Rash and pulmonary events. The drugs given to one patient (case 3) at the time of presentation, though possibly associated with eosinophilia associated with the isolation of Clostridium difficile from the stool, were fenbufen given shortly after the onset of the illness and were unlikely to have contributed to the eosinophilia.

Although investigations for organisms that cause BMJ: first published as 10.1136/bmj.300.6717.82 on 13 January 1990. Downloaded from G H Burton intestinal and other tropical infectious disorders were not performed in all cases, it seems unlikely that such organisms could have been responsible for the clinical Medicines Control Agency, , , , , events considering that none of the patients had Department of Heatth, and are the only non-steroidal anti-inflam- recently travelled to a foreign country and the rapid London SW8 5NQ matory drugs that have been implicated in the develop- recovery after withdrawal of fenbufen. G H Burton, MRCP, senior ment ofan allergic pulmonary eosinophilia. 4I describe The Committee on Safety of Medicines has received medical officer four cases in which fenbufen was associated with the a further report implicating fenbufen in an acute Correspondence to: Dr G H development of rash and pulmonary eosinophilia. The episode of rash followed about two weeks later by Burton, Regulatory Affairs Committee on Safety of Medicines was notified of all malaise, acute cough, and worsening of breathlessness and Product Surveillance, cases through the yellow card adverse drug reaction in a 72 year old man who had a pre-existing clinically Upjohn Limited, West reporting system between 1982 and 1988. stable idiopathic pulmonary fibrosis. The illness was Sussex RH10 2NJ. accompanied by additional patchy shadowing in the chest radiograph and a peripheral blood eosinophilia of BrAIed7 1990;300:82-3 Case reports 2 1 x 10/I. No infective organism was identified. The table gives details of the patients and a summary Clinical recovery and resolution of the eosinophilia of investigations and outcomes. All four patients were ensued over the following four weeks. non-smokers; none had recently travelled abroad; and All of the patients reported on in the present study all had been in good general health up to the time of presented with a generalised systemic illness accom- presentation. All had taken fenbufen for non-inflam- panied by dry cough, breathlessness, pulmonary matory osteoarthritis, and two patients (cases 3 and 4) infiltrates on chest radiography, and appreciable had taken the drug for about two weeks before the peripheral blood eosinophilia consistent with a diag- onset ofsymptoms. The two other patients (cases 1 and nosis of pulmonary eosinophilia. The time interval 2) had taken the drug intermittently. All patients from the start of treatment to the onset of the illness in recovered clinically over one to two weeks after two of the cases was two weeks, which is consistent withdrawal of fenbufen; two patients (cases 1 and 3) with previous reports.')2The appearance at the outset were given immunosuppressants. Radiological of the illness of a florid, widespread erythematous rash recovery took somewhat longer except in case 4. was interesting: only one of the published reports Rechallenge with fenbufen was not undertaken. implicating other non-steroidal anti-inflammatory drugs in pulmonary eosinophilia has mentioned the occurrence of a preceding rash.' Review of all the Comment yellow card reports of associations between non- To my knowledge this is the first report of a steroidal anti-inflammatory drugs and interstitial lung pulmonary eosinophilia occurring in association with disorders disclosed a single case of pulmonary eosino- fenbufen. It is unlikely that other drugs that the philia preceded by an acute urticaria associated with patients had taken could account for the clinical . http://www.bmj.com/

Results ofinvestigations in and details offour patients with rash and pulmonary toxicity after takingfenbufen Case I Case 2 Case 3 Case 4

Scx, age years) MNI, 72 F, 73 Ml, 85 M, 68 Indication f or use Hip loint pain Osteoarthrosis of hands, Osteoarthrosis of knees Back and neck pain hips, and knees

Duration of use Intermittent for three Intermittent for several Two to three weeks Two weeks on 30 September 2021 by guest. Protected copyright. months months Daily dose (mg) 600 900 900 900 Other medical conditions None Palpitations Mild untreated hypertension None Concomitant None Propranolol for 12 months, Ampicillin and frusemide at tema-zepam time of presentation Presenting symptoms Rash then dry cough, Ers thematous rash then dry Rash then thirst, malaise, Blistering rash with dry breathlessness, malaise, cough, fever, malaise breathlessness, fever, cough, breathlessness, fever diarrhoea fever, wheezing, and nausea Radiology Dense bilateral alveolar Widespread bilateral Bilateral basal shadows with Ill-defined, migratory shadowing infiltrates progression to generalised shadows on serial bilateral shadowing radiography (mainly mid and lower zones) Haematology: Haemoglobin (g/l) 147 145 162 128 White cell count (x 10'/1) 22 0 10 7 32 0 16-5 Eosinophil count (x 10'/1) 3-5 1 7 8 0 4-0 Lung function Reduced lung volumcs and Results of airways tests, Not tested Normal gas transfer transfer factor lung volumes, and gas transfer normal Othcr tests Antinuclear factor: negatiVe Aspergillus precipitzns: Clostridium difficile isolated Aspergillus precipitins: Rheumatoid factor: negative negative from stool at time of negative Legionella antibody titre: IgE normal eosinophilia. No parasites Sputum culture: no growth not raised Stool: no parasites or ova Progress after stopping taking All radiological, Complete cliinical and Full clinical, radiological, Complete recovery in two fenbufen haematological, and radiological recovery at six and haematological weeks lung function variables weeks, with eosinophil recovery occurred over one improved within one wcek count of 0-66 x 10'/l week taking 15 mg taking azathtoprine and prednisolone daily prednisolone and were normal four months later

82 BMJ VOLUME 300 13 JANUARY 1990 The cases described in the present report implicate I O'Brien WM, Bagbv GF. Rare adverse reactions to non-steroidal anti- inflammatory drugs. J Rheumatol 1985;12:13-20. fenbufen in causing an adverse reaction that presents as 2 Flint KC, Johnson NM. Pulmonary eosinophilia associated with naproxen a florid rash followed by a severe, short lived rever- therapy. 7 R Soc Med 1987;80:120-1. 3 Stromberg C, Palva E, Alhasa E, Aranko K, Idanpaan-Haikkila J. Pulmonary sible illness whose characteristic feature is a pulmonary infiltrations induced by tolfenamic acid. Lancet 1987;ii:685. eosinophilia. 4 Ghevsens B, Van Miegham W. Pulmonary infiltrates with eosinophilia due to glafenine. Eurj Respir Dis 1984;65:456-9. I thank the Committee on Safety of Medicines for permis- 5 Chuck AJ, Wilcox M, Bossingham DH. Fenbufen-associated pneumonitis. BMJ: first published as 10.1136/bmj.300.6717.82 on 13 January 1990. Downloaded from sion to report these cases and the doctors who originally Brj Rheumatol 1987;26:475-6. submitted the reports for providing further clinical details. (Accepted 12 September 1989)

Effect of intranasal oxygen on fell below 80%, but only one of these patients was cyanosed. These patients received naloxone. hypoxia and tachycardia during No significant difference in pulse rate or blood endoscopic pressure was detected between the controls and the group given oxygen, but those patients with oxygen cholangiopancreatography saturations below 90% had a significantly higher pulse rate than those who were not hypoxic (p<005). Increasing age correlated with a lower oxygen saturation S Michael Griffin, S C Sydney Chung, in both groups (group given oxygen p<0-01, r=0 512; Joseph W C Leung, Arthur K C Li controls p<001, r=0 464). Increasing dose of pethi- dine correlated with desaturation only in the control Prince of Wales Hospital, Most patients tolerate endoscopic procedures well, group (p<002, r=0 574). Body weight, length of Chinese University ofHong although complications occur occasionally. Cardio- procedure, and dose of diazepam given did not show a Kong, Shatin, New pulmonary complications are the commonest and positive correlation. Territories, Hong Kong occur once in every 1600 investigations.' We found S Michael Griffin, FRCS, that 44% of patients who had endoscopic retrograde visiting lecturer in surgery Comment S C Sydney Chung, MD, cholangiopancreatography became hypoxic., Factors senior lecturer in surgery that contribute to hypoxia include advancing age, pre- Compared with gastroscopy endoscopic retrograde Joseph W C Leung, MD, existing lung disease, increasing intravenous sedation, cholangiopancreatography poses several problems. senior lecturer in medicine and the diameter of the endoscope.4"We carried out a Patients undergoing the procedure are normally older; Arthur K C Li, MD, professor randomised controlled study to assess whether adminis- a duodenoscope is larger than a gastroscope; the ofsurgery tration of oxygen intranasally prevents such hypoxia. procedure is more complicated, is more unpleasant, and may take longer; heavier sedation may be necess- Correspondence to: ary; and the patient is obscured by the x ray machine, Professor Li. Patients, methods, and results making monitoring more difficult. We studied 80 consecutive patients undergoing We found that hypoxia was common during endo- BrAfedJ 1990;300:83-4 elective endoscopic cholangiopancreatography, who scopic retrograde cholangiopancreatography and were randomised to receive intranasal oxygen or serve occurred within the first 15 minutes; tachycardia as controls. Pulse oximetry was performed during the occurred during these hypoxic episodes, suggesting procedure with an Oxyshuttle oximeter and Transend that the oxygen desaturation is clinically important. http://www.bmj.com/ strip recorder (Sensorimedics) with the sensor attached Oxygen desaturation associated with tachycardia can to an index finger. Blood pressure was measured at one result in myocardial ischaemia in patients with border- minute intervals with a Dinamap monitor. Patients lay line coronary perfusion.5 The hypoxia during endo- supine for one minute to establish baseline values and scopic retrograde cholangiopancreatography was pre- then turned to the prone position. The throat was vented in all but four of 39 patients given intranasal sprayed with local anaesthetic, and intravenous oxygen at a dose of 4 5 1/min. Monitoring of pulse diazepam and pethidine were administered at doses left

to the endoscopist's discretion. Endoscopy was started on 30 September 2021 by guest. Protected copyright. one minute later. Oxygen (4 5 1/min) was administered 100 Group given oxygen intranasally to the patients randomised to receive it as - Control group j soon as they were prone. Age, weight, doses of diazepam and pethidine, and length of procedure were ** *** * subjected to multiple regression analysis to determine 98- predictors of hypoxia, tachycardia, and hypotension. Forty one patients (19 men, 22 women) served as _4 ~~*** controls and 39 (21 men, 18 women) were given .296- oxygen. The two groups were similar in age (mean (SEM) 60 5 (2 2) v 61-0 (2 6)), weight (55 8 (1 3) v 52 6 (1-6) kg), duration of procedure (28 2 (2 8) v 27 8 (2 4) 94- `. min), and doses of pethidine (25 8 (1-6) v 23 4 (1 5) x mg) and diazepam (4 6 (0 3) v 4 1 (0-3) mg). The 0 incidence of cardiac disease (five v seven patients) and 92- ofprevious respiratory infection (19 v 20 patients) were comparable between the two groups. The figure shows mean oxygen saturation in the two 90i r -T -T rT groups. At the start of endoscopy oxygen saturation in q, ? 5 10 15 20 25 30 35 45 60 both groups fell sharply and then rose. Mean oxygen saturation was higher throughout the procedure in the Time (minutes) patients who received oxygen. Oxygen saturation dropped below 90% for more than 60 seconds at some point during the procedure in 21 controls and four Mean (SEM) oxygen saturation in 39 patients given intranasal patients given oxygen (p<0-001, Fisher's exact test). oxygen and 41 controls during endoscopic retrograde cholangio- In three patients (two controls and one given oxygen) it pancreatographv. *p<0.QI; **p

BMJ VOLUME 300 13 JANUARY 1990 83