INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Both asthma and respiratory symptoms are increasing among teenagers (1, 2), but very little is known about the natural history of asthma through childhood and adolescence (3). Although studies of selected patients have shown that most children with respiratory symptoms become asymptomatic in later childhood (4), further follow-up suggests that these symptoms may reappear in adulthood in a significant proportion (7). Most studies of the natural history of respiratory symptoms from childhood to adult life are based on selected groups of patients with moderate or severe asthma (5, 7, 9) and, as such, do not provide information on the full spectrum of the disorder. The few published longitudinal, population-based studies suggest that wheeze in early childhood is associated with different factors than persistent wheeze at age 16 yr (12) and that atopy, impaired lung function, and parental history of asthma are childhood risk factors for adult asthma (13).

This report presents the findings of a follow-up study of a population-based cohort of children aged 14-16 yr, previously studied when aged 6-8 yr, designed to examine the natural history of respiratory symptoms. The prevalence of current, persistent, and late-onset respiratory symptoms is reported, as is the prevalence of doctor-diagnosed asthma. The association between familial, environmental, and social factors and symptoms has been studied to assess which factors predict the natural history of respiratory symptoms in adolescents.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In 1987 one of the authors (J.B.C.) undertook a questionnaire survey of respiratory symptoms in a group of 3,542 children aged 6-8 yr (all children born between July 1, 1978, and June 30, 1980, and who were registered with one of 86 family practitioners in the Southampton area). The results of this survey have been previously reported (14- 16). In December 1994, the current addresses of all respondents to the 1987 survey were sought through the records of the Hampshire Family Health Services Authority (FHSA) and the UK Office of Population Consensus Surveys (OPCS).

In order to determine current prevalence of respiratory symptoms among the cohort, a postal questionnaire using questions identical to those employed in 1987 was sent to the parents of all children for whom it was possible to obtain a current address. In addition, this questionnaire contained questions concerning history of atopic disorders (asthma, hay fever, and eczema) in the child in question; history of asthma and atopy in his or her immediate family; the presence of smokers, other children aged less than 16 yr, and pets within the household; the use of gas cookers and stoves within the household; the birth weight of the child in question; and the current parental occupations. Each questionnaire was sent out with a reply-paid envelope. Although it was recognized that children of this age might complete a symptom-based questionnaire more accurately than their parents, the questionnaire also contained several retrospective sections pertaining to data from 1987 and thus was addressed to parents. Each child was, however, encouraged in an accompanying letter to help complete the questionnaire.

To assess current smoking habits of the children within the cohort, a separate confidential smoking questionnaire was sent to each child in question, together with a separate reply-paid envelope. Both questionnaires were distributed by second-class mail between January and March 1995; a reminder questionnaire was sent to all nonrespondents 6 wk after the original posting date.

Ethical approval for this study was granted by the Southampton University and Hospitals Local Joint Ethics Committee.

Statistical Analysis

Responses to both questionnaires were coded and double-entered onto a personal computer. All subsequent analyses were performed using SPSS statistical software (SPSS for Windows, Release 6.0; SPSS Inc., Chicago, IL). Social class for each child was determined from paternal occupation as defined in Government Standard Occupational Classifications (17) or, for households in which the father was absent, maternal occupation. For subsequent analyses the following definitions were employed:

Current wheeze/coughPresence of reported wheeze/cough in the 12 mo before the 1995 questionnaire.

Doctor-diagnosed asthmaPositive response to the question. "Has a doctor/nurse ever told you your child has asthma?"

Persistent wheeze/coughPresence of current wheeze or cough in 1995 in a child reported to have had wheeze cough in the 1987 survey.

Late-onset wheeze/coughPresence of current wheeze or cough in 1995 in a child reported never to have had wheeze or cough in the 1987 survey.

Active regular smokingSmoking of at least 1 cigarette/wk by the child in question in the 12 mo preceding the 1995 questionnaire.

Smoking householdHousehold with at least one smoker, other than the child in question, present in 1995. (Other smokers were identified on a "Yes/No" basis and no measure was made of the number of cigarettes smoked by each individual.)

Univariate analyses were performed using cross tabulations, ² tests, and logistic regression models to investigate the association between various risk factors and

a) the occurrence of current wheeze and cough,

b) the occurrence of doctor-diagnosed asthma,

c) the occurrence of persistent wheeze and cough, and

d) the occurrence of late onset wheeze and cough.

Variables included in initial univariate analysis were presence of eczema and hay fever in the child; presence of asthma, hay fever, and eczema in parents and siblings; gender; birth weight; smoking by the child; presence of another smoker in the household; presence of a gas cooker in the household; presence of furry pet(s) in the household; number of other children aged less than 16 yr in the household; single-parent households; and social class. The independent effect of each variable on symptoms was then assessed using mutually adjusted odds ratios derived from multivariate logistic regression models. Each initial model included those variables shown to be important from univariate analyses. Subsequently, variables shown not to be statistically significant were removed, using procedures outlined by Homer and Lemeshow (18), to obtain the final parsimonious models. In addition, for the subgroup for whom data concerning bronchial responsiveness to methacholine and skin-prick-test positivity to common aeroallergens were available from the 1987 study (14, 15), logistic regression models for persistent wheeze and cough including these two variables were derived.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 3,033 (95.2%) of the 3,187 respondents to the 1987 survey were traced and sent the two 1995 study questionnaires. Replies to the symptom questionnaires were received from 2,289 (75.5%) parents, while 2,164 (71.3%) smoking questionnaires were returned by the children themselves. A total of 2,150 (70.9%) households returned both questionnaires. Current wheeze was reported in 414 (18.2%) adolescents and current cough in 285 (12.4%). The corresponding figures for current wheeze and cough in the 1987 survey were 14.6% and 16.2%, respectively. Among the 334 children whose parents had reported wheeze in 1987 and for whom 1995 data was available, 184 (55.1%) were reported to have persistent wheeze. From the total of 1,670 children who had never wheezed by 1987, 169 (10.1%) had reported current wheeze in 1995, and thus had late-onset wheeze. Corresponding figures for persistent cough and late-onset cough were 131 (35.0%) and 102 (6.2%), respectively.

In 1995, 504 (22.3%) children had a doctor diagnosis of asthma, including 331 (80.1%) children with current wheeze, 172 (94.0%) with persistent wheeze and 117 (69.2%) with late-onset wheeze. Corresponding figures for current, persistent, and late-onset cough were 200 (70.4%), 134 (74.0%), and 65 (64.3%), respectively. Among the 192 symptomatic children who participated in the 1987 longitudinal study, responses were obtained from 150 (78.1%), of whom 46 had persistent wheeze and 40 had persistent cough.

Tables 1 and 2 show the results of the multivariate analyses examining the association between variables and symptom subgroups. Table 1 demonstrates the distribution of each variable within each subgroup, while Table 2 shows the odds ratios for each variable found to be significantly associated with a symptom subgroup in the multivariate analyses.

A personal history of atopy was linked with the presence of current symptoms, persistent symptoms, late-onset symptoms, and a doctor diagnosis of asthma. In the analysis of the subgroup of children studied in detail in 1987, skin-prick-test positivity was not, however, independently associated with persistent symptoms. In this group, presence of bronchial hyperresponsiveness was associated with persistent wheeze (odds ratio 1.27; 95% confidence interval, 1.04-1.56), but not with persistent cough. A family history of atopy or asthma was also significantly associated with current, persistent, and late-onset symptoms and with doctor-diagnosed asthma. In most analyses the association was with maternal asthma/atopy, although paternal hay fever was linked with persistent cough. The association between maternal asthma and persistent wheeze was one of two significant associations that were sex-linked, this one only reaching significance in females.

In the multivariate analyses, significant associations were also seen between tobacco exposure and each symptom subgroup. Active smoking was associated with current cough, late-onset cough, wheeze and, in male children only, with persistent wheeze. Paternal smoking was linked with late-onset wheeze, maternal smoking was linked with doctor-diagnosed asthma; and the presence of any smoker other than the child in question was significantly associated with current symptoms and with persistence of cough.

Besides atopy, family history of atopy, and smoking, few factors were shown to be significantly independently associated with symptoms. Coming from a lower social class was associated with late-onset wheeze, while single-parent households were linked with current and persistent coughs. Females were significantly more likely than males to suffer from late-onset wheeze, as were children of higher birth weight. The number of other children in the household was positively associated with late-onset cough and negatively with persistent wheeze; gas cookers were negatively associated with current and persistent wheeze; and pets were negatively linked with current wheeze.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has successfully followed up a cohort of children from early childhood into adolescence and demonstrated factors associated with the natural history of respiratory symptoms. Response rates to the 1987 and 1995 questionnaires employed compare favorably with previous longitudinal population-based studies of childhood respiratory disease (10, 12, 13) and questionnaire studies of adolescent smoking (12, 19). Although studies of this design are subject to potential responder bias, the prevalence of wheeze and cough in 1987 was very similar among both respondents and nonrespondents in the 1995 study (wheeze 14.8% versus 14.3%; cough 27.6% versus 26.8%), suggesting that the current cohort was representative of that studied in 1987.

Within this cohort, the prevalence of current wheeze had increased from 14.6% in 1987 to 18.2% in 1995, while the prevalence of current cough had fallen from 16.2% to 12.5%. More than half of the children with wheeze in 1987 continued to have symptoms in 1995, and more than one-third of children with cough in 1987 had persistent symptoms. Previous population-based studies have shown smaller proportions of subjects with continuing symptoms (12, 13), although the latter of these was reporting persistence much later in adult life. Other studies of selected populations have shown similar or higher levels of symptoms continuing in adolescence and into adult life (5, 6, 7, 9, 20), but by selecting children under hospital care for their asthma, these studies, by design, include children with more severe symptoms, a variable which itself has been linked with a poorer prognosis (6, 10). The prevalence of late-onset wheeze is similar to that of late-onset asthma in the only other population-based study that addresses this question (13). No previous study has specifically studied the prevalence of persistent or late-onset cough in a cohort of children.

A personal history of atopy was shown to be independently associated with current, persistent, and late-onset symptoms. Many previous studies have shown no association between presence of other atopic disorders and persistence of childhood asthma or wheeze (6, 21) but our finding of a positive association between atopy and persistent symptoms is consistent with the findings of Jenkins and colleague's large population-based study (13). Although the Tasmanian study did not find any relationship between late-onset symptoms and atopy, when one examines the data from that study in more detail, 62 of 81 subjects with late-onset asthma had "current atopic asthma," suggesting that atopy developing after the age of 7 yr was associated with late onset of asthma.

This current study also revealed a relationship between a maternal history of asthma and late-onset wheeze within the cohort. In addition, maternal asthma was associated with persistent wheeze, but this trend only reached significance among females. Previous studies of childhood allergic disorders and asthma have highlighted a specific link between symptoms and maternal asthma of atopy (24). In the current study, the possible inheritance of a predisposition to persistent wheeze was seen only in females, suggesting that females might be more susceptible to the effects of inheriting genetic influences associated with wheeze and that such genes might exert increasing effects in later childhood. This might then explain both the sex-associated link between persistent wheeze and maternal asthma, seen in the current study, and also the link between female sex and late-onset wheeze, shown in this and other studies (5, 13), which in turn leads to the poor prognosis for female adolescents in most longitudinal studies of childhood asthma (5, 7, 11, 22).

Other than personal atopy and family history of atopy, smoking, both active and passive, appeared to be the most important factor associated with the natural history of symptoms. In this cohort, active smoking was linked with current cough and late-onset wheeze and cough, while passive smoking was associated with current symptoms, doctor-diagnosed asthma, persistent cough, and late-onset wheeze. The association between symptoms and passive smoking was related only to the presence of another smoker, with no additional association with the number of other smokers present being shown. Among males there was also a significant association between persistent wheeze and regular active smoking, suggesting that wheeze in some males may primarily be a product of environmental factors, beginning in early life with the adverse effects of passive smoking on smaller caliber airways and continuing in later childhood as a result of both active and passive smoking. This greater effect of smoking in boys might be a function of more direct exposure either by smoking larger numbers of cigarettes per week or having smoked over a longer period of time. Although data on number of cigarettes smoked and duration of smoking was not collected in this study, it is unlikely that boys were currently smoking more often than girls, as more females (8.5% versus 6.3%) admitted to smoking every day.

No environmental factors other than active and passive smoking were shown to be positively associated with current symptoms. The negative association between the presence of furry pets in the household and current wheeze has previously been demonstrated (27) and has been explained as the reversal of a positive association by parent's tendency to remove pets from the home after a child or any other family member develops atopy or wheeze (28).

Several other factors were associated with the natural history of symptoms within the cohort. Children from lower socioeconomic classes were more likely to have late-onset wheeze, in contrast to other studies reporting asthma as a disease of the higher social classes (29). Persistent cough was more common in children from single-parent households, and late-onset cough was associated with an increased number of other children in the household. Conversely, a negative association was demonstrated between the number of household children and persistent wheeze. This latter association may reflect an immunologically mediated protective mechanism against atopy/asthma provided by increased exposure to viral infections in early life, which might prevent the proliferation of Th2 helper cell clones by predominantly activating Th1-like helper cells (30).

Conclusions

This study has shown that wheeze is more prevalent in this cohort at ages 14-16 yr than 6-8 yr, and that more than half of the children who were symptomatic in early childhood continued to suffer symptoms in adolescent life. Persistent wheeze in males and females may, in part, arise by different pathways and, as found in previous studies, females were more likely than males to develop wheeze in later childhood. A personal history of atopy was a strong predictor of current, persistent, and late-onset symptoms and of doctor-diagnosed asthma, while active and passive smoking were significantly associated with respiratory morbidity. Presence of other children in the household may protect against persistence of wheeze into adolescence, but is associated with increased risk of late-onset cough. Further study of this and other cohorts into adulthood is necessary to further understand the long-term natural history of childhood respiratory symptoms, thus assisting in the planning of future potential interventional studies and providing parents with more reliable information on prognosis.