sign in sign up
<<
Home , Antiarrhythmic agent

372 Thorax 1992;47:372-376 Pulmonary complications after long term treatment

Thorax: first published as 10.1136/thx.47.5.372 on 1 May 1992. Downloaded from

J Roca, M Heras, R Rodriguez-Roisin, J Magrifin, A Xaubet, G Sanz

Abstract Amiodarone hydrochloride is a benzofurane Background Amiodarone hydrochlor- derivative that was used widely as a coronary ide is an antiarrhythmic agent useful in vasodilator in Europe from 1967 to the late refractory to standard 1970s.' It started to be used as an experimental therapy. Although interstitial pneumon- antiarrhythmic agent in the United States in itis is known to be its most serious side the early 1980s,2 and is effective in 75-80% of effect, several aspects ofamiodarone lung cases of supraventricular and ventricular ar- toxicity are still controversial. rhythmia refractory to standard antiarrhyth- Methods Pulmonary side effects were mic agents.' The effectiveness of amiodarone examined in a sample of 61 symptomless is, however, offset by various side effects that patients (mean (SD) age 55 (7) years) who limit its long term therapeutic use. Most had had long term treatment with prominent among these are cutaneous and amiodarone (daily maintenance dose 400 neurological manifestations, thyroid dysfunc- mg), selected from 482 men attending the tion, hepatotoxicity, muscle weakness, ocular University of Barcelona myocardial disturbances, and pulmonary complications.5 infarction project. To allow for the con- Interstitial pneumonitis associated with founding effects of coronary artery dis- amiodarone was first reported by Rotmensch in ease and tobacco history on lung function, 19806 and is often its most serious side 46 patients who had taken amiodarone for effect.5~"" Several aspects of amiodarone lung more than one year were matched with a toxicity are still controversial-namely, its control group from the same population. prevalence (which has ranged from 1 to Subjects underwent measurement of 20%1012) and the precise pathophysiological lung volumes, arterial gas analysis mechanism, which remains poorly under- http://thorax.bmj.com/ and an incremental bicycle exercise test. stood."" Routine monitoring of pulmonary Results Most lung function values function, particularly single breath transfer were close to predicted values, though factor for carbon monoxide (TLCO), has been there was a small increase in resting recommended for early detection ofpulmonary alveolar-arterial oxygen tension differ- complications by Kudenchuk and associates.'0 ence (A-aDo2) at rest (4-8 (1P4) kPa in both Their results have not, however, been groups). There were no differences in the reproduced in other series.81112 results offorced spirometry or static lung The present cross sectional study was on September 28, 2021 by guest. Protected copyright. volumes between the two groups, or in the specifically aimed to assess pulmonary side fall in A-aDo2 from rest to exercise. There effects ofamiodarone in patients with coronary was a small difference between the disease having long term treatment with the amiodarone and the control group in . Special care was taken in selecting the transfer factor for carbon monoxide sample populiation and in the protocol design to corrected for lung volume (Kco 1P67 (0 3) ensure compliance with treatment and to avoid and 1-83 (0-3) mmol min-' kPa-' 1` res- any confounding effect of subclinical cardiac pectively) and in exercise capacity (140 failure or smoking habit on pulmonary func- (25) and 120 (30) w). Only three patients tion. showed lung function impairment consis- tent with pneumonitis. No relation be- Methods Servei de tween lung function measures and POPULATION STUDIED Pneumologia The original sample consisted of 61 subjects J Roca cumulative doses of amiodarone or R Rodriguez-Roisin desethylamiodarone was found. who had received amiodarone for a cardiac A Xaubet Conclusions The prevalence of clin- ; 46 of them had been treated with Servei de Cardiologia ically evident pulmonary side effects was amiodarone for at least one year and these were M Heras 4*9%, which is lower than that reported in matched with a control group of 46 subjects. J Magri-na studies in which higher daily mainten- Both groups of subjects were selected from the G Sanz University of Barcelona Hospital Clinic, ance doses ofamiodarone were given. The Universitat de slightly lower Kco values and lower work project, reported in full elsewhere."7 Briefly, Barcelona, Villarroel load achieved by the patients taking patients with a definite myocardial infarction 170,08036 Barcelona, amiodarone suggest a small effect of were admitted to the coronary care unit of our Spain amiodarone in doses of 400 mg on lung hospital, where they were monitored for at least Reprint requests to: function. A role for individual suscep- five days. Patients underwent right and left Dr J Roca coronan- Accepted for publication tibility to pulmonary complications of catheterisation and selective 19 December 1991 amiodarone treatment is suggested. giography four weeks later, the percutaneous Pulmonary complications after long term amiodarone treatment 373

femoral approach being used. They had an formed at the time of the study as follows: (1) exercise test six months after the infarct and intrathoracic gas volume (Vtg) and specific each year thereafter. All patients were seen airways conductance (sGaw) were measured regularly by the same physician every three to with a constant volume plethysmograph (Body Thorax: first published as 10.1136/thx.47.5.372 on 1 May 1992. Downloaded from six months in the outpatient clinic. Their Pneumo-test; E Jaeger, Wurzburg, Germany); clinical and angiogiaphic records have been (2) single breath transfer factor for carbon reported."7 monoxide (TLCO) (PK Morgan, model A, Patients treated with amiodarone received an Chatham, UK); (3) arterial blood gas tensions initial loading dose of 600 mg per day during at rest (BMS3 MK2, Radiometer, Copen- the first week, followed by a daily maintenance hagen, Denmark) in 41 of 46 patients in each dose of 400 mg for five days a week. The mean group; and (4) slow and forced spirometry (SD) cumulative dose was 325 (160)g (range before and after a (salbutamol) 27-648 g) and mean duration of treatment 35-1 (HP 47804A Pulmonary System, Waltham, (16-8) months (range 2-76 months). Plasma Montana, Massachusetts). An incremental was taken for measurement of concentrations cycling exercise was performed by 40 of the 46 of amiodarone and desethylamiodarone but patients in each group. Six patients treated because of technical problems these were with amiodarone and their controls were determined in only 26 patients. The mean excluded from the exercise study because of plasma concentration of amiodarone was 1-23 lack of cooperation (four cases) or technical (1-02) pg/ml (range 0 13-3A4 ug/ml) and of problems (two cases). The exercise protocol desethylamiodarone 1 19 (0-82) Mg/ml (range consisted of a workload increase of 20 watts 0 1-2-68 pg/ml). In the remaining 35 every minute to the maximal tolerated work individuals we relied on clinical methods to load. Arterial blood gas tensions at rest and confirm that patients took amiodarone, includ- during maximal exercise were obtained ing and examination for through an indwelling radial or brachial artery corneal microdeposits with a split lamp. All catheter (Seldicath, Plastimed, Saint-Leu-La- patients had either a plasma amiodarone deter- Foret, France) placed in the non-dominant mination or a corneal examination to ensure wrist or arm. compliance with amiodarone treatment. A diagnosis of pneumonitis was established Control group A control group was included according to three or more of the following to address the confounding effect of coronary criteria: bilateral end expiratory crepitations; disease on the results of the pulmonary func- chest radiograph with infiltrates in the lung tion tests. The 46 patients who had received parenchyma;'9 restrictive ventilatory pattern amiodarone for more than one year were paired (FVC < 80% predicted, FEV,/FVC ratio > individually with 46 non-treated individuals 80%, TLC < 80%)21; TLCO < 80% predic- http://thorax.bmj.com/ from the same project.'7 Both groups were ted22; arterial oxygen tension (Pao2) at rest < matched according to age (five year interval), 10-7 kPa; and/or an increase in the alveolar- presence or absence of smoking (they were arterial oxygen tension difference (A-aDo2) of classified as smokers, lifetime non-smokers, more than 0-6 kPa during the exercise test in and ex-smokers), chronic cough or phlegm (yes the absence offunctional limitations suggesting or no), left ventricular ejection fraction risk cardiac failure.23 In all the patients suspected of groups (up to 20%, 21-49%, and 50% or having pneumonitis pulmonary infiltrates with

more), and left ventricular end diastolic pres- no evidence of left ventricular failure were on September 28, 2021 by guest. Protected copyright. sure (5 0 mm Hg interval).'7 Patients with reported by an independent radiologist reinfarction were excluded. unaware of the patient's treatment. The diag- nosis was supplemented by a positive gallium- 67 lung scan and an abnormal count of inflam- matory cells in bronchoalveolar lavage fluid.24 PROTOCOL Pulmonary haemodynamics were measured Patients were given an extended questionnaire with a Swan-Ganz catheter in one of the three based on the epidemiology standardisation patients (No 3) with suspicion of pneumonitis project of the American Thoracic Society.'8 only. Specific questions were appended to identify non-pulmonary complications related to the long term use of amiodarone, including cutan- eous manifestations; muscular weakness; and DATA ANALYSIS neurological, ocular, gastrointestinal, and Results are presented as mean (SD) values. thyroid disturbances. A physical examination Paired Student's t tests and one way analysis of and chest radiography were performed in each covariance, with age as the covariate, were used case. Chest radiographs were reviewed and to assess differences between amiodarone compared with the ILO/UICC standard treated and control groups (n = 46 each). radiographs used for assessment of pneumo- Relations between lung function and conioses.'9 Hepatic and thyroid function were cumulative dose of amiodarone were examined measured.20 Patients were matched according by using a stepwise multiple regression to how long after the myocardial infarction the analysis. In the regression formula each lung lung function measurements were made; for function index was the dependent variable; the the amiodarone group the mean was 4-7 (SD cumulative dose of amiodarone, age, height, 1-9) years and for the control group 5-9 (1-7) and body weight were covariates. A p value years (p < 0-003). below 0-05 was considered statistically sig- Routine pulmonary function tests were per- nificant. 374 Roca, Heras, Rodriguez-Roisin, Magrinid, Xaubet, Sanz

Table I Lungfunction results in the three patients with blurred vision in two; a morning sensation of presumed pulmonary side effects of amiodarone sand in the eye in three; muscle weakness in two; and fine distal tremor in one. There were Patient No: 1 2 3 no gastrointestinal symptoms or clinical signs Thorax: first published as 10.1136/thx.47.5.372 on 1 May 1992. Downloaded from FVC (% pred) 101 82 68(*) of thyroid disturbance. Amiodarone treated FEVy (% pred) 99 89 61 (*) patients had higher plasma levels of thyroxine FEV,/FVC ratio 72 82 70 and reverse tri-iodothyronine than the control TLC (% pred) 94 66 (*) 90 RV (% pred) 87 42(*) 107 group (thyroxine 8 7 (3 3) v 6-2 (1-8) ug/100 ml; RV/TLC ratio 36 29 42 p < 0-001; reverse tri-iodothyronine 388 (155) TLCO (% pred) 88 59 (*) 47 (*) v 167 (1-8) pg/ml; p < 0-0001), findings Pao2 (kPa) 7.9(*) 9.0(*) 7-3(*) PaCo2 (kPa) 4 9 5-1 4-8 consistent with enzymatic inhibition of the peripheral conversion of thyroxine to tri- *Results below the lower reference limit. iodothyronine by amiodarone.'6 FVC-forced vital capacity; FEV,-forced expiratory Only three of the 61 patients fulfilled the volume in one second; FEV,/FVC-ratio of actual values of FEV, and FVC; TLC-total lung capacity; RV- criteria of possible pneumonitis as defined residual volume; RV/TLC ratio expressed as actual value; under "Methods" (table 1). One of the three TLCO-single-breath CO transfer capacity; Pao,-arterial had a recent history of dyspnoea on mild partial pressure of oxygen; Paco2-arterial partial pressure exercise, dry cough, and muscle weakness, and of carbon dioxide. had a right pleural rub; another had bilateral crackles. The chest radiograph in two patients Results showed an interstitial infiltrate and the third PREVALENCE OF AMIODARONE INTERSTITIAL showed a peripheral alveolar infiltrate. All PNEUMONITIS three patients had arterial hypoxaemia at rest Ofthe initial 61 patients receiving amiodarone, (table 1) and two had a reduction in static lung 53 were ex-smokers, three lifetime non-smok- volumes and TLCO. All improved after ers, and five current smokers. Seven patients amiodarone withdrawal. Two were also treated had had mild cardiac failure and were receiving with oral prednisone (initial dose 1 mg/kg a and ; the remaining 54 were day for six weeks followed by a reducing dose symptomless, in functional class 1.23 None of for two months in one and six months in the the patients had clinical or radiographic other) until both amiodarone and desethyl- features of cardiac failure at the time of the amiodarone were undetectable. The radio- study. No differences in pulmonary function graphs, gallium lung scans, and lung function were found between the 46 patients who had showed progressive improvement at follow up received amiodarone for more than one year in two patients.

and the 15 who had been treated for a shorter http://thorax.bmj.com/ time. None of the 61 patients had severe non- COMPARISON OF LUNG FUNCTION BETWEEN pulmonary side effects from amiodarone. Mild AMIODARONE AND CONTROL GROUPS to moderate cutaneous manifestations, mainly There was no difference between the photosensitivity or increased cutaneous pig- amiodarone and the control groups for height, mentation or both, were present in 28 patients; weight, mean systemic blood pressure, or car- diac functional class23 (table 2). Patients in the amiodarone group were slightly younger than Table 2 Comparison ofanthropometric data, matching indices, and lungfunction the controls. Mean values for maximal resultsfor amiodarone (n 46) and control (n = 46) groups* expiratory flow rates, static lung volumes, on September 28, 2021 by guest. Protected copyright. Amiodarone Control p airways conductance, and TLCO were close to predicted values in both the amiodarone and Age (y) 54-3 (7-8) 56-6 (6-5) < 0-05 the control group. Both groups had mild Height (cm) 166 (7-1) 166 (6-5) - hypoxaemia at rest and an increased A-aDo, Weight (kg) 72-0 (9-5) 72-8 (83) - (table 2). Tobacco (pack years) 30-3 (21-0) 28-1 (16-0) - Ejection fraction (%) 47-1 (13-2) 48-1 (13-6) - There was no difference between the EDVP (mm Hg) 12-0 (5 5) 13-4 (59) - amiodarone and control groups in dynamic and PAP (mm Hg) 13-8 (3-3) 13-9 (3-7) - static lung volume, specific airways conduc- Years after amiodarone 4-7 (1-9) 5-9 (1-7) <0-003 FVC (ml) 4060 (619) 3951 (721) - tance, TLCO, or helium single breath total lung FEV, (ml) 2907 (561) 2826(609) - capacity (VA). The amiodarone group had a FEV,/FVC ratio 71 (7) 71 (7) - slightly lower Kco (1 -67 (0-3) v 1-83 (0-3) mmol FEF2,1, (1/min) 2-25 (1-05) 2-29 (0-89) - TLC (ml) 6569 (775) 6403(1030) - CO min' kPa-' 1-) (p < 0-04). The alveolar- RV (ml) 2907 (561) 2826(609) - arterial oxygen difference at maximal work load RV/TLC ratio 34 (7) 35 (6) - showed a similar fall from the values obtained sGaw (SI unit) 1-63 (0-6) 1-59 (0-6) - TLCO (SI unit) 9-03 (1-8) 9-60 (1-7) - at rest in the amiodarone and control groups Kco (SI unit) 1-67 (0-3) 1-83 (0-3) < 0-04 (-1-6 (1-3) and -1-5 (1-0) kPa). The maxi- Pao2(kPa) 9-3(1-5) 9-5(1-0) mum tolerated work load was significantly PaCo2 (kPa) 4-8 (0-6) 4-8 (0 4) A-aDo, at rest (kPa) 4-8 (1-4) 4-6 (1-0) lower in the amiodarone group than in the A-aDo2E - r(kPa) -1-6(1-3) -1-5(1-0) control group (120 (30) v 140 (25) w; p < Work load (w) 120 (30-2) 140 (24-9) <0-004 0-004). The differences between the two groups persisted when the three patients with *For arterial blood gas analysis n = 41; for exercise data n = 40. EDVP-end diastolic ventricular pressure; PAP-pulmonary artery pressure; FEF2,,,,- presumed amiodarone pulmonary fibrosis were forced expiratory flow between 25% and 75% of FVC; sGaw-specific airways removed and the remaining 43 individuals were conductance (s ' kPa '); TLCO (mmol CO min ' kPa '); Kco (mmol CO min ' kPa- ' I ', compared with 43 controls (Kco 1-72 (0-2) v BTPS); A-aDor-alveolar-arterial oxygen difference; E - R-difference between exercise and rest. 1-83 (0-3) mmol CO min-'kPa-'1 -'(p < 0-04); For other abbreviations see table 1. decrease in A-aDo, at maximal workload - 1-7 Pulmonary complications after long term amiodarone treatment 375

(1 3) v - 1-5(1 0) kPa; maximum workload 120 finding and the lower work load achieved by the (30) v 138 (25) w; p < 0-004). Finally, there patients taking amiodarone might be due to a were no significant correlations between any small but genuine effect ofthe drug at the doses

lung function variable and cumulative dose of used in the present study. The similar results Thorax: first published as 10.1136/thx.47.5.372 on 1 May 1992. Downloaded from amiodarone (daily dose times days of treat- for the A-aDo2 during exercise in the two ment) when multiple regression analysis was groups suggests that the lower Kco in the carried out. patients having amiodarone does not explain Amiodarone had to be withdrawn in 17 the difference in maximum work load. The patients (32%) originally treated with the drug design ofthe exercise protocol cannot provide a because it failed as an antiarrhythmic agent. clear explanation for the lower maximum work The therapeutic efficacy (68%) is slightly lower load observed in the amiodarone group (table than that reported from studies"5 11'2 with 2). It could be due to a direct cardiac effect of higher daily maintenance doses. Amiodarone amiodarone. Both groups were studied about was effective in controlling cardiac arrhythmias five years after the acute coronary episode, so in the 61 patients in the present study. the slightly longer time since myocardial infarction in the amiodarone group (table 2) is unlikely to account for the difference. In summary, our patients receiving Discussion amiodarone for a year showed a low prevalence This study examined a population ofsymptom- oflung function disturbances that is likely to be less patients followed by the same cardiologist due to amiodarone. Lung function impairment in the outpatients clinic.'7 The prevalence of compatible with amiodarone pneumonitis was clinically evident pulmonary side effects was observed in only three patients. Despite 49% (three cases), which is similar to that differences in study design, our results are reported by Rotmensch et al (3-2%)2 in a consistent with a carefully conducted study multicentre follow up study in which a similar done by Magro et al.'2 They showed that average daily dose of amiodarone had been routine pulmonary function tests failed to given. A higher prevalence, ranging from 6% predict amiodarone pulmonary side effects in to 20%, has been reported in studies from the asymptomatic patients, but were useful in United States,58112 where higher doses of patients with mild manifestations of the dis- amiodarone are used (loading dose 1400 mg for ease. They reported that a fall in TLCO of 15% two to three weeks followed by daily mainten- or more in a patient with clinical suspicion of ance dose of 600-800 mg). Our daily mainten- pulmonary toxicity gives a 100% sensitivity ance dose of 400 mg provided amiodarone and an 89% specificity for the diagnosis of plasma concentrations above the therapeutic pulmonary complications of amiodarone treat- http://thorax.bmj.com/ threshold (1 Mg/l) and, in general, below the ment. Such figures could not be extrapolated to toxic threshold of 2-5 Mg/1.25 symptomless patients.12 In the present study two ofthe three patients A direct cytotoxic effect of amiodarone on with presumed pulmonary side effects showed the lung has been strongly supported by studies mild to moderate clinical manifestations and in animal models,26 clinical studies in patients,12 the third had a subacute clinical onset. and epidemiological findings.5 The small Radiographic and lung function data were in differences in pulmonary function (Kco) be- keeping with the diagnosis ofpneumonitis in all tween the amiodarone and the control group three patients and the diagnosis was supported contrast with the large difference in thyroid on September 28, 2021 by guest. Protected copyright. by a positive gallium lung scan and an abnor- function indices,20 which perhaps suggest a role mal cell composition of the bronchoalveolar for individual susceptibility in pulmonary lavage fluid.24 The abnormalities tended to complications of amiodarone treatment. resolve after amiodarone was withdrawn. Pulmonary function tests showed moderate We thank C Gistau, F Burgos, T Lecha, M Simo, C Argaiia, and arterial hypoxaemia and an increase in the B Ochoa for their skilful technical assistance and themedical staff who participated in the University of Barcelona myocardial A-aDo2. This impairment in respiratory gas infarction project for their cooperation and care of the patients. exchange at rest cannot be associated with The work was supported in part by grants from the Fondo de Investigaci6n Sanitaria (84/864), ACARD (1984), and the amiodarone, however, because similar abnor- Comision de Investigacion de Ciencia y Tecnologia (PA86- malities occurred at rest and during exercise in 0345). the control group (table 2). The reduction in the A-aDo2 during exercise in both groups is 1 Vastesaeger M, Gillot P, Rasson G. Etude clinique d'une nouvelle anti-angoreusse. Acta Cardiol 1967; consistent with ventilation-perfusion mis- 22:483-500. matching at rest that improves during exercise. 2 Rosenbaum MB, Chiale PA, Halpern MS. Clinical efficacy of amiodarone as an antiarrythmic agent. Am J Cardiol This could be due to underlying coronary 1976;38:934-44. disease, despite the fact that the patients did not 3 Heger JJ, Prytowky EN, Jackman WM, et al. Amiodarone: clinical efficacy and electrophysiology during long term have functional limitations to suggest heart therapy for recurrent ventricular or ven- failure. Altematively, as many ofthese patients tricular fibrillation. N Engi J Med 1981;305:539-45. 4 Kopelman HA, Horowitz LN. Efficacy and toxicity of had smoked heavily, small airway dysfunction amiodarone for the treatment of supraventricular could contribute to these mild abnormalities of tachyarrhythmias. Progr Cardiovasc Dis 1989;31:355-66. 5 Mason JW. Amiodarone. N Engl J Med 1987;316:455-66. lung function. 6 Rotmensch HH, Liron M, Tupilski M, Laniado S. Possible Although the difference was statistically sig- association of Pneumonitis with amiodarone therapy [letter]. Am Heart J 1980;100:412-3. nificant, the fact that patients taking 7 Marchlinski FE, Gansler TS, Waxman HL, Josephson ME. amiodarone had a lower Kco than the control Amiodarone pulmonary toxicity. Ann Intern Med 1982; 97:839-45. group is likely to be of limited clinical sig- 8 Rakita L, Sobol SM, Mostow N, Vrobel T. Amiodarone nificance as it was a small difference. This pulmonary toxicity. Am Heart J 1983;106:906-14. 376 Roca, Heras, Rodriguez-Roisin, Magrinia, Xaubet, Sanz

9 Darmanata JI, van Zandwijk N, Duren DR, et al. 1982;306:1065-70. Amiodarone pneumonitis: three further cases with a 18 Ferris GB. Epidemiology standardization project. Am Rev review of published reports. Thorax 1984;39:57-64. Respir Dis 1978;118 (suppl 2):55-1 10. 10 Kudenchuk PJ, Pierson DJ, Greene HL, Graham EL, Sears 19 International Labour Office and Union Internationale Con- GK, Trobaugh GB. Prospective evaluation ofamiodarone tre le Cancer. International classification of radiographs of toxicity. Chest 1984;86:541-8. pneumoconiosis. Geneva: International Labour Office, 11 Adams GD, Kehoe R, Lesch M, Glassroth J. Amiodarone- 1980. (Occupational Safety and Health Series, No 22.) Thorax: first published as 10.1136/thx.47.5.372 on 1 May 1992. Downloaded from induced pneumonitis: assessment of risk factors and 20 Heras M, Roca J, Sanz G, et al. Alteraciones subclinicas possible risk reduction. Chest 1988;93:254-63. producidas por el tratamiento prolongado con amio- 12 Magro SA, Clinton PE, Wheeler SH, Krafchek J, Lin H, darona. Rev Esp Cardiol 1987;40:44-50. Wyndham C. Amiodarone pulmonary toxicity: prospec- 21 Roca J, Sanchis J, Agusti-Vidal A, et al. Spirometric tive evaluation of serial pulmonary function tests. J Am reference values from a mediterranean population. Bull Coll Cardiol 1988;12:781-8. Eur Physiopathol Respir 1986;22:217-24. 13 Akoun GM, Milleron BJ, Gauthier-Rahman S, Mayaud Ch 22 Roca J, Rodriguez-Roisin R, Cobo E, Burgos F, Perez J, M. Amiodarone-induced hypersensitivity pneumonitis. Clausen JL. Single-breath carbon monoxide diffusing Evidence of immunological cell-mediated mechanism. capacity (TL,,,) prediction equations for a mediterranean Chest 1984;85:133-5. population. Am Rev Respir Dis 1990;141:1026-32. 14 Israel-Biet D, Venet A, Caubarrere I, et al. Bronchoalveolar 23 Goldman L, Mashimoto B, Cook F, Loscalzo A. Com- lavage in amiodarone pneumonitis. Cellular abnormalities parative reproducibility and validity ofsystems for assess- and their relevance to pathogenesis. Chest 1987;91: ing cardiovascular functional class. Advantages of a new 214-20. specific activity scale. Circulation 1981;64:1227-34. 15 Crystal RG, Gadek JE, Ferrans VJ, Fulmer JD, Line BR, 24 Xaubet A, Roca J, Rodriguez-Roisin R, et al. Broncho- Hunninghake GW. Interstitial lung disease: current con- alveolar lavage cellular analysis and gallium lung scan in cepts of pathogenesis, staging and therapy. Am J Med the assessment of patients with amiodarone-induced 1981;70:542-68. pneumonitis. Respiration 1987;52:272-80. 16 Olsen T, Laurberg P, Weeke J. Low serum triiodothyronine 25 Rotmensch HH, Belhassen B, Swanson BN, et al. Steady- and high serum reverse triiodothyronine in old age: an state serum amiodarone concentrations: relationships with effect of disease not age. J Clin Endocrinol Metab 1978; antiarrhythmic efficacy and toxicity. Ann Intern Med 47:1111-5. 1984;101 :462-9. 17 Sanz G, Castaiier A, Betriu A, et al. Determinants of 26 Martin WJ II, Howard DM. Amiodarone-induced lung prognosis in survivors of myocardial infarction. A pros- toxicity. In vitro evidence for the direct toxicity of the pective clinical angiographic study. N Engl J Med drug. Am J Pathol 1985;120:344-50. Adventitia Bungalow hospital and coral reef Saddened by the prospect of rapid demoralising to the enemy. His part in the following compulsory retirement, I applied delivery of intelligence to clandestine radio without success to a well known charity for stations in the mountains by native runners, work in the third world, possibly in starving who might cover 30 miles in a night, con- East Africa. But within a few days, wheels tributed substantially to military successes in having clearly worked within wheels, a tele- the area. Sadly he was of mixed New Zealand phone call came from the Overseas Develop- and Melanesian parentage, and thus inherited ment Administration, an organisation ofwhich the genetic makeup which led to his sudden http://thorax.bmj.com/ I then knew little. A pleasant interview fol- death from cardiac infarction. lowed with the senior medical adviser, the Pidgin English was a source of delight, redoubtable Dr Penny Key, herselfa veteran of though it was not easy to acquire fluency. I fell several years' single handed practice in the in with an Australian naval padre who went depths of the jungles of Papua New Guinea. I round visiting the villages in a helicopter, arrived as temporary physician at Honiara, known locally, and affectionately, as the "big- Guadalcanal, capital of the Solomon Islands, fellah mixmaster blong God"-one of many

within a month. fine examples of the genre. Interestingly, the on September 28, 2021 by guest. Protected copyright. The geographical surroundings were Service handout on tropical diseases, describ- glorious, in particular the vast high vermilion ing malaria, says ". . . the patient becomes hot and indigo cumulus in the evening sky. My and dry. He may vomit a lot and become quality oflife was further enhanced by a heavily delicious.. .," which seems appropriate in a populated coral reef in the crystal clear sea, country that in the past was said to incline at some 20 paces from my quarters. times to cannibalism. The work, in a bungalow hospital with basic Refreshingly, carcinoma ofthe bronchus was facilities, was caring for mainly young patients also virtually absent, my only case being the with treatable infective illnesses-sepsis, chief of a hill village who presented with malaria, and tuberculous in roughly advanced and distressing superior vena caval equal mix, with a scattering of leprosy. The obstruction from an adenocarcinoma in the medical department had been for some months right upper lobe. On being told that the only under the sole care of an ethnic Chinese fifth treatment available was , likely year undergraduate from Melbourne, who was to have unpleasant side effects and to be of understandably showing signs of stress. Nurs- doubtful benefit, he declined and asked with ing care by locally trained women was excellent deep courtesy whether he might, without caus- and very practical, and highly competent ing offence, return home to seek the help of the laboratory technicians were very supportive in village shaman. I visited his village a couple of the absence of a pathologist. Few people live months later, on the occasion of a pig feast for long enough to develop degenerative disease, the opening ofa long house for the local church, and ischaemic heart disease was absent, save for and was invited to his home, where I was the case of Billy Bennett. warmly welcomed. Squatting on the earth floor This delightful man had a unique record of of his grass hut, I was able to observe that the personal valour in clandestine wartime jungle congested superficial veins had disappeared and operations, his particular skills being with the that the facial oedema was strikingly reduced. dagger and the hand grenade. He specialised in His equanimity was complete. I came home the immediate burial of the remains down to with an enhanced ofhumility with regard the last button, so that the regular disap- to Western medicine.-JOHN EDGE