Perinatal Factors and the Risk of Asthma in Childhood A Population-based Register Study in Finland

Transcription

1 American Journal of Epidemiology ª The Author Published by the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please Vol. 168, No. 2 DOI: /aje/kwn105 Advance Access publication May 28, 2008 Original Contribution Perinatal Factors and the Risk of Asthma in Childhood A Population-based Register Study in Finland Johanna Metsälä 1,2, Annamari Kilkkinen 3, Minna Kaila 4,5, Heli Tapanainen 1, Timo Klaukka 6, Mika Gissler 7, and Suvi M. Virtanen 1,2,8 1 Department of Health Promotion and Chronic Disease Prevention, National Public Health Institute, Helsinki, Finland. 2 Tampere School of Public Health, University of Tampere, Tampere, Finland. 3 Department of Health and Functional Capacity, National Public Health Institute, Helsinki, Finland. 4 Finnish Office for Health Technology Assessment, National Research and Development Center for Welfare and Health, Tampere, Finland. 5 Pediatric Research Center, Tampere University Hospital, Tampere, Finland. 6 Social Insurance Institution, Helsinki, Finland. 7 Information Division, Alcohol and Drug Statistics and Reproduction Statistics, National Research and Development Center for Welfare and Health, Helsinki, Finland. 8 Research Unit, Tampere University Hospital, Tampere, Finland. Received for publication October 31, 2007; accepted for publication March 27, The aim of the study was to assess whether perinatal factors are associated with the risk of asthma in childhood in a register-based, nested case-control study in Finland. All children born between January 1, 1996, and April 30, 2004, who were entitled to a special reimbursement for antiasthmatic drugs (i.e., had diagnosed asthma by 2006 and had purchased inhaled corticosteroids or montelukast at least once), were identified (n ¼ 21,038). For each case, one matched control child was selected. The associations between perinatal factors, derived from the Finnish Medical Birth Register, and the risk of asthma were analyzed by conditional logistic regression. In the final multivariate model, maternal asthma, young age, smoking, previous miscarriages, and a high number of previous deliveries, as well as cesarean section, low gestational age, and low ponderal index, were associated with an increased risk of asthma in children diagnosed before the age of 3 years. Among children diagnosed at the age of 3 years or later, maternal asthma, low gestational age, and low ponderal index were associated with an increased risk, and a high number of previous deliveries was associated with a decreased risk of asthma. In conclusion, perinatal factors play a role in the development of asthma in childhood, but the etiology may differ in early and late-onset asthma. asthma; cesarean section; child; cohort studies; gestational age; registries; risk factors Abbreviations: ATC, Anatomical Therapeutic Chemical; CI, confidence interval; OR, odds ratio; SGA, small for gestational age. The prevalence of asthma and allergic diseases in childhood has substantially increased during the last decades in several Westernized countries (1). This rise indicates that, besides the strong hereditary component, environmental factors play an important role in the development of allergic immune responses. Increasing evidence suggests that these environmental factors operate in early life or even in utero (2, 3). Perinatal factors have been hypothesized to be associated with the risk of asthma and allergic disorders by several possible mechanisms (2 4). These mechanisms focus on the development of the immune system, especially changes Correspondence to Johanna Metsälä, Department of Health Promotion and Chronic Disease Prevention, National Public Health Institute, Mannerheimintie 166, FIN Helsinki, Finland ( johanna.metsala@ktl.fi). 170

2 Perinatal Factors and Asthma 171 in the production of type 1 and type 2 T-helper (Th-1, Th-2) cell-type cytokines. Changes in cytokine profiles in cord blood as a response to neonatal exposures may be related to the development of asthma and allergic diseases later in childhood (3, 4). An appropriate environmental stimulus, for example, microbial exposure, is crucial in the development of a healthy type 1 T-helper cell dominant immune response after birth (5). Diminished microbial exposure and factors influencing microbial exposure after birth may affect the development of an allergic immune response. In epidemiologic studies, several perinatal factors have been suggested to be associated with the risk of asthma, but the evidence is not conclusive. Maternal smoking (6 9) and young age (6, 10, 11) have consistently been reported to be associated with an increased risk of wheeze or asthma in the offspring. Several studies have also shown that complications during pregnancy and delivery are associated with increased risk of asthma (12 14). Furthermore, both cesarean section (15) and prematurity (16) have been associated with increased risk of asthma in recent meta-analyses. Results regarding birth weight are contradictory (7, 9, 12, 17 21). To clarify the role of maternal background and perinatal factors in the development of asthma in childhood, we conducted a register-based, nested case-control study including all children born between January 1, 1996, and April 30, 2004, in Finland and diagnosed with asthma by MATERIALS AND METHODS Study population and data sources The study population was derived from three nationwide registers. All children who were born in Finland between January 1, 1996, and April 30, 2004, and had received a special reimbursement for antiasthmatic drug costs by the end of 2005 were identified (n ¼ 22,548). In Finland, a special reimbursement for certain prescribed drugs including antiasthmatic drugs is provided and recorded in the Special Reimbursement Register kept by the Social Insurance Institution of Finland. The reimbursement is divided into three compensation categories (basic; lower special, 75 percent; and higher special, 100 percent) with the lower special reimbursement category covering approximately 10 diseases, including asthma (22). To be eligible for this special reimbursement, a patient must have a certificate from a physician, in the case of a child mostly from a pediatrician, stating that the diagnostic criteria of asthma are fulfilled. For each case, one nonasthmatic control matched to the case by age (28 days), gender, and the hospital district where the child was born was randomly selected from the Population Register of the Social Insurance Institution. Controls were allowed to be selected only once. Information on the use of antiasthmatic drugs was obtained from the Drug Prescription Register of the Social Insurance Institution, and it includes information on drug class (Anatomical Therapeutic Chemical (ATC) Classification System), date of purchasing, and personal identity code. In Finland, all drugs prescribed by physicians and reimbursed by the National Sickness Insurance Scheme are registered in the Drug Prescription Register, established in 1993 and being almost complete in In the present study, those children who were entitled to a special reimbursement for antiasthmatic drugs and purchased them at least once after the diagnosis were considered as asthma cases (n ¼ 21,038). Selection of the antiasthmatic drugs was based on the Finnish Current Care Guidelines (23) and recommendations by the Social Insurance Institution. These drugs included inhaled corticosteroids (ATC code R03BA), a fixed combination containing inhaled corticosteroids and salmeterol (code R03AK06), a fixed combination containing inhaled corticosteroids and formoterol (code R03AK07), and montelukast (code R03DC03). Longacting beta-agonists were included in the case definition because, despite the fact that they are neither licensed nor recommended, especially for children under the age of 4 years, they are increasingly used. In each of these drug groups, the concomitant use of a short-acting beta-agonist was allowed. In Finland, cromones (code R03BC) and teophyllines (code R03DA04) are used only as add-on therapy and, therefore, they were not included as antiasthmatic drugs. The total number of different drug purchases in was calculated separately for two groups of cases: those diagnosed before the age of 3 years and at the age of 3 years or later. A total of 415 control children had purchased inhaled corticosteroids (ATC code R03BA), a fixed combination containing inhaled corticosteroids and salmeterol (code R03AK06), a fixed combination containing inhaled corticosteroids and formoterol (code R03AK07), cromoglycate (code R03BC01), nedocromil (code R03BC03), montelukast (code R03DC03), xanthines (code R03DA) or xanthines, and adrenergics (code R03DB) before the asthma diagnosis of the case. They were excluded from the present study as we did not have information on whether they were just transient users of antiasthmatic drugs or asthma patients without a physician s confirmed diagnosis. Finally, 20,623 case-control pairs were included in the present study. The data were completed by linkage with the Finnish Medical Birth Register, kept by the National Research and Development Center for Welfare and Health. The register includes information on maternal background and interventions during pregnancy and delivery and on the newborn s medical interventions and outcome up to the age of 7 days. Data from the Medical Birth Register were missing for 457 subjects (158 cases and 299 controls) included in the study population. Most of these children were not born in Finland. The variables used in this study were maternal previous miscarriages (no/yes); number of previous deliveries (0, 1, 2, 3, 4, 5); smoking during pregnancy (no/yes, including those who quit smoking during the first trimester); complications during pregnancy (none, maternal high blood pressure during pregnancy that had to be treated in the hospital, placental complications including placenta previa and placenta avulsion, abnormal presentation including breech and other abnormal presentation, and fetal asphyxia); and age at delivery (<25, 25 29, 30 34, and 35 years), as well as mode of delivery (normal vaginal delivery, assisted vaginal delivery, planned cesarean section, and emergency cesarean section), Apgar score at 1 minute (0 6, 7 8, 9 10), prematurity (gestational age, <37 weeks), low birth weight (birth weight, <2,500 g), and gestational age-adjusted birth weight, defined according to Finnish population-based

3 172 Metsälä et al. growth curves: small for gestational age (SGA), 2 standard deviations or less of the mean birth weight for gestational age; appropriate for gestational age, within the mean 2 standard deviations; large for gestational age, 2 standard deviations or more of the mean (24). In addition, children were stratified according to gestational age and ponderal index (birth weight/birth length 3 ) at birth on the basis of the quartiles of the controls. Those mothers who were entitled to a special reimbursement for antiasthmatic drugs were considered as having asthma. No information on the father was available in the Medical Birth Register. Linkages between the registers were based on unique personal identity codes assigned to all Finnish citizens shortly after birth (25). The study was approved by the national data protection authority, the institutions keeping the registers, and the Institutional Review Board of the National Public Health Institute. Statistical analysis For examination of the associations between maternal background factors and perinatal factors among controls, the Pearson chi-squared test for two categorical variables, the Mann-Whitney U test for the combination of a continuous and a dichotomous variable, and the Spearman rank correlation for two continuous variables were used. The associations between perinatal factors and the risk of asthma were analyzed by conditional logistic regression, and the results are displayed as odds ratios with 95 percent confidence intervals. A model including only one explanatory variable at a time was fitted first; the ponderal index, the Apgar score at 1 minute, the mode of delivery, and complications were, however, adjusted for gestational age in univariate analysis. Interactions among maternal age, smoking, asthma, and perinatal factors were investigated. The multivariate model included possible interaction terms and all the variables used in the univariate analysis, except for prematurity, low birth weight, gestational age-adjusted birth weight, the Apgar score, and pregnancy complications. In the final multivariate model, the nonsignificant variables were eliminated one by one. Variables were entered into the multivariate models in the categorical mode. Statistical significance was set at the 5 percent level, and two-sided p values were used. All analyses were done separately for children diagnosed with asthma before the age of 3 years (early onset) and at the age of 3 years or later (late onset), because the diagnosis of asthma is more challenging in the younger age group, and the diagnostic criteria of asthma in children younger than 3 years differ from those of older children (23, 26). Further analyses were conducted by excluding premature children and cases with only one or two purchases of antiasthmatic drugs during the 3 years after diagnosis. All analyses were performed using STATA, version 9.1, software (StataCorp LP, College Station, Texas). RESULTS The mean age of asthma cases at diagnosis was 2.2 years, ranging from 0 months to 9 years; two thirds (n ¼ 13,679) of TABLE 1. Total number of purchases of antiasthmatic drugs after the diagnosis during in cases born in Finland between January 1, 1996, and April 30, 2004 (n ¼ 20,623), by different drug groups Drug group Cases diagnosed at <3 years (n ¼ 13,679) Cases diagnosed at 3 years (n ¼ 6,944) No. % No. % Inhaled corticosteroids 137, , Short-acting beta-agonists 69, , Inhaled corticosteroids in combination with long-acting beta-agonists 26, , Montelukast 3, ,374 3 Others* 1,361 <1 205 <1 Total 238, , * Including cromones and xanthines. the cases were diagnosed before the age of 3 years. Asthma was more common in boys (64 percent of cases) than in girls (36 percent) (p < 0.001). Inhaled corticosteroids and shortacting beta-agonists were the most often purchased antiasthmatic drugs (table 1). Short-acting beta-agonists were always used concomitantly with other antiasthmatic drugs; the majority of cases purchasing corticosteroids had also purchased short-acting beta-agonists (84 percent of cases diagnosed before the age of 3 years and 81 percent of cases diagnosed at 3 years of age or later). Maternal background factors were associated with several perinatal factors. Among controls, cesarean section was prevalent among older mothers, mothers with no previous deliveries, and low gestational age (table 2). Maternal smoking during pregnancy was related to young age and no previous deliveries. Previous miscarriages were prevalent among women with older age and a low number of previous deliveries. The only clinically relevant correlation observed was between maternal age and number of previous deliveries (r ¼ 0.4, p < 0.001). Maternal asthma was the strongest predictor of asthma in children diagnosed before the age of 3 years (odds ratio (OR) ¼ 3.41, 95 percent confidence interval (CI): 3.08, 3.77) (table 3). Young maternal age, smoking during pregnancy, previous miscarriages, and high number of previous deliveries were also associated with an increased risk of asthma. We observed an interaction between maternal age and number of previous deliveries (figure 1); in the analysis stratified by number of previous deliveries of the mother, the risk of asthma decreased with increasing maternal age among children with older siblings, while no association was observed among mothers without previous deliveries. Emergency cesarean section and low gestational age were also associated with an increased risk of asthma (table 3). Compared with the risk for full-term and normal-weight children, a substantially increased risk of asthma was observed for premature children (OR ¼ 2.77, 95 percent CI: 2.53, 3.03), for children with low birth weight (OR ¼ 1.40, 95 percent CI: 1.20, 1.60), and for children who were SGA (OR ¼ 2.00, 95 percent CI: 1.73, 2.31), respectively. A low

4 Perinatal Factors and Asthma 173 TABLE 2. Cesarean section, maternal smoking during pregnancy, and previous miscarriages related to maternal age, previous deliveries, and gestational age among controls born in Finland between January 1, 1996, and April 30, 2004 (n ¼ 20,623) Section,* yes (n ¼ 3,371) Apgar score at 1 minute (for scores 0 6 vs. 9 10, OR ¼ 1.32, 95 percent CI: 1.17, 1.49) and placental complications (for none vs. placental complications, OR ¼ 1.61, 95 percent CI: 1.10, 2.36) were also associated with an increased risk of asthma. In the multivariate model, maternal age, asthma, smoking during pregnancy, and previous miscarriages as well as previous deliveries, both planned and emergency cesarean sections, gestational age, and ponderal index were associated with the risk of asthma in children diagnosed before the age of 3 years (table 3). All variables remained significant also in the final multivariate model. Maternal asthma was also the strongest predictor of asthma in children diagnosed at the age of 3 years or later (OR ¼ 3.70, 95 percent CI: 3.22, 4.33) (table 4). Young maternal age, a low number of previous deliveries, and short gestational age, as well as emergency cesarean section, were associated with an increased risk of asthma. Ponderal index was at borderline significance. Prematurity (OR ¼ 1.34, 95 percent CI: 1.16, 1.54) and low Apgar score at 1 minute (OR ¼ 1.19, 95 percent CI: 1.00, 1.42) were associated with an increased risk of asthma, while no association was observed for maternal smoking, previous miscarriages, planned cesarean section, low birth weight, or SGA status. In the final multivariate model, maternal asthma, number of previous deliveries, gestational age, and ponderal index were associated with the risk of asthma. Maternal smoking,y yes (n ¼ 2,873) The prevalence of prematurity was higher among cases than controls (11 percent vs. 5 percent). The majority of premature cases were diagnosed before the age of 3 years (80 percent). The exclusion of premature children from further analysis did not substantially change the results (data not shown). The only exception was ponderal index, which dropped out from the final multivariate model in both age groups. Of those cases followed up for at least 3 years after diagnosis (i.e., cases born during , n ¼ 13,505), 1,716 had only one or two purchases of inhaled corticosteroids or montelukast during the 3 years after the diagnosis. The majority of these children were diagnosed before the age of 3 years (79 percent). Exclusion of these children from the analyses did not substantially change the results (data not shown). The number of those cases diagnosed before the age of 7 months was 1,486. Exclusion of these children from further analyses did not substantially change the results (data not shown). DISCUSSION Previous miscarriages, yes (n ¼ 4,094) % p valuez % p valuez % p valuez Maternal age (years) < < < < Maternal previous deliveries None 50.9 < < < Missing data Gestational age (days) <273 (quartile 1) 39.5 < < < (quartile 2) (quartile 3) (quartile 4) * Including planned (n ¼ 1,606) and emergency (n ¼ 1,765) sections. y Including those who have quit smoking during pregnancy (n ¼ 339). z p values from the Mann-Whitney U test. In the present large, population-based study, maternal asthma, short gestational age, and low ponderal index were associated with increased risk of childhood asthma. In addition, young maternal age, smoking, a high number of

5 174 Metsälä et al. TABLE 3. Maternal background and perinatal factors related to asthma in childhood in children born in Finland between January 1, 1996, and April 30, 2004, and diagnosed before the age of 3 years Factors Cases (n ¼ 13,679) Controls* (n ¼ 13,679) No. % No. % Odds ratio Univariate model 95% confidence interval Odds ratio Multivariate modely 95% confidence interval Maternal age (years) <25 2, , , , , , , , , , , , , , 0.58 Maternal asthma No 11, , Yes 1, , , 3.80 Maternal smoking during pregnancy No 10, , Yesz 2, , , , 1.42 Missing data Maternal previous miscarriages 0 10, , , , , , 1.25 Missing data Maternal previous deliveries 0 4, , , , , , , , , , , , , , , , 4.31 Missing data Mode of delivery Normal vaginal delivery 9, , Assisted vaginal delivery , , 1.34 Planned cesarean section 1, , , , 1.31 Emergency cesarean section 1, , , , 1.58 Missing data Gestational age (days) <273 (quartile 1) 4, , (quartile 2) 3, , , , (quartile 3) 2, , , , (quartile 4) 2, , , , 0.66 Missing data Ponderal index (g/cm 3 ) < (quartile 1) 3, , (quartile 2) 3, , , , (quartile 3) 3, , , , (quartile 4) 3, , , , 0.95 Missing data * Controls were matched to cases by gender, date of birth (628 days), and the hospital district. y Odds ratios were adjusted for all the other variables in the table and for the interaction term, maternal age 3 number of previous deliveries. z Including those who have quit smoking during pregnancy (n ¼ 261 in cases, n ¼ 223 in controls). Odds ratios were adjusted for gestational age in the univariate model.

6 Perinatal Factors and Asthma 175 FIGURE 1. Maternal age at delivery and the risk of childhood asthma by number of maternal previous deliveries (0, 1, 2, 3) in children diagnosed before the age of 3 years, Finland, Data points are crude odds ratios in maternal age groups <25, 25 29, 30 34, and 35 years, with points placed at the mean maternal age within each age group. Bars, 95% confidence intervals. previous deliveries, and cesarean section were associated with an increased risk of asthma in children diagnosed before the age of 3 years. In children diagnosed at the age of 3 years or later, a high number of previous deliveries was associated with decreased risk of asthma. There is general consensus on the importance of a genetic component in the development of asthma (27). In the present study, maternal asthma was the strongest predictor for asthma. This is consistent with previous epidemiologic studies reporting that maternal asthma is a risk factor for both early and late-onset wheeze (11, 28 30). It has been suggested that maternal asthma may modify the associations between early environmental exposures, such as day-care attendance (31), and risk of asthma. However, we did not observe any effect modification by maternal asthma. In the present study, the number of maternal previous deliveries was associated with an increased risk of asthma in children diagnosed before the age of 3 years but with a decreased risk in children diagnosed at the age of 3 years or later. Several studies have reported a decreased risk of hay fever among children with older siblings or attending day care with other children, while results for asthma and wheeze have been more inconsistent (32). Recent large studies have demonstrated that the presence of older siblings is associated with a higher risk during the first years of life (6, 7, 32) but a lower risk of late-onset asthma or wheeze (6, 12, 32). It is possible that respiratory infections from older siblings increase the risk of early transient wheezing symptoms. Development of wheezing or asthma after the age of 3 years may involve atopic mechanisms, when the explanation focuses more on immune maturation or in utero programming (32). Both planned and emergency cesarean sections were associated with an increased risk of asthma in children diagnosed before the age of 3 years. Previous findings regarding cesarean section and the risk of asthma are contradictory; while some studies have showed that cesarean section is a risk factor for subsequent asthma (33 37), others have failed to find an association (28, 38, 39), and only a few studies have reported results for planned and emergency sections separately (36, 37, 39). It has been hypothesized that an increased risk of asthma in children born by cesarean section could be due to different microbial exposures related to cesarean section and vaginal delivery. The gut microflora in children born by cesarean section differ from those of children born vaginally (40). This might affect the development of allergic diseases, as intestinal microflora play an important role in the development of the immune system (41, 42). However, our results do not fully support this microflora theory, as we did not observe an association between cesarean section and asthma diagnosed at the age of 3 years or later. Another possible mechanism by which cesarean section increases the risk of asthma is that asthma may be predisposed by respiratory problems identified in some newborns born by section (43, 44). In the present study, prematurity and short gestational age were associated with an increased risk of asthma. These findings are consistent with some (11, 16, 45), although not all (20, 21), previous studies. The prematurity-asthma association has been explained by the fact that prematurity causes reduced lung growth and reduced airway caliber, which may increase wheezing symptoms during respiratory infections and in turn increase asthma diagnoses (16). In addition, prematurity may predispose the child to respiratory tract infections, which would further strengthen these associations. However, caution must be taken when interpreting the results, because distinguishing asthma from the chronic lung disease of infancy caused by prematurity is most challenging. The observation that most premature asthma cases were diagnosed during their first years of life probably reflects these same difficulties in clinical practice. However, exclusion of premature children did not substantially change the results. In children diagnosed before the age of 3 years, low birth weight was associated with increased risk of asthma. Previous studies on birth weight and the risk of asthma are even more contradictory with both inverse (12) and direct (17, 18), as well as null (7, 19 21), associations being reported. These inconsistent results may in part be due to differences in methodology among the studies, but they may also indicate that the association is only weak or is caused by bias. The personal identity code system used in Finland gave us the opportunity to link several national registers shown to have nearly complete population coverage and high validity and accuracy (46), thus providing comprehensive information from a large population-based sample. By using the registers, we were also able to avoid recall bias related to questionnaire-based data. Furthermore, the definition of asthma cases can be considered rather reliable, because to have a special reimbursement for antiasthmatic drug costs, it is required that asthma be diagnosed by a physician or a pediatrician. The main limitation of our study is that we could not identify different wheezing phenotypes, especially in the youngest children. Early onset transient wheezing is

7 176 Metsälä et al. TABLE 4. Maternal background and perinatal factors related to asthma in childhood in children born in Finland between January 1, 1996, and April 30, 2004, and diagnosed at the age of 3 years or later Factors Cases (n ¼ 6,944) Controls* (n ¼ 6,944) No. % No. % Odds ratio Univariate model 95% confidence interval Odds ratio Multivariate modely 95% confidence interval Maternal age (years) <25 1, , , , , , , , , , , , , , 1.08 Maternal asthma No 6, , Yes , , 4.48 Maternal smoking during pregnancy No 5, , Yesz 1, , , 1.12 Missing data Maternal previous miscarriages No 5, , , , , , 1.18 Missing data Maternal previous deliveries 0 3, , , , , , , , , , , , , , , 0.49 Missing data Mode of delivery Normal vaginal delivery 5, , Assisted vaginal delivery , , 1.13 Planned cesarean section , , 1.22 Emergency cesarean section , , 1.17 Missing data Gestational age (days) <273 (quartile 1) 1, , (quartile 2) 1, , , , (quartile 3) 1, , , , (quartile 4) 1, , , , 0.93 Missing data Ponderal index (g/cm 3 ) < (quartile 1) 1, , (quartile 2) 1, , , , (quartile 3) 1, , , , (quartile 4) 1, , , , 1.04 Missing data * Controls were matched to cases by gender, date of birth (628 days), and the hospital district. y Odds ratios were adjusted for all the other variables in the table. z Including those who have quit smoking during pregnancy (n ¼ 123 in cases, n ¼ 116 in controls). Odds ratios were adjusted for gestational age in the univariate model.

8 Perinatal Factors and Asthma 177 characteristic of children under 3 years of age (26) and, despite the definition of asthma cases based on physiciandiagnosed asthma and information on purchases of certain more potent asthma drugs, we cannot rule out the possibility that some of our asthma cases had only transient wheezing symptoms rather than persistent wheeze or true asthma. However, children with early onset persistent wheeze have been shown more often to have physician-diagnosed asthma in early life than children with early onset transient wheeze (47). Restricting the analysis only to those with potentially more true asthma, that is, several purchases of antiasthmatic drugs, did not substantially change the results. We attempted to investigate the effect of age at onset of asthma on the relation of perinatal factors and risk of asthma by doing separate analyses for children diagnosed before 3 years of age and 3 years of age or later. The diagnosis age-dependent effects in our study may be due to heterogeneity of respiratory symptoms in children under the age of 3 years and difficulty in making a diagnosis of asthma in children at that age compared with the situation in older children. In addition, limitations of relying on the pharmacy records include a rough estimation of drug use; that is, because of the absence of information on actual usage of the dispensed drugs, subjects who failed to take their drugs have been incorrectly classified as users. However, this only dilutes the association and biases the relative risks toward unity. We conclude that perinatal factors, especially maternal asthma, number of previous deliveries, and gestational age, play a role in the development of asthma in childhood. Although asthma diagnoses before the age of 3 years and at 3 years or later have some common early environmental exposures, our results indicate some differences in the etiologies of asthma diagnosed at these times. ACKNOWLEDGMENTS Grants were received from the Yrjö Jahnsson Foundation (grant no. 5663) and the Juho Vainio Foundation. The authors thank Dr. Raili Salmelin for her contribution to this study. Conflict of interest: none declared. REFERENCES 1. Eder W, Ege MJ, von Mutius E. The asthma epidemic. N Engl J Med 2006;355: Bernsen RM, van der Wouden JC, Nagelkerke NJ, et al. Early life circumstances and atopic disorders in childhood. Clin Exp Allergy 2006;36: Chung EK, Miller RL, Wilson MT, et al. Antenatal risk factors, cytokines and the development of atopic disease in early childhood. Arch Dis Child Fetal Neonatal Ed 2007;92: Prescott SL. Maternal allergen exposure as a risk factor for childhood asthma. Curr Allergy Asthma Rep 2006;6: Prescott SL, Macaubas C, Smallacombe T, et al. Development of allergen-specific T-cell memory in atopic and normal children. Lancet 1999;353: McKeever TM, Lewis SA, Smith C, et al. Siblings, multiple births, and the incidence of allergic disease: a birth cohort study using the West Midlands general practice research database. Thorax 2001;56: Yuan W, Fonager K, Olsen J, et al. Prenatal factors and use of anti-asthma medications in early childhood: a populationbased Danish birth cohort study. Eur J Epidemiol 2003;18: Haberg SE, Stigum H, Nystad W, et al. Effects of pre- and postnatal exposure to parental smoking on early childhood respiratory health. Am J Epidemiol 2007;166: Jaakkola JJK, Gissler M. Maternal smoking in pregnancy, fetal development, and childhood asthma. Am J Public Health 2004; 94: Martinez FD, Wright AL, Holberg CJ, et al. Maternal age as a risk factor for wheezing lower respiratory illnesses in the first year of life. Am J Epidemiol 1992;136: Sherriff A, Peters TJ, Henderson J, et al. Risk factor associations with wheezing patterns in children followed longitudinally from birth to 3 1/2 years. Int J Epidemiol 2001;30: Dik N, Tate RB, Manfreda J, et al. Risk of physician-diagnosed asthma in the first 6 years of life. Chest 2004;126: McKeever TM, Lewis SA, Smith C, et al. The importance of prenatal exposures on the development of allergic disease a birth cohort study using the West Midlands General Practice Database. Am J Respir Crit Care Med 2002;166: Nafstad P, Samuelsen SO, Irgens LM, et al. Pregnancy complications and the risk of asthma among Norwegians born between 1967 and Eur J Epidemiol 2003;18: Thavagnanam S, Fleming J, Bromley A, et al. A meta-analysis of the association between Caesarean section and childhood asthma. Clin Exp Allergy 2008;38: Jaakkola JJ, Ahmed P, Ieromnimon A, et al. Preterm delivery and asthma: a systematic review and meta-analysis. J Allergy Clin Immunol 2006;118: Flaherman V, Rutherford G. A meta-analysis of the effect of high weight on asthma. Arch Dis Child 2006;91: Sin DD, Spier S, Svenson LW, et al. The relationship between birth weight and childhood asthma a population-based cohort study. Arch Pediatr Adolesc Med 2004;158: Caudri D, Wijga A, Gehring U, et al. Respiratory symptoms in the first 7 years of life and birth weight at term the PIAMA birth cohort. Am J Respir Crit Care Med 2007;175: Xu BZ, Pekkanen J, Jarvelin MR. Obstetric complications and asthma in childhood. J Asthma 2000;37: Yuan W, Basso O, Sorensen HT, et al. Fetal growth and hospitalization with asthma during early childhood: a follow-up study in Denmark. Int J Epidemiol 2002;31: Martikainen J, Rajaniemi S. Drug reimbursement systems in EU member states, Iceland and Norway. Helsinki, Finland: The Social Insurance Institution, (Social security and health reports no. 54). ( nsf/abid/ pn/$file/drug_reimbursement. pdf?openelement). 23. The Finnish Current Care Guidelines. Asthma. (In Finnish). Helsinki, Finland: Duodecim Medical Publications, Ltd, ( 6&p_url¼ hoi06030). (Accessed March 27, 2008). 24. Pihkala J, Hakala T, Voutilainen P, et al. Characteristics of recent fetal growth curves in Finland. (In Finnish). Duodecim 1989;105: Personal identity code. Helsinki, Finland: Population Register Center, ( home.nsf/pages/951e0bdff6e9fca7c fbbe1). (Accessed March 27, 2008).

9 178 Metsälä et al. 26. Martinez FD, Wright AL, Taussig LM, et al. Asthma and wheezing in the first 6 years of life. N Engl J Med 1995;332: Blumenthal MN. The role of genetics in the development of asthma and atopy. Curr Opin Allergy Clin Immunol 2005;5: Rusconi F, Galassi C, Forastiere F, et al. Maternal complications and procedures in pregnancy and at birth and wheezing phenotypes in children. Am J Respir Crit Care Med 2007;175: De Sario M, Di Domenicantonio R, Corbo G, et al. Characteristics of early transient, persistent, and late onset wheezers at 9 to 11 years of age. J Asthma 2006;43: London SJ. Family history and the risk of early-onset persistent, early-onset transient, and late-onset asthma. Epidemiology 2001;12: Celedon JC, Wright RJ, Litonjua AA, et al. Day care attendance in early life, maternal history of asthma, and asthma at the age of 6 years. Am J Respir Crit Care Med 2003;167: Karmaus W, Botezan C. Does a higher number of siblings protect against the development of allergy and asthma? A review. J Epidemiol Community Health 2002;56: Kero J, Gissler M, Gronlund MM, et al. Mode of delivery and asthma is there a connection? Pediatr Res 2002;52: Renz-Polster H, David MR, Buist AS, et al. Caesarean section delivery and the risk of allergic disorders in childhood. Clin Exp Allergy 2005;35: Bager P, Melbye M, Rostgaard K, et al. Mode of delivery and risk of allergic rhinitis and asthma. J Allergy Clin Immunol 2003;111: Hakansson S, Kallen K. Caesarean section increases the risk of hospital care in childhood for asthma and gastroenteritis. Clin Exp Allergy 2003;33: Smith GC. Neonatal respiratory morbidity at term and the risk of childhood asthma. Arch Dis Child 2004;89: McKeever TM, Lewis SA, Smith C, et al. Mode of delivery and risk of developing allergic disease. J Allergy Clin Immunol 2002;109: Werner A, Ramlau-Hansen CH, Jeppesen SK, et al. Caesarean delivery and risk of developing asthma in the offspring. Acta Paediatr 2007;96: Gronlund MM, Lehtonen OP, Eerola E, et al. Fecal microflora in healthy infants born by different methods of delivery: permanent changes in intestinal flora after cesarean delivery. J Pediatr Gastroenterol Nutr 1999;28: Bjorksten B. Effects of intestinal microflora and the environment on the development of asthma and allergy. Springer Semin Immunopathol 2004;25: Kalliomaki M, Isolauri E. Role of intestinal flora in the development of allergy. Curr Opin Allergy Clin Immunol 2003; 3: Levine EM, Ghai V, Barton JJ, et al. Mode of delivery and risk of respiratory diseases in newborns. Obstet Gynecol 2001;97: Schaubel D, Johansen H, Dutta M, et al. Neonatal characteristics as risk factors for preschool asthma. J Asthma 1996;33: Gessner BD, Chimonas MAR. Asthma is associated with preterm birth but not with small for gestational age status among a population-based cohort of Medicaid-enrolled children < 10 years of age. Thorax 2007;62: Gissler M, Teperi J, Hemminki E, et al. Data quality after restructuring a national medical registry. Scand J Soc Med 1995;23: Kurukulaaratchy RJ. Characterization of wheezing phenotypes in the first 10 years of life. Clin Exp Allergy 2003;33:573 8.