1036

Guanarito virus from the spleen of a cotton rat (Sigmodonhispidus), and antibodies to the virus were identified in theserum of a rice rat (Oryzomys spp) (unpublished). Thesimilar sex distribution among our cases and the infection ofa 6-year-old child (table I) indicate that transmission occursin and around houses, as with Lassa fever and Bolivianhaemorrhagic fever. 1,2,18 By contrast, Argentinianhaemorrhagic fever occurs mainly in adult male agriculturalworkers, who usually acquire the disease in rodent-infestedfields.l’

We thank Mr Juan Carlos Navarro, Mr Humberto Montanez, and Mr JoseManuel Rios, Malariology Service, Ministry of Health and Social Assistance,for help in collecting some of the human sera and rodent samples and Ms IrisRojas, Ms Celia Rivas, and Ms Magdalia Rojas, National Institute of Hygiene"Rafael Rangel" for their technical assistance. We also wish to acknowledgethe help of all the medical and paramedical personnel of the Miguel OraaHospital, in the care of the haemorrhagic fever patients. Our work wassupported in part by the Pan American Health Organization (travelfellowship to R. S.), the National Institutes of Health (grant AI-10984 toR. B. T.), and the US Army Medical Research and DevelopmentCommand (DAMD 17-90-Z-0020 to R. E. S.).

REFERENCES

1. Monath TP. Lassa fever: review of epidemiology and epizootiology. BullWHO 1975; 52: 577-92.

2. Johnson KM, Webb PA, Justines J. Biology of Tacaribe-complexviruses. In: Lehmann-Grube F, ed. Lymphocytic choriomeningitisvirus and other arenaviruses. Heidelberg: Springer-Verlag, 1973:241-58.

3. Peters CJ, Kuehne RW, Mercado RR, LeBow RO, Spertzel RO, WebbPA. Hemorrhagic fever in Cochabamba, Bolivia. Am J Epidemiol 1974;99: 425-33.

4. Monath TP. Lassa fever-new issues raised by field studies in WestAfrica. J Infect Dis 1987; 155: 433-36.

5. Rugiero HR, Parodi AS, Gotta H, Boxaca M, Olivari AJ, Gonzalez E.

Fiebre hemorrhagica epidemica. Infeccion de laboratorio y pasajeinterhumano. Rev Asoc Med Argent 1962; 76: 413-19.

6. Communicable Disease Program. Dengue hemorrhagic fever inVenezuela. PAHO Epidemiol Bull 1990; 11 (special issue no 2): 1-2.

7. Hernandez T. Entre ratas y garrapatas viaja la fiebre hemorrhagica. ElNacional Caracas, April 7, 1991.

8. Tesh RB. A method for the isolation and identification of dengue viruses,using mosquito cell cultures. Am J Trop Med Hyg 1979; 28: 1053-59.

9. Beaty BJ, Calisher CH, Shope RE. Arboviruses. In: Schmidt NJ,Emmons RW, eds. Diagnostic procedures for viral, rickettsial andchlamydial infections. Washington: American Public HealthAssociation, 1989: 797-856.

10. White HA. Lassa fever: a study of 23 hospital cases. Trans R Soc TropMed Hyg 1972; 66: 390-98.

11. Walker DH, McCormick JB, Johnson KM, et al. Pathologic and virologicstudy of fatal Lassa fever in man. Am J Pathol 1982; 107: 349-56.

12. Monath TP, Maher M, Casals J, Kissling RE, Cacciapuoti A. Lassa feverin Eastern Province of Sierra Leone, 1970-1972. II. Clinicalobservations and virological studies on selected hospital cases. Am JTrop Med Hyg 1974; 23: 1140-49.

13. McCormick JB, King IJ, Webb PA, et al. A case-control study of theclinical diagnosis and course of Lassa fever. J Infect Dis 1987; 155:549-53.

14. Mertens PE, Patton R, Baum JJ, Monarth TP. Clinical presentation ofLassa fever cases during the hospital epidemic at Zorzor, Liberia,March-April 1972. Am J Trop Med Hyg 1973; 22: 780-84.

15. Stinebaugh BJ, Schloeder FX, Johnson KM, Mackenzie RB, EntwisleG, DeAlba E. Bolivian hemorrhagic fever: a report of four cases. Am JMed 1966; 40: 217-30.

16. Child PL, Mackenzie RB, Valverde LR, Johnson KM. Bolivianhemorrhagic fever: a pathologic description. Arch Pathol 1967; 83:434-45.

17. Maiztegui JI. Clinical and epidemiologic patterns of Argentinehaemorrhagic fever. Bull WHO 1975; 50: 567-75.

18. McCormick JB, Webb PA, Krebs JW, Johnson KM, Smith ES. Aprospective study of the epidemiology and ecology of Lassa fever.J Infect Dis 1987; 155: 437-44.

19. Mackenzie RB, Beye HK, Valverde L, Garron H. Epidemic hemorrhagicfever in Bolivia. 1. A preliminary report of the epidemiologic andclinical findings in a new epidemic area in South America. Am J TropMed Hyg 1964; 13: 620-25.

Efficacy of nifedipine and isosorbide mononitrate incombination with atenolol in stable angina

ADDRESS: Cardiac Department, Guy’s Hospital, St ThomasStreet, London SE1 9RT, UK (F. Akhras, MD, G Jackson, FRCP).Correspondence to Dr Graham Jackson

Many patients with angina pectoris whose

symptoms are not completely controlled by beta-blockers are treated with several types of drugs, but itis not clear whether addition of a calcium-channel

antagonist and/or a nitrate confers any advantageover beta-blockade alone.18 patients receiving atenolol for stable angina

pectoris completed a double-blind, randomised,crossover trial of atenolol treatment plus placebo,isosorbide mononitrate, nifedipine, and mononitrateand nifedipine (triple therapy). The patients wereassessed subjectively and by treadmill exercise

testing and 24 h ambulatory electrocardiographicrecordings at the end of each 4-week treatmentperiod. There were no significant differences amongthe treatment periods in angina attack rates,glyceryl trinitrate consumption, exercise durationto onset of angina or 1 mm ST depression, or

duration of symptomless ischaemia. Total exerciseduration was longer on atenolol plus mononitratethan on atenolol alone (mean difference 46 [95%confidence interval 18-88] s; p=0·005), atenololplus nifedipine (36 [2-71] s; p=0·04), or triple

therapy (28 [6-61 ] s; not significant). In 12 patientsthe exercise time was shorter on triple therapy thanon atenolol plus mononitrate alone.Although "maximum" antianginal treatment with

two or three drugs is commonly accepted, thisapproach confers no substantial advantage overoptimum beta-blockade as monotherapy. If a seconddrug is needed, there is a slight advantage in favour ofisosorbide mononitrate, but if this is not effective,treatment should be changed rather than added.Many patients with angina pectoris seem to bepharmacologically overtreated.

Introduction

Nitrates and beta-adrenoceptor antagonists, either aloneor in combination, have been the main therapeutic optionsin management of chronic stable anginal The calcium-channel antagonists, such as nifedipine, are another class of

1037

drugs which act by different mechanisms from nitratesand beta-blockers. What was previously considered

"maximum" therapy may now be extended, with the use ofall three drugs. Nifedipine is a potent coronary vasodilator atrest and reduces afterload.2 These actions could add to the

antianginal effects of mononitrates,3.4 which lower

myocardial oxygen demand primarily by reducing preload,and of beta-blockade,5,6 which reduces heart rate, bloodpressure, and contractility. Nitrates and nifedipine are

complementary and are widely coprescribed with beta-adrenoceptor antagonists in stable angina thought to beinadequately controlled by beta-blockade alone. Manystudies have shown a clinical benefit from use of the drugsindividually, and with some combinations.1.7 However,there is little information to show whether nitrates orcalcium-channel antagonists represent the "better" second-step drug or whether all three drugs prescribed togetherprovide objective and clinical improvement over two inpatients with anginal symptoms not completely controlledby beta-blockade alone. In this prospective study weexamined the effects of nitrate and nifedipine incombination with the beta-blocker.

Patients and methods

This randomised, double-blind, crossover trial comparedatenolol 100 mg daily alone with atenolol plus nifedipine (’Adalatretard’) 40 mg daily and/or isosorbide mononitrate 40 mg daily.Study medication was predispensed for each patient in identicalcapsules containing either 50 mg atenolol or 50 mg atenolol plus 20mg nifedipine. Isosorbide mononitrate was dispensed as a 20 mgtablet with matching placebo. Each drug was taken by mouth twicedaily at 0900 h and 2100 h (one capsule and one tablet being taken ateach time) for 4 weeks. Each patient then changed treatment asdetermined by the random scheme. Randomisation to treatmentsequence was done in blocks of four for 24 potential patients.Patients were assessed before entry to the study and at the end ofeach of the four treatment periods by maximum symptom-limitedtreadmill exercise testing, 24 h ambulatory electrocardiography(ECG) recordings, and an analysis of angina diaries. Venous bloodsamples were taken at each visit for routine laboratory tests, andplasma drug concentrations were measured. The study protocolwas approved by the hospital ethics committee.

Fully informed written consent was obtained from 22 patients (1 1woman, 21 men) of mean age 60 (range 47-67) years. All patientshad proven coronary artery disease (luminal narrowing of 70% ormore in at least one of the three major coronary arteries), positiveexercise ECG, and exertional angina pectoris of not less than threeepisodes a week during adequate beta-blockade. 11 patients weretaking nitrates and 10 nifedipine (7) or diltiazem (3) with thebeta-blocker. All agents other than atenolol were withdrawn at least24 h before entry to the study. 13 patients had had a myocardialinfarction and all but 1 had multivessel coronary artery disease.

At the end of each 4-week treatment period and at the same timeof the day, patients underwent symptom-limited maximumexercise testing according to the Bruce protocol. ST segmentdepression was defmed as 1 mm or more horizontal or downslopingshift 80 ms beyond the J point that persisted for at least 30 s.

Exercise was stopped if ST depression of 3 mm or more occurred,

TABLE I-GLYCERYL TRINITRATE CONSUMPTION AND ANGINAATTACKS DURING EACH TREATMENT PERIOD

Fig 1-Exercise duration during each treatment period.

A=atenolol ; N = nifedipine; M = mononitrate. p values apply to totalexercise duration.

or in the event of chest pain, hypotension, or fatigue. The bloodpressure, heart rate, and rate-pressure product were measured atrest, at submaximum exercise (onset of 1 mm ST segmentdepression and onset of angina), and at peak exercise. Maximum STdepression and total exercise duration were measured. The doses ofnifedipine and isosorbide mononitrate used were based on previousresults.3,4,8,9 The dose of atenolol was selected to achieve adequatebeta-adrenergic receptor blockade as shown by control of rest andexercise heart rates before entry to the study."

All patients underwent 24 h ambulatory ECG recordings bymeans of frequency-modulated ’Tracker’ recorders type TR I(Reynolds) which gave reproducible low frequency signals. Thebipolar leads CM2 and CM5 were recorded continuously for 24 h,initially with the patients receiving no medication except beta-blockade therapy, and then at the time of the exercise tests for eachof the four active treatment periods. The tapes were analysed at 300times normal speed on the Reynolds ’Pathfinder III’ analyser typeTM PA3. The time, duration, and maximum ST segmentdepression during each episode were measured. The heart rates 60 sbefore and at the onset of ST segment depression during eachepisode were measured, as well as minimum and maximum heartrate and numbers of ventricular extrasystoles. The numbers ofpainful and symptomless episodes were also recorded in eachtreatment period.The main efficacy variables for each treatment period were

compared by means of Latin square analysis, which fitted treatmentperiod and period effects as well as any possible treatmentinteractions: estimates of the direct effects of treatments, comparingeach treatment with all three other treatments, were obtained fromthe model. These yielded estimates of treatment differences andtheir standard errors. Such analysis generated p values for thesignificance of the pairwise comparisons, though not adjusted formultiple testing.An accurate comparison of the clinical efficacy of nifedipine and

isosorbide mononitrate in patients receiving atenolol is difficultbecause complete dose-response curves for all these agents are notavailable. A factorial analysis was therefore undertaken andassumed a factorial design in which atenolol was baseline treatmentand nifedipine and mononitrate the two factors of interest. Allpatients received all combinations of the drugs during differentperiods. This approach allowed the study to test for any interactionbetween nifedipine and mononitrate, and provided support forestimates of the effects of results obtained from the Latin squareanalysis. These results were consistent with the Latin squareanalysis so are not reported here.The subjective variables were analysed by calculation of both

means and medians as measures of central position, and thenon-parametric Friedman’s test was used to compare the four

1038

Fig 2-Ambulatory cardiography: effect of treatment on totalduration of ischaemia, duration of symptomless ischaemia,and heart rate within 60 s of symptomless ischaemia.

treatment periods with respect to the original attack rate andglyceryl trinitrate consumption.

Statistical analysis was carried out independently by means of thecomputer package SAS (version 5, 1986. SAS Institute, Cary,North Carolina).

4 patients (including the only woman) did not complete the study:3 had intolerable headaches when taking mononitrate with orwithout nifedipine; and in 1 angina grew worse and an uneventfulmyocardial infarction occurred during the last treatment period(atenolol alone). Side-effects in the 18 patients who completed thetrial were minor and well tolerated. 1 patient reported headacheduring atenolol alone; 2 patients had dizziness, ankle oedema, andpalpitations on nifedipine; and 1 patient had headache and skin rashduring the triple medication period.

ResultsThe mean duration of angina before entry to the study

was 39-8 (SD 25-7) months, the mean number of anginaattacks per week was 9-1 (5-9), and the mean number ofglyceryl trinitrate tablets taken per week was 13’1 (4-6).

Glyceryl trinitrate consumption was slightly, but notsignificantly, higher during atenolol alone than during thecombination treatment periods (table i). There were nosignificant differences in angina attack rate among the fourtreatment periods (table 1), but within-patient variabilitywas greater for angina attack rate than for glyceryl trinitrateconsumption, with the largest difference between atenololand atenolol plus nifedipine.The Latin square analysis showed a significant treatment

effect on total exercise duration (p = 0-041) which wasshown by the pairwise comparison to be due mainly to thebetter performance on atenolol plus mononitrate (fig 1).Addition of mononitrate to atenolol increased tested exerciseduration by 46 s (95% confidence interval [CI] 18-88 s;p=0005). The difference in exercise duration betweenatenolol plus mononitrate and atenolol plus nifedipine was36 s (2-71 s; p = 0-0467). Exercise duration was also longer,but not significantly so, with atenolol plus mononitrate thanwith the combination of all three drugs (difference 28[6-61] s; p=0-109). The combination of atenolol plusmononitrate was used as reference to determine whether thecombination of all three drugs gave any greaterimprovement. Exercise tolerance improved in only 5

patients, by an average of 87 (range 36-210) s; 1 patientshowed no change; and 12 patients did worse with tripletherapy than with atenolol plus mononitrate.

There were no significant differences in time to 1 mm STdepression, time to onset of angina, or maximum STsegment depression between atenolol alone and any of thecombinations (fig 2, table II). Most patients stopped exercisebecause of pain and no significant differences were foundamong the treatment periods in the proportion of patientswho were stopped by chest pain.

Adequate beta-blockade was confirmed by suppression ofthe heart rate response at peak exercise during all treatmentperiods. The addition of nifedipine and/or mononitrateproduced no significant increase in heart rate either at rest orduring exercise (table II). Values of systolic blood pressure at

TABLE II-REST AND EXERCISE HAEMODYNAMICS AND ECG CHANGE DURING TREATMENT PERIODS

*Significantly lower than at onset of pain (p = 0 02)

1039

TABLE III-BLOOD OR PLASMA CONCENTRATIONS OF

TREATMENT DRUGS

submaximum or peak exercise and of double product (heartrate x systolic blood pressure) at rest or peak exercise weresimilar in all four treatment periods (table n). However, thedouble product was significantly lower at time to 1 mm STdepression than at time to onset of angina on atenolol aloneor on all three drugs combined (p = 0-02).24 h ambulatory ECG recordings showed no significant

treatment differences with respect to heart rate, ventricular

arrhythmias, and ischaemic ST segment change (fig 2).Since not all patients had ischaemic episodes at baseline, thedistributions of the number or duration of ischaemic

episodes and total duration of ischaemia (symptomless andpainful) were skewed. The results for a subgroup of 12patients with ischaemic ST segment change at baselineshowed no significant difference among the four treatmentperiods (p > 0- 15 all tests) on any of the ischaemic variablescalculated in the analysis. The time of Holter monitoringwas also analysed in relation to the ischaemic episodes andshowed no time effect for these variables (p > 0-2, all tests).The ranges of blood or plasma concentrations of each

drug are shown in table ill for all patients and for eachtreatment period. There were no differences in the ranges ofdrug concentrations with various combinations and noapparent pharmacokinetic interaction of the drugs.

Discussion

In many patients with coronary artery disease, angina iseffectively controlled with &bgr;-adrenoceptor-blocking drugs,calcium antagonists, and nitrate preparations. Whensymptoms escalate, these agents are often used in

combination, though it is not always clear whichcombination should be chosen first. Atenolol, a betal-selective receptor antagonist with no intrinsic

sympathomimetic activity, is used in many such patientsbecause of its proven efficacy.10 Isosorbide mononitrate ismore reliably and effectively absorbed from the

gastrointestinal tract than isosorbide dinitrate.11 It is an

effective, well-tolerated, antianginal agent with provenclinical efficacy as monotherapy.3,12 Furthermore, in a

comparison study, nifedipine and isosorbide mononitratewere shown to be equally effective as monotherapy inachieving both subjective and objective improvements inpatients with chronic stable angina.4 However, informationon the advantages or disadvantages of combining all threedrugs remains scanty.13We have compared the roles of combination therapy with

nifedipine, mononitrate, or both drugs in patients withangina who received maintenance atenolol beta-blockade.Our exercise evaluation was carried out a mean of 7-4 h afterdose for all treatment periods. All patients had exertionalangina on atenolol and 12 had symptomless ischaemia on theambulatory ECG recordings. The majority (72 %) had had amyocardial infarction and all but 1 had multivessel coronaryartery disease. In such patients, the effects of antianginalagents when used in combination cannot always be

predicted from the individual response to each administeredseparately. The drugs we used have different mechanisms ofaction, and each has many direct and indirect effects on thecardiovascular system. The potential benefit of combinationtherapy will, therefore, depend on the extent to whichfavourable effects on myocardial oxygen supply are

augmented and adverse effects are cancelled.The only significant finding in our study was the longer

duration of exercise during atenolol plus mononitrate thanduring atenolol alone or in combination with nifedipine.Triple therapy conferred no further advantage on exercisetesting and no real clinical advantage, with no effect onsubjective variables, time to pain, time to 1 mm ST

depression, or the amount of exercise-induced ST

depression.The theoretical advantages of the nitrates and nifedipine

centre on reducing demand by predominant effects onpreload and afterload, respectively. Clearly, these effects arenot sufficient to justify the routine widespread use of theseagents in combination with atenolol. Group studies maskthe potential for individual benefit, and drug combinationsmay in some cases be effective. However, such treatmentshould be monitored and reviewed regularly. The

assumption that the addition of agents will confer furtherbenefitl4 is not valid, and it may be better to change therapyrather than to add to it. We suspect that patients with anginapectoris are overtreated: a logical drug combination may notconfer clinical benefit.

REFERENCES

1. Russek HJ. Propranolol and isosorbide dinitrate synergism in anginapectoris. Am J Cardiol 1968; 21: 44-54.

2. Stone PH, Antman EM, Muller JE, Braunwald E. Calcium channelblocking agents in the treatment of cardiovascular disorders: II

Haemodynamic effects and clinical applications. Ann Intern Med 1980;93: 886-904.

3. Akhras F, Jefferies S, Jackson G. Isosorbide-5-mononitrate effectivemonotherapy in chronic stable angina. Z Kardiol 1985; 74 (suppl 4):16-20.

4. Akhras F, Chambers J, Jefferies S, Jackson G. A randomised double-blind crossover study of isosorbide mononitrate and nifedipine retardin chronic stable angina. Int J Cardiol 1989; 24: 191-96.

5. Gianelly RE, Goldman H, Treister BL, Harrison DC. Propranolol inpatients with angina pectoris. Ann Intern Med 1967; 67: 1216-25.

6. Wolfson S, Heinle RA, Herman MV, Kemp HG, Sullivan JM, Gorlin R.Propranolol and angina pectoris. Am J Cardiol 1966; 18: 345-53.

7. De Buitleir M, Rowland E, Krikler D. Haemodynamic effects ofnifedipine given alone and in combination with atenolol in patients withimpaired left ventricular function. Am J Cardiol 1985; 55: 15E-20E.

8. Tweddel AC, Beattie JM, Lawrie TDV, Hutton I. Effects of nifedipineon physical performance in patients with angina pectoris on &bgr;-blockers.In: Puech P, Krebs R, eds. Proceedings of 4th International AdalatSymposium. Amsterdam: Excerpta Medica, 1980: 143-46.

9. Stein G. Antianginal efficacy of different doses of Adalat in anginapectoris patients in double-blind trial. In: Puech P, Krebs R, eds.Proceedings of 4th International Adalat Symposium. Amsterdam:Excerpta Medica, 1980: 233-39.

10. Jackson G, Harry JD, Robinson C, Kitson D, Jewitt DE. Comparison ofatenolol with propranolol in the treatment of angina pectoris withspecial reference to once daily administration of atenolol. Br Heart J1978; 40: 998-1004.

11. Abshagen H, Sporl-Radun S. First data on effects and pharmacokineticsof isosorbide mononitrate in normal man. Eur J Clin Pharmacol 1981;19: 423-29.

12. Akhras F, Chambers J, Jackson G. An interim report on the efficacy ofisosorbide-5-mononitrate in a sustained release formulation in patientswith stable angina. In: Julian DG, Rittingbousen R, Überbacher H,eds. Mononitrate II. Berlin: Springer-Verlag, 1987: 184-87.

13. Nesto RW, White HD, Ganz P, et al. Addition of nifedipine to maximalbeta-blocker-nitrate therapy: effects on exercise capacity and global leftventricular performance at rest and during exercise. Am J Cardiol 1985;55: 3E-8E.

14. Tolins M, Weir EK, Chesler E, Pierpont GL. ’Maximal’ drug therapy isnot necessarily optimal in chronic angina pectoris. JACC 1984; 3:1051-57.