Trends in pneumonia and empyema in Scottish children in the past 25 years

Background: The incidence of childhood empyema, a complication of pneumonia, is increasing, and the underlying mechanisms are not understood. Whether a rise in pneumonia incidence could account for the increase in empyema remains to be seen. Objective: To report trends for empyema admissions in the context of pneumonia and croup admissions in Scottish children over a 25-year period to 2005. Design: Whole-population study with retrospective analysis using diagnosis codes (International classification of diseases, 9th and 10th revisions). Setting: All non-obstetric and non-psychiatric hospitals in Scotland. Participants: Patients,15 years admitted with a diagnosis of empyema, pneumonia or croup (the latter included for reference) between 1 January 1981 and 31 December 2005. Results: There were 217 admissions for empyema in children (76 1-4-year olds), 24 312 admissions for pneumonia (11 299 1-4-year olds), and 31 120 (20 332 1 -4-year olds) for croup. Empyema admissions increased after 1998 from,10 per million children per annum to reach a peak of 37 per million in 2005. In the 1-4-year age group, empyema admissions rose in the late 1990s and 2000s from an average of 6.5 per million per year between 1981 and 1998 to 66 per million in 2005. Overall annual admission rates for pneumonia remained unchanged in most age groups. However, among 1-4year olds, admissions rose steadily by an average of 50 per million per year between 1981 and 2005. Admission rates for croup in Scottish children (<15 years) remained stable over the preceding 25 years. Conclusions: This whole-population study shows that the incidence of childhood empyema has risen since 1998 and continues to rise independently of pneumonia. Croup admissions remained stable, suggesting that changes in coding or admission policies are not likely to explain the observed trends. The observations suggest that the rise in empyema is not related to an increase in pneumonia. Changes in bacterial pathogenicity and/or host susceptibility may be important.