Haemophilus influenzae may cause serious illness in childhood, particularly capsulated Pittman type b organisms.1-3 In the respiratory tract of adults the organism is usually non-capsulated4 and is frequently isolated during acute episodes or exacerbations of respiratory diseases. 5 We have observed that Hparainfluenzae is clinically indistinguishable from non-capsulated H influenzae, suggesting that H parainfluenzae is pathogenic in the respiratory tract of adults.6 Subdivision of H influenzae and H parainfluenzae into biotypes is now possible' 8 and, although this has been performed in childhood illnesses,2 our aim was to determine the epidemiological and clinical value of haemophilus biotyping in adult respiratory diseases.
MethodsUsing the laboratory records for 1983 we undertook a retrospective study of the case notes of patients who had had Haemophilus species isolated from the respiratory tract. Sputum specimens were routinely inoculated on to horse blood agar and heated horse blood agar containing 10 units of bacitracin/ml and incubated for 18 hours at 37°C in 7% carbon dioxide. Specimens from patients with suspected anaerobic infections were incubated anaerobically for 48 hours. Non-haemolytic haemophilus organisms were identified by their morphological and colonial appearance and the demonstration of satellitism with Staphylococcus aureus. Differentiation between H influenzae and H parainfluenzae was then achieved by the ability of the organism to metabolise 6-amino laevulinic acid (HP +, HI -). H influenzae and H parainfluenzae were further subdivided into their biotypes by the method of Kilian.7 8The following variables were recorded and analysed: (i) distribution of H influenzae and Hparainfluenzae biotypes in different diagnostic groups; (ii) antibiotic sensitivities, determined by a standard disc susceptibility technique; (iii) frequency of isolation of multiple organisms; (iv) pus cell counts in undiluted sputum-samples being taken from the most purulent part of the specimen-were scored (1-10 cells/high power field (HPF) = 1, [11][12][13][14][15][16][17][18][19][20] After exclusion of 105 isolates from general practice (showing the same biotype distribution as isolates from hospital) and repeat specimens, 574 separate clinical episodes were studied. The diagnostic groups were: chronic bronchitis (n = 201), asthma (n = 63), bronchial carcinoma (n = 107), bronchiectasis (n = 52), pneumonia (n = 38), and "others' (n = 113). In the last group 37% of patients had pre-existing pulmonary disease and 36% had undergone thoracic surgery. All diagnostic groups showed a similar biotype distribution (H influenzae I 15%, II 33%, III 17%, IV 6%, V 4%, VI 2%, VII < 1%; H parainfluenzae 1 6%, 11 1%, III 6%). Recurring clinical episodes were not always attributable to the same biotype, though multiple specimens from a single clinical episode usually (but not invariably) had the same single biotype. Biotype had no bearing on the symptoms recorded in all diagnostic groups, and biotype distribution was similar both ...