Main Findings
In this study, we investigated temporal changes in the prevalence of asthma and allergies in the two main districts of the Republic of Cyprus based on two cross-sectional surveys eight years apart. We have also assessed whether any changes that occurred in this period were differential between the urban and rural areas of the island. With the exception of current symptoms of rhinoconjuctivitis, the prevalence of all outcomes investigated appeared significantly increased in 2008 compared to 2000. In fact, the prevalence of lifetime diagnosis of asthma, eczema and rhinitis have almost doubled. Rises were generally observed in both urban and rural areas but, with the exception of eczema symptoms and diagnosis, these consistently appeared more pronounced in the rural areas of the country.
Limitations
In contrast to the high participation rate observed in the first survey, participation was quite low in the subsequent 2008 survey. Anecdotal evidence among research teams in Cyprus points to an increased reluctance in recent years to participate in research studies resulting in lower response rates compared to a decade ago. It is likely that a combination of factors are responsible for this including (a) an increasing reluctance of parents to offer personal information and (b) a recent steady rise in the number of surveys among the school population, which in a country the size of Cyprus may be important.
We observed a lower participation rate in urban as opposed to rural areas (44.5% Vs 59.5% respectively). In the absence of information on non-responders, or any data from official sources more recent than the 2001 census, for the purpose of contrasting the socio-demographic profile of the study participants, we identified a 2004 study on childhood obesity among Cypriot children aged 2-6,. The time frame and age band of the participants that would have participated four years later at age 8 in our 2008 survey were captured by that study with a participation rate of over 75% [
27]. Parental level of education as reported in that study (as well as several other participant characteristics such as birth weight) compares favourably with the figures observed in our study sample. This is the case both in terms of the overall sample as well as when broken down to urban-rural areas with the same approach as in our study. This serves as an indication that the low response rate in 2008 did not seem to compromise the representativeness of our study population. Although it is likely that affected children may have an incentive to participate in the survey, we have no reason to believe that this should be differential in urban Vs rural areas (i.e. there is no reason why proportionally more asthmatics than non-asthmatics from the urban areas would have participated in the study but the opposite to have been the case in rural areas). In any case, sensitivity analyses have indicated that, even though not impossible, such differential participation would have to be quite marked for the observed rise in rural areas across several study outcomes to be an artifact. To some extent, this is also reinforced by the fact that the two samples appear internally consistent both in terms of stable characteristics (e.g. the proportion of children born in Cyprus is significantly higher in the rural compared to the urban areas in both surveys) as well as observed changes between the two survey years (e.g. a decrease in parental smoking in the home occurred in both urban and rural areas). Finally, in 2001 (when the participation rate was over 80%) the observed difference between urban and rural areas was clearly striking. This was not the case by the year 2008, both before and after adjusting for differences in the composition of our two samples, serving at least as an indication that the previously wide urban-rural gap may have now narrowed.
Limited by its cross-sectional design, no inferences can be made on the causal association of risk factors assessed in the two surveys with the study outcomes. For instance, the observed increase in the report of family history of atopy between the two surveys may represent a "real" change, as has been seen in other studies [
21], or may simply be the result of sampling. However, we adjusted for family history of atopy in our assessment of changes in prevalence of asthma and allergy outcomes among the children so that this potential source of error has been accounted for. While it was important to adjust for the possible confounding effect due to compositional differences in the study populations, we have made no attempt to associate changes in any environmental or lifestyle factors with the increase in the asthma and allergy outcomes in the urban and rural areas of Cyprus. Of course, residual confounding by other important factors, such as socioeconomic status or exposure to animals and/or farming practices, could not be corrected for in the analysis due to lack of comparable data from the two surveys. Nevertheless, as also mentioned above, while adjustments are necessary to correct for sample differences, it was not the purpose of the study to identify the specific factors related to rural living, or even characteristics of rural populations, which underline these changes. Lastly, we should note that the Greek version of the ISAAC questionnaire has not been officially validated. However, there is only one accepted version which has been in use as a standard tool ever since in several studies in Greece [
24] and asthma reporting using this questionnaire has been found to correlate well with objective measures [
28].
Temporal changes in asthma and allergy outcomes between years 2000 and 2008
This is the first study to have assessed temporal changes in the prevalence of asthma and allergies in Cyprus. We have found that the prevalence of current wheezing symptoms has significantly increased by 0.23% annually to 8.7% in eight years. While still at a level much lower than the global average of 11.6%, this translates to an annual increase greater than the global average of 0.13% [
29]. The rise in the prevalence of current wheeze is consistent with findings from studies conducted using the ISAAC questionnaire in neighboring countries such as Greece [
20], Turkey [
21], Malta [
30] and other Eastern European and Mediterranean countries [
29]. The annual increase of 0.23% however is lower than the regional average (Eastern Mediterranean) of 0.79% as was shown by the ISAAC Phase III study group which nonetheless only included centers from three countries in this region. In addition, the prevalence of current eczema symptoms increased significantly between the two study periods (from 3.7% to 4.6%); consistent with ISAAC Phase III studies in the same age group which have shown increases in current eczema symptoms in most countries [
5]. Current eczema prevalence in Cyprus is close to the Eastern Mediterranean average of 4.8% [
31]. The prevalence of current symptoms of allergic rhinitis in Cyprus is very low but comparable nonetheless with Greece and Eastern European countries [
32].
The significant increase in the prevalence of lifetime asthma, eczema and allergic rhinitis in Cyprus in eight years is consistent with the rise in prevalence of these outcomes in most parts of the world, even in countries with declining prevalence in current symptoms [
29]; to some extent, this can be attributed to improved awareness and diagnostic labeling of these conditions. However, considering the fact that the prevalence of allergic symptoms has also increased considerably in Cyprus, changes between the two time periods may at least partly reflect a true increase. Furthermore, the current prevalence in Cyprus of ever asthma diagnosis (17.4%) and ever eczema (13.5%) is much higher than respective regional averages of 9.1% and 7.2% [
5]. In contrast, allergic rhinitis diagnosis prevalence at 5.2% is much lower than the regional average of 13.9% [
32]; this may represent lower awareness of this disease in Cyprus or a true low prevalence given the similarly low prevalence of current allergic rhinoconjuctivitis symptoms.
Temporal changes in asthma and allergy outcomes by urban and rural residence
We have found that the increase in symptoms of asthma and allergic rhinitis in this period may have been specific to rural areas (with the previous urban-rural differences narrowing) whereas current eczema symptoms appeared to have increased significantly in urban areas only. In this study, only the place of residence at the time of the survey was recorded; hence this may not necessarily represent a child's early life exposure, which is important in the development of allergic diseases. However, the purpose of this study was not to investigate whether "exposure" to a rural environment (and which particular aspects of it) poses an independent effect on the development of the study outcomes but to describe the current urban-rural patterns of allergic diseases. In any case, it is unlikely that rises in rural areas are a product of a recent movement of population out of metropolitan areas. At 0.8%, population mobility in Cyprus (defined as the proportion of people living in a different address a year before the census) is particularly low [
25].
Studies from urban areas elsewhere in the Eastern Mediterranean region such as the city of Patras Greece [
20] and Istanbul, Turkey [
33], have indicated continuing rises in current symptoms of all allergic diseases; nevertheless both of these studies were conducted in urban areas during an earlier time period (i.e. between the early 1990s to early 2000), and possibly in populations at an earlier stage of urbanization both at baseline as well as follow-ups, and were perhaps still in the process of adopting lifestyle changes that play a role in allergic diseases development.
Only a small number of studies have investigated trends in the prevalence of allergic conditions within the same country population in urban and rural areas separately [
10,
21,
34]. A Swedish study over an earlier period of three decades (1952-1981) has shown a continuing rise in asthma, eczema and rhinitis diagnosis prevalence among conscripts in both rural and urban areas [
10]. Another study from Italy, investigating asthma and allergy trends between 1994-2002 has found the increase in the prevalence of allergic rhinitis and eczema symptoms to be higher in adolescents living in large metropolitan areas compared to areas with lower population density [
34]. Furthermore, a more recent study from Edirne Turkey which investigated prevalence of allergies differentially in urban and rural areas between 1994-2004, showed the prevalence of current wheeze to have increased significantly in both urban and rural areas [
21]. In contrast to our findings, the Edirne study showed the rise in the asthma and allergic rhinitis outcomes to be greater in the urban rather than the rural areas. Unlike the Edirne study, rural areas in our study are not defined in terms of farming practices (but in terms of their population size and remoteness from big centres of population) and include places in transition from a more rural to a more urban/westernized way of living.
In Sweden [
10] and in several earlier studies [
8,
9], living on a farm has been shown to be associated with reduced risk of symptoms relating to upper and lower airways but not eczema. Therefore, it wouldn't be unreasonable to assume that as a result of declining rates of farming practice across rural areas of Cyprus, any protective effect of farming on asthma and rhinitis has weakened, explaining the increase in the prevalence of symptoms relating to these diseases but not eczema. Of course, other environmental and lifestyle factors are likely to be implicated in the aetiology of the observed changes. For example, the role of socio-economic factors (and any underlying changes in the socio-economic status of the population in this period) is crucial. Since no socio-economic data were recorded in the 2000 survey, we could not assess whether any significant changes in the socio-economic profile of the population occurred during this period. According to the Cyprus Statistical Service survey on Income and Living Conditions, population socio-economic indicators (income, poverty, education, etc) have been very stable between 2005-2008 [
35]. In any case, not adjusting for any potential differences in the socio-economic status of the children might have introduced uncontrolled confounding. We used data on parental level of education (available from the 2008 survey) in order to assess the extent to which socio-economic differences between participants in urban and rural areas explain the observed 2008 urban-rural patterns. Despite the fact that, as expected, educational attainment was higher in urban areas (i.e. 59.6% Vs 40.3% of participants with one parent with tertiary education respectively), this did not seem to affect our findings, indicating that the observed diversion (e.g. in the case of eczema) or narrowing (e.g. in the case of wheezing) in 2008 is not accounted by socio-economic differences between urban and rural parts of the country.
Generally, in the absence of detailed information on lifestyle and environmental factors which may be implicated in the differential epidemiology of allergic diseases in Cypriot children, we can only speculate that the higher increase in allergies in the rural areas may be the result of a general trend of reduced exposure to farming and/or, in parallel, an increased adoption of a more "urban" lifestyle among families in rural areas related to diet patterns, reduced exercise levels and a more hygienic environment.