Br Heartj 1994;71:141-145 Department of Pharmaology and Therapeutis, University of Leiester C Roffe S Flether KLWoods Correspondene to: Dr Kent L Woods, Department of Pharmaology and Therapeutis, Clinial Sienes Building, Leiester Royal Infirmary, Leiester LE2 7LX. Aepted for publiation 22 September 1993 Investigation of the effets of intravenous magnesium sulphate on ardia rhythm in aute myoardial infartion C Roffe, S Flether, K L Woods Abstrat Objetive-To examine the effet of doubling serum magnesium onentration on the inidene of arrhythmias in patients with suspeted aute myoardial infartion. Design-Randomised double blind linial trial. Setting-Coronary are unit of a teahing hospital. Patients-Clinial data were olleted on 2316 randomised patients with suspeted aute myoardial infartion. Holter monitoring was performed in a subgroup of 70 patients and analysed in 48 patients in whom aute myoardial infartion was onfirmed. Interventions-By random alloation, patients reeived either an intravenous loading dose of 8 mmol magnesium sulphate over five minutes plus 65 mmol over the next 24 hours, or equal volumes of saline. Main outome measures-(a) Clinially doumented arrhythmias; (b) use of antiarrhythmi treatments, ardioversion, and insertion of a paemaker; () inidene of all abnormal rhythms during Holter monitoring. Results-In the main trial the inidene of rhythm disturbane while in the oronary are unit (expressed as the odds ratio (OR) for magnesium: plaebo and its 95% onfidene interval) was not signifiantly different between treatment groups for ventriular fibrillation (OR 0 74; 0-46 to 1.20), ventriular tahyardia (OR 0-87; 0-63 to 1-20), supraventriular tahyardia (OR 0-69; 0-38 to 1.26), atrial fibrillation (OR 0-92; 0-69 to 1.23), or heart blok of any degree (OR 1-17; 0*83 to 1.65). Sinus bradyardia was signifiantly more ommon in the magnesium group (OR 138; 1-03 to 1-85; p = 0 02). These findings were orroborated by the use of treatments for rhythm disturbane and the data from Holter monitoring. Conlusion-The regimen of intravenous magnesium sulphate used here had no signifiant effet on arrhythmia in aute myoardial infartion. The redution in mortality that has been shown with this form of treatment is not attributable to suppression of life threatening rhythm disturbanes. (Br Heart _J 1994;71:141-145) 141 Intravenous magnesium salts have been used as antiarrhythmi agents for over 50 years' and have been given (but never rigorously evaluated) in a wide range of rhythm disorders. In reent years there has been partiular interest in their possible prophylati value in reduing the frequeny of arrhythmi ompliations of aute myoardial infartion. Experimentally, intravenous magnesium sulphate redued the inidene of ventriular fibrillation after oronary ligation in dogs.23 A onentration dependent suppression of ventriular arrhythmias ourred over the range of 1 2-4-8 mm [Mg2 +] during reperfusion of the isolated rat heart after temporary left main oronary artery olusion4; suppression of ligation indued (but not reperfusion indued) ventriular fibrillation in the same rat model had earlier been reported at 7-3 mm [Mg2 +].5 In the isolated guinea pig myoyte depolarised by high external [K +], 4 mm external [Mg2 +] signifiantly inreased the threshold onentration of isoprenaline required to produe automatiity.6 Several small linial trials of intravenous magnesium sulphate or hloride in aute myoardial infartion have given mixed results. Attained [Mg2'+] has been substantially lower than in the experimental studies desribed earlier, in the range 1P0-IS5 mmol/l (normal range approximately 0 70-0 95 mmol/l). Four double blind, randomised trials in a total of nearly 900 patients with suspeted or definite aute myoardial infartion showed no signifiant redution in any lass of ventriular arrhythmia.7-'0 One study (48 patients with onfirmed aute myoardial infartion) showed a highly signifiant redution in inidene of ventriular tahyardia in the magnesium group but a higher inidene of other ventriular arrhythmias."1 Two randomised trials, eah of fewer than 50 patients, reported signifiant redutions in the inidene of ventriular arrhythmias."2"3 A further randomised trial in whih 130 patients with onfirmed aute myoardial infartion reeived either intravenous magnesium hloride or plaebo showed a highly signifiant redution in the inidene of supraventriular tahyardias in the magnesium group but no effet on ventriular arrhythmias."4 Pooled analysis of the arrhythmia data in these trials is unhelpful beause of the differing definitions of rhythm end points. We have therefore analysed all available data on the frequeny of arrhythmias in a large double blind trial (the seond Leiester
142 Roffe, Flether, Woods intravenous magnesium intervention trial LIMIT-2) study) in whih 2316 patients with suspeted aute myoardial infartion were randomly alloated to reeive intravenous magnesium sulphate or saline for 24 hours after admission. In this study 28 day mortality was the primary endpoint and frequeny of arrhythmia within the oronary are unit a seondary endpoint defined in the protool. Three methods (eah blind to treatment alloation) were used to identify a possible antiarrhythmi effet of magnesium: an analysis of all arrhythmias reorded linially in trial patients during routine monitoring while in the oronary are unit; analysis of the use of antiarrhythmi treatments speified by the protool; and a Holter monitoring study in a representative subgroup of trial patients to quantify all abnormal rhythms. A summary of the data from the first two analyses were presented in brief in the first trial report."5 Patients and methods The protool of LIMIT-2 and data on 28 day mortality and early morbidity have been desribed elsewhere.'5 Briefly, onsenting patients entering the oronary are unit of Leiester Royal Infirmary between September 1987 and February 1992 were eligible for randomisation if they were suspeted of having aute myoardial infartion of less than 24 hours in duration. Exlusion riteria were omplete heart blok or serum reatinine > 300,umol/l. The trial treatment was 8 mmol magnesium sulphate infused intravenously over five minutes, then 65 mmol over 24 hours. The plaebo treatment was an equal volume of saline. The magnesium regimen raised serum [Mg2 +] to a mean of 1-5 mmol/l for the duration of treatment; the onentration then delined to the normal range during a further 24 hours. Trial treatments were supplied in idential oded paks and the alloation ode was not broken until the end of the trial. Analysis was by intention to treat. While in the oronary are unit (about 48 hours on average) all patients underwent ontinuous eletroardiographi monitoring. While in the unit, a standard linial reord is made for eah patient (whether or not in a trial) of all signifiant events and of treatments given. These data are reorded by linial staff during the admission and subsequently entered into a omputer database. The ourrene of speifi lasses of arrhythmia is reorded but not quantified. Treatments are given aording to a omprehensive, written management poliy for the unit. Patients entering the LIMIT-2 study were managed onventionally in all other aspets of treatment. No antiarrhythmi agents are used prophylatially, with the exeption of lignoaine given for 24 hours after emergeny ardioversion for ventriular tahyardia. Holter monitoring was performed for the first 24 hours after randomisation in a subgroup of 70 patients entering the trial. The diagnosis of aute myoardial infartion was subsequently onfirmed in 51 of these, in three of whom the reordings were tehnially inadequate. The analysis was arried out on the remaining 48 patients with onfirmed aute myoardial infartion, among whom the reording was inomplete for two patients in the plaebo group beause of a premature end when emergeny ardioversion was given for ventriular fibrillation. These patients were inluded in the analysis of inidene of arrhythmia but ould not be inluded in the analysis of frequeny of events over 24 hours. Extrapolation from the available reordings from these two patients was onsidered to be invalid as there was a strong general trend to lower hourly rates of events over the reording period. Tapes were analysed by omputer, with visual verifiation of all abnormal rhythms. All data were olleted without knowledge of the patients' treatment group. Ventriular tahyardia was defined as the ourrene of > six ventriular extra systoles at a rate >100 beats/min and >120% of the sinus rate. Runs not meeting these riteria were lassified as salvos. Hypotheses were tested byz2 and Mann- Whitney statistis. Results The inidene of linially doumented tahyardias in the trial groups was not signifiantly different for ventriular fibrillation, ventriular tahyardia, supraventriular tahyardia, or atrial fibrillation (fig 1) although eah was slightly less ommon in 0 Ol' a) a) -0 10C 12 - D1 Plaebo U Magnesium 6 4 2 n n,.2 o o Co C 1.- *_=*- (l C.Q -CO C Q >>0 >0 Figure 1 Inidene while in the oronary are unit of tahyardia in patients randomised to intravenous magnesium sulphate (n = 1159) or plaebo (n = 1157).
Investigation of the effets of intravenous magnesium sulphate on ardia rhythm in aute myoardial infartion 143 Figure 2 Use of treatments for tahyardia in the randomised groups of LIMIT-2. Table 1 Inidene of heart blok and use of temporary paing in randomised groups while in the oronary are unit Magnesium Plaebo (n = 1159) (n = 1157) Degree of blok % % p Value First 2-1 1-3 0 11 Seond 2 7 2 2 0.51 Third 2.1 2-5 0 57 Any 48 42 049 Temporary paemaker 2-2 1 6 0 37 Table 2 Odds ratios (magnesium: plaebo) for inidene of arrhythmias and interventions (n = 2316) Arrhythmia or intervention Odds ratio (95% Cl) p Value Ventriular tahyardia 0-87 (0-63 to 1.20) NS Ventriular fibrillation 0 74 (0-46 to 1.20) NS Supraventriular tahyardia 0 69 (0-38 to 1 26) NS Atrial fibrillation 0 92 (0 69 to 1-23) NS Any tahyardia 0 82 (0 66 to 1 02) 0 06 Lignoaine 0 76 (0 51 to 1.11) NS Amiodarone 0 74 (0 49 to 1 10) NS Digoxin 0 93 (0-66 to 1 31) NS DC ardioversion 1.02 (0-66 to 1 57) NS Heart blok (any degree) 1 17 (0 83 to 1 65) NS Sinus bradyardia 1 38 (1-03 to 1 85) 0 02 Atropine 1 69 (1l25 to 2 30) < 0 001 Temporary paemaker 1 32 (0-70 to 2-51) NS the magnesium group. Heart blok, whether analysed by grade of blok or pooling all grades, was of similar inidene in both groups (table 1). The only differene in ardia rhythm that reahed signifiane between the groups was the doumentation of sinus bradyardia, whih was reorded in 108% of the magnesium group and 8-0% of the plaebo group (p = 002). Table 2 shows the effets as odds ratios (ORs) and their 95% onfidene intervals (95% CIs). These findings are supported by the data on treatments used and proedures performed. There was a slightly lower inidene of the use of lignoaine, amiodarone, and digoxin in the magnesium group but none of the omparisons reahed signifiane (fig 2). 0 a a1) -0 10 _ Plaebo * Magnesium Table 3 Charateristis ofpatients inluded in the Holter monitornng subgroup Magnesium Plaebo (n = 22) (n = 26) Age (yr, SD) 65.7 (12 7) 60 2 (9 7) Men (%) 15 (682) 19 (73 1) Site of infartion (%): Anterior 12 (54 5) 14 (53 8) Inferior 7 (31 5) 10 (38 5) Indeterminate 3 (13 6) 2 (7 6) Mean delay from onset 4 2 3 5 to admission (h) Peak reatine kinase (u/i, SD) 2438 (1773) 1937 (1469) Thrombolysis (%) 14 (63 6) 17 (65 3) Insertion of a temporary paemaker and diret urrent ardioversion were similar in the two groups. Consistent with the signifiant exess of sinus bradyardia in the magnesium group, there was a orrespondingly higher rate of use of atropine (p < 0-001). A detailed analysis of mortality and morbidity in the trial shows no evidene that the ourrene of sinus bradyardia was deleterious"5; left ventriular failure in the oronary are unit and 28 day mortality were both signifiantly lower in the magnesium group. Among the analysable patients inluded in the Holter monitoring study who had a onfirmed aute myoardial infartion, 22 reeived magnesium sulphate and 26 plaebo. Table 3 shows their linial harateristis. The inidene of the various lasses of disturbane of ventriular rhythm deteted during Holter monitoring (fig 3) was not signifiantly different in the two groups. Beause the numbers of events in eah lass were not normally 2- GL) 0C0 100 90 80 _ 70 e- 60 50 40K 30 F- 20 10 Plaebo Magnesium T,T - C -X en e e Do La o X o > Q > > o Figure 3 Cumulative inidene of ventriular rhythm disturbanes during 24 hour Holter monitoring among patients with onfirmed myoardial infartion randomised to intravenous magnesium sulphate (n = 22) or plaebo (n = 24).
144 Roffe, Flether, Woods Table 4 Median number of ventriular arrhythmias n124 h (two reordings were inomplete in the plaebo group and were exluded) Magnesium Plaebo (n = 22) (n = 24) p Value Salvos 4 0 2-5 0-73 Triplets 5-5 3-5 0.99 Couplets 21 0 13-5 0-43 Ventriular extrasystoles 879 512 0-23 distributed, they were summarised (table 4) by their median values and omparisons were tested by a non-parametri method (Mann- Whitney test). No signifiant differenes were found in the amount of any lass of ventriular arrhythmia that ourred during the 24 hours of reording. Disussion The LIMIT-2 study is by far the largest randomised trial in whih the possible prophylati value of an intravenous magnesium salt against ishaemi arrhythmias has been examined. Although the identifiation of arrhythmias from routine ontinuous monitoring is known to be inomplete, the likelihood of detetion is in proportion to the linial signifiane of the arrhythmia. Moreover, the double blind design of the protool and the orroboration provided by the analysis of antiarrhythmi treatments given support the validity of the findings from the main study. The Holter monitoring study in a representative group of patients with onfirmed myoardial infartion, designed to ahieve omplete and aurate asertainment of all disturbanes of rhythm, also showed no signifiant antiarrhythmi effet of magnesium. The apparent disrepany between the findings of our study and the experimental models of myoardial ishaemia2-6 may reflet the differene in the extemal [Mg2 +] reahed or may be model dependent. The published linial trials of magnesium salts in myoardial infartion have generally been small and their findings inonsistent. The inonsisteny may partly arise from the retrospetive analysis of many different lasses and measures of arrhythmias. It is a strength of the present investigation that the ategories of arrhythmia analysed were either defined in advane in the linial data system (the main study) or else were omprehensive and mutually exlusive lasses desribing the totality of abnormal rhythms ourring during the study period (the Holter study). Data derived signifiane testing was therefore preluded. Outside the ontext of aute myoardial infartion, intravenous magnesium salts have been reported to be variably effetive in a wide range of arrhythmias, inluding sustained monomorphi ventriular tahyardia,'6 polymorphi ventriular tahyardia (torsade de pointes) of various aetiologies,' re-entrant supraventriular tahyardia,'820 and multifoal atrial tahyardia.2' In none of these have randomised trials of effiay been arried out. Torsade de pointes is the arrhythmia that seems to be most onsistently responsive to intravenous magnesium. The mehanism of ation does not depend on preexisting hypomagnesaemia or on shortening of an abnormally prolonged QT interval. The effets of doubling of serum [Mg2 +] on generation and propagation of the ardia ation potential have been examined in several linial eletrophysiologial studies.22-25 There is a small but onsistent prolongation of the atrium-his interval due to slowing of ondution through the upper AV node. This ation may aount for the reported effiay of magnesium salts in the treatment of AV re-entrant supraventriular tahyardia; it seems to arise from ompetitive blokade of the slow Ca2+ hannel and resembles the ation of the organi alium antagonist verapamil. Beause of the reported effet of Mg2 + on ondution through the AV node, evidene was sought in our study for any assoiation between magnesium treatment and the development or exaerbation of heart blok; none was found. The inreased inidene of sinus bradyardia was not predited by eletrophysiologial studies of intravenous Mg2 + in humans, whih have indiated a lak of effet on the length of the sinus node yle.22-2426 In these studies a diret inhibition of the sinus node by Mg2 + through an effet on the slow Ca2 + hannel or the aetylholine ativated K + hannel27 might have been offset by the reflex response to peripheral vasodilatation indued by Mg2 +, but if so it is unlear why our group of patients should have responded differently. A more likely explanation is that the relatively small eletrophysiologial studies laked the statistial power to detet a modest Mg2 + effet. These findings annot be generalised to patients with hypomagnesaemia, suh as ommonly ours during ardiopulmonary bypass; a randomised plaebo ontrolled trial has shown that partial repletion of [Mg2 +] with intravenous magnesium sulphate signifiantly redues the inidene of postoperative ventriular arrhythmias in this situation.28 Further systemati study of magnesium salts is required to larify the extensive ase reports suggesting effiay against a variety of arrhythmias. In torsades de pointes, for example, suppression of early afterdepolarisations and triggered ativity by extraellular [Mg2 +] offers a ellular mehanism in support of the empirial linial findings.29 In aute myoardial infartion, however, we have found no evidene that doubling serum [Mg2 +] redues the inidene of early arrhythmias. The signifiant redution of mortality seen in LIMIT-2 and in the pooled analysis of earlier studies'5 annot be aounted for by prevention of lethal arrhythmias. The LIMIT-2 study was supported by projet grants from the Leiestershire Health Authority and the British Heart Foundation. We thank the staff of the oronary are unit of the Leiester Royal Infirmary for their help in the olletion of data, and the staff of the eletroardiography departnent for their help with the Holter study.
Investigation of the effets of intravenous magnesium sulphate on ardia rhythm in aute myoardial infartion 145 1 Boyd L, Sherf D. Magnesium sulfate in paroxysmal tahyardia. Am J Med Si 1943;206:43-8. 2 Barros L, Da-Luz P, Siveira M, Chagas A, Pileggi F. Ventriular fibrillation in aute experimental myoardial ishaemia: protetion by magnesium sulphate. Braz J Med Biol Res 1988;21:791-9. 3 Billman G, Hoskins R. Prevention of ventriular fibrillation with magnesium sulphate. Eur J Pharmaol 1988; 158:167-71. 4 Bril A, Rohette L. Prevention of reperfusion-indued ventriular arrhythmias in isolated rat heart with magnesium. Can J Physiol Pharmaol 1990;68:694-9. 5 Crampton R, Clark C. Varying extraellular [Mg +] alters ishaemi and reperfusion ventriular tahyarrhythmias. Cirulation 1983;68(supp 111):146. 6 Hasegawa J, Matsumoto T, Takami T, Fujimoto Y, Kotake H, Mashiba H. Suppression of ateholamineindued abnormal paemaker ativities by magnesium ion in guinea pig ardia musle ells. Magnesium 1989;8:94-9. 7 Morton B, Nair R, Smith F, MKibbon T, Poznanski W. Magnesium therapy in aute myoardial infartion-a double blind study. Magnesium 1984;3:346-52. 8 Smith LF, Heagarty AM, Bing RF, Barnett DB. Intravenous infusion of magnesium sulphate after aute myoardial infartion: effets on arrhythmias and mortality. IntJCardiol 1986;12:175-80. 9 Feldstedt M, Boesgaard S, Bouhelouhe P, et al. Magnesium substitution in aute ishaemi heart syndromes. EurHeartJ 1991;12:1215-8. 10 Shehter M, Hod H, Marks N, Behar S, Kaplinsky E, Rabinowitz B. Benefiial effet of magnesium sulfate in aute myoardial infartion. Am 7 Cardiol 1990;66: 271-4. 11 Ceremuzynski L, Jurgiel R, Kulakowski P, Gebalska J. Threatening arrhythmias in aute myoardial infartion are prevented by intravenous magnesium sulphate. Am HeartJ 1989;118:1333-4. 12 Bershat F, Ising H, Gunther T, Sorgenfrei J, Wollitz M, Ibe K. Antiarrhythmi effets of magnesium infusions in patients with aute myoardial infartion. Magnesium Bulletin 1989;11:155-8. 13 Abraham AS, Rosenmann D, Kramer M, et al. Magnesium in the prevention of lethal arrhythmias in aute myoardial infartion. Arh Intern Med 1987; 147:753-5. 14 Rasmussen H, Suenson M, Norregard P, Balslev S. Magnesium infusion redues the inidene of arrhythmias in aute myoardial infartion. A double-blind plaebo-ontrolled study. Clin Cardiol 1987;10:351-6. 15 Woods KIb Flether S, Roffe C, Haider Y. Intravenous magnesium sulphate in suspeted aute myoardial infartion: results of the seond Leiester Intravenous Magnesium Intervention Trial (LIMIT-2). Lanet 1992;339: 1553-8. 16 Allen B, Brodsky M, Capparelli E, Lukett C, Iseri L. Magnesium sulphate therapy for sustained monomorphi ventriular tahyardia. Am J Cardiol 1989;64: 1202-4. 17 Tzivoni D, Banai S, Shuger C, et al. Treatment of torsade de pointes with magnesium sulphate. Cirulation 1988;77:392-7. 18 Etienne Y, Blan J, Boshat J, et al. Effets antiarythmiques du sulphate de magnesium intraveineux dans les tahyardies supraventriulaires paroxystiques. Ann Cardiol Angeiol 1988;37:535-8. 19 Wesley R, Haines D, Lermon B, DiMaro J, Crampton R. Effet of intravenous magnesium sulphate on supraventriular tahyardia. AmJ Cardiol 1989;63:1129-3 1. 20 Sager P, Widerhom J, Petersen R, et al. Prospetive evaluation of parenteral magnesium sulphate in the treatment of patients with reentrant AV supraventriular tahyardia. Am Hearty 1990;119:308-16. 21 Iseri L, Fairshter R, Hardemann J, Brodsky M. Magnesium and potassium therapy in multifoal atrial tahyardia. Am HeartJ 1985;110:789-94. 22 Kulik D, Hong R, Ryzen E, et al. Eletrophysiologi effets of intravenous magnesium in patients with normal ondution systems and no linial evidene of signifiant ardia disease. Am HeartJ 1988;115:367-73. 23 Rasmussen H, Thomsen P. The eletrophysiologial effets of intravenous magnesium on human sinus node, atrioventriular node, atrium, and ventrile. Clin Cardiol 1989;12:85-90. 24 Rogiers P, Vermeier W, Kesteloot H. Effet of the infusion of magnesium sulfate during atrial paing on ECG intervals, serum eletrolytes, and blood pressure. Am Hearty 1989;117:1278-83. 25 Pertione F, Ceravolo R, Costa R, Mattioli P. Eletrophysiologi effets of magnesium sulfate infusion in patients with ardia ondution defets. J Am Coil Nutr 1992;11:405-9. 26 DiCarlo L, Morady F, De Buitleir M, Krol R, Shurig L, Annesley T. Effets of magnesium sulfate on ardia ondution and refratoriness in humans. Y Am Coll Cardiol 1986;7:1356-62. 27 White R, Hartzell H. Magnesium ions in ardia funtion. Regulator of ion hannels and seond messengers. Biohem Pharmaol 1989;38:859-67. 28 England M, Gordon G, Salem M, Chemow B. Magnesium administration and dysrhythmias after ardia surgery. A plaebo-ontrolled, double-blind, randomized trial. JAMA 1992;268:2395-402. 29 Davidenko J, Cohen L, Goodrow R, Antzelevith C. Quinidine-indued ation potential prolongation, early after depolarizations, and triggered ativity in anine Purkinje fibers. Cirulation 1989;79:674-86. Br Heart J: first published as 10.1136/hrt.71.2.141 on 1 February 1994. Downloaded from http://heart.bmj.om/ on 8 Otober 2018 by guest. Proteted by opyright.