1435

insulin is immunochemically distinguishable from porkinsulin although both have the same aminoacid sequence.From these data, it would seem that these insulins differin their conformation despite their sharing of an identicalprimary sequence (Berson and Yalow 1966). This furthersuggests that the precise spatial conformation of a proteinis perhaps not solely determined by aminoacid sequenceand that there may be genetic control over the conforma-tion of proteins as well (Smith 1966).The abnormal resistance of diabetic insulin to insulinase

may very well be indicative of some structural change inthe insulin molecule making the disulphide bonds lessreadily available to the action of the enzyme and moreresistant to proteolysis. Because of the high frequency ofthe diabetic gene, it is not inconceivable that insulin is a

polymorphic protein. The overlap between our threepopulations and the scatter within each group are perhapsrelated to the relative proportions of various forms ofinsulin present in the same individual. The finding ofsimilar insulin-destruction rates on two occasions in

isogenic twins favours the hypothesis that individualvariations might be on a genetic basis and suggests thatthe test of resistance of insulin to insulinase may measurea primary gene effect which in unrelated individuals mightbe modified by other genetic, possibly allelic variants.The inheritance of diabetes remains undetermined

, (Steinberg 1965): elucidation of its mode of inheritanceawaits a more precise clarification of the disease (Frezal1965). If a large number of predisposed genotypes arenever detected, it is probably related to the fact that thediscriminants used are far removed from the primary geneeffect (Steinberg 1965). In fact, the situation with respectto diabetes reminds one of attempts to analyse the haemo-globinopathies before chemical techniques were availablefor the identification of haemoglobins to discriminatebetween possible genotypes. To our knowledge, theabnormal resistance of juvenile diabetic insulin to insulinaseconstitutes the first but not the only (Shapcott, O’Brien,and Roy 1966) evidence, that diabetic insulin is abnormal.This resistance seems to be present to a variable degree ina significant proportion of parents of juvenile diabeticsand in perhaps a small percentage of healthy individualswith no family history of diabetes. Despite the short-comings inherent to the limited sampling and the poordiscriminant power in individual cases, we suggest thetest of resistance of insulin to insulinase may be a geneticdiscriminant since it is concerned with a behaviouralcharacteristic of the insulin itself.

SummaryThe destruction of insulin by insulinase has been investi-

gated in seventeen healthy adults, sixteen parents ofjuvenile diabetics, and fifteen juvenile diabetics. The mean(±s.D.) destruction-rates per 30-minute period, expressedas % of mean values for standards, were 78-8 (±21-8),47-6 (±20-9), and 13-5 (±13-3), respectively. Whenplotted as- a frequency-distribution polygon there are

. indications that the parental distribution may be otherthan unimodal.

This test for insulinase resistance may be a measure of a

primary gene effect. The inheritance of diabetes may bemediated through a mutation which results in a struc-turally different insulin.

This work was supported by U.S. Public Health Service Grants,nos. HD-2578-01 and FR-69. We thank Miss Clara Trombley forher invaluable technical help; Dr. Edmond A. Murphy, Dr. ArthurRobinson, and Dr. Strother Walker for their helpful suggestions; thehouse staffs of both the Children’s Hospital of Denver and of the

Colorado General Hospital for providing us with serum-samplesfrom untreated juvenile diabetics.

Requests for reprints should be addressed to C. C. R., Departmentof Pediatrics, University of Colorado Medical Center, Denver,Colorado 80220, U.S.A.

REFERENCES

Berson, S. A., Yalow, R. S. (1961) Nature, Lond. 191, 1392.- - (1966) Am. J. Med. 40, 676.

Cassie, R. M. (1954) Aust. J. mar. Freshwat. Res. 5, 513.Elliott, R. B., O’Brien, D., Roy, C. C. (1965) Diabetes, 14, 780.Frezal, J. (1965) in On the Nature and Treatment of Diabetes in Man

(edited by B. S. Leibel and G. A. Wrenshall); p. 609. Amsterdam.Katzen, H. M., Tietze, F. (1966) J. biol. Chem. 241, 3561.Kenny, A. J. (1960) Br. med. Bull. 16, 202.Kruskal, W. H., Wallis, W. A. (1952) J. Am. statist. Ass. 47, 583.Morgan, C. R., Lazarow, A. (1962) Proc. Soc. exp. Biol. Med. 110, 29.Murphy, E. A. (1964) J. chron. Dis. 17, 301.Shapcott, D. J., O’Brien, D., Roy, C. C. (1966) Unpublished.Smith, L. F. (1966) Am. J. Med. 40, 662.Steinberg, A. G. (1965) in On the Nature and Treatment of Diabetes in Man

(edited by B. S. Leibel and G. A. Wrenshall); p. 601. Amsterdam.Stimmler, L., Elliott, R. B. (1964) Lancet, i, 956.Varandani, P. T., Tomizawa, H. H. (1965) Biochim. biophys. Acta, 97, 498.Whitney, D. R. (1948) Ann. math. Statist. 22, 274.Williams, R. H., Wood, F. C. (1965) in On the Nature and Treatment of

Diabetes in Man (edited by B. S. Leibel and G. A. Wrenshall); p. 759.Amsterdam.

Yagi, Y., Maier, P., Pressman, D. (1965) Science, N.Y. 147, 617.

PROPRANOLOL IN ACUTE MYOCARDIALINFARCTION

A Multicentre Trial *

IN a survey of twenty-four series comprising 11,153cases of acute myocardial infarction between 1937 and1964, the average mortality during the period in hospitalwas 33-1% (Stephen 1966); and it is now well recognisedthat arrhythmias after acute myocardial infarction are

commoner than was once thought (Imperial et al. 1960,Spaan et al. 1964, Julian et al. 1964, Mower et al. 1964,Day 1966, Meltzer and Kitchell 1966). Mortality in casesof acute myocardial infarction has been reduced by theintroduction of acute coronary-care units and may now beas low as 16% (Day 1965). But such units are not readilyavailable to all, and an effective method of treatment forpatients in ordinary medical wards is urgently required.The established value of 3-receptor blocking agents,pronethalol and propranolol, in various cardiac arrhythmias(Stock and Dale 1963, Payne and Senfield 1964, Tayloret al. 1964, Rowlands et al. 1965, Sloman et al. 1965,Harrison et al. 1965, Bath 1966, Szekely et al. 1966, Harris1966, Schamroth 1966, Stock 1966, Turner 1966) and theencouraging preliminary report of reduction in mortalityin acute myocardial infarction (Snow 1965) prompted thearrangement of a multicentre, double-blind trial to assessthe prophylactic value of propranolol after acute

myocardial infarction.Patients and Methods

Physicians from ten centres participated in the trial. Patientswere included who presented with a history of acute myocardial* The following participated: Dr. J. C. J. L. BATH (Western General

Hospital, Edinburgh); Dr. S. BLAKE (Mater Misericordiæ

Hospital, Dublin); Dr. R. A. BLOOM and Dr. R. BROWN (St.Andrew’s Hospital, Billericay, Essex); Dr. H. A. FLEMING(Addenbrooke’s Hospital, Cambridge); Dr. A. J. FRANKLIN

(Astley Hospital, Astley, Lancs); Dr. R. M. FULTON (SteppingHill Hospital, Stockport); Dr. E. JACKSON (Nottingham GeneralHospital); Dr. W. J. H. LECKIE (County Hospital, Hereford);Dr. A. G. MELROSE (Southern General Hospital, Glasgow);Dr. J. R. B. TURNER (Kingston General Hospital, Hull); Dr.R. W. D. TURNER (Western General Hospital, Edinburgh); andDr. R. J. TWORT (Nottingham General Hospital). The trial wascoordinated by Dr. S. A. STEPHEN (I.C.I. Research Laboratories,Alderley Park, Macclesfield, Cheshire).

1436

TABLE I-REASONS FOR WITHDRAWAL FROM TRIAL

infarction within the preceding 24 hours (later extended to48 hours). The diagnosis, which was confirmed as soon aspossible after admission, required the presence of any two of thefollowing criteria: characteristic clinical presentation; electro-cardiogram changes of recent infarction (pathological Q wavesand ST/T changes) or evidence of acute ischaemia (ST/T waveonly); elevation of serum-glutamic-oxaloacetic-trans-aminase(S.G.O.T.), serum-lactate-dehydrogenase (S.L.D.H.), or serum-hydroxybutyric-dehydrogenase (S.H.B.D.) above the normalfor the hospital.Prognostic Index

Peel’s coronary prognostic index (Peel et al. 1962) was usedto assess the severity of the infarction, and the immediateprognosis for each patient. The distribution in each group isset out in table iv.

Non-acceptancePatients were excluded from the trial if: the diagnostic

criteria were not fulfilled; there was evidence of bronchospasmTABLE II-MORTALITY

or a clinical history of bronchial asthma; the heart-rate wasless than 60 per minute persisting throughout a 24-hour period;or the systolic blood-pressure was less than 80 mm. Hg afteradmission.

InvestigationsThe following observations were recorded: pulse-rate

(4-hourly); temperature (twice daily); blood-pressure (daily);Hb, erythrocyte-sedimentation rate (within 24 hours of

admission); serum-enzymes (between 24 and 48 hours afteradmission); E.C.G. (on admission, twice in first week, andthereafter at weekly intervals, more frequent records if

necessary); and chest X-ray (if thought necessary).Treatment

Propranolol (’ Inderal’) 20 mg. or an identical placebo wasgiven 6-hourly by mouth for 28 days and the study was double-blind with random allocation of patients to the placebo ortreated group. Anticoagulants were given according to thenormal practice for each physician. Digitalis, quinidine,procainamide, diuretics, and pain-relieving drugs were givenwhen indicated to both groups.

TABLE III-AGE AND SEX DISTRIBUTION

Results

Of the 226 patients who entered the trial, 31 werewithdrawn for reasons shown in table I, leaving 195 foranalysis. There were 155 men and 40 women. Agesranged from 27 to 83 years (average 58 years). Out of 100patients receiving propranolol, there were 15 deaths

(15%) and of 95 patients receiving the placebo, there were12 deaths (12-6%) (table 11). Deaths in relation to sexand age distribution are shown in table ill. Groups wereevenly matched except for a preponderance of placebo-treated patients under the age of 45, but this was nota significant difference.There is no statistically significant difference between

the distribution of coronary prognostic indices (C.P.I.) inthe drug and placebo groups (table iv). To test the

TABLE IV-CORONARY PROGNOSTIC INDEX

accuracy of this classification, the mortalities for drug andplacebo groups in this study were compared with expectedmortality using Peel’s classification. Although the numberof severely ill patients is small, the general impression isthat the mortality in both drug and placebo groups andthe expected mortality from Peel’s classification are

similar (table v). The large proportion of patients in the" mild " groups (1-8) was due to selection, patients withpersistent hypotension or bradycardia having beenexcluded from the trial.The average time between onset of symptoms and start

of treatment was 16-2 hours in the propranolol group and13-4 hours in the placebo group (table vi). The averagetime between onset of treatment and death was 8-5 days in

TABLE V-EXPECTED MORTALITY-RATES BASED ON THE RESULTS OF

PEEL ET AL. (1962)

the propranolol group and 8-9 days in the placebo group(table vn). These differences are not significant.Complications of Infarction

Episodes of cardiac failure and hypotension are reportedin table vm. Twice as many patients due to receive

propranolol were admitted in heart-failure or had heart-failure within 24 hours. It seems unlikely that propranololcontributed to heart-failure at this stage. Similarly,episodes of hypotension (5 on propranolol and 4 onplacebo) were unlikely to be influenced by treatment.During the period 24 hours to 28 days, however, it isevident that both cardiac failure and hypotension occurredmore than twice as often in the propranolol-treatedpatients and it is hard to escape the conclusion that

1437

TABLE VI-TIME BETWEEN ONSET OF SYMPTOMS AND START OF TREATMENT

&bgr;-receptor blockade contributed to these episodes(table vill).

Qardiac arrest occurred in 9 patients in the propranololgroup and in 7 in the placebo group. Because monitoringof the electrocardiogram was not generally carried out, nouseful conclusions about the incidence of arrhythmias canbe made.Nature of DeathsDeath was sudden or unexpected in 7 propranolol-

treated patients and in 9 in the placebo group. Cardiacfailure contributed to 3 deaths in each group, and thusthe higher incidence of cardiac failure noted previously inthe treated group (24 hours-28 days) does not seem tohave contributed to mortality.

TABLE VII-TIME BETWEEN START OF TREATMENT AND DEATH

Discussion

This study in 195 patients with acute myocardialinfarction has demonstrated no difference in mortalitybetween a propranolol-treated group and a control group.This is in agreement with other series (Balcon et al. 1966,Clausen et al. 1966) and differs from the results obtainedby Snow (1965, 1966) which showed a significant reduc-tion in mortality using &bgr;-receptor blockade. The reasonfor the difference is not clear. Snow (1965) used alternatepatients for treatment with propranolol, and no placebotablets were given. Thus the trial was not truly random-ised nor was it blind. Since this was the first trial of

propranolol in a potentially dangerous situation, we feelthat these methods were justified. As it turned out, histrial ended with a comparable number in each group withregard to age, sex, and clinical severity of infarct. Mostimportant, Snow (1965) accepted patients of all degreesof severity and noted a mortality in the control group of35% and in the treated group of 16%. In our series,seriously ill patients were excluded and mortality mighttherefore be expected to be low. In fact 73% of treated

TABLE VIII-COMPLICATION OF INFARCTION

patients and 75-8% of placebo patients had a C.P.I. of 12or less (table v), and the mortality in the two groups was15% (propranolol) and 12-6% (placebo). The predictedmortality, based on the c.P.i., was 15% (propranolol) and14% (placebo). It is possible that any favourable effects,of j3-receptor blockade with propranolol would not beapparent in such a low-risk group of patients.

Series in which patients were monitored (Julian et al.1964, Brown 1965, Meltzer and Kitchell 1966) haveshown that cardiac arrhythmias are common complica-tions of acute myocardial infarction and that the prognosisin serious arrhythmias is poor. Julian et al. (1964) found20 deaths in 56 patients with serious arrhythmias com-pared with 11 deaths in the remaining 44 patients in theirseries. Propranolol can be effective in controllingventricular fibrillation after acute cardiac infarction

(Sloman et al. 1965); and Shanks (1966b) has shown indogs that the exaggerated ectopic response to adrenalineadministered after ligation of the left carotid artery can beblocked by propranolol. It might, therefore, be expectedthat propranolol would reduce the incidence of arrhythmiasand hence the mortality after acute myocardial infarction.

It is also well known that increased cardiac contractileforce and increased heart-rate follow cardiac sympatheticor catecholamine stimulation and these effects are readilyabolished by propranolol in animals (Shanks 1966a) andin man (Hamer and Sowton 1965, Sowton and Hamer1966) resulting in a decrease in left-ventricular work anda decrease in oxygen uptake by the myocardium (Sowtonand Hamer 1966). Measurements of myocardial oxygenconsumption in individuals with normal hearts and inthose with coronary-artery disease during diagnosticcardiac catheterisation have demonstrated that pro-

pranolol decreases oxygen consumption by an average of25% (Wolfson et al. 1966).The results of catheterisation studies on conscious

unpremedicated patients (Mendel and Byrne-Quinn 1966)suggest that the decrease in coronary blood-flow caused

by the administration of propranolol is in proportion tothe reduced amount of work that the heart has to perform.There seems to be no reliable evidence at present that thereduction of coronary blood-flow after propranolol is

disproportionate, as suggested by Balcon et al. (1966).The reduction in ventricular work and thus the demandson the coronary circulation after propranolol might beexpected to prove beneficial in myocardial infarction

although no confirmation of this has been forthcoming inour study.

SummaryIn a multicentre, double-blind trial of propranolol in

acute myocardial infarction 195 patients received eitherthe drug or a placebo for 28 days. Mortality was 15%in the treated group and 12-6% in the placebo group. Thelow mortality-rate seems to be due to selection of patients.Cardiac failure and hypotension occurred more commonlyin patients receiving propranolol.We thank our colleagues who kindly allowed us to study patients

under their care and the registrars, house-physicians, and nursingstaff, for their help; also Mr. C. J. Clark and Mr. D. Williams,statistical department, I.C.I.; and finally Miss Susan Leask forsecretarial assistance.

Requests for reprints should be addressed to S. A. S.

REFERENCES

Balcon, R., Jewitt, D. E., Davies, J. P. H., Oram, S. (1966) Lancet, ii, 917.Bath, J. C. J. L. (1966) Am. J. Cardiol. 18, 415.Brown, K. W. G. (1965) Cited by Yu, P. W. et al. (1965) Mod. Concepts

cardiovasc. Dis. 34, 23.

References continued overleaf

1438

DR. STEPHEN AND OTHERS: REFERENCES—continued

Clausen, J., Felsby, M., Schønan Jørgensen, F., Lyager Nielsen, B., Roin,J., Strange, B. (1966) Lancet, ii, 920.

Day, H. W. (1965) Am. J. Cardiol. 15, 51.- (1966) ibid. 17, 121.

Hamer, J., Sowton, G. E. (1965) Br. Heart J. 27, 892.Harris, A. (1966) Am. J. Cardiol. 18, 431.Harrison, D. C., Griffin, J. R., Fiene, T. J. (1965) New Engl. J. Med. 273, 410.Imperial, E. S., Carballo, R., Zimmerman, H. A. (1960) Am. J. Cardiol. 5, 24.Julian, D. G., Valentine, P. A., Miller, G. G. (1964) Am. J. Med. 37, 915.Mendel, D., Byrne-Quinn, E. (1966) Lancet, ii, 1026.Meltzer, L. E., Kitchell, J. B. (1966) Progr. cardiovasc. Dis. 9, 50.Mower, M. M., Miller, D. I., Nachlas, M. M. (1964) Am. J. Cardiol. 67, 437.Payne, J. P., Senfield, R. M. (1964) Br. med. J. i, 603.Peel, A. A. F., Semple, T., Wang, I., Lancaster, W. M., Dall, J. L. G. (1962)

Br. Heart J. 24, 745.Rowlands, D. J., Howitt, G., Markman, P. (1965) Br. med. J. i, 891.Schamroth, L. (1966) Am. J. Cardiol. 18, 438.Shanks, R. G. (1966a) ibid. p. 308.- (1966b) Cardiovasc. Res. (in the press).

Sloman, G., Robinson, J. S., McLean, K. (1965) Br. med. J. i, 895.Snow, P. J. D. (1965) Lancet, ii, 551.- (1966) Am. J. Cardiol. 18, 458.

Sowton, E., Hamer, J. (1966) ibid. p. 317.Spaan, J. F., Moellering, R. C., Haber, E., Wheeler, E. O. (1964) New Engl.

J. Med. 271, 427.Stephen, S. A. (1966) Am. J. Cardiol. 18, 460.Stock, J. P. P. (1966) ibid. p. 444.- Dale, N. (1963) Br. med. J. ii, 1230.

Szekely, P., Jackson, F., Wynne, N. A., Vohra, J. K., Batson, G. A., Dow,W. I. M. (1966) Am. J. Cardiol. 18, 426.

Taylor, R. R., Johnston, C. I., Jose, A. D. (1964) New Engl. J. Med. 271, 877.Turner, J. R. B. (1966) Am. J. Cardiol. 18, 450.Wolfson, S., Heinle, R. A., Herman, M. V., Kemp, H. G., Sullivan, J. M.,

Gorlin, R. (1966) ibid. p. 345.

POTASSIUM, GLUCOSE, AND INSULIN INTREATMENT OF HEART BLOCK AFTER

MYOCARDIAL INFARCTION

B. MITTRAM.B., Ph.D. Belf., M.R.C.P.E., D.T.M. & H.

ACTING LECTURER, DEPARTMENT OF MEDICINE,QUEEN’S UNIVERSITY, BELFAST 12

HIGH-GRADE atrioventricular block is a rare but seriouscomplication of myocardial infarction because of the highmortality associated with it. Death in such cases is due,not to the conduction defect directly, but to the increasedsusceptibility to such arrhythmias as ventricular asystoleand fibrillation and to the adverse haemodynamic effectsof bradycardia.Many workers have commented on the transient nature

of complete heart block in myocardial infarction (Pentonet al. 1956, Gilchrist 1958), but since the immediate

mortality of this condition is so high it is reasonable to tryand restore normal conduction as quickly as possiblerather than wait for spontaneous cessation of thedisturbance.

Corticotrophin and corticosteroids (Prinzmetal andKennamer 1954, Dall and Buchanan 1962, Aber and WynJones 1965), isoprenaline and epinephrine (Nathanson andMiller 1952, Zoll et al. 1958), chlorothiazide diuretics(Tobian 1961), and molar sodium lactate (Bellet et al.1955, Vandam and McLemore 1957), have been used forthe treatment of heart block, with variable success.

Artificial cardiac pacing (Zoll et al. 1954, Davies andSiddons 1965) is used only as a last resort because it

requires complex equipment which is not always readilyavailable and also because it has certain other dis-

advantages such as the risk of provoking ventricularfibrillation (Julian et al. 1964, Harris et al. 1965).While conducting a clinical trial of potassium, glucose,

and insulin (P.G.I.) therapy in myocardial infarction

(Mittra 1965), it was observed that the regimen causedstriking improvement in atrioventricular conduction inpatients with heart block. It seemed desirable, therefore,

to draw attention to this effect of the therapy because thecommonly used therapeutic agents are not without theirshortcomings, whereas this regimen is practically devoidof any serious side-effects.

Patients and MethodsBetween August, 1964, and June, 1965, a controlled clinical

trial of P.G.I. in myocardial infarction was in progress at theRoyal Victoria Hospital, Belfast; the results have been pub-lished elsewhere (Mittra 1965). High-grade atrioventricularblock was noted in nine of the eighty-five treated patients andthe data on these nine cases are summarised in table i. Theresults achieved in these cases were compared to those observedamong eight of the eighty-five control patients with heartblock (table II).Of the nine treated patients, five (four men and one woman)

had complete heart block. The mean age was 65 years (range54-74 years) and the mean duration of heart block before thestart of P.G.I. therapy was 25-3 hours (range 30 minutes-4 days).Three of the five patients (cases 1, 3, and 4) presented withcomplete block on admission and the other two cases (nos. 2and 5) gradually progressed from first to second degree blockand then to complete heart block. None of the patients withcomplete block had received P.G.I. therapy before this com-plication developed. The four remaining patients had onlysecond-degree heart block (2:1 response and Wenckebach

phenomenon) and three of them had this defect while on P.G.I.therapy; in the fourth patient the regimen was given 24 hoursafter the onset of the disturbance, which was noted just afteradmission. The mean age in this group was 64-5 years (range59-71 years).

P.G.I. therapy was given intravenously to all patients withcomplete block and to one patient (case 7) with partial block.The rest received the potassium and glucose orally and theinsulin subcutaneously. The details of the regimen have beendescribed elsewhere (Mittra 1965). None of the nine patientsreceived any other medication directed to correct the con-duction defect.

All the patients (nine treated and eight controls) were

monitored in an intensive-care unit while they had the heartblock, and 12-lead electrocardiograms were taken frequently torecord their progress.

Results

The results obtained in the control and treated patientswith high-grade atrioventricular block are compared intable 11. Although the incidence of complete and second-degree heart block was similar in the two groups, themortality was significantly lower (p < 0-05) in the treatedpatients when compared with controls.

Details of the results achieved in the treated patientswith high-grade atrioventricular block are given in table I.Eight of the nine patients had posterior infarct, the

remaining one (case 9) showed evidence of subendocardialinfarction. Sinus rhythm was restored in all the patientsalthough grade-I block (P-R interval more than 0-2 second)persisted in three of them (cases 2, 4, and 6) for a longtime. None of the patients required any treatment, otherthan the regimen, for correction of heart block. There wasa striking relation between the duration of complete heartblock before commencement of therapy and the time takenfor resolution of the conduction defect after treatment.Sinus rhythm was restored within 12 hours in cases 1, 3,and 5 when treatment was started within 75 minutes ofonset of the heart block, whereas it took up to 42 hours incases 2 and 4 for resolution of the conduction defect whenthere was a delay in instituting therapy (table i).

Stokes-Adams syncopal attacks (cases 1-3) were

promptly abolished on commencement of P.G.I. therapywell before the restoration of sinus rhythm. Multiple andmultifocal ventricular ectopic beats (cases 3 and 4) werealso easily controlled by this regimen.