Wheeze and asthma prevalence and related health‐service use in white and south Asian pre‐schoolchildren in the United Kingdom

Background Epidemiological data for south Asian children in the United Kingdom are contradictory, showing a lower prevalence of wheeze, but a higher rate of medical consultations and admissions for asthma compared with white children. These studies have not distinguished different asthma phenotypes or controlled for varying environmental exposures.


Introduction
Epidemiological data about asthma in minority ethnic groups are conflicting [1]. A recent systematic review on ethnic variations in asthma frequency, morbidity and health-service use in the United Kingdom concluded with the paradox that the prevalences of wheeze and diagnosed asthma were lower in south Asian than white children aged 5 years and older, but that conversely medical consultations and hospital admissions were more common [2]. The review summarized only crude prevalence rates and did not consider the effects of treatment or potential heterogeneity between different south Asian ethnic groups. In addition, there is more than one phenotype of childhood asthma [3][4][5][6], and there might be ethnic differences in their relative proportions. Episodic viral wheeze is a common transient non-atopic disorder in young children [3,5], whereas multiple-trigger wheeze, a phenotype that shares features with chronic atopic asthma, is more common in older children [6].
The lower observed prevalence [2] contrasts with the hypothesis that migrants from tropical to temperate environments might be at a higher genetic risk of allergic disease in general and atopic asthma in particular [7]. South Asians from the Indian subcontinent and East Africa together form the largest minority ethnic group in the United Kingdom [8]. We have shown that the prevalence of wheeze and asthma in young south Asian women is strongly associated with young age at immigration, providing further evidence for the importance of early life events on the development of asthma [9]. We now report on the offspring of these women, for whom we had prospectively collected data on respiratory symptoms and environmental exposures from the first years of life. The objectives of this study were to compare the prevalence of and morbidity due to viral wheeze and multiple wheeze in south Asian and white pre-schoolchildren, and to examine the influence of environmental and socioeconomic factors and the utilization of clinical care.

Setting and study population
We used the Leicestershire Health Authority Child Health Database, which includes the birth notification with mother's self-reported ethnic origin, religion, country of birth and language, to select an age-stratified random sample of white and south Asian (mother self-identified as Indian, Pakistani, Bangladeshi or other) children aged 1-4 years with complete birth records [9][10][11][12]. We oversampled south Asian children and under 2 years to increase precision.

Questionnaires
We mailed questionnaires, printed in English, to the children's parents and asked about the ethnic origin of the child and household language using a tick list. Current wheeze was assessed with a standard validated question [13]. We defined 'viral wheeze' as attacks triggered exclusively by viral infections and 'multiple wheeze' if children also wheezed with other triggers [3,10]. There were questions on the frequency, severity and triggers of wheeze, asthma treatment and healthcare utilization, environmental exposures such as air pollutants (tobacco smoke, cooking and heating devices), breastfeeding, pets, nursery care, number of siblings and other household members, and parental smoking and history of atopy. We measured socio-economic conditions at individual (duration of parental education, number of people per room and single parenthood) and area levels (Townsend score, 1991 census data). Four families accepted the offer of a phone interview in another language. The Leicestershire Health Authority Research Ethics Committee approved the study.

Statistical analysis
Analyses were performed with Stata, version 8.2 (Stata Corporation, Austin, TX, USA). There were o2% missing answers for most questions on symptoms. Missing and 'don't know' answers were coded as 'no' as sensitivity analyses showed that this did not influence the magnitude or direction of the findings.
We found evidence of heterogeneity between the effects of ethnicity and age on the prevalence of asthma phenotypes, both in unadjusted and in fully adjusted models (likelihood ratio tests for interaction terms P o 0.01). We therefore stratified our analyses by age (under 2 vs. 2 years and older) and used multinomial logistic regression to compare the two phenotypes (viral and multiple wheeze) with non-wheezing children. All variables that were associated with the prevalence of wheeze in univariable analysis (P o 0.05), plus age, sex and ethnicity, were entered into multivariable models. For consistency of presentation, all factors associated with viral or multiple wheeze in either of the two age groups were included in both models. We also investigated whether the association between different exposures and wheeze phenotypes differed by ethnic group, by performing stratified analyses and interaction tests. The results are presented as odds ratios (ORs) with 95% confidence intervals (95% CIs). The following sensitivity analyses were performed for all final models: including children with inconsistently reported ethnicity; not recoding missing data; including only the more homogeneous Gujarati-speaking subgroup; including only those who reported no problem understanding the questionnaire [12]; including only severe wheezers; and weighting analyses for oversampling of south Asians. All gave results very similar to the main analysis.

Results
We invited 6100 white and 2600 south Asian families to participate. After discounting 200 invalid addresses, 84% (4986/5954) of the families where the mother was defined as white and 72% (1825/2546) where the mother was defined as south Asian replied (this equates to 16% of white and 29% of south Asian children aged 1-4 years born and living in Leicestershire at the time). The mother's ethnicity on the Child Health Database corresponded with that of the child in the questionnaire in 94% of the south Asians (1714) and in 88% of the whites (4366). Only these 6080 children were included for further analysis. All south Asian children were born in the United Kingdom. Of their mothers, 25% (407) were born in the United Kingdom, 28% (445) in Africa and 47% (756) in the Indian subcontinent (611 from India, 68 from Bangladesh and 77 from Pakistan).
White children were more likely to be exposed to environmental factors such as maternal tobacco smoke, short duration of breastfeeding and pets (Table 1). A family history of atopic disease was also more common in whites. Low birthweight was more common in south Asians. More south Asian families lived in overcrowded households in deprived inner-city areas, but there were more white children in single-parent households.
Overall, more white (25.9%) than south Asian (21.6%) children were reported to have wheezed in the previous 12 months (P = 0.003, Table 2). The prevalence of current wheeze in south Asian children did not vary by language, religion or maternal country of birth. The pattern of wheeze, however, varied with age: both viral wheeze and multiple wheeze were less frequent in 1-year-old south Asian than white children, but there was no difference in the reported crude prevalence of either phenotype in 2-4-year-olds ( Fig. 1a and Table 2). There were no consistent ethnic differences in the reported severity of wheeze: parents of white children reported more attacks in the past year and those of south Asian children more activity restriction (both P o 0.001). The frequency of sleep disturbance and shortness of breath with wheeze did not differ between the groups.
In univariable analyses of 1-year-olds, the following variables were strongly associated with both asthma phenotypes: living in a single-parent household, having a mother who smoked, being breastfed for less than 6 months and a family history of asthma or wheeze (Table 3a). Multiple wheeze but not viral wheeze was strongly associated with increasing material deprivation. After controlling for all factors, these associations persisted, but the effects of maternal smoking and having a single parent were attenuated. The lower risk of wheeze in south Asian compared with white 1-year-olds remained after controlling for all other factors (adjusted OR for viral wheeze 0.81, 95% CI 0.64-1.03, and for multiple wheeze 0.66, 95% CI 0.47-0.92).
In 2-4-year-olds (Table 3b), after controlling for environmental and family risk factors there was strong evidence of a higher risk of multiple wheeze in south Asian compared with white children (OR 2.21, 95% CI 1. 19-4.09). Associations between material deprivation and wheeze in this age group differed by ethnicity (P for interaction test = 0.003): in south Asian children, neither viral nor multiple wheeze was clearly associated with increasing deprivation while the prevalence of multiple wheeze but not viral wheeze was much higher in white children living in the most deprived areas. Maternal smoking, older siblings and duration of breastfeeding were not strongly associated with either phenotype at this age. The association between south Asian ethnicity and multiple wheeze in 2-4-year-olds was unchanged when the analysis was restricted to children with severe wheeze (more than three attacks per year accompanied at least sometimes by shortness of breath [OR 2.24 (95% CI 0.99-3.08), n = 207], to children of Gujarati-speaking families [OR 2.69 (1.37-5.28), n = 447] and to children whose parents reported no problem understanding the questionnaire [OR 2.00 (1.05-3.81), n = 471]. The factors that had most influence in unmasking the higher odds of multiple wheeze in south Asian children were Townsend score, single parenthood, parental atopy and maternal smoking ( Table 4). The person answering the questionnaire (mother vs. father) did not confound the association between ethnicity and wheeze.
South Asian children in both pre-school age groups were more likely than white children to have seen a general practitioner for wheeze or cough, even after adjustment for severity of wheeze (OR 2.29 in 1-year-olds and 2.99 in 2-4-year-olds; Table 5). Parents of 2-4-year-old south Asian children also reported more hospitalizations for asthma or wheeze compared with parents of white children (OR 1.58). In contrast, inhaled corticosteroid use was lower in south Asian children (OR 0.46 in 1-year-olds, and 0.63 in 2-4-year-olds) after controlling for severity.

Discussion
This large population-based study showed that the risk of multiple wheeze (a condition that has many features of chronic atopic asthma) was higher in south Asian than white 2-4-year-olds after taking into account differences   in exposures to environmental, social and genetic risk factors (the latter approximated by a family history). In 1-year-olds, wheezing in general (and both wheezing phenotypes) was less common in south Asian children.
Conversely, south Asian children were more likely to be taken to their general practitioner, but less likely to use inhaled corticosteroids. South Asian 2-4-year-olds were also admitted to hospital for wheeze or asthma more often.

Methodological issues
The strengths of this study include its population-based sampling strategy, large sample size and good response rate, and so our results should be generalizable within the United Kingdom, particularly for the largest ethnic minority sub-group, Gujarati speakers of Indian origin. Studies of ethnic comparisons that try to answer aetiological questions have been criticized for using crude measures of ethnicity [14]. In our study, we obtained two reports of a self-defined ethnic group and included only children whose ethnic group was the same as their mothers'. We also collected additional information about language, religion and mother's country of birth and confirmed our results in a homogenous subgroup. It is possible that south Asian families under-reported symptoms due to misunderstanding of the term 'wheeze' [15]. The questionnaires were deliberately not translated, as we could not be sure that the term 'wheeze' would retain the same meaning in six languages. We do not think that this explains our results because the findings were similar in children with severe wheeze, which is more reliably reported by parents and when we included only families from both ethnic groups who reported to have had no problems understanding the questionnaire [12]. There was little heterogeneity in the findings according to household language, maternal country of birth or religion. Also, misunderstanding or cultural attitudes are unlikely to explain the differences that we found by age and asthma phenotype. Although we could not validate reported treatment and healthcare utilization with medical records, the responses to these questions were highly repeatable in both ethnic groups when the questionnaire was reapplied 3 months later to 413 children (Cohen's k between 0.57 and 0.80). This gives us some confidence in the reliability of our findings. One limitation of the study is the lack of objective measures of asthma-associated traits, such as atopy and bronchial responsiveness (BR). An interesting issue that emerged from the stepwise adjustment for different confounders was that socioeconomic factors, which had a different effect in south Asian and white children, especially the Townsend score, were largely responsible for the differences between the adjusted and unadjusted models. These deprivation measures represent a large number of correlated exposures related to lifestyle, health and psychosocial issues. The fact that they seem to have a different effect in the two ethnic groups is intriguing and needs further investigation. Other exposures, which in the unadjusted analysis had partly hidden the higher propensity of south Asian children for multiple-trigger wheeze, were the lower exposure to maternal smoking and the lower prevalence of atopic diseases in the family.

Comparison with other studies
Our results are consistent with a meta-analysis of studies in the United Kingdom that reported a lower crude prevalence of wheeze in south Asian than white children [2]. However, meta-analyses of observational studies cannot adequately incorporate the effects of confounding factors [16]. We found that age, wheezing phenotype and differential exposure to risk factors need to be taken into consideration when interpreting crude prevalence figures. A family history of atopy, maternal smoking, keeping pets, short duration of breastfeeding and a family history of wheeze were all more common in whites, while socioeconomic conditions also had an additional important confounding effect. After controlling for all these influences, the odds of multiple wheeze in 2-4-year-old south  Asian children were more than twice as high as in white children.

Interpretation of ethnic differences in asthma
The higher prevalence of multiple wheeze in south Asian compared with white toddlers, which became apparent after large differences in exposure to risk factors were taken into account, clearly shows the importance of the environment in determining the prevalence and phenotype of wheeze. One explanation is that young children of south Asian ancestry are at a higher intrinsic risk from the environmental determinants of atopic disorders such as asthma [7]. Le Souëf and co-authors proposed that proinflammatory genotypes based on IgE and eosinophilmediated mechanisms are more common in populations in tropical countries because these genotypes provide an evolutionary advantage against prevalent helminth and parasitic infections. In those who live in temperate climates but have recent tropical ancestry, asthma might be a manifestation of these strong immune responses. The observation that south Asian children in the United Kingdom have a higher prevalence of positive skin prick tests (SPTs) [17] and a higher prevalence of food-related wheeze [18] than white children lends further support to this hypothesis. We are currently conducting SPTs on the cohort children (ongoing study, unpublished observations), and have found that 61% of the 139 south Asians and 36% of the 326 white children who have already visited the lab are SPT positive (P o 0.001), the difference being significant both for never wheezers (45% vs. 26%, P = 0.006) and for current wheezers (81% vs. 64%, P = 0.047).
The differences between ethnic groups in the findings for viral and multiple wheeze in 2-4-year-olds in this study confirm that it is important to distinguish these two phenotypes [3,4]. This difference was not seen in 1-year olds, perhaps because most wheezing in under 2-year-olds is probably of non-atopic origin. The lower prevalence of all forms of wheeze in under-2-year-olds in south Asian than in white children might be explained by a lower prevalence of innate bronchial hyper-responsiveness (BHR). The sparse data available suggest that BR in infants, measured shortly after birth and before the onset of any respiratory illness, is associated with subsequent wheezing during the first 1-2 years of life [19,20]. There are no data available for south Asian infants. However, in our cohort, south Asian pre-schoolchildren with current wheeze were less likely than white children to report wheeze in response to the common non-allergic triggers of bronchoconstriction; laughter (23% vs. 34%, P = 0.001); and running or playing (17% vs. 24%, P = 0.047). Although, we controlled for some environmental exposures that have been associated with non-allergic BHR, such as parental smoking, the lower risk of wheeze in south Asian under 2-year-olds persisted: genetic and environmental factors might therefore protect south Asians against non-atopic BHR and thus against early (viral) wheeze.

Use of health services
South Asian children have higher hospital admission rates [21][22][23][24] and higher general practitioner consultation rates [25,26] than white children. We confirmed that south Asian pre-schoolchildren were more likely to visit their general practitioner even after adjusting for severity of wheeze. Hospital admissions in our cohort were also more common in south Asian than white 2-4-year-olds. These parent-reported data are confirmed by hospital-episode statistics for the Leicester Royal Infirmary, the only secondary care children's centre in the county. In 2004, 243 white and 64 south Asian children aged 1-12 years were hospitalized for asthma or wheeze, yielding hospitalization rates of 1408/100 000 (95% CI 1334-1479) in south Asian and 307/100 000 (95% CI 272-339) in white children, a rate ratio of 4.6 (95% CI 3.5-6.0, P o 0.001, data obtained from East Midlands Public Health Observatory). Cross-sectional data such as those presented in our paper cannot be conclusive about the reasons for this discrepancy, because it is impossible to disentangle asthma severity, asthma control and medication. However, the lower reported use of inhaled corticosteroids by south Asian children, as found in this and other populations [11,27], suggest that there might be ethnic differences in the availability and access to care for asthma. This deserves further sensitive investigation. Cultural differences in attitudes to asthma were shown in a recent hospital-based study from London [28]; south Asian parents were more reluctant to give drugs to their children and to tell others about their children's asthma. This would not explain the greater use of GP services but the lower use of inhaled corticosteroids in our group.

Conclusions
Crude prevalence data for ethnic groups [2] may be useful for healthcare planning, but may mask important differences in environmental exposures and genetic risk [29]. These limitations cannot be overcome by meta-analysis, but underline the need for carefully conducted cohort studies with good measures of ethnicity, socio-economic conditions and environmental exposures. If ongoing studies confirm the largely atopic nature of multiple wheeze and the transient, non-atopic nature of viral wheeze, these results will support the hypothesis that the genetic risk of atopic asthma might be higher in recent migrants from a tropical environment [7] and should prompt research into genetic and environmental risk factors for both phenotypes of early childhood asthma.