To the Editors:
In a recent issue of the European Respiratory Journal, Teepe et al. 1 published their interesting observations of determinants of community-acquired pneumonia (CAP) in children in primary care. The authors included 107 children with either radiologically or clinically diagnosed CAP treated as outpatients in four Dutch healthcare centres in 1999–2008, and compared the potential determinants of CAP between the cases and 321 controls from the same area with no CAP during the study period. In adjusted analyses, lower age (OR 1.14), asthma history (OR 3.57) and the number of previous visits for upper respiratory tract infections (URTIs) (OR 1.80 for one or two episodes and 2.46 for more than three episodes) were independently associated with CAP. The authors concluded that the association between CAP and the number of URTIs can be explained by infection susceptibility of the individuals 1.
In the discussion, the authors mentioned that their study was the first to explore the determinants of CAP in children in primary care. Actually, their study was the second one.
As part of the Savo Pneumonia Study performed in 1981–1982 in Eastern Finland 2, 3, we also analysed risk factors for CAP in children aged 3 months to 15 yrs 4. The design of the study was prospective and strictly population-based. During a surveillance period of 12 months, all CAP cases were registered in a small manufacturing town and three rural municipalities. Chest radiographs were studied in all clinically presumptive cases, and only radiologically confirmed CAP cases were included in the analyses. The incidence of CAP was 36 per 1,000 per yr for <5 yrs and 16 per 1,000 per yr for 5–15-yr-old children 2. 51% in the younger and 11% in the older age group were treated in hospital. Pneumococcus caused 28% of the cases overall 3, and Mycoplasma <10% at <5 and >50% at >5 yrs of age, over 80% of mycoplasmal cases being treated at home 5.
To evaluate the possible risk factors for paediatric CAP, identical standardised questionnaires were sent to the parents of the 201 children with CAP and to 250 controls from the same four municipalities. In all, 176 (88%) cases and 233 (93%) controls answered.
In adjusted analyses, significant risk factors for CAP were a history of recurrent (at least three) URTIs within 12 months (OR 5.5), a history of wheezing at any age (OR 5.3) and a history of otitis media and tympanocentesis before 2 yrs of age (OR 3.6) in <5-yr-old children. The significant risk factors in 5–15-yr-old children were a history of recurrent URTIs within 12 months (OR 5.5) and a history of wheezing at any age (OR 5.3).
In line with the study by Teepe et al. 1, wheezing tendency and susceptibility to respiratory infections were the only significant determinants of paediatric CAP in our prospective, population-based study 4. Young age is an indisputable determinant of paediatric CAP, as also seen in our incidence figures. Interestingly, the urban versus rural place of residence and passive smoking were not associated with the risk of paediatric CAP 4.
In conclusion, pneumonia and other respiratory infections seem to cluster in the same children in the populations of high-income Western countries. Therefore, further studies should focus on the role of host factors in respiratory infections, including CAP in children. The CAP studies should be powered enough to allow the monitoring of the numerous environmental confounding factors, as well as age- and microbe-specific stratified analyses. The designs of the studies should be prospective and population-based to represent the whole spectrum of the disease, and in optimal cases should continue for several years with different epidemiological features. The selection and number of controls is of utmost importance. In an optimal study, both healthy controls and controls with non-pneumonic respiratory infections matched for age and sex from the same area should be considered. Genetic epidemiology offers the frame of reference for such studies.
Footnotes
Statement of Interest
None declared.
- ©ERS 2010
