Wheezing is a common symptom in early life, affecting up to one-third of young children.1^–3 There are several factors associated with early wheeze including abnormalities of pulmonary function.1 4^–6 Wheezy infants have obstructed pulmonary function1 4 5 and increased airway responsiveness5 6 prior to the onset of respiratory symptoms. Factors that may be important to abnormal pulmonary function in early life include male gender,7 a family history of asthma,8 maternal smoking during pregnancy9 and low birth weight.10

In a recent publication, Lucas et al11 showed that lung function (maximal flow at functional residual capacity, V′maxFRC) in young infants was negatively associated with postnatal weight gain, independent of length and birth weight. A negative relationship between weight and infant lung function has also been reported in a population of infants born prematurely.12 One suggestion11 is that these associations may be due to “catch up” in somatic growth in infants of low birth weight after a period of in utero stress, and that the postnatal somatic growth exceeded growth in pulmonary function. An alternative mechanism could be that some babies of normal birth weight experience rapid postnatal weight gain due to an excessively calorie-rich diet, and growth in pulmonary function is either impaired due to less favourable airway wall or airspace growth or is unable to match somatic growth.

In the present study we sought to determine whether the change in V′maxFRC between ages 1 and 12 months was negatively influenced by the change in weight over infancy. Our secondary aim was to relate change in V′maxFRC to infant feeding practices and the development of asthma symptoms. The data were taken from a longitudinal cohort study (Perth Infant Asthma Follow-Up Study) where serial measures of pulmonary function were taken during infancy, a contemporaneous record of infant feeding was made and a detailed follow-up was completed at 11 years of age.4

METHODS

Subjects and study protocol

Study participants were recruited from expectant parents attending a local antenatal clinic. The details of recruitment have been described previously.13 There was no selection for parental asthma or atopy and exclusion criteria included delivery before 36 weeks gestation and the presence of respiratory symptoms in the first month of life. A history of maternal or paternal smoking during the pregnancy and of asthma (ever) was recorded at enrolment. At ages 1, 6 and 12 months, weight, length and pulmonary function were measured, skin prick reactivity to common allergens was determined and urinary cotinine levels were measured. During infancy, mothers returned monthly postal questionnaires that reported details of feeding and respiratory symptoms. Annual postal questionnaires were returned between ages 2 and 6 years which detailed the presence of respiratory symptoms. At 11 years of age the children completed an assessment which included a respiratory questionnaire, spirometry, inhaled histamine challenge and skin prick reactivity to common allergens.

The Institutional Ethics Committee of Princess Margaret Hospital approved this study. Written consent was obtained from parents and verbal assent was obtained from the children.

Classification of respiratory symptoms

Current doctor diagnosed asthma (DDA) at all ages was defined as an affirmative response to the question: “Does your child have asthma which has been diagnosed by a doctor?” Doctor diagnosed asthma by 3 years of age was defined as the presence of DDA at either 2 or 3 years of age, DDA between 4 and 6 years was defined as the presence of DDA at 4, 5 or 6 years of age.

Infant pulmonary function

The techniques used have been described in detail elsewhere.13 After induction of sleep with chloral hydrate, a tightly fitting non-distensible jacket was applied to the chest. A balloon within the jacket was rapidly inflated at the end of tidal inspiration and V′maxFRC was derived from the resultant partial forced expiratory manoeuvre. The mean of at least five technically acceptable manoeuvres was reported.

Urinary cotinine analysis

Urine collection bags were placed by the parents a few hours prior to attending the hospital. Urine samples were stored for a maximum of 5 days at 4°C until transfer to a freezer at −20°C. The urine was analysed later in batches. Cotinine was measured by radioimmunoassay, standardised to creatinine concentration to adjust for urine dilution and expressed in units of ng/mg creatinine. The distribution of urinary cotinine was skewed with a long right-hand tail and was log transformed prior to analysis to achieve a near normal distribution

Skin prick testing

Skin reactivity of infants to cows’ milk, egg white, rye grass and Dermatophagoides farinae was determined using the skin prick test as described by Pepys.14 For the assessment at 11 years, skin reactivity to the following additional six allergens was also assessed: mixed grass (no 7); Dermatophagoides pteronyssinus; cat dander; dog dander; Alternaria alternans; and Aspergillus fumigatus. All allergens were supplied by Hollister-Stier, Elkhart, Indiana, USA. Histamine sulphate (10 mg/ml) was used as the positive control and 0.9% saline as the negative control. A positive skin test was defined as a weal to any allergen ⩾3 mm in its longest dimension or ⩾3 mm greater than the negative control.

Childhood pulmonary function

A portable spirometer (Pneumocheck Spirometer 6100; Welch-Allyn, Skaneateles Falls, New York, USA) was used in accordance with published guidelines.15 The rapid inhalation technique16 was used to determine bronchial hyperresponsiveness (BHR) to histamine; BHR was defined as a fall of at least 20% in forced expiratory volume in 1 s (FEV₁) after inhalation of a dose of ⩽7.8 μmol histamine. Short-acting bronchodilators were withheld for at least 6 h and long-acting β agonists were withheld for 12 h before testing; no children were treated with leucotriene antagonists.

Statistical analysis

V′maxFRC was log transformed prior to linear regression analysis to achieve a near normal distribution and divided by length (in centimetres) to adjust for length. A Student t test (two-tailed, equivariance assumed), χ² analysis or analysis of variance were used as appropriate to compare differences between groups. The primary outcome variable (which was normally distributed) was the percentage change in length-adjusted V′maxFRC which was calculated as follows:

The primary explanatory variable was percentage change in weight (weight at age 12 months/weight at age 1 month ×100). A multiple linear regression model was constructed to adjust for potential confounding variables (including birth weight, V′maxFRC at 1 month, exclusive breast feeding for ⩾6 months and gender). In a forward stepwise manner, these variables were introduced into the model and were retained if they changed the r2 value of the model by >30%; change in weight was the first variable entered. Logistic regression models were created to adjust the relationship between asthma and predictive variables for confounders which were introduced in a forward stepwise manner and only retained in the model if they were significant. A standard statistical program was used for the analyses (SPSS V.13.0) and significance was assumed at 5%.

RESULTS

Study subjects

The study population included 253 individuals (142 boys) in whom V′maxFRC was measured in 243 at the age of 1 month, 194 at the age of 6 months and 165 at the age of 12 months. Paired measurements of V′maxFRC were available in 154 infants at 1 and 12 months of age, in whom the presence or absence of DDA between ages 2 and 3 years was known in 145, the presence or absence of DDA between ages 4 and 6 years was known in 136 and duration of breast feeding for ⩾6 months was established in 127. Paired measurements of V′maxFRC were also available in 182 infants at 1 and 6 months of age and in 159 infants at 6 and 12 months of age. The 53 infants lost to follow-up after 1 month of age did not differ from the remaining cohort (table 1).

Lung function, weight, feeding practices, urinary cotinine and atopy during infancy

The median (interquartile range, IQR) values for V′maxFRC at 1, 6 and 12 months of age were 93 (64, 124), 150 (120, 196) and 188 (139, 277) ml/s, respectively. The mean (SD) change in length-adjusted V′maxFRC between 1 and 12 months of age was 87 (15)%; the corresponding change between 1 and 6 months was 92 (13)% and between 6 and 12 months was 94 (13)%. V′maxFRC/length was reduced in boys compared with girls at each assessment and this difference was significant at 1 month of age (p = 0.017). The mean (SD) weight at each assessment was as follows: 4.9 (0.7) kg at 1 month, 8.3 (1.0) kg at 6 months and 10.4 (1.2) kg at 12 months of age. The mean (SD) percentage change in weight was 215 (30)% between 1 and 12 months, 173 (24)% between 1 and 6 months and 128 (12)% between 6 and 12 months. Boys were heavier on each assessment but the percentage change in weight between assessments did not differ between boys and girls. Seventy infants received only breast milk for the first 6 months and 90 received formula milk within the first 6 months of life. Solids were introduced before 4 months of age in 87 and after this age in 101 infants. The median (IQR) urinary cotinine levels were 101 (51, 352) ng/mg creatinine at 1 month (n = 104), 60 (25, 303) at 6 months (n = 82) and 48 (26, 98) at 12 months (n = 72). Thirty-two individuals had at least one positive skin prick test during infancy.

Percentage change in weight and percentage change in V′maxFRC

There was a negative relationship between percentage change in V′maxFRC and percentage change in weight between ages 1 and 12 months (r = −0.18, r2 = 0.13, p<0.001, fig 1); this relationship was consistent over the first and second 6 months of infancy (r = −0.18, r2 = 0.11, p = 0.001 and r = −0.34, r2 = 0.15, p<0.001, respectively). The relationship between the change in V′maxFRC and weight between 1 and 12 months remained significant when confounders were considered: r = −0.14 when V′maxFRC at 1 month was considered; r = −0.18 when gender, birth weight urinary cotinine at 1 month or atopy was considered; and r = −0.20 when breast feeding was considered. When individuals were dichotomised by weight gain and V′maxFRC/length at 1 month, the greatest percentage change in V′maxFRC between 1 and 12 months of age was in those with a lower V′maxFRC at 1 month and a lower weight gain between ages 1 and 12 months (mean change 99%; fig 2). The remaining mean percentage changes in V′maxFRC were as follows: 88% for low V′maxFRC and high weight gain; 81% for high V′maxFRC and low weight gain; 77% for high V′maxFRC and high weight gain (p<0.001 for trend, ANOVA).

Weight gain, infant lung function and asthma

Asthma symptoms were not associated with the percentage change in V′maxFRC per se or the percentage change in weight between ages 1 and 12 months. The proportion with DDA by 3 years of age was highest (38%) in individuals with a low V′maxFRC at 1 month and a smaller change in V′maxFRC between 1 and 12 months and lowest (13%) in those with a high V′maxFRC at 1 month and a greater change in V′maxFRC (p = 0.017 for trend adjusting for percentage change in weight, table 2).

Weight gain, infant feeding and asthma

Infants in whom formula milk was introduced by 6 months of age were no heavier at birth or 1 month than those breast fed to at least 6 months but they had greater weight gain between 1 and 12 months (mean increase 0.65 kg (95% confidence interval (CI) 0.27 to 1.04), p = 0.001) and increased weight at 6 months (mean increase 0.34 kg (95% CI 0.02 to 0.70), p = 0.036) and 12 months of age (mean increase 0.65 kg (95% CI 0.25 to 1.05), p = 0.002). Diagnosed asthma by 3 years was reduced in association with exclusive breast feeding until at least 6 months of age (odds ratio (OR) 0.44 (95% CI 0.19 to 1.00), p = 0.043), independent of gender, maternal asthma and maternal smoking during pregnancy and weight gain between 1 and 12 months; this relationship did not persist for asthma during the second 3 years (OR 0.80 (95% CI 0.37 to 1.70)) or at 11 years of age (OR 0.92 (95% CI 0.37 to 1.70)). The percentage change in V′maxFRC between 1 and 12 months of age was not associated with duration of exclusive breast feeding. The introduction of solids before or after 4 months of age was not associated with change in weight and V′maxFRC or asthma.

Relationship between factors measured in infancy and outcomes at 11 years of age

Of the 154 individuals in whom V′maxFRC was determined at 1 and 12 months of age, 127 had a detailed assessment at 11 years of age. Mid expiratory flow (FEF_25–75) at 11 years of age was independently and positively associated with the percentage change in V′maxFRC; this was also independent of birth weight, V′maxFRC/length at 1 month, breast feeding and atopy and BHR at 11 years (table 3). Eighteen children assessed at 11 years of age (14%) had current DDA, 65 (51%) were atopic and 41 (32%) had BHR; none of these was related to the percentage change in V′maxFRC or weight, or to the duration of exclusive breast feeding.

DISCUSSION

This analysis studied the influence of postnatal weight gain on lung function during infancy. Our study has shown potentially important relationships between somatic and pulmonary growth during infancy, infant feeding practices and early transient asthma symptoms, and also pulmonary function in later childhood. Our hypothesis was that weight gain would be negatively related to change in V′maxFRC between 1 and 12 months of age and the data supported this. A series of interrelated associations became apparent when exploring the relevance of changes in postnatal weight gain and pulmonary function to respiratory symptoms, and this indicated a rather complex potentially causative mechanism (fig 3). Infants with lower birth weight and greater somatic growth during infancy had relatively reduced growth in lung function which was associated with transient asthma symptoms when combined with reduced lung function at 1 month. Exclusive breast feeding to 6 months had a potentially modifiable effect on excessive weight gain during infancy and was associated with reduced transient asthma symptoms. These associations could be one explanation for a transient wheezing syndrome in the infant who becomes a “fat happy wheezer”.

Relationship between changes in postnatal weight gain and V′maxFRC

A mechanism by which increased postnatal weight gain is associated with a relative reduction in lung growth—as evidenced by change in V′maxFRC—is not clear. This study has shown that the mechanism is independent of gender, atopy and antenatal and postnatal exposure to products of tobacco smoke. In a previous report based on observations taken from this cohort, we have observed that low V′maxFRC at 1 month of age persists at 12 months in some individuals but not others, and that those individuals in whom low V′maxFRC did persist were more likely to wheeze in the first and second years of life.17 A mechanism whereby low V′maxFRC occurs throughout infancy in some individuals but not others has not previously been reported, and the findings of the present study suggest that increased postnatal weight gain, possibly in infants of reduced birth weight, may be one factor associated with relatively low growth in pulmonary function during infancy.

Relationship between lung function in infancy and childhood

Previous studies, including those based on the present cohort, have reported tracking of pulmonary function from early life into later childhood;4 18 however, in the present study we have shown that the relationship between pulmonary function in very early life (“baseline”) and that in later life may be modified. Pulmonary function in very early life is mostly influenced by antenatal factors but, during infancy, the level of “baseline” pulmonary function may be modified by postnatal factors and this may include weight gain; thus, lower weight gain during early life could ameliorate adverse later respiratory outcomes associated with reduced “baseline” pulmonary function. Alternatively, the increased weight gain observed among those with low baseline and percentage change in V′maxFRC may be inevitable “catch up” growth after in utero stress, and this may not be modifiable.

Infant feeding practices

In the present study the introduction of formula feeds by 6 months of age was associated with increased postnatal weight gain and early asthma symptoms, and a similar relationship has previously been described among children from Western Australia.19 20 In their follow-up of a large birth cohort, Oddy et al19 20 report associations between less exclusive breast feeding and increased respiratory symptoms in infancy and at 6 years of age; this group also found that a higher body mass index was a risk factor for asthma at the age of 6 years. In the present study we found no relationship between duration of breast feeding and increased asthma beyond 3 years of age, which may be due to the relatively small numbers studied or the effect being limited to early childhood. The relationship between asthma and duration of breast feeding is complex and there are many inconsistencies between studies;21 some find a positive relationship, others no relationship and still others find a negative relationship.21 The results of the present study could indicate more than one mechanism associating exclusive breast feeding to 6 months with respiratory outcomes; the first could be via altered postnatal weight gain and involve abnormalities in pulmonary function and the second acting independently of pulmonary function.

Study limitations

There are issues which could be considered when interpreting our results. First, there is invariably regression to the mean for V′maxFRC and weight and this may partly explain the inverse relationship observed for change in V′maxFRC and change in weight. Infants where regression to the mean was not present (ie, low baseline and percentage change in V′maxFRC or high baseline and percentage change in V′maxFRC, table 2) were different in terms of early symptoms, birth weight and postnatal weight gain, and these differences cannot be explained by regression of V′maxFRC or weight to the mean. Second, a number of analyses were performed during this study and this increases the chance of false positive results. However, our findings were of sufficient consistency to suggest that multiple testing has not influenced our results. Finally, in the present study we present a series of associations which might indicate causative mechanisms but do not prove them; proof could only be inferred from a longitudinal study of infant lung function where infant weight gain was modified in a randomised controlled manner.

Conclusions

In summary, this study reports two novel findings: (1) the change in V′maxFRC between 1 and 12 months of age was negatively related to weight gain over that period; and (2) infants with reduced change in V′maxFRC between 1 and 12 months of age, in association with reduced V′max FRC at 1 month of age, have increased early transient asthma symptoms and reduced FEF_25–75 at 11 years of age. The relationships described are complex, but these observations will help understand possible mechanisms and useful therapeutic interventions.