or specific http://jnnp.bmj.com/ T Mignogna A Ciervo Several adverse effects of amiodarone have treatments). Patients treated with potentially P Tonali been reported, including neurological toxi- neurotoxic agents'5 were also excluded. All Clinica Cardiologica city.12 The commonest neurological symp- the eligible patients were invited by appoint- F Bellocci toms and signs include tremor, ataxia, and ment to undergo a detailed neurological eval- Ospedale "Casa polyneuropathy with or without spinal cord uation, which included a standardised active Sollievo della involvement.34 Clinical and electrophysio- search for symptoms of polyneuropathy and a Sofferenza", San Giovanni Rotondo, logical data supporting polyneuropathy have formal clinical and electrophysiological on September 24, 2021 by guest. Protected copyright. Italy been confirmed by pathological findings in assessment. A clinical diagnosis of poly- Divisione di animals and humans, showing demyelination neuropathy was made when there was evi- Neurologia or axonal loss with lysosomal inclusions in dence of bilateral impairment of strength, or P Di Viesti data to case sensation, or deep tendon reflexes, or a com- P Simone different cell types.'5 Most refer M Zarrelli reports and selected clinical series, of which bination in the upper, or lower, or both Divisione di only two screened as many as 50 to 54 extremities with fairly symmetrical distribu- Cardiologia patients.3 tion. P Lanna Despite the recognition of common elec- The electrophysiological screening was D Potenza trophysiological effects of the antiarrhythmic done on the right limbs according to a stan- Centro Ricerche have been dard Briefly, motor and sensory Cliniche in drugs,'011 only a few reports pub- procedure.'6 Biomagnetismo- lished on the potential neurotoxicity of agents nerve conduction velocity, distal latency, and Fisiologia, Roma, Italy other than amiodarone."2-'4 It is therefore not potential amplitude of median, ulnar, com- R Fenici possible to estimate the risk of polyneuro- mon peroneal, and sural nerves were mea- G Melillo pathy in patients on chronic amiodarone sured with surface stimulating and recording Correspondence to: Ettore Beghi, MD, Istituto di treatment, the specificity of that risk, or the techniques. Room temperature was main- Ricerche Farmacologiche factors that may be accompanied by a higher tained at 22-24°C. Skin temperature was "Mario Negri", Via Eritrea, 62-20157 Milano, Italy. risk. measured in the presence of cool limbs. Limb Received 31 December 1992 We screened a fairly unselected group of warming was performed when specifically and in revised form patients treated with different antiarrhythmic indicated. All the electrophysiological vari- 8 June 1993. Accepted 21 June 1993 drugs, including amiodarone, to assess the ables were compared with normal reference Antiarrhythmic drugs and polyneuropathy 341

values obtained from each centre separately Thirty six patients (19%) were not given J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.3.340 on 1 March 1994. Downloaded from and measured according to the same proce- neurological assessment. The commonest dure. reasons for exclusion were refusal and resi- In Monza there were 54 normal subjects dence far from the hospital. These patients (27 men, 27 women) aged 18 to 69 (mean 42 were similar to those completing the neuro- years). In Rome there were 60 patients (35 logical assessment in terms of age, sex, and men, 25 women) aged 22 to 67 (mean 44 centre. Ninety six men and 55 women under- years), in Turin 60 (31 men, 29 women) aged went complete neurological assessment. In 20 to 69 (mean 44 years), and in San these cases the underlying cardiac disorders Giovanni Rotondo 72 (47 men, 25 women) were atrial arrhythmia (74 cases), Wolff- aged 17 to 69 (mean 47 years). No significant Parkinson-White disease (22), coronary differences or age trends were detected within artery disease (21), ventricular arrhythmia centres for any of the values. The normal lim- (19), arterial hypertension (five), sinus tachy- its for conduction velocity and distal latency cardia (four), and miscellaneous cardiac dis- were set at 2-5 SD from the mean for the orders (six). The distribution of treated controls. The action potential was abnormal patients and the commonest drugs by centre if the amplitude was below the lowest value were: Monza 54 (amiodarone 30, polytherapy found in the controls. An electrophysiological six); Turin 47 (propafenone 15, verapamil diagnosis of polyneuropathy was made when and flecainide 10); Rome 36 (propafenone conduction velocity, or distal latency, or 10, amiodarone eight); San Giovanni potential amplitude, or a combination were Rotondo 14 (flecainide four, amiodarone and abnormal in at least two nerves. Specific atenolol three). Untreated patients were strategies were adopted to exclude mononeu- evenly distributed among centres. Table 1 ropathies or multineuritides of diverse origin: shows the general characteristics of the sam- (1) Patients with clinical findings suggesting ple by treatment group. Untreated patients mononeuropathy or profoundly asymmetric and subjects receiving flecainide were gener- polyneuropathy were excluded; (2) abnormal ally younger and those receiving mexiletine electrophysiological variables were to be were older. Treatment duration ranged from recorded from different nerves; (3) when 6 months to 10 years and was more than 3 entrapment neuropathies (for example, carpal years in 38% of the cases. Mean treatment tunnel syndrome) were suspected, the contra- duration varied for the different drugs. The lateral (left) limbs were examined; (4) mixed daily doses of the drugs were within the stan- (sensory-motor) nerves (median, ulnar) were dard therapeutic range for all the patients. considered a single entity. Plasma concentrations of the antiarrhythmic Statistical analysis was carried out with the drugs were not routinely measured. Statistical Package for the Social Sciences Forty six patients (30%) complained of (SPSS). The prevalence of clinical and elec- one or more symptoms of polyneuropathy, trophysiological polyneuropathy was esti- the commonest being paresthesiae (27 cases), mated in the whole sample, in different followed by muscle cramps (18), restless legs treatment groups, and in the untreated sub- (nine), standing or gait disturbances (five), jects. The presence of polyneuropathy was burning feet, muscle pain, and troubles with

correlated with age, sex, and treatment object handling (four cases each). Abnormal http://jnnp.bmj.com/ modalities. clinical findings were present in 12 cases (amiodarone three, flecainide and verapamil two, propafenone and atenolol one, poly- Results therapy three). They were mostly impairment From October 1990 to the end of December of tendon reflexes (11 cases), then autonomic 1991 a total of 187 patients (117 men and 70 dysfunction (eight), hypotrophic muscles women) fulfilled the inclusion criteria. (three), muscle weakness, hypotonia, and Patients' ages ranged from 21 to 69 (mean 55 sensation impairment (one case each). on September 24, 2021 by guest. Protected copyright. (SE 12) years). Age and sex distribution of Electrophysiological polyneuropathy was pre- the cases within centres were fairly similar. sent in 38 cases (25%) with similar distribu- One hundred and seventy five patients tion for the commonest drugs. Of these, five were given one drug only. Amiodarone was (3% of the entire sample) had concurrent the commonest drug (55 cases) then clinical abnormalities (amiodarone, pro- propafenone (41), verapamil and flecainide pafenone and flecainide one, polytherapy (25), atenolol (18), quinidine (seven), and two). In none of the patients was treatment mexiletine (four). withdrawn because of symptoms or signs of

Table 1 Antiarrhythmic drugs andpolyneuropathy: general characteristics ofthe sample by treatment group Drug treatment Variable None Amiodarone Propafenone Atenolol Verapamil Mextletne Flecainide Quinidine Multiple drugs Total 22 46 32 14 19 3 20 5 12 Age (mean (SE))(y) 45(4) 57(1) 54(2) 58(2) 56(3) 62(3) 45(3) 54(4) 56(2) Sex: Men 16 31 19 10 11 3 11 3 10 Women 6 16 14 4 8 - 9 2 2 Treatnent duration (mean (SE))(months) - 50(8) 31(4) 30(7) 43(8) 24(12) 30(5) 63(30) - Daily dose (mean (SE))(mg) 196(6) 474(33) 80(15) 217(23) 800(53) 190(17) 695(118) - 342 2The collaborative groupfor the study ofpolyneuropathy

Table 2 Antiarrhythmic dnrgs andpolyneuropathy: electrophysiological (EP) and candidates for antiarrhythmic treatment but J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.3.340 on 1 March 1994. Downloaded from clinical (CP) polyneuropathy by patient, treatment, and disease characteristics were still untreated when assessed. Cardiac EP CP disorders included sinus tachycardia (eight Vaiable No ofcases No (%) No (%/) cases), ventricular arrhythmia (five), coronary artery disease (four), Wolff-Parkinson-White Sex: Men 96 25 26-0 4 4-2 disease (two), and miscellaneous disorders Women 55 13 23-6 1 1-8 (three). Of these, four patients (18%) pre- Age (y) < 45 30 7 23-3 - - sented abnormalities confirming a diagnosis 46-55 37 7 18-9 2 5-4 of electrophysiological polyneuropathy. One 56-65 58 15 25-9 1 1-7 > 65 26 9 34-6 2 7-7 patient had electrophysiological abnormalities Treatment duration in three nerves; two had significant (> 40%) (months): impairment of nerve potential amplitude < 12 36 6 16-7 2 5-6 13-36 57 15 26-3 1 1-8 (either sensory or motor). When performed, > 36 58 17 29-3 2 3-4 the clinical examination failed to detect Drugs: Amiodarone 46 10 21-7 1 2-2 abnormalities indicating peripheral nerve Propafenone 32 7 21-9 1 3-1 impairment. Atenolol 14 3 21-4 - - Verapamil 19 7 36-8 Mexiletine 3 2 66-7 - - Flecainide 20 5 25-0 1 5 0 Quinidine 5 - - - - Discussion Multiple 12 4 333 2 16-7* In this study 25% of the patients treated with Untreated 22 4 18-2 - -t amiodarone and other antiarrhythmic drugs *p < 0 05 v single drug treatment. presented electrophysiological findings in CP = electrophysiological and clinical polyneuropathy. keeping with a diagnosis of polyneuropathy tClinical examination not done in 12 cases. but only 3% also had clinical polyneuropathy. Impairment of deep tendon reflexes and polyneuropathy. Clinical signs of peripheral symptoms of polyneuropathy were the most nerve involvement were generally mild, and frequent features. The commonest antiar- muscle weakness, when present, did not rhythmic drugs seemed involved to a similar cause standing or gait difficulties. extent. The most commonly impaired nerve was Polyneuropathy among patients receiving the median (46 cases (30%)), then the ulnar artiarrhythmic drugs in clinical practice nerve (35 cases (23%)), the peroneal nerve seems a common, but clinically irrelevant (30 cases (20%)), and the sural nerve (28 event. In fact, despite the frequent reports of cases (19%)). Twenty three subjects had paresthesiae here, which might be the result abnormal electrophysiological recordings in of the active search for the symptom, clinical two nerves, 12 in three nerves, and three in polyneuropathy was generally mild and none four nerves. of the patients needed treatment changes Nerve potential amplitude was the most because of neurological toxicity. The use of often impaired measure, maximally in the low daily doses compared with other series sural nerve (24 cases) followed by median may explain why no severe signs of neurotoxi- motor (19 cases) and sensory (16 cases), city were detected, even among amiodarone treatment. Our data ulnar motor (13 cases) and sensory (seven users, despite prolonged http://jnnp.bmj.com/ cases), and peroneal (10 cases). Impairment contrast with those of Charness and col- was more than 20% above normal in 27 cases leagues3 who found neurological side effects and more than 40% in 22. All the cases with in 54% of their patients taking amiodarone. electrophysiological polyneuropathy pre- In those patients, however, the mean daily sented impairment of potential amplitude. maintenance dose of amiodarone was 580 mg Distal latency was abnormal in the median as against 196 mg in the present study. In a nerve in 16 cases, in the ulnar nerve in 15, recent experimental study a correlation was and in the peroneal nerve in 13. Abnormal found between the amount of drug injected in on September 24, 2021 by guest. Protected copyright. nerve conduction velocity values were slightly the peripheral nerve and the extent of electro- less frequent and were detected in similar physiological and pathological changes.'7 proportions in the nerves investigated: Electrophysiological abnormalities in median sensory (12 cases) and motor (nine patients receiving antiarrhythmic drugs other cases), peroneal and sural (11 cases each), than amiodarone may be explained by the ulnar motor (nine cases), and sensory (three partly overlapping mechanisms of action of cases). Severe (> 40%) impairment of nerve these drugs, which contrast with the theoreti- conduction velocity and distal latency was cal selectivity of their effects on the action rare. No correlation was found between drugs potential of cardiac cells (sodium channel and number of impaired nerves or between blockade, fl-adrenergic blockade, prolonged the extent of the electrophysiological impair- repolarisation, calcium channel blockade)." ment and age, presence of clinical abnormali- Alternatively, abnormal folate metabolism, ties, or type of drug. believed to predispose patients treated with Table 2 gives the correlations between the anticonvulsant phenytoin'8 to peripheral polyneuropathy and demographic and clinical neuropathy, may be implicated here too. As features in the whole sample. Treatment with phenytoin also has antiarrhythmic properties multiple drugs was the only variable that and most antiarrhythmic drugs undergo seemed correlated with the risk of polyneu- extensive hepatic metabolism" this may lead ropathy. Twenty two patients (16 men, six to folate deficiency. Tocainide and mexiletine women) aged 17 to 69 years were considered increase the excitation threshold and reduce Antiarrhythmic drugs andpolyneuropathy 343

sodium channel permeability in peripheral perhexiline maleate25 and chloroquine,26 the J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.57.3.340 on 1 March 1994. Downloaded from nerves.'9 toxic effects of amiodarone, like perhexiline Another possible interpretation of our find- and chloroquine, may act through an induc- ings is that antiarrhythmic drugs at chronic tion of enzyme deficiency in predisposed maintenance dosages may induce functional patients.27 Such findings might not be present impairment of the nerve with minimal or no with other antiarrhythmic drugs, which thus anatomical damage. Indirect evidence tends merit fuirther investigation. to support this. Firstly, in our study the high Financial support from FIDIA Research Laboratories is grate- frequency of electrophysiological abnormali- fully acknowledged. ties contrasts with a low prevalence of clinical 1 Reader EA, Podrid PJ, Lown B. Side effects and compli- signs of polyneuropathy. Secondly, in an cations of amiodarone therapy. Am Heart Jf 1985; ongoing prospective study of amiodarone 109:975-83. 2 Schwartz JB, Keefe D, Harrison DC. Adverse effects of neuropathy some patients receiving high antiarrhythmic drugs. Drugs 1981;21:23-45. doses presented abnormalities of nerve con- 3 Charness ME, Morady F, Scheinman MM. Frequent neurologic toxicity associated with amiodarone therapy. duction that were detected after a few days of Neurology 1984;34:669-71. treatment (A Bono, personal communica- 4 Hugon J, Vallat JM, Dumas M, Outrequin G, Vallat M. Myeloneuropathy, neurologic toxicity, and amiodarone. tion). Thirdly, in several patients taking Neurology 1985;35:282. amiodarone the electrophysiological abnor- 5 Costa-Jussa FR, Jacobs JM. The pathology of amiodarone neurotoxicity. I. Experimental studies with reference malities were reversed by reduction of treat- to changes in other tissues. Brain 1985;108: ment or withdrawal.9 2021 735-52. 6 Jacobs JM, Costa-Jussa FR. The pathology of amiodarone The electrophysiological polyneuropathy neurotoxicity. II. Peripheral neuropathy in man. Brain seen in 18% of untreated patients with car- 1985;108:753-69. 7 Pellissier JF, Pouget J, Cros D, De Victor B, Serratrice G, diac disorders arouses the suspicion that Toga M. Peripheral neuropathy induced by amiodarone peripheral nerve impairment may not always chlorhydrate. A clinicopathological study. Jf Neurol Sci 1984;63:251-66. be related to antiarrhythmic treatment (elec- 8 Dudognon P, Hauw JJ, Baecque C, Derrida JP, trophysiological assessment before treatment Escourolle R, Nick J. Neuropathie au chlorhydrate d'amiodarone. Rev Neurol 1979;135:527-40. was not available in our patients). The dis- 9 Martinez-Arizala A, Sobol SM, McCarty GE, Nichols BR, ease may also be the complication of other Rakita L. Amiodarone neuropathy. Neurology 1983;33: 643-5. known or less known factors, including the 10 Vaugham Williams EM. A classification of antiarrhythmic underlying heart disease. Given the preva- actions reassessed after a decade of new drugs. J Clin Pharmacol 1984;24: 129-47. lence of the disease in an untreated, otherwise 11 Bigger JT Jr, Hoffman BF. Antiarrhythmic drugs. In: similar, population the proportion of poly- Goodman and Gilman's The pharmacological basis of therapeutics, 8th edn. New York: Pergamon Press, 1990: neuropathy attributable to drug treatment 840-73. may thus be even lower than that reported in 12 Talbot RG, Julian DG, Prescott LF. Long-term treatment of ventricular arrhythmias with oral mexiletine. Am the whole sample. HeartJ 1976;91:58-65. Our patients presented diffuse electrophys- 13 Keefe DLD, Kates RE, Harrison DC. New antiarrhythmic drugs: their place in therapy. Drugs 1981;22: iological signs of sensorimotor neuropathy. 363-400. Previous reports provide conflicting results 14 Getzkow GD, Sullivan JY. Extracardiac adverse effects of flecainide. Am Jf Cardiol 1984;53: 101B-105B. showing motor,6 sensory,7 or sensorimotor 15 Le Quesne P. Neuropathy due to drugs. In: Dyck PJ, signs.720 Our findings suggest a mixed Thomas PK, Lambert EH, Bunge RP eds. Peripheral neuropathy. Philadelphia: Saunders, 1984:2162-79. polyneuropathy with axonal and demyelinat- 16 Beghi E, Delodovici ML, Bogliun G, et al. Hypo- ing components. Predominant axonal damage thyroidism and polyneuropathy. J Neurol Neurosurg Psychiatry 1989;52: 1420-3. http://jnnp.bmj.com/ was detected on nerve biopsy by some 17 Santoro L, Barbieri F, Nuciotti R, et al. Amiodarone- authors7 22 whereas others found prevailing induced experimental acute neuropathy in rats. Muscle Nerve 1992;15:788-95. demyelination with only mild axonal loss,623 18 Horwitz SJ, Klipstein FA, Lovelace RE. Relation of and others found combined axonal and abnormal folate metabolism to neuropathy developing during anticonvulsant drug therapy. Lancet 1968;i: myelin involvement.8 The different findings 563-5. of the neuropathological studies may be 19 Tippe A, Linderer EM. Action of tocainide and mexiletine on the excitation threshold of myelinated nerves. Eur Jf explained by the different treatment sched- Pharmacol 1987;141:225-33. ules, which produce various electrophysio- 20 Fraser AG, McQueen INF, Watt AH, Stephens MR. on September 24, 2021 by guest. Protected copyright. Peripheral neuropathy during long-term high-dose logical and pathological changes, ranging amiodarone therapy. J Neurol Neurosurg Psychiatry 1985; from pure demyelination to complete axonal 48:576-8. 21 Heger JJ, Prystowsky EN, Jackman WM, et al. degeneration in the experimental animal.'7 Amiodarone. Clinical efficacy and electrophysiology The lack of pathological reports in our study during long-term therapy for recurrent ventricular tachycardia or ventricular fibrillation. N Engl J Med prevents comparisons or definite conclusions 1981; 305:539-45. on this issue and we could not establish why 22 Meier C, Kauer B, Muller U, Ludin HP. Neuromyopathy during chronic amiodarone treatment. J Neurol 1979; only some of the patients currently exposed 220:231-9. to antiarrhythmic drugs develop signs of 23 Kaeser HE, Uerich J, Wuthric R. Amiodarone- neuropathie. Praxis 1976;65:1121-2. peripheral nerve impairment. Except for the 24 Mason JW. Amiodarone. N Englj Med 1987;316:455-66. use of drug combinations, which may imply a 25 Hauw JJ, Mussini JM, Boutry JM, Escourolle R, Pollet S, Albouz S, et al. Perhexiline maleate induced lipidosis in potentiation of the effect on nerve conduction human peripheral nerve and tissue culture: ultrastruc- on the basis of well-known drug inter- tural and biochemical changes. Clin Toxicol 1981;18: 1405-9. actions,24 none of the more common charac- 26 Lemaire JF, Autret A, Biziere K, Romet-Lemone JL, Gray teristics of patients and treatment seemed to F. Amiodarone neuropathy: further arguments for human drug-induced neurolipidosis. Eur Neurol 1982; predispose to polyneuropathy. As there are 21:65-8. anecdotal reports of lysosomal inclusions 27 Shah RR, Oates NS, Idle JR, Smith RL, Lockhart JDF. Prediction of a subclinical perhexiline neuropathy in a among patients treated with other drugs with patient with inborn error of debrisoquine hydroxylation. pronounced lipophilic properties, such as Am Hearty 1983;105:159-61.