We studied the ultrastructural findings in biopsies from the main carina of seven school-aged children who had had chronic cough for at least 3 months and who all had a history of early lower respiratory illness (LRI). They had their first LRI between birth and 7 yr of age (range, 5 to 11 yr). The cross-sectional area of the epithelium was quantified by point counting for the percentage area of intercellular spaces (ICS) denoting edema, and the numbers of both inflammatory cells (leukocytes, including eosinophils, and mast cells) and ciliated cells. The children (excluding the one using inhaled steroids) demonstrated nearly 17- and more than sevenfold increases in the mean area of ICS and number of inflammatory cells per epithelial area, respectively, and a nearly three-fold decrease in the mean number of ciliated cells per epithelial area compared with the biopsy specimens from the orifice of the right upper lobe bronchus of two healthy adults. In the children, the increase in inflammatory cells (greater than 91% were lymphocytes) was more prominent in the children with two LRI before the age of 1 yr. Our findings imply a close association of early LRI and later epithelial inflammation during chronic cough. Allergic mechanisms in the epithelial inflammation could not be ruled out as six of the patients had, either alone or in combination, signs of atopia, positive family history of allergic rhinitis or asthma, and eosinophils or mast cells in the epithelium. All eight subjects (six children, two adults) showed a very high and significant inverse correlation between the number of ciliated cells and the percentages of both ICS (edema index: r = -0.877, p < 0.004) and inflammatory cells (inflammatory cell index: r = -0.826, p < 0.011) in the upper half of the epithelium. There was an upward spreading of epithelial inflammation in the patients with concomitant reduction in the number of ciliated cells; with less inflammation, the number of ciliated cells approached that of the adult reference subjects. Using multiple correlation analysis, the joint impact of both indexes on the number of ciliated cells was even higher (R = -0.960, p < 0.002), and omission of the adult reference subjects did not alter this correlation, which remained very high and significant (R = -0.988, p < 0.003). The fit of this model, by which both indexes together account for the number of ciliated cells, is extremely good.
