Introduction
Community acquired pneumonia (CAP) is a common illness in children.1,2 Its incidence in primary care (PC) has been established at 4%-6% new cases/100 susceptible/year.2
The incidence of recurrent CAP (RP) has been estimated at 8% in hospital based studies and its presence is associated with bronchial asthma, among other diseases.1 Clark et al reported that children hospitalised due to CAP had a higher probability of subsequently developing asthma.3 After an exhaustive literature search, no studies in PC could be found which determined whether RP constituted a possible risk factor for the subsequent development of asthma, as compared to children who only have one episode of pneumonia (non-recurrent pneumonia, NRP).
Children diagnosed with asthma or wheezing episodes could have a higher probability of presenting with CAP.4-6 These studies, carried out in hospitals, did not look at the association between asthma and the subsequent development of RP.
RP has been associated with different diseases besides asthma, the majority severe, and whose early diagnosis is essential.1 CAP can also have serious repercussions on the pulmonary parenchyma,7 particularly in RP,8 therefore it requires suitable treatment, as well as the identification and treatment of its predisposing factors.
It is important to know the extent of RP in the community environment and establish whether it constitutes a possible risk factor for the subsequent development of childhood asthma (CA), the most prevalent chronic illness in childhood.9
The present study was carried out to determine if RP is a risk factor for the subsequent development of CA as compared with the presence of only 1 episode of CAP (NRP) in PC, and if patients with CA are more predisposed to present with RP.
Pacients and Methods
Study Design
Historical cohort study.
Population
All the episodes of CAP diagnosed in children <15 years old in the Girona Basic Health Area-4, between 1st January 1996 and 30th June 1999 were gathered (80 episodes in 65 children). The method for searching the CAP episodes has been published earlier.10
All children <15 years old who had at least one episode of CAP diagnosed and controlled in PC were included. The diagnosis was carried out by 2 radiologists, independently and blind.10
Episodes of CAP diagnosed, treated and monitored in the hospital environment and children with underlying chronic diseases (cystic fibrosis, heart disease, immunodeficiency) were excluded.
Definitions
Reviews of the computerised clinical histories were made to identify the cases according to the following definitions:
RP: children diagnosed with more than one episode of pneumonia during the period of the study.
NRP: children diagnosed with pneumonia on only one occasion during the period of the study.
CA: children who comply with the criteria which defines asthma based on a medical diagnosis (clinical and/or functional), according to the criteria of the Asturia Regional Working Party on Childhood Asthma Consensus Group11,12 (Table 1).
Statistical Analysis
Means, standard deviations (SD), ratios, and 95% confidence intervals (95% CI) were estimated. The *2 and the Student-Fisher t test were used, along with the estimation of the relative risk (RR) and its 95% CI.
Results
During the period of the study, 65 children presented with 80 episodes of CAP. The mean age±SD was 52.8±33.8 months. Of these, 38 were girls (58.5%). There were no significant differences between the age of those affected according to sex (53.1 vs 52.6 months; P=.9). Of the cases identified, 18 children (27.7%; 95% CI, 16.8-38.6) had more than 1 episode of CAP (RP). The prevalence of CA in the whole sample was 49.2% (32 children) (95% CI, 37.1-61.4).
RP was associated with the development of previously undiagnosed CA (Table 2). The risk of CA was greater in children with RP than in those who had NRP (RR=4.1; 95% CI, 1.9-8.9; P=.0002). The prevalence of CA among children with RP was 64.7% versus 15.4% in those who had NRP.
There were no significant differences between the percentages of RP and NRP in children previously diagnosed with CA who had not presented with any previous episode of CAP (Table 3); 26% had 1 episode of CAP and 33.3% more than 1 episode (RR=1.28; 95% CI, 0.5-3.0; P=.6).
Of the 18 children with RP, 15 were previously diagnosed or were subsequently diagnosed with CA (P=.0007).
Discussion
This study has some limitations. The definition of RP which has been used was less strict than that found in the literature (at least 2 episodes in 1 year or more than 3 episodes throughout life13), whilst in the present study RP was considered when more than 1 episode was presented during the period of the study (approximately 3,5 years). Despite this fact, the definition of RP in this study has been sufficiently sensitive to be able to demonstrate significant and clinically relevant differences as regards the subsequent development of asthma after presenting with RP as compared to children diagnosed with NRP. Another important limitation of this work is its small sample size, a fact which prompts caution in the interpretation of the statistically non-significant associations, since this can happen in a beta type error. The 95% CI are too wide for the same reason. The sample size was not determined beforehand, but was arrived at by the number of patients who were detected with RP during the period of the study. It is important that prospective studies are carried out in the future with sufficient statistical power to be able to provide more accurate data.
This study suggests that RP can be a risk factor for the subsequent development of CA, as compared to the presence of only 1 episode. There are no PC studies carried out to be able to compare these results. In the hospital environment, it has been shown that children admitted due to CAP have a higher prevalence of CA (45%) in subsequent years3. However, the repercussions of RP on the subsequent incidence of CA have not been studied separately. In our study, the prevalence of CA among patients with RP was unusually high. This fact should serve as a warning to the PC paediatrician, since the probability of being confronted with a still undiagnosed asthmatic child, as such, is very high. In the present study, the prevalence of CA in our sample group was very high. There was no patient control group without CAP to be able to establish more accurate comparisons. However, epidemiological population studies carried out in Spain are available, and using the same definition of CA, have reported a prevalence of CA in the general paediatric population which varies between 10.1% and 11.5%.11,12. The percentages found in our study are much higher than these.
The hypothesis of CA as a predisposing factor for presenting with CAP has been the subject of more studies, the majority in hospital situations, case-controlled and carried out in developing countries. All have established, to a greater or lesser extent, the association of the previous existence of asthma episodes or wheezing with a higher probability of presenting with a CAP. In one study carried out in the community, it also confirmed this association.14 Uniquely, 1 study carried out in a hospital which had as its objective to determine whether the pre-existence of CA was associated with a higher risk of presenting with RP and/or persistent pneumonia, confirmed this association in hospitalised patients.15 In our study, the percentage of children with RP was very high in comparison with children who only had one episode of CAP. The difference between both groups was not statistically different, perhaps owing to the small sample size. It cannot be ruled out that CA, besides being a risk factor for the development of CAP,4-6 also increases the probability of presenting with RP. Future prospective studies with a larger sample and in the community should help to resolve the problem.
In the present work, the strong association between CA and RP in our area is obvious. In the absence of larger studies which could confirm this data, the association between the 2 clinical entities, reported in hospital studies, also appears to be confirmed in the PC environment.
This study is the first to analyse the relationship between RP and childhood asthma in PC. Despite the limitations already mentioned, its results are relevant to the PC paediatrician: it is necessary to be alert and make a special follow-up on every child diagnosed with RP (at least 2 episodes of CAP), since the possibility of developing CA is multiplied by four; more studies in the community are needed to accurately determine the prevalence of CA in children previously diagnosed with CAP, to study whether the incidence of RP is higher than that of NRP in children previously diagnosed with CAP and to confirm if RP is a risk factor in PC for the subsequent development of asthma.
