1096 hyperresponsiveness in asthmatics without significant sedation.5 Double-blind placebo-controlled studies of inhaled levomepromazine are needed to decide whether there is a place for his drug in acute asthma. Until those studies have been done, we suggest that inhalation of levomepromazine may be tried in patients with asthma refractory to conventional therapy. J E Pedersen, A Olsen, A Banning Department of Anaesthesiology and Intensive Care, Herlev Hospital, University of Copenhagen, Herlev DK-2730, Denmark 1 Saulnier FF, Durocher AV, Deturck RA, Lefevre MC, Wattel FE. Respiratory and hemodynamic effects of halothane in status asthmaticus. Intensive Care Med 1990; 16: 104-07. 2 Conti G, Dell’Utri D, Vilardi V, et al. Propofol induces bronchodilation in mechanically ventilated chronic obstructive pulmonary disease (COPD) patients. Acta Anaesth Scand 1993; 37: 105-09. 3 Groggins RC, Milner AD, Stokes GM. The bronchodilator effects of chlorpheniramine in childhood asthma. Br J Dis Chest 1979; 73: 297-301. 4 Henry RL, Hodges IGC, Milner AD, Stokes GM. Bronchodilating effect of the H1-receptor antagonist: clemastine. Arch Dis Child 1983; 58: 304-05. 5 Madsen F, Faurschou P, Banning A-M, Engel A-M, Sjøgren P, Rosetzsky A. The protective effect of inhaled levomepromazine (Nozinan®) on histamine-induced bronchial constriction. Pulm Pharmacol 1993; 6: 129-36. Myositis with tretinoin SiR-Tretinoin is a standard therapy for remission induction in acute promyelocytic leukaemia (APL).’ Common side- effects include cheilitis, hyperkeratosis, headache, myalgia, bone pain, hypertriglyceridaemia, raised aminotransferases, pseudotumour cerebri, and a hyperleucocytic syndrome that can lead to acute respiratory distress syndrome.2 We report severe myositis associated with tretinoin in a patient with APL. A 33-year-old man presented with petechiae, bleeding gums, leucocytosis (228X109/L) with 79% blast cells, and thrombocytopenia (17X10"/L). Bone marrow examination revealed APL. Cytogenetic analysis showed typical t(15;17). Treatment was started with tretinoin 45 mg/m2 daily with a good initial response. After 18 days of treatment the patient presented with a painful nodule on his right leg without other inflammatory signs. Ultrasonography was inconclusive. 3 days later he developed high fever (40°C), multiple painful nodules in both legs, and bilateral maleolar oedema. A right pleural effusion was found at day 25. The white cell count was 16-3X10VL with 80% granulocytes and 1% blast cells. Platelets were 149 X 109/L. He had increased lactate dehydrogenase (386 U/L), creatine kinase (458 U/L), and creatinine (221 jumoI/L). A muscle biopsy specimen showed focal necrosis of the muscle fibres, oedema, infiltration with inflammatory cells of the endomyseum and the perimyseum, without changes in the skin, vasculitis, or leukaemic involvement; this pattern was consistent with acute myositis. Because of possible retinoic acid syndrome, tretinoin was discontinued and intravenous dexamethasone 10 mg twice daily was started. Standard chemotherapy for AML was initiated 3 days later (day 28) with daunorubicin 45 mg/m’ for 3 days and cytarabine 200 mg/m2 for 7 days by continuous infusion. We observed gradual resolution of the nodules in parallel with complete normalisation of biochemical abnormalities by day 44. Severe myositis in association with tretinoin has not been described before to our knowledge. Hakimian et aP reported a case of erythema nodosum with a similar course, but this diagnosis could be ruled out in our patient based on the histopathological findings. We are indebted to Roche Farmaceutica Portuguesa and to Hoffman La Roche, Switzerland, who kindly supplied tretinoin. N Miranda, P Oliveira, M J Frade, J Melo, M S Marques, A Parreira Department of Haematology, Instituto Português de Oncologia, Lisbon 1093, Portugal 1 Fenaux P, Le Deley MC, Castaigne S, et al. Effect of all transretinoic acid in newly diagnosed acute promyelocytic leukemia: results of a multicenter randomized trial. Blood 1993; 82: 3241-49. 2 Frankel SR, Eardley AE, Lauwers G, Weiss M, Warrell RP Jr. The "retinoic acid syndrome" in acute promyelocytic leukemia. Ann Intern Med 1992; 117: 292-98. 3 Hakimian D, Tallman MS, Zugerman C, Caro WA. Erythema nodosum associated with all-trans-retinoic acid in the treatment of acute promyelocytic leukemia. Leukemia 1993; 7: 758-59. Relation between human papillomavirus type 16 and potential for progression of minor-grade cervical disease SiR-Study of the natural history of cervical disease has to rely on analysis by non-invasive techniques up to the point of termination of the study. Downey and colleagues (Aug 13, p 432) do not dispute the widely held view that biopsy is likely to affect the natural course of the disease, and their criticism of Campion and colleagues’ is unfounded. Of course, if no initial biopsy is done the existence of undetected high-grade disease cannot be ruled out, but there is no alternative. This being the case Downey and colleagues are attempting to answer a meaningless, or at least unanswerable, question-ie, does human pappiloma virus (HPV) 16 infection predict a high risk of future high-grade disease in those women who can be proven not to have existing high-grade disease? In addition to affecting the course of the disease a punch biopsy is likely to give a misleading sense of confidence in the accuracy of the initial categorisation of patients because (a) to decide exactly where to take the specimen in lesions of low-grade appearance can be troublesome, (b) having decided, accuracy in the execution is uncertain, and (c) histological assessment is more difficult than with loop or cone biopsies. Downey et al allude to some of these points but do not give due weight to them when drawing their conclusions. Another important factor is that, of 49 women with mild disease (cervical intraepithelial hyperplasia [CIN] 1 or HPV 1), at least 18 developed precancer (CIN 2 or CIN 3) within 12 months. This frequency represents a progression rate of at least 37% and lends credibility to the hypothesis (contemplated but rejected by Downey and co-workers) that their false-negative rate for initial detection of CIN 2 and CIN 3 was very high. Even more surprising is the overall result of the study, since Bavin et aF originally reported that of the total study population of 200 women, 66 had existing CIN 2 or CIN 3; together with the present data this gives a progression rate of at least 84/200 (42%). Why the incidence of imminent high-grade disease in women referred with mild dyskaryosis should be so high in the population studied when 15-20% would be a more generally accepted figure3 is not discussed. Four alternative hypotheses were suggested by Downey and colleagues to explain their unique results but the reasoning behind some of them is somewhat obscure. First, the diagnosis of CIN 1 is based solely on the appearance and