OBJECTIVE. The use of antibiotics in infancy and subsequent changes in the intestinal bacterial flora have been discussed as risk factors for the development of asthma. However, it has been difficult to exclude the possibility that antibiotics have been given in early episodes of wheezing. As a result, there has been a risk of reverse causation. To minimize the risk of reverse causation, we have focused on the effect of antibiotics that are already administered on the neonatal ward.
METHODS. In a cohort study of infants born in western Sweden in 2003, we studied the development of wheezing. The families of the infants were randomly selected and sent a questionnaire at child ages 6 and 12 months. The response rate was 68.5% to the 6-month questionnaire and 68.9% to the 12-month questionnaire.
RESULTS. At 12 months, 20.2% of infants had had 1 or more episodes of wheezing, and 5.3% had had 3 or more episodes. Inhaled corticosteroids had been taken by 4.1% of the infants. Independent risk factors for wheezing disorder treated with inhaled corticosteroids were neonatal antibiotic treatment, male gender, gestational age of <37 weeks, having a mother with asthma, having a sibling with asthma or eczema, and breastfeeding for <5 months.
CONCLUSIONS. Treatment with antibiotics in the neonatal period was an independent risk factor for wheezing that was treated with inhaled corticosteroids at 12 months of age. These results indirectly support the hypothesis that an alteration in the intestinal flora can increase the risk of subsequent wheezing.
Several studies have suggested that antibiotic exposure in the first year of life is a risk factor for the development of childhood asthma.1–3 However, it has been difficult to exclude so-called reverse causation, that is, the tendency for children with asthmatic symptoms to receive more antibiotics than their peers, presumably for the treatment of early episodes of wheezing.2,4 To date, this question has remained unresolved.
The aim of this study was to analyze the risk factors for wheezing at 12 months of age with special reference to antibiotic treatment. To minimize the risk of confounding due to reverse causation, we have focused on the influence of antibiotics given during the neonatal period.
MATERIALS AND METHODS
Infants of Western Sweden is an ongoing study in the western Sweden region. The region has 1.5 million inhabitants, which is one sixth of the Swedish population. It comprises urban, rural, and coastal areas, and the largest city is Göteborg, which has 500000 inhabitants.
Of the total birth cohort of 16682 infants born in the region in 2003, 8176 families (50%) were randomly selected and sent an invitation to participate in the study. Respondents answered questions regarding the infant and the family, which were to be filled out when their child was 6 months of age (Fig 1). If the family was willing to participate, it was sent another questionnaire when their child was 12 months of age. The first questionnaire, Q1, was filled in by 5605 (664 + 4941) parents (response rate: 68.5%) and the second, Q2, was sent to 7241 parents and filled in by 4987 parents (response rate: 68.9%). Details regarding the questionnaires have been published previously.5
In the statistical analysis, 2 × 2 tables with a χ2 test and binary logistic regression were used, and the risks were estimated by using odds ratios (ORs) with 95% confidence intervals (CIs). Factors that were significant with a P value of <.01 in the univariate analysis were then analyzed in multivariate models, 1 with “wheezing ever” (1 or more episodes of wheezing) and the other with “wheezing disorder treated with inhaled corticosteroids” as dependent variables. To avoid reverse causation, variables interpreted as secondary to the wheezing disorder were not included in the multivariate logistic regression model. The SPSS (SPSS Inc, Chicago, IL) statistical package6 was used to perform the calculations. The study was approved by the ethics committee at Göteborg University.
At 12 months of age, 20.2% of the infants studied had had 1 or more episodes of wheezing, and 5.3% had had 3 or more wheezing episodes during the first year of life. Of the 12-month-old infants, 4.1% had received inhaled corticosteroids.
Univariate risk factors (P < .01) for wheezing ever are shown in Table 1. Major risk factors were a short breastfeeding period, male gender, having a mother or sibling with asthma, rhinoconjunctivitis or eczema, having a father with asthma, gestational age of <37 weeks, admission to a neonatal ward, receiving neonatal antibiotics, sleeping problems, pacifier (dummy) use, receiving antibiotics in the first year (neonatal antibiotics excluded), admission to a hospital, 3 or more awakenings nightly, and maternal smoking during pregnancy.
The analysis of risk factors for wheezing disorder treated with inhaled corticosteroids is shown in Table 2. Major risk factors (P < .01) were a short breastfeeding period, male gender, maternal asthma or eczema, having sibling with asthma, rhinoconjunctivitis or eczema, gestational age of <37 weeks, admission to a neonatal ward, receiving neonatal antibiotics, sleeping problems, no consumption of food containing lactobacilli (eg, yogurt), receiving antibiotics in the first year (neonatal antibiotics excluded), admission to a hospital, a diagnosis of food allergy, and 3 or more awakenings nightly. Maternal smoking was not significant in this analysis.
In the resulting multivariate models, independent significant risk factors for wheezing ever were neonatal antibiotic treatment, male gender, gestational age of <37 weeks, mother with asthma, sibling with asthma or eczema, rare use of pacifiers, and breast feeding for <5 months (Table 3).
Independent significant risk factors for wheezing disorder treated with inhaled corticosteroids were neonatal antibiotic treatment, male gender, gestational age of <37 weeks, having a mother with asthma, having a sibling with asthma or eczema, and breastfeeding for <5 months (Table 3).
Maternal smoking in pregnancy was not an independent risk factor in the multivariate analyzes.
The variables antibiotics in first year, ever admitted to a hospital, troubled sleep, and >3 awakenings nightly are closely related to wheezing, because infants with wheezing are more likely to receive antibiotics, to be admitted to a hospital because of wheezing, and to have troubled sleep. Accordingly, we excluded these variables from the multivariate logistic regression model.
Treatment with antibiotics was more common among extremely preterm infants. However, neonatal antibiotic treatment increased the risk of later wheezing in both term and preterm infants. In a series of stratified analyses, the OR for infants with a gestational age of ≥33 weeks was 2.9 (95% CI: 1.8–4.7), and for infants with a gestational age of ≥37 weeks was 2.9 (95% CI: 1.7–4.9).
In this article we report that treatment with antibiotics in the neonatal period is the most potent independent risk factor for wheezing treated with inhaled corticosteroids during the first year of life. Furthermore, we present new high-prevalence figures for early wheezing disorder, including asthma, from a large region of Sweden.
The aim of the study was to analyze the risk factors for wheezing during the first year of life. To address this issue, we asked for background information on the environment and various events. The strengths of the study were the large size of the cohort (50% of the infants born in western Sweden in 2003) and that the background information was tainted with no more than 6 to 12 months of retrospection.
Although there is always a need to be concerned about the response rate in studies with large postal questionnaires, we regard our response rate of ∼70% as satisfactory, especially in view of the large size of the cohort. In addition, there is always a risk that less-privileged social strata will be underrepresented in the responses to postal questionnaires. To further evaluate such a selection bias, we compared statistics concerning gestational age, birth weight, maternal age, smoking during early pregnancy, and percentage of cesarean deliveries between the full cohort in Sweden in 2003 (n = 97539) and our study material (n = 4921). There were no significant differences regarding these variables (Table 4), which suggests that our subjects were representative of the national Swedish cohort in 2003. Given the strong relationship between smoking and several disadvantageous factors, it is reasonable to assume that the material from western Sweden is also representative of the population from a sociodemographic point of view. However, even if a slight overrepresentation of middle and higher social strata was present, such selection mechanisms would hardly affect the internal validity of the relationship between risk factors and disease.
When the infants were 12 months old, 20.2% had had 1 or more episodes of wheezing. This figure corresponds very well to early data from the Tucson Children's Respiratory Study,7 which reported an incidence of 19.6% of bronchiolitis during the first year. At 2 years of age, 25.7% of the infants in the Swedish BAMSE: Child, Allergy and Milieu in Stockholm, an Epidemiological Study (in Swedish) birth cohort study had had wheezing episodes at some time.8 These figures are of approximately the same magnitude and show how common this symptom is among infants.
It is interesting to note that 4.1% of the infants had been treated with inhaled corticosteroids at 12 months of age. We did not find any other studies that reported the prevalence of treatment with inhaled corticosteroids in the first year of life. The group of infants who received inhaled corticosteroids, the usual treatment for recurrent wheezing in this age group in Sweden today,9 could be seen as corresponding to the idea of doctor-diagnosed asthma. This assumption is further supported by the finding of a highly significant difference in mean number of obstructive episodes between infants treated with inhaled corticosteroids (7.0 episodes) and infants not treated with them (2.4 episodes) (P < .001).
The main hypothesis tested in this article was whether treatment with antibiotics early in life is a risk factor for subsequent wheezing. Several studies have reported an increased prevalence of asthma after antibiotics were given during the first year of life.1,3,10–12 However, this association has been questioned in recent studies.13–15 One problem with studies that show an association between antibiotics during the first year of life and an increased risk of asthma is that many of them have not been able to exclude the possibility that the antibiotics were given for early episodes of wheezing, which might confound the results by reverse causation.4 To minimize the risk of reverse causation, we have focused on the effect of antibiotics administered on the neonatal ward.
However, it could be argued that reverse causation cannot be excluded if antibiotics were given to infants with diagnoses such as bronchopulmonary dysplasia (BPD), respiratory distress syndrome, or pneumonia. In such infants, the subsequent development of wheezing could be a consequence of the neonatal respiratory problems, not the administration of antibiotics.
Our parental questionnaire data did not allow us to determine the type of antibiotics given, duration of treatment, or what underlying diagnoses led to their administration. However, to try to answer the question to what extent infants who were given antibiotics had a respiratory diagnosis that could increase the risk of future wheezing, we have analyzed material from the western Sweden region between the years 2004 and 2007, which consists of 6228 infants from the 4 neonatal clinics in the area, who were registered in a Swedish national perinatal quality register (PNQ MedSciNet, Stockholm, Sweden).16,17 Data for 2003 were not available. In this register, we found that of the 6228 infants, 2147 had received antibiotics. Of these antibiotic courses, 86.6% had a duration of 3 days or more, and 70% had a duration of 4 days or more. Of the infants who received antibiotics, 108 (5%) had BPD, 396 (18.4%) had respiratory distress syndrome, and 176 (8.2%) had pneumonia. Thus, there were few infants with BPD, which probably is the most common diagnosis for which an obvious link with future wheezing can be seen.
If these figures can be extrapolated back to our material, it can be concluded that the majority of infants who received antibiotics did not have a diagnosis for which it would be likely that subsequent development of wheezing was linked to the neonatal diagnosis. Furthermore, it is probable that such an effect is diminished by the fact that gestational age is adjusted for in the model. In addition, we believe that the inclusion of short-term treatments (terminated after 24–48 hours), if anything, dilute our results, because short-term treatments with antibiotics probably have less effect on the gut flora.
The suggested pathway between antibiotics and a subsequent change in the gut flora and asthma originated in the proposed hygiene hypothesis18 behind the allergy epidemic of the 20th century. Even if the cause of the increase in allergic diseases is still unknown, the fact that exposure to farm animals, pets, and older siblings in early childhood reduces the risk of allergy development suggests that early exposure to microbes may be involved.19,20 In particular, infections by microbes via the oral route seem to be associated with a reduced risk of allergy, whereas airway infections are not.21 The timing of exposure is probably of great importance. The first weeks of life are probably the most important for changing the intestinal flora. In the Dutch KOALA: Child, Parent and Health: Lifestyle and Genetic Constitution (in Dutch) birth-cohort study, the intestinal flora at 1 month of age was influenced by antibiotic treatment22 with a pattern predisposing to both eczema and wheezing at 2 years of age.23 There is probably a concentration of children who subsequently develop allergic asthma in the group that fit the “wheezing treated with inhaled corticosteroids” category. However, it should be noted that the link demonstrated in our study is between neonatal antibiotics and wheezing treated with inhaled corticosteroids but not between early antibiotics and allergy. A future follow-up of the cohort should demonstrate whether there is also a link to allergic sensitization or to allergic asthma.
The fact that the use of a pacifier at 6 months of age emerged as an independent risk factor for the wheezing-ever category in the analysis cannot be fully explained. To our knowledge, only the Avon Longitudinal Study of Pregnancy and Childhood, which originated from Bristol, United Kingdom, has published similar results.24,25 We found no sign of a dose-response relationship in our results, which argues against a causal relationship. The result may be a random finding. Preterm birth was associated with an increased prevalence of wheezing. For several reasons, preterm birth may also be associated with an increased risk of receiving antibiotics. However, in both multiple logistic regression and stratified analysis for preterm birth, preterm birth, and neonatal antibiotics were independent risk factors for the “wheezing ever” and “wheezing disorder treated with inhaled corticosteroids” categories, which counters the assumption that the finding that neonatal antibiotic treatment was a risk factor for wheezing depends on the group of preterm infants.
We were surprised that we did not see a stronger effect on wheezing disorder caused by maternal smoking. Rylander26 reported a strong effect of maternal smoking on the prevalence of wheezing bronchitis (relative risks of the magnitude: 1.8–2.6), whereas the BAMSE birth cohort study reported ORs of ∼1.5 for the ever-wheezing and asthma groups.8 Similarly, the Avon Longitudinal Study of Pregnancy and Childhood reported ORs of ∼1.4 to 1.6 for wheezing due to maternal smoking during pregnancy.27 In our data, the effect of maternal smoking was visible among infants in the wheezing-ever group but not in the group of those with wheezing disorder treated with inhaled corticosteroids, in which there probably is a concentration of infants who will subsequently develop allergic asthma. Our results seem to be compatible with the findings in the National Asthma Campaign Manchester Asthma and Allergy Study, in which maternal smoking increased the risk of wheezing during the first year but was not associated with an increased risk of atopy.28
We have shown that treatment with antibiotics in the neonatal period is an independent risk factor for wheezing that was treated with inhaled corticosteroids at 12 months of age. The result indirectly supports the hypothesis that a change in intestinal flora can increase the risk of subsequent wheezing. This may be another argument for avoiding unnecessary use of antibiotics in the neonatal period but must, of course, be balanced against the indication for antibiotics. To finally resolve this question, additional research with attention to the specifics of neonatal antibiotic administration is needed.
- Address correspondence to Bernt Alm, MD, PhD, Department of Pediatrics, Göteborg University, Queen Silvia Children's Hospital, SE-416 85 Göteborg, Sweden. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Antibiotic exposure in the first year of life is correlated to childhood asthma. It is not clear whether this is a causal relationship or merely a reflection of asthmatic children receiving more antibiotics.
What This Study Adds
Antibiotics given during the neonatal period, before start of asthmatic symptoms, is an independent risk factor for wheezing treated with inhaled corticosteroids at 12 months. The result could indirectly support the hypothesis that a change in intestinal flora can increase the risk of subsequent wheezing.
- ↵Alm B, Möllborg P, Erdes L, et al. SIDS risk factors and factors associated with prone sleeping in Sweden. Arch Dis Child.2006;91 (11):915– 919
- ↵Norušis M. SPSS 13.0 Guide to Data Analysis. Upper Saddle River, NJ: Prentice Hall; 2005
- ↵Wright AL, Taussig LM, Ray CG, Harrison HR, Holberg CJ. The Tucson Children's Respiratory Study II: lower respiratory tract illness in the first year of life. Am J Epidemiol.1989;129 (6):1232– 1246
- ↵Pharmacological treatment in asthma: recommendations [in Swedish]. Information from the Swedish Medical Products Agency.2002;13 :4– 70
- ↵Cohet C, Cheng S, MacDonald C, et al. Infections, medication use, and the prevalence of symptoms of asthma, rhinitis, and eczema in childhood. J Epidemiol Community Health.2004;58 (10):852– 857
- ↵Illi S, von Mutius E, Lau S, et al. Early childhood infectious diseases and the development of asthma up to school age: a birth cohort study. BMJ.2001;322 (7283):390– 395
- ↵Persson B. National healthcare quality registries in Sweden. Stockholm, Sweden: Swedish Association of Local Authorities and Regions (SALAR); 2005. Report No.: ISBN 91-7164-096-7. Available at: http://brs.skl.se/brsbibl/kata_documents/doc36980_1.pdf. Accessed December 18, 2007
- ↵Strachan DP. Hay fever, hygiene, and household size. BMJ.1989;299 (6710):1259– 1260
- ↵Matricardi PM, Rosmini F, Riondino S, et al. Exposure to foodborne and orofecal microbes versus airborne viruses in relation to atopy and allergic asthma: epidemiological study. BMJ.2000;320 (7232):412– 417
- ↵Penders J, Thijs C, Vink C, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics.2006;118 (2):511– 521
- ↵Penders J, Thijs C, van den Brandt PA, et al. Gut microbiota composition and development of atopic manifestations in infancy: the KOALA Birth Cohort Study. Gut.2007;56 (5):661– 667
- ↵North K, Fleming P, Golding J. Pacifier use and morbidity in the first six months of life. Pediatrics.1999;103 (3). Available at: www.pediatrics.org/cgi/content/full/103/3/e34
- ↵North Stone K, Fleming P, Golding J. Socio-demographic associations with digit and pacifier sucking at 15 months of age and possible associations with infant infection: the ALSPAC Study Team—Avon Longitudinal Study of Pregnancy and Childhood. Early Hum Dev.2000;60 (2):137– 148
- ↵Lux AL, Henderson AJ, Pocock SJ. Wheeze associated with prenatal tobacco smoke exposure: a prospective, longitudinal study—ALSPAC Study Team. Arch Dis Child.2000;83 (4):307– 312
- Copyright © 2008 by the American Academy of Pediatrics