INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The maximum achieved level of lung function during early adult life has important implications for the onset of obstructive airways diseases in later adult life (1, 2). The level of this plateau is determined by the course of growth in lung function during childhood. Cross-sectional population studies in children and adolescents have established that level of lung function is related to height and is reduced in the presence of asthma. However, there are areas of uncertainty in existing knowledge.

It has been established that airway hyperresponsiveness (AHR) predicts the subsequent rate of decline in lung function in adults (3). However, little is known about the temporal relations between AHR and lung function growth in childhood and young adulthood. Although a number of longitudinal studies of asthma and lung function in childhood and young adulthood have been conducted (4), they do not allow a conclusion about the direction of the association between these characteristics. Interpretation is limited by the relatively small number of repeated observations and the absence of analysis of the temporal sequence of events; that is, the effect of asthma on the subsequent growth in lung function or, alternatively, the effect of impaired lung function on the subsequent incidence of asthma.

The relation between growth in height and growth in lung function also requires further investigation. It has been shown that the peak growth velocity in lung function occurs after the peak growth velocity in height (8). However, the relation between cessation of growth in height and cessation of growth in lung function requires further investigation. Describing this relation has important implications for understanding the determinants of the maximum level of lung function achieved in early adult life.

Longitudinal population studies, in which AHR, respiratory symptoms, and lung function are measured on several occasions, make it possible to address the question of whether the presence of asthma (manifest as AHR and reported recent wheeze) is likely to be a cause for reduced lung function growth. Such studies, especially in a cohort enrolled in childhood, followed through adolescence and to young adulthood, also allow assessment of whether lung function continues to increase after height growth has stopped.

The aim of this study was to examine the effect of asthma, measured as AHR and a history of recent wheeze, on subsequent lung function in a cohort of children who were studied prospectively into early adulthood. The growth pattern of lung function and its relation to growth in height were also examined.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

In 1982, a large random sample of third- and fourth-grade schoolchildren, aged 8-10 yr, living in the coastal town of Belmont, New South Wales, Australia, was studied. The study methods have been described previously in detail (9). In brief, the study was conducted at 8 of the 10 primary schools in Belmont. Two small schools with fewer than 30 children in the appropriate grades were excluded. A total of 718 children participated in the initial study, representing a response rate of 87%. From 1984 to 1992, five follow-up studies were conducted at 2-yr intervals. At each follow-up, every effort was made to contact each subject previously enrolled. All studies were undertaken in the winter months.

Of the cohort of 718 subjects, those who participated in at least one of the follow-up studies between 1988 (the third study) and 1992 (the sixth study) are included in the data analyses presented in this article. A total of 557 subjects (78% of the initial sample) satisfied this selection criterion.

Questionnaires

The questionnaires sought information on symptoms (wheeze, wheeze after exercise, and cough), diagnosed asthma, medication use, respiratory illness history, parental smoking, and environmental exposures. In the studies between 1982 and 1988 (when the subjects were younger than 16 yr of age), the questionnaire was completed by the parents of the subjects. From 1990, a self-administered questionnaire was completed by each subject.

Lung Function and Airway Hyperresponsiveness

The lung function of each subject was measured at each study time. From 1982 to 1990, Vitalograph spirometers (Vitalograph, Buckingham, UK) were used and tracings were read manually. In 1992 S&M dry rolling seal spirometers (Mijnhardt BV, Bunnik, The Netherlands) linked to a personal computer were used. Spirometry was performed with the subject standing and without a nose clip. Both spirometers complied with ATS standards (10). All spirometry tests were conducted by trained technicians in accordance with the ATS guideline. The spirometers were calibrated daily and checked for any malfunction. Forced expiratory maneuvers were repeated until readings of FEV₁ and FVC were reproducible to within 100 ml, and the largest value of the two reproducible readings was used in the analyses.

A bronchial challenge test with histamine, using the rapid method (11), was administered to all subjects who had a baseline FEV₁  60% predicted. After baseline lung function was measured, the FEV₁ was recorded again after inhalation of saline. Histamine diphosphate was then administered, using No. 40 glass hand-held nebulizers from 1982 to 1986 and using DeVilbiss No. 45 hand-held nebulizers from 1988 to 1992 (DeVilbiss, Heston, UK), in doubling doses ranging from 0.03 to 3.92 µmol. The test was stopped if the FEV₁ fell by 20% or more or if all histamine dose steps to 3.92 µmol had been administered. Salbutamol aerosol was given to aid recovery when necessary. For subjects who had a fall in FEV₁ of 20% or greater, the dose of histamine that caused a 20% fall (PD₂₀FEV₁) was calculated. Subjects with a PD₂₀ FEV₁ less than or equal to 3.92 µmol were classified as having airway hyperresponsiveness (AHR). Subjects who had used a -agonist within 6 h of presenting were asked to withhold medication before returning for later testing.

Definitions

AHR and recent wheeze were defined at each study time. Subjects who reported wheeze or wheeze after exercise in the previous 12 mo of the study time were classified as having recent wheeze at study time. Subjects who had recent wheeze and AHR at study time were classified as having current asthma. Subjects who did not have AHR and recent wheeze were classified as normal. At each follow-up visit, AHR and recent wheeze were reclassified using the same definitions.

Statistical Methods

Data were analyzed with the statistical package SAS (SAS Institute, Cary, NC). An autoregressive model (12) was used to examine the pattern of lung function growth and the effect of various factors on lung function growth. This is a suitable statistical model with which to analyze longitudinal data, because it can produce serial correlations among measurements performed on several occasions. The autoregressive model could be used with unequally spaced observations due to missing observations within subjects. The dependent variable (outcome variable) in the autoregressive model is lung function (FEV₁ or FVC) at the current study time, while the explanatory variables (predictors) are lung function (FEV₁ or FVC) at the previous study time, together with height, change in height, age, sex, AHR at the previous study time, and recent wheeze at the previous study time. In the autoregressive model, AHR and recent wheeze are treated as time varying predictors similar in nature to age and height. In this model, the classification of each subject with respect to recent wheeze and AHR status is permitted to change from one study time to a subsequent study time.

The term "expected value" was used to describe the fitted value calculated from the autoregressive model with all the predictors included in the model. Therefore, "expected FEV₁" refers to the estimated FEV₁ for a subject that was calculated from the autoregressive model by using FEV₁ at the previous study time, height, change in height, age, sex, AHR at the previous study time, and recent wheeze at the previous study time as predictors.

To assess more accurately the possible causal effect of AHR and/ or recent wheeze on the growth in lung function, AHR and recent wheeze at the previous study time, rather than those at the current study time, were included as predictors in the autoregressive model. To study the sex difference in lung function growth at different stages of growth, an age and sex interaction was included in the autoregressive model. Other possible interactions, such as AHR and age, recent wheeze and age, and AHR and recent wheeze, were also considered.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Among the 718 subjects in the initial cohort, there were 557 (78%) who participated in at least one follow-up visit between 1988 (visit 4) and 1992 (visit 6). Of the 557 subjects, 222 (40%) attended all the six visits, 148 (27%) attended five visits, 100 (18%) attended four visits, 53 (9%) attended three visits, and 34 (6%) attended only two visits. At baseine, there were no significant differences in the prevalence of atopy, recent wheeze, AHR, and the level of lung function between the 557 subjects included in this analysis and the remaining 161 subjects, indicating that the follow-up group was representative of the initial cohort (Table 1).

Growth in Height

Height has been established as the best predictor of lung function both in cross-sectional and longitudinal studies (13, 14). Figure 1 shows the pattern of height growth in this cohort. At age 8 yr, boys had a slightly higher median height than girls (130 versus 127 cm), but from age 9 to 13 yr, the median height of boys and girls was similar. The relationship between height and age in both sexes was linear from 8 to 12 yr, but had a curved relation thereafter. From age 13 yr, boys had a steeper growth velocity in height than girls. The peak growth velocity in height occurred between age 12 and 13 in both sexes, but the magnitude was different: 10 cm/yr in boys and 8 cm/yr in girls. Growth in height stopped between age 15 and 16 yr in girls, when the median height was 160 cm and 1 yr later in boys, between age 16 and 17 yr, when the median height was 172 cm.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Growth pattern in height of the whole sample, separated according to sex. The median height of each age group is shown against age.

Growth in FEV₁ and FVC in Normal Subjects

The growth patterns of FEV₁ and FVC in normal subjects of both sexes are shown in Figures 2 and 3. The initial FEV₁ values at age 8 yr in boys and girls were slightly different. The mean FEV₁ was 1.75 L in boys and 1.65 L in girls (p = 0.09). However, girls had a greater increase in expected FEV₁ before age 12 yr. At age 12 yr, the expected FEV₁ in normal subjects was similar for both sexes, but after age 13 yr, normal boys had a larger growth in expected FEV₁ than girls. The peak growth velocity in FEV₁ in boys (400 ml/yr) occurred at age 14 yr, 1 yr later than the peak growth velocity in height. The peak growth velocity in FEV₁ in girls (300 ml/yr) occurred at age 13 yr, the same age as the peak growth velocity in height. Both in boys and in girls, FEV₁ continued to increase at age 17 yr, when growth in height had stopped. However, the change in expected FEV₁ in boys after age 17 yr was about 200 ml/yr, whereas the change in expected FEV₁ in girls was about 100 ml/yr.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Growth pattern in FEV1 and FVC in normal subjects. Expected FEV1 and FVC values were calculated from the autoregressive model.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Change in expected FEV1 and FVC in normal subjects and change in median height separated according to sex.

Effects of AHR and Recent Wheeze on FEV₁ Growth

Figure 4 shows the relationship between having AHR and/or recent wheeze at the previous visit and the growth pattern of FEV₁. After taking into account the effects of height, growth in height, age, sex, age and sex interaction, and FEV₁ at the previous visit, having AHR at the previous visit was a significant predictor of FEV₁ at the subsequent visit (p < 0.001). Having recent wheeze at the previous visit was also a significant predictor of FEV₁ at the subsequent visit, but the effect was less than that of AHR (Table 2). The effect of AHR on FEV₁ was age specific as shown by a significant negative interaction between AHR and age. For a subject with AHR throughout the whole study period, the expected FEV₁ at age 19 yr was 12% lower than that for a subject of the same sex who did not have AHR, provided both had the same FEV₁ at age 8 yr and the same growth in height. For a subject with recent wheeze throughout the whole study period, the expected FEV₁ at age 19 yr was 3% lower than that for a subject of the same sex who did not have recent wheeze, assuming both had the same FEV₁ at age 8 yr and the same growth in height.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Effects of AHR and recent wheeze on the growth in FEV1. Expected FEV1 was calculated from an autoregressive model, with AHR and recent wheeze at the previous visit as predictors.

Effects of AHR and Recent Wheeze on FVC Growth

The effects of AHR and recent wheeze on FVC growth were different from the effects on FEV₁ growth. After taking into account the effects of height, growth in height, age, sex, interaction between age and sex, and FVC at previous visit, having AHR at the previous visit was not a significant predictor of FVC at the subsequent visit (p = 0.16). Recent wheeze at the previous visit was also not a significant predictor of subsequent FVC (p = 0.99).

Growth Pattern of FEV₁/FVC Ratio

The expected FEV₁/FVC ratio of normal subjects and of subjects with AHR is shown in Figure 5. The initial FEV₁/FVC ratio at age 8 yr was 0.9 in boys. The expected FEV₁/FVC ratio was unchanged until age 15 yr in normal boys and then declined slightly. However, for boys with AHR throughout the study period, the expected FEV₁/FVC ratio declined consistently and was much steeper after age 14 yr. At age 19 yr, the expected FEV₁/FVC ratio was 0.79 in boys with AHR, which was 9% lower than in normal boys. In normal girls, the expected FEV₁/FVC ratio increased slightly before age 13 yr and remained at a similar level at age 19 yr. However, the expected FEV₁/FVC ratio declined after age 14 yr in girls with AHR. At age 19 yr the expected FEV₁/FVC ratio in girls with AHR was 0.82, which was 9% lower than in normal girls.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Growth pattern of expected FEV1/FVC ratio in boys and girls with and without AHR. Expected FEV1/FVC ratio was calculated as the ratio of expected FEV1 and expected FVC.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

By studying a cohort, from age 8 to 20 yr, we were able to demonstrate that growth in lung function in normal adolescents continues after the cessation of growth in stature. The presence of AHR and/or recent wheeze, which are two central features of asthma, is associated with a subsequent reduced rate of growth in airway caliber, measured as FEV₁ and FEV₁/ FVC ratio. However, these two features of asthma did not reduce the rate of growth in lung size, measured as FVC. The implications are either that asthma causes the reduced rate of growth in lung function or that factors that promote asthma also impair lung function growth. In either case, this would lead to people with asthma having a lower level of lung function in early adult life. The observed temporal sequence does not support the premise that asthma is a consequence of reduced lung function growth.

We used reliable methods of data collection in this study. The initial study sample was selected randomly from a general population of schoolchildren and the sample members who participated in the follow-up studies did not have a significant selection bias in terms of respiratory characteristics as measured at the initial study. Because lung function, height, AHR, and recent wheeze were all measured at each visit by the same protocols, the effects of AHR and wheeze could be assessed before lung function became impaired. Although not every subject participated in all six visits during the follow-up period, we used an autoregressive model to adjust for unequally spaced observations. The two possible sources of bias of this study were that questionnaires were completed by parents in the first four visits and were self-completed in the last two visits and that the nebulizers and spirometry used for histamine challenge test were changed (15). Although these methodological changes may have influenced measurements of the prevalence of AHR and symptoms, they were less likely to have biased the results in assessing AHR and recent wheeze as risk factors for lung function impairment.

The pattern of growth in lung function differed between normal boys and girls in this study. Peak growth rates in lung function (both FEV₁ and FVC) lagged behind peak growth rates in height in boys, but not in girls. Other studies have shown that the age of peak growth rate in height precedes the age of peak growth rate in lung function, but there are some inconsistencies in reported sex differences. Wang and colleagues found that peak FEV₁ and FVC velocity occurred on average about 0.6 to 0.9 yr later than peak height velocity, but there was no sex difference in the relationship between growth spurt of lung function and height (16). However, Sherrill and coworkers found that there was a significant difference between the age of peak growth in height and the age of peak growth in FEV₁ and FEF_25-75 in boys but not in girls (17). This is consistent with our findings and suggests that the early growth spurt in girls includes lung volume and lung function.

We found that pulmonary function continued to grow after the cessation of height growth, and this was more evident in boys. Wang and colleagues found that at age 16 yr, FVC growth appeared to reach a plateau in girls but that boys continued to grow at least up to age 18 yr (16). Hibbert and coworkers found that in girls, growth in lung volumes ceased at about the same time as growth in height stopped, but in boys it continued to grow after height growth was completed (8). In our study, we found that the annual change in FEV₁ was approximately 200 ml/yr from age 17 to 19 yr in boys and approximately half this rate in girls. This implies that there are some factors, other than growth in stature, that influence growth in lung function in late adolescent boys. We could speculate that it may relate to an increase in muscle power during this period. Our findings highlight the difficulty in using height as the single predictor for lung function when calculating normal values in young adulthood.

This longitudinal study also gave us the opportunity to examine the effect of asthma on growth in lung function. We have shown that AHR and recent wheeze at the previous study time predict lower growth in lung function over the succeeding interval. Our findings in this regard extend the observation of Sherrill and coworkers (6), who demonstrated, in a cohort of children from age 9 to 15 yr, that the presence of AHR and/or a history of wheeze at any given time was associated with lower growth in FEV₁ over the preceding interval.

Our study did not show significant sex differences in the effect of asthma on lung function growth. However, others have observed differences. Gold and colleagues found that children who reported asthma and wheezing in the preceding year had lower levels of FEV₁ than those who did not have these symptoms. These researchers also found that the effect was greater in boys (5.7% lower) than in girls (3.4% lower) (18). Weiss and coworkers found that active wheezing symptoms were associated with a deficit in growth in FEV₁ in girls but was not a significant predictor of FVC growth (19).

Asthma influences growth in airway caliber rather than lung size. Our observation of significant effects of AHR and recent wheeze on growth of FEV₁ and FEV₁/FVC ratio but not on FVC is similar to the findings of other studies (5, 6, 18). Redline and colleagues measured FEF_25-75, a sensitive measure of airway caliber, in a cohort study of children and found that, compared with normal subjects, those with AHR had lower levels of FEF_25-75 but their FVC values were not significantly different (5). As noted above, Sherrill and coworkers found that level of FEV₁ were significantly lower in children with AHR than in the control population, but they also observed that the effect of AHR on VC growth was less obvious (6). This differential effect on airway caliber and lung size, which is consistent across several studies, supports the hypothesis (20) that asthmatic children have nonproportional growth of airways and lung parenchyma.

We found that a history of recent wheeze and the presence of AHR had independent, additive effects on the subsequent growth in FEV₁. This supports our contention from previous work that the combined presence of recent wheeze and AHR represents a clinically relevant entity and we have used this criterion as an epidemiologically useful definition for asthma (21).

Subjects who have wheeze and/or AHR (that is, asthma) at any time during the course of the study will be more likely to have bronchoconstriction, and hence lower levels of FEV₁, at any study assessment. This would cause such subjects to have an overall lower level of lung function (FEV₁) but would not explain the observed reduced rate of growth in lung function since there is no reason to suspect that the degree of bronchoconstriction would increase over time. Hence, we believe that these data do provide a robust test for the hypothesis that asthma is associated with a reduced rate of growth in FEV₁.

In summary, we found that lung function continues to grow in late adolescence after the cessation of growth in height. This effect is more marked in males than in females and may be related to growth in muscle mass and power during this period. We have also shown that, during childhood and adolescence, airway caliber grows at a slower rate in those with asthma than in normal subjects. We would predict that this will lead to a lower maximum level of lung function and place them at greater risk of having important impairment of lung function in late adult life (1). We will test the first component of this prediction by further follow-up of this cohort.