Abstract
The task of deciphering the genetics of asthma is very complex. Recent studies of the familiar segregation of asthma showed that no single gene accounts for a major part of the expression of the disease, and that a polygenic model with some evidence of an oligogenic influence (i.e., a handful of loci being responsible for most of the genetic control) provided the best fit to the data. Although a final common pathway of recurrent bronchial obstruction is present in most cases of asthma, the disease shows marked phenotypic variability, suggesting etiologic heterogeneity and strong environmental influences. In an effort to circumvent these obstacles, linkage studies for genes controlling for apparently simpler phenotypes have been attempted. Total serum immunoglobulin E (IgE) levels, for example, show strong familiar aggregation and are known to be strongly correlated with asthma risk. Recent epidemiologic studies have suggested, however, that the inherited component of total serum IgE may be of little relevance as a determinant of asthma. Sensitization to certain aeroallergens is also associated with increased prevalence of asthma and is likely to have a genetic component, but the aeroallergens involved vary markedly with locale. In addition, sensitization to aeroallergens occurring at an early age is more strongly associated with asthma risk than late allergic sensitization, suggesting genetic heterogeneity. Therefore, studies of the genetics of phenotypes known to be strongly associated with asthma may clarify the causal role (if any) of the genes regulating their expression in the pathogenesis of asthma. Martinez FD. Complexities of the genetics of asthma.
The development of asthma is determined by complex interactions between genetic and environmental factors. Using data from monozygotic and dizygotic twins, estimates of heritability of asthma liability have been reported to be as high as 60–70% (1), but the mechanisms responsible for the inheritance of asthma are not well understood. We recently provided evidence suggesting that these mechanisms are indeed very intricate (2). We performed familial aggregation and segregation analysis of physician-diagnosed asthma in a large, unselected sample of over 900 nuclear families with 3,369 individuals living in Tucson, Arizona. Of these families, 29.6% had at least one family member with physician-diagnosed asthma. Intrafamily correlations (ρ) were compatible with a genetic influence: there was no increased risk of having asthma for the spouse of an asthmatic subject when compared to that of the spouse of a nonasthmatic subject (ρSp = –0.06, not significant), but there was a significant correlation between parents and offspring (ρPO = 0.09, p < 0.001) and between siblings (ρSS = 0.23, p < 0.001). Clearly, the correlation for asthma between siblings was higher than that between parents and offspring, suggesting either a significant environmental component or variance due to dominance. These studies thus confirmed the existence of significant familial aggregation of asthma, but both genetic and environmental influences could explain the results. Segregation analysis was thus performed to better characterize the observed aggregation patterns.
Segregation analysis is an elaborate statistical technique whose main objective is to determine if the familial aggregation of a phenotypic trait is compatible with the existence of a putative major gene that is transmitted from generation to generation. In our case, we used regressive algorithms (3, 4) to examine the fit of several genetic and nongenetic models to the data. These models test for the existence of types, which describe an unobserved characteristic that affects an individual's susceptibility (γ) to develop a trait—in our case, asthma. Genotypes are a special class of types, and most segregation analyses test for the existence of three genotypes, and therefore, for the presence of a single autosomal locus with two alleles, called A and B. This method of analysis permits distinction between genetic transmission and environmental influences, which also determine asthma susceptibility. The main difference between genetic and nongenetic types is the manner in which the types are transmitted from generation to generation. If transmission is genetic, then subjects with the AA type will only transmit A alleles to their offspring; i.e., the probability τAA of transmitting allele A is 1.0. It can be easily shown that τAB = 0.5 and τBB = 0. Environmental models, on the other hand, will show no specific transmission pattern, thus τAA = τAB = τBB. In addition, the models can be made compatible with a dominant, codominant, or recessive model of inheritance by varying the relative susceptibility γ of the types: thus, in a recessive model for A, γAA > γAB and γAB = γBB; in a codominant (also called Mendelian “arbitrary”) model γAA > γAB > γBB; and in a dominant model for A, γAA = γAB and γAB > γBB. Models that include a multifactorial (genetic and/or environmental) influence together with a putative major gene can also be tested for by adding parameters representing residual family effects (parent-offspring, sibling-sibling, and spouse effects). Genetic and nongenetic models can also be compared with the best fitting model, i.e., one in which all parameters are left unrestricted. If the data fits a known transmission pattern as well as the best fitting mode, and better than any other genetic or nongenetic model, then one can say that there is evidence of a major gene that is transmitted according to that known pattern. To evaluate these different models, statistical fit is compared by taking the difference between twice the ln likelihood of the models of interest. This difference has a χ2 distribution and can be used to give a probability level.
Results of our segregation analysis of asthma showed no evidence for a major gene when the data were tested without residual family effects. As can be seen in Table 1, no genetic and environmental models fit the data better than the unrestricted model. However, the genetic models fit the data more closely than the environmental model (compare line 2 with any of the last three lines in Table 1). This suggests that multifactorial/polygenic factors other than those that could be modeled as a single major gene may be at work. We thus used two different ways of testing for these factors, and Table 2 shows the results of one such trial. The addition of residual family effects markedly decreased the –2 ln likelihood for all models (compare Tables 1 and 2), suggesting the presence of a significant familial component contributing to the inheritance of asthma. Again, all genetic models fit the data significantly worse than the unrestricted model but significantly better than the environmental model, and the recessive model and the “arbitrary” model fit the data better than the dominant model, suggesting that, beyond the multifactorial/polygenic influence, an oligogenic “recessive” influence may be present.
| Model | Gene Frequency qa | Transmission Probabilities | Age Parameters | Susceptibilities | –2 Ln (Likelihood) | χ2 | df* | p Value | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Baseline β | Coefficient | |||||||||||||||||||||||||||
| τaa | τab | τbb | Male | Female | α | γaa | γab | γbb | ||||||||||||||||||||
| No residual familial effects | ||||||||||||||||||||||||||||
| Unrestricted | 0.33 | [1.0] | [1.0] | [0.0] | –2.37 | –2.86 | 0.17 | [1.0] | [0.00] | [0.00] | 4,367.6 | — | — | — | ||||||||||||||
| No major type | ||||||||||||||||||||||||||||
| (environmental) | — | — | — | — | –1.54 | –2.11 | 0.17 | 0.16 | 0.16 | 0.16 | 4,616.7 | 249.1 | 0–6 | < 0.001 | ||||||||||||||
| Mendelian arbitrary | 0.32 | (1.0) | (0.5) | (0.0) | –1.49 | –2.06 | 0.17 | [1.0] | [0.00] | 0.09 | 4,570.8 | 203.2 | 0–1 | < 0.001 | ||||||||||||||
| Mendelian dominant | 0.12 | (1.0) | (0.5) | (0.0) | –1.50 | –2.05 | 0.17 | 0.51 | 0.51 | 0.05 | 4,579.1 | 212.5 | 0–2 | < 0.001 | ||||||||||||||
| Mendelian recessive | 0.31 | (1.0) | (0.5) | (0.0) | –1.49 | –2.05 | 0.17 | 0.97 | 0.06 | 0.06 | 4,574.1 | 206.5 | 0–2 | < 0.001 | ||||||||||||||
| Model | Gene Frequency qa | Transmission Probabilities | Residual Parent Effects | Residual Spouse Effects | Susceptibilities | –2 Ln (Likelihood) | χ2 | df‖ | p Value | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| τaa | τab | τbb | δPu* | δPa† | δSpu‡ | δSpa§ | γaa | γab | γbb | |||||||||||||||||||||
| Plus residual parent-offspring and spouse effects and conditioning on parents | ||||||||||||||||||||||||||||||
| Unrestricted | 0.39 | 0.97 | [1.0] | [0.0] | [0.0] | 0.94 | [0.0] | –2.19 | [1.0] | [0.0] | 0.13 | 3,033.0 | — | — | — | |||||||||||||||
| No major type | ||||||||||||||||||||||||||||||
| (environmental) | — | — | — | — | –0.01 | 1.12 | [0.0] | –2.59 | 0.68 | 0.68 | 0.68 | 3,106.2 | 73.2 | 1–6 | < 0.001 | |||||||||||||||
| Mendelian arbitrary | 0.54 | (1.0) | (0.5) | (0.0) | [0.0] | 0.96 | [0.0] | –3.04 | 0.91 | [0.0] | 0.56 | 3,090.3 | 57.3 | 0–3 | < 0.001 | |||||||||||||||
| Mendelian dominant | 0.31 | (1.0) | (0.5) | (0.0) | [0.0] | 0.76 | [0.0] | –3.02 | 0.78 | 0.78 | 0.03 | 3,095.0 | 62.0 | 0–4 | < 0.001 | |||||||||||||||
| Mendelian recessive | 0.67 | (1.0) | (0.5) | (0.0) | [0.0] | 0.72 | [0.0] | –3.06 | 0.88 | 0.03 | 0.03 | 3,092.6 | 59.6 | 0–4 | < 0.001 | |||||||||||||||
Results of segregation analysis thus suggest that many genes are almost certainly involved in the pathogenesis of asthma, and that none of these genes has an effect on susceptibility that is strong enough to emerge as a major contributor. Recent preliminary reports from a genome-wide search performed as part of the U.S. Collaborative Study of the Genetics of Asthma (CSGA) confirm this finding (5). These researchers genotyped 261 affected sibling pairs and their parents for 360 microsatellite markers and found, with a rather lenient level of statistical significance (p < 0.01), suggestive evidence for linkage in 11 different chromosomal regions for three different ethnic groups (caucasians, hispanics, and African-Americans). Markers in only one region in chromosome 12q reached the selected level of significance in more than one ethnic group.
These results are not unexpected. The pathogenesis of asthma is known to be extremely complex, and it is quite likely that alterations in the regulation of many different immune pathways may give rise to the same common final pathway, namely recurrent, reversible (at least partially reversible) bronchial obstruction associated with chronic airway inflammation and bronchial hyperresponsiveness. Several recent epidemiologic studies suggest that, at least in the case of childhood asthma, the disease may have different phenotypic expressions at different ages, as assessed by their risk factors and prognosis (6). Therefore, although it is likely that a small set of genes common to all forms of asthma (the “asthma genes” detected by our oligogenic segregation models) may control the expression of the disease's purported common final pathway, the pleomorphic, often pleonastic genetic and environmental influences responsible for the expression of the different risk factors for asthma may considerably influence the penetrance of these “asthma” genes.
Given the complexities described above, it is not surprising that many researchers have opted to study the genetics of phenotypes known to be strongly associated with asthma that have also been shown to aggregate in families. These attempts are justified by the assumption that finding the gene(s) responsible for the expression of these so-called intermediate phenotypes will enhance our knowledge of the genetic mechanisms involved in asthma. These efforts, however, are not devoid of significant difficulties. I will use the examples of total serum IgE and skin-test reactivity to allergens to illustrate my point.
In 1989 Burrows and colleagues showed that log total serum immunoglobulin E (IgE) levels were linearly related to the prevalence of asthma in a large population sample of children and adults (7). This was true for subjects with positive or negative skin tests to common aeroallergens. Moreover, no prevalent asthma cases were found among subjects with very low total serum IgE levels. This gave rise to the idea that the determinants of total serum IgE were involved in the pathogenesis of asthma, and that all forms of asthma were IgE-mediated. Segregation analyses published before (8) and after (9) Burrows and colleagues' findings have suggested that log total serum IgE levels may be controlled by a major gene, although the exact mechanism (dominant, codominant, or recessive) of transmission is controversial. Recent candidate gene approaches (10) and genome-wide screens (11) have yielded evidence for linkage of total serum IgE levels with markers in chromosomes 5 and in chromosomes 6, 7, 11, and 16, respectively. This has suggested the possibility that the putative genes underlying these linkage signals may be involved in the pathogenesis of asthma.
Recent, more detailed analyses of the intrafamily relations between total serum IgE and asthma challenge a simple, cause-effect association between these two variables. Burrows and colleagues (12) examined the extent to which the strong familial aggregation of asthma could be explained by the known association between parental IgE levels and those of their children. In the same sample used to assess the relation between prevalence of asthma and total serum IgE levels, the researchers found that whereas 11.5% of children with no parental asthma had asthma, approximately one-third of children with one asthmatic parent and almost half of children with two asthmatic parents had asthma, respectively. They also found the expected strong parent-offspring correlation of log total serum IgE levels. To determine the relative importance of parental asthma versus parental IgE in the transmission of asthma to their offspring, Burrows and colleagues also studied asthma rates in children in relation to tertiles of serum IgE in each of their parents after stratifying for asthma in the same parent (Table 3). They found no statistically significant association between prevalence of asthma in children and serum IgE levels in their parents when the mother or the father did not have asthma. High rates of asthma were found in the children of asthmatic parents who had high serum IgE levels when compared with children of parents with asthma and low serum IgE levels. Both maternal and paternal serum IgE levels were found to be highly significant determinants of total serum IgE levels in their children (Figure 1), but for any level of parental serum IgE, asthmatic children had up to three times higher geometric mean serum IgE than nonasthmatic children. This association was independent of skin-test reactivity to allergens in the children.
| Children with Asthma according to Tertile of IgE z-Score in the Same Parents | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Low | Mid | High | ||||||||||
| Parental Asthma | n | % | n | % | n | % | ||||||
| Mother without asthma | 381 | 12.9 | 376 | 11.4 | 290 | 14.8 | ||||||
| Father without asthma | 320 | 13.8 | 290 | 17.2 | 286 | 14.3 | ||||||
| Mother with asthma | 34 | 17.6 | 52 | 40.4 | 128 | 40.6 | ||||||
| Father with asthma | 27 | 22.2 | 59 | 25.4 | 61 | 45.9 | ||||||
Fig. 1. Relation of children's IgE z-score to their asthma status and the tertiles of their parents' IgE z-scores. The tertile combinations are shown in the order of the serum IgE z-scores of the parents' nonasthmatic children. Reproduced from Reference 12 by permission.
[More] [Minimize]These results suggested that inheritance of a tendency to develop high total serum IgE levels is only one factor related to the inheritance of asthma susceptibility and that, by itself, it has limited ability to predict asthma inheritance. This was confirmed by the finding that the patterns of inheritance described in the previously quoted segregation analyses of asthma remained basically unchanged when serum IgE levels were introduced as covariates into the models (2). Moreover, parental serum IgE seemed to increase the likelihood of developing asthma only when the parents themselves had asthma. Burrows and colleagues (12) thus speculated that there were at least two components of the familiar aggregation of total serum IgE levels: one such component would be a generic, perhaps noncognate capacity to develop IgE-mediated responses; and the second one, a capacity to produce IgE in response to environmental stimuli, which is inherited together with asthma and not independent of it. Although this second component seemed to be independent of skin-test reactivity to allergens in their data, Burrows and colleagues argued that the right allergens may not have been included in their battery of skin tests (12). It is still possible, therefore, that the association of childhood asthma with higher total serum IgE levels in the child than would have been expected on the basis of parental IgE levels could be explained by a tendency for both asthmatic parents and their offspring to become sensitized preferentially to certain aeroallergens (13), which would in turn increase their total serum IgE. This cognate, asthma-related mechanism of IgE production could be inherited independently of total serum IgE levels.
Interestingly, Xu and colleagues (14) recently reported strong statistical evidence in favor of such a two-locus model of IgE inheritance. These authors studied data from 92 Dutch families ascertained through a parent with asthma and showed evidence for linkage between total serum IgE levels and markers located in chromosome 5q, with lod scores of 3.0 (9). Based on these findings, the authors performed two-locus segregation and linkage analysis. They found that the two-locus model fit the data significantly better than the one-locus model, increasing the lod score for their previously reported 5q locus to 4.67. This 5q locus explained a very large part of the variance in total serum IgE in these asthma-enriched families, whereas the second locus explained 19% of the variance. In a subsequent study (15), the same researchers reported that bronchial hyperresponsiveness also showed linkage with the same markers in chromosome 5q in their Dutch families selected on the basis of an asthmatic proband. It is thus possible that the locus in chromosome 5q detected by this group may be one of the “asthma genes” described above—i.e., one of the genes responsible for the asthmatic “final common pathway,” of which bronchial hyperresponsiveness is an important component. Whether this locus is associated with skin-test reactivity to local asthma-related allergens is yet unknown.
Another phenotype considered for years to be intermediate between genetic background and expression of asthma is allergic sensitization to common environmental antigens. The basic paradigm has been that asthmatic subjects first become sensitized to certain allergens and, as a consequence, develop asthma. This paradigm has been used with particular insistence in regards to the allergens of the house dust mite (HDM). The strong association between sensitization to HDM and prevalence of asthma in coastal areas has been interpreted as a proof that HDM sensitization is the cause of asthma in these subjects (16, 17). The corollary of this assumption has been that allergen avoidance would prevent many cases of asthma (16). In addition, if the causal association between HDM and asthma were true, discovering the gene or genes responsible for sensitization against this allergen would certainly be a major step in finding a cure for the disease.
Unfortunately, much as for total serum IgE levels, the nature of the association between sensitization to HDM and asthma has proven to be more complex than expected. Reports of marked improvement in severity of asthma symptoms in children transferred to HDM-free mountain resorts were initially perceived as somehow proving that HDM was causing their asthma (18), but those reports have been recently reinterpreted with much more caution and circumspection (19). The same investigators who first proposed that HDM causes asthma have more recently reported that in certain geographical areas, namely desert regions (20) and regions at high altitude (21) where HDM exposure is very low or absent, the prevalence of asthma is as high or even higher than that in regions where subjects are heavily exposed to HDM. We recently confirmed this finding in our longitudinal study of asthma in Tucson, Arizona (22). Exposure to HDM in this area of the U.S. is much less conspicuous than in coastal regions, yet the prevalence of asthma is very high, up to 10% of schoolchildren having active disease. As has been shown for desert regions of Australia (20), the allergen most strongly associated with childhood asthma in Tucson is the mold Alternaria, with few asthmatic subjects being sensitized to HDM. Because selective immigration into coastal or desert regions of subjects genetically predisposed to become sensitized against certain allergens is implausible, the most likely explanation of these findings is that asthmatics may have a nonspecific predisposition to become sensitized to certain aeroallergens. Which allergens these may be will depend on the biological characteristics of the allergen, the intensity of exposure (23, 24), and perhaps the timing of the exposures (25). Becoming sensitized to these allergens, on the other hand, may certainly contribute by itself to the development of asthma symptoms, but sensitization would be not the cause but the consequence of the asthmatic predisposition. It is tempting to speculate, for example, that the gene or genes detected in chromosome 5q (9) could increase total serum IgE and bronchial hyperresponsiveness by predisposing people to increased sensitization to local asthma-related allergens. Although still unproven, this hypothesis may explain the observed associations between asthma and allergic sensitization in populations.
In attempting to avoid the perplexing difficulties of tackling directly the genetics of asthma, researchers have studied the genetics of phenotypes known to be strongly associated with asthma, on the assumption that they may be intermediate causes of the disease. One important byproduct of these studies will be to clarify the causal role (if any) of these phenotypes and of the different genes regulating their expression in the pathogenesis of asthma.
This work was funded by a Research Development Award for Minority Faculty (HL03154) from the National Heart, Lung, and Blood Institute.
| 1. | Duffy D. L., Martin N. G., Battistutta D., Hoffer J. L., Mathews J. D.Genetics of asthma and hay fever in Australian twins. Am. Rev. Respir. Dis.142199013511358 |
| 2. | Holberg C. J., Elston R. C., Halonen M., Wright A. L., Taussig L. M., Morgan W. J., Martinez F. D.Segregation analysis of physician-diagnosed asthma in Hispanic and non-Hispanic white families. Am. J. Respir. Crit. Care Med.1541996144150 |
| 3. | Bonney G. E.Regression logistic models for familial disease and other binary traits. Biometrics421986611625 |
| 4. | S.A.G.E.: Statistical Analysis for Genetic Epidemiology, Release 2.1 and 2.2, 1991 and 1994. Computer program package available from the Department of Epidemiology and Biostatistics, Case Western Reserve University School of Medicine, Cleveland, OH. |
| 5. | Meyers D. A., Banks-Schlegel S., Bleecker E. R., Blumenthal M. N., Cox N. J., Maestri N. E., Marsh D. G., Ober C., Rich S. S.the Collaborative Study of the Genetics of Asthma (CSGA)A genome-wide search for asthma susceptibility loci in ethnically diverse populations (abstract). Am. J. Hum. Genet.591996A228 |
| 6. | Martinez F. D., Wright A. L., Taussig L. M., Holberg C. J., Halonen M., Morgan W. J.the Group Health Medical AssociatesAsthma and wheezing in the first six years of life. N. Engl. J. Med.3321995133138 |
| 7. | Burrows B., Martinez F. D., Halonen M., Barbee R. A., Cline M. G.Association of asthma with serum IgE levels and skin-test reactivity to allergens. N. Engl. J. Med.3201989271277 |
| 8. | Marsh D. G., Bias W. B., Ishizaka K.Genetic control of basal serum immunoglobulin E levels and its effect on specific reaginic sensitivity. Proc. Natl. Acad. Sci. U.S.A.71197435883592 |
| 9. | Martinez F. D., Holberg C. J., Halonen M., Morgan W. J., Wright A. L., Taussig L. M.Evidence for Mendelian inheritance of serum IgE levels in Hispanic and non-Hispanic white families. Am. J. Hum. Genet.551994555565 |
| 10. | Meyers D. A., Postma D. S., Panhuysen C. I., Xu J., Amelung P. J., Levitt R. C., Bleecker E. R.Evidence for a locus regulating total serum IgE mapping to chromosome 5q. Genomics231994464470 |
| 11. | Daniels S. E., Bhattacharrya S., James A., Leaves N. I., Young A., Hill M. R., Faux J. A., Ryan G. F., le Söuef P. N., Lathrop G. M., Musk A. W., Cookson W. O. C. M.A genome-wide search for quantitative trait loci underlying asthma. Nature3831996247250 |
| 12. | Burrows B., Martinez F. D., Cline M. A., Lebowitz M. D.The relationship between parental and children's serum IgE and asthma. Am. J. Respir. Crit. Care Med.152199514971500 |
| 13. | Burrows B., Sears M. R., Flannery E. M., Herbison G. P., Holdaway M. D.Relations of bronchial responsiveness assessed by methacholine to allergy skin test reactivity, lung function, respiratory symptoms, and diagnoses in thirteen-year-old New Zealand children. J. Allergy Clin. Immunol.951995548556 |
| 14. | Xu J., Levitt R. C., Panhuysen C. I. M., Postma D. S., Taylor E. W., Amelung P. J., Holroyd K. J., Bleecker E. R., Meyers D. A.Evidence for two unlinked loci regulating total serum IgE levels. Am. J. Hum. Genet.571995425430 |
| 15. | Postma D. S., Bleecker E. R., Amelung P. J., Holroyd K. J., Xu J., Panhuysen C. I., Meyers D. A., Levitt R. C.Genetic susceptibility to asthma-bronchial hyperresponsiveness coinherited with a major gene for atopy. N. Engl. J. Med.3331995894900 |
| 16. | Peat J. K., Tovey E., Toelle B. G., Haby M. M., Gray E. J., Mahmic A., Woolcock A. J.House dust mite allergens: a major risk factor for childhood asthma in Australia. Am. J. Respir. Crit. Care Med.1531996141146 |
| 17. | Chapman M. D., Heymann P. W., Sporik R. B., Platts-Mills T. A.Monitoring allergen exposure in asthma: new treatment strategies. Allergy5019952933 |
| 18. | Spieksma F. T. M., Zuidemo P., Leupen H. J.High altitude and house dust mites. Br. Med. J.1197182 |
| 19. | Carswell F.Asthma and altitude (Editorial). Clin. Exp. Allergy231993973975 |
| 20. | Peat J. K., Toelle B. G., Gray E. J., Haby M. M., Belousova E., Millis C. M., Woolcock A. J.Prevalence and severity of childhood asthma and allergic sensitization in seven climatic regions of New South Wales. Med. J. Austr.16319952226 |
| 21. | Sporik R., Ingram J. M., Price W., Sussman J. H., Honsinger R. W., Platts-Mills T. A.Association of asthma with serum IgE and skin test reactivity to allergens among children living at high altitude: tickling the dragon's breath. Am. J. Respir. Crit. Care Med.151199513881392 |
| 22. | Halonen M., Stern D., Wright A. L., Taussig L. M., Martinez F. D.Alternaria as a major allergen for asthma in children raised in a desert environment. Am. J. Respir. Crit. Care Med.155199713561361 |
| 23. | Sporik R., Holgate S. T., Platts-Mills T. A. E., Cogswell J. J.Exposure to house-dust mite allergen (Der p I) and the development of asthma in childhood: a prospective study. N. Engl. J. Med.3231990502507 |
| 24. | Marks G. B., Tovey E. R., Toelle B. G., Wachinger S., Peat J. K., Woolcock A. J.Mite allergen (Der p I) concentrations in houses and its relation to the presence and severity of asthma in a population of Sydney schoolchildren. J. Allergy Clin. Immunol.961995441448 |
| 25. | Peat J. K., Salome C. M., Woolcock A. J.Longitudinal changes in atopy during a 4-year period: relation to bronchial hyperresponsiveness and respiratory symptoms in a population sample of Australian school children. J. Allergy Clin. Immunol.8519906574 |
