Lifetime prevalence of childhood asthma based on articles using the ISAAC written questionnaire varied from 1.1 to 11.0% with the lowest prevalence found among Tibetan children. The pattern with lowest prevalence in Tibet was also seen in articles on asthma related symptoms based on the ISAAC video questionnaire and in articles based on the Chinese diagnostic criteria. There were some variation and uncertainties about how asthma prevalences were assessed in some of the articles following the Chinese diagnostic criteria. Even so, most of these articles reported asthma prevalences below 3%. One article presented asthma prevalences for children living in 43 different cities in 2000. Current prevalence varied between 0.1−3.3% and lifetime prevalence varied between 0.3−4.6%. The highest prevalence was found in Chongqing and the lowest in Xining. Lhasa had the second lowest prevalence [26
]. Another article presented the current asthma prevalence for large populations of children living in 27 different cities. The average current prevalence estimates were 0.9% in 1990 and 1.5% in 2000 [13
In this review, the assessment of population-based estimates of asthma prevalence and assessment of regional variation and changes over time turn out to be difficult due to, for instance, differences in diagnostic criteria, data collection methods, sampling of populations and how prevalences are measured. Studies based on the ISAAC protocol relied on self-reports or parental reports of asthma, while this was not sufficient information for the articles following the Chinese diagnostic criteria. In these latter articles physicians had to evaluate the collected information for each potential asthma case and make sure that a set of necessary criteria were fulfilled. The differences described above point towards the need to have reasonably similar disease definitions and data collection methods in order to be able to compare prevalences between studies. Another potential source of uncertainty for the comparisons was the fact that in some articles it was unclear whether lifetime prevalence or current prevalence was presented. The diagnosis of asthma in children younger than three years old was made by clinical criteria, but one could question the reliability of diagnosis of asthma in children at that age. Information on age distribution was often insufficiently presented in studies following Chinese diagnostic criteria, another source of uncertainty when comparing the studies. Taking the above mentioned sources of uncertainty into account we tried to compare articles that seemed to have reasonably comparable outcome definition and data collection methods. For some of the articles this turned out to be more difficult than for others and they were therefore given less attention and results from several articles are only presented in a supplement for completeness.
The five versions of diagnostic criteria for asthma used in articles following the Chinese diagnosis criteria were all quite similar, and we did not consider that these minor differences would substantially distort the comparisons of asthma prevalences between these studies. The size of the population samples were large, especially for the articles based on Chinese diagnostic criteria reducing the problem of random errors.
Secular trend in disease occurrence is difficult to document, especially for diseases like asthma, in which self-report of symptoms has to be a major part of disease ascertainment [89
]. It is a common belief that the prevalence of asthma has been increasing in many societies around the world, even if some authors have raised questions about how to adjust for increasing awareness in the populations and among health workers [90
]. One of the articles in this review was convincingly able to show that there was an average increase in registered cases of current asthma between 1990 and 2000. The study population included a huge sample of children and the data collection methods were comparable for the subsamples of the population. The increase in current asthma prevalence is further supported by the finding of increased occurrence in the majority of the cities and in the age stratified comparisons in the study. The findings support an increased occurrence of asthma within the study period even if one cannot exclude effects of potential changes in awareness and minor changes in diagnostic criteria from 1990 and 2000.
Results from the two national surveys presenting asthma prevalences showed that the prevalences were higher in cities in the eastern part compare to cities in other parts of China. However, the prevalences were low all over China with small absolute differences. Furthermore, the prevalences had more often been assessed in urban than in rural areas, making it difficult to compare the difference in prevalence between urban and rural areas. This was also the case for the ISAAC studies as most of the surveys had been conducted in larger and more modern cities, like Hong Kong and Beijing. Furthermore, the prevalence of asthma in ISAAC studies from China were lower than in ISAAC studies carried out in many other places in the world and especially in developed countries like Austria (32%) [93
], United States of America (24.4%) [94
], United Kingdom (14.9%) [95
] and Singapore (27.4%) [96
]. These findings are consistent with the idea that the degree of modernization or westernization is relate to higher prevalence of asthma.
The effect of living altitude was difficult to assess since most studies were conducted in populations living at rather low altitude. One exception was studies from Tibet, which presented some of the lowest prevalence figures regardless of the type of asthma definition used. It is tempting to speculate that factors like the high living altitude, the extreme climate or other living conditions as well as the high child mortality could have contributed to this [97
]. However, these are speculations, as so far as few studies have been carried out in Tibet to corroborate them. The finding of higher asthma prevalence in rural Tingri and Sakya (4,300 meters above sea level) than in urban Lhasa (3,700 meters above sea level) additionally complicates such speculations. It could only be a by chance finding and both prevalences should be considered as extremely low.