Evidence for Genetic Associations between Asthma ... - ATS Journals

8 downloads 0 Views 142KB Size Report
The study was undertaken in two phases. ... more alike than ordinary brothers and sisters. .... The recruitment involved two stages, and our analysis allowed for.
Evidence for Genetic Associations between Asthma, Atopy, and Bronchial Hyperresponsiveness A Study of 8- to 18-Yr-Old Twins JANE R. CLARKE, MARK A. JENKINS, JOHN L. HOPPER, JOHN B. CARLIN, CAROL MAYNE, DAVID G. CLAYTON, MARITA F. DALTON, DIANE P. HOLST, and COLIN F. ROBERTSON Respiratory & Cystic Fibrosis Unit, The Birmingham Children’s Hospital NHS Trust, Birmingham, United Kingdom; Centre for Genetic Epidemiology, The University of Melbourne, Carlton, Victoria, Australia; Clinical Epidemiology & Biostatistics Unit, Murdoch Children’s Research Institute, Parkville, Victoria, Australia; Epidemiology Unit, Queensland Institute of Medical Research, Brisbane, Queensland, Australia; Medical Research Council Biostatistics Unit, Institute of Public Health, Cambridge, United Kingdom; and Department of Respiratory Medicine, Royal Children’s Hospital, Parkville, Victoria, Australia

We measured asthma in the last 12 mo, diagnosed by a respiratory physician at interview; atopy, defined by a positive skin prick test to any of eight common allergens; and bronchial hyperresponsiveness (BHR) to hypertonic saline, in 381 twin pairs aged 8 to 18 yr selected from the Australian Twin Registry—183 monozygous (MZ) and 198 dizygous (DZ). The associations between twins, as measured by an odds ratio, were greater in MZ pairs compared with DZ pairs for asthma: 25.6 (95% confidence interval 11.3– 57.8) versus 1.9 (1.0–3.5); atopy: 14.6 (7.1–30.1) versus 2.5 (1.4– 4.5); and BHR: 14.1 (6.4–31.0) versus 4.2 (2.1–8.6) (all p ⬍ 0.002). The associations between each pair of traits within an individual were slightly greater than the association between one trait in a twin and the other trait in the cotwin (cross-trait cross-pair) in MZ pairs. Further, the associations in MZ pairs were greater than in DZ pairs (p ⬍ 0.05). Under the assumptions of the classic twin model, these data suggest that the strong cross-sectional associations between these three traits are due to an overlap between the genetic factors involved in each of these three traits.

Asthma is a common respiratory condition in childhood, with more than one in five Melbourne children aged 7 to 15 yr old experiencing symptoms of asthma in the past 12 mo (1). People with asthma are more likely to exhibit bronchial hyperresponsiveness (BHR) and, especially in childhood, are more likely to be atopic. For example, in a 1996 study of 187 Victorian school children, those with current asthma (defined as parent reported wheezing in the past 12 mo) were 52% more likely to be atopic (0.79 versus 0.52) and 160% more likely to exhibit BHR to hypertonic saline (0.78 versus 0.30) compared with children who had never had asthma (2). A study of 483 New Zealand secondary school students, observed that those with self-reported current asthma were 20 times more likely to have BHR to methacholine (0.64 versus 0.03) and 3 times more likely to have atopy (0.69 versus 0.25) (3). Results from a study of 2,503 Southampton 7–11 yr olds suggested that both conditions may be importantly associated with asthma. Those with parent reported current wheeze were 2.3 times more likely to exhibit BHR to methacholine but only if they were also atopic (0.83 versus 0.36) and twice as likely to be atopic but only if they

(Received in original form April 13, 1999 and in revised form July 14, 2000) J.R.C. was a clinical research fellow in the Department of Thoracic Medicine, Royal Children’s Hospital, Melbourne, supported by a Kings Fund Travelling Bursary and a Scadding-Morriston Davies Travelling Fellowship. Correspondence and requests for reprints should be addressed to Dr. Mark Jenkins, Centre for Genetic Epidemiology, University of Melbourne, 200 Berkeley St, Carlton, Victoria 3053, Australia. E-mail: [email protected] Am J Respir Crit Care Med Vol 162. pp 2188–2193, 2000 Internet address: www.atsjournals.org

also exhibited BHR (0.92 versus 0.44) compared with children without current wheeze (4). Several studies of familial aggregation have suggested a genetic etiology for each of asthma, atopy, and BHR, and it is likely that several or even many genes, as well as interactions with environmental factors, are involved (5). Further, there is some evidence from studies that have examined the familial aggregation of pairs of these traits in combination to suggest that there may be an overlap of these genetic factors in the etiology of the traits. For self-reported asthma and atopy, a study of 14,000 adult twins in Sweden found that cross-trait crosspair concordance (between one trait in a twin and the other trait in the cotwin) was greater in monozygotic (MZ) than dizygotic (DZ) pairs (6). A similar finding was observed in an Australian study of asthma and hay fever in 7,616 adult twins (7, 8). Under the assumptions of the classic twin model, these studies imply that genetic factors common to both asthma and allergy traits may be implicated (9). There have been no twin studies examining the overlap of genetic factors in the etiology of the three traits, asthma, atopy, and BHR, where the determination of the presence of the traits and zygosity is not reliant on self- or parent-report. The aims of the current study were (i) to measure the associations within MZ and DZ twin pairs, aged 8 to 18 yr old, of physician-diagnosed asthma, atopy defined as allergic reaction to common allergens, and BHR to hypertonic saline, (ii) to analyze the associations within and between traits under the assumptions of the classic twin model, so as to assess the evidence for genetic factors in influencing the etiology of each of these conditions, and (iii) to assess whether these genetic factors could explain the association between these traits in the same individual. That is, this study asks whether there is a common genetic etiology to asthma, BHR, and atopy?

METHODS The study was undertaken in two phases. In Phase 1—the screening survey—parents of twins were mailed a questionnaire to determine the parent-reported asthma status and zygosity of pairs in the initial sampling frame (see below). In Phase 2—the clinical study—all twins responding to Phase 1 were invited to attend the Royal Children’s Hospital where they were interviewed by a respiratory physician for diagnosis of asthma in the past 12 mo, tested for BHR to hypertonic saline, and tested for atopy to common allergens using skin prick tests.

Subjects The initial sampling frame comprised all twin pairs born between 1974 and 1986 inclusive, voluntarily registered by their parents on the Australian Twin Registry at or before June 30, 1994, and living in Victoria when a postal questionnaire was mailed to the parents of all eligible twin pairs in July 1994.

2189

Clarke, Jenkins, Hopper, et al.: Twin Study of Asthma, Atopy, and BHR Phase 1—The Screening Survey Current wheeze was determined as a positive response by the parent(s) to the question: “Has the twin had wheezing or whistling in the chest in the last 12 mo.” Zygosity was determined by asking: “Nonidentical twins are no more alike than ordinary brothers and sisters. Genetically identical twins, on the other hand, look so much alike (that is they have such a strong resemblance to each other in stature, coloring, facial features, etc.) that people often mistake one for the other. Do you think that your twins are genetically identical?”

Phase 2—The Clinical Study Clinical diagnosis of current asthma. Each twin was interviewed by a pediatric respiratory physician (J.R.C.) and asked if he or she had ever experienced any “wheezing or whistling in the chest.” Twins were asked to explain the word “wheeze.” For those who confused wheeze with other respiratory symptoms, such as dyspnea after exertion or stridor from the upper airway, the physician simulated wheezing. Subjects were then asked about the circumstances in which wheeze occurred, whether it occurred at any particular time of the day, or after any triggering factor, such as exercise. Any history of nocturnal cough or sputum production, and smoking habits, were recorded, as were details of any associated atopic diseases, such as hay fever and eczema. The subjects who reported wheeze were asked about any treatment they had received, and their response to therapy. A physician diagnosis of “current asthma” was then defined as a history of wheeze suggestive of a clinical diagnosis of asthma within the past 12 mo. Measurement of BHR. Subjects were asked to withhold the following asthma medications prior to the challenge test: antihistamines for 48 h, theophyllines for 24 h, and inhaled anticholinergics, ␤-agonists, and cromoglycate for 8 h. Baseline forced expiratory volume in 1 s (FEV1) measurements were performed using a spirometer (Flowmate; Jaeger, Germany) until two successive FEV1 readings did not differ by more than 5%. Bronchial challenge was precluded if baseline FEV1 was less than 65% of that predicted for age, sex, height, and weight. The hypertonic saline (HS) challenge tests were carried out according to the protocol developed by Smith and Anderson (10). The nebulizer (Timeter Compuneb Ultrasonic Nebuliser Model MP 500) was connected to a two-way nonrebreathing valve (Hans Rudolph, Kansas City; no. 1400 for children less than 15 yr old, otherwise no. 2700) via 76 cm of corrugated aerosol tubing (internal diameter 2.2 cm, smooth internal surface). The nebulizer canister was filled with 200 ml of 4.5% saline, and the output set to maximum. Previous studies have shown that the output of this nebulizer system, at tidal volumes of 300–500 ml and respiratory rates of 12–20/min, ranges from 1.9 to 2.5 ml/min, with a particle size distribution of mass median aerodynamic diameter between 2.33 and 2.87 ␮m (11). During tidal breathing, commencing with an inhalation time of 30 s, subjects breathed increasing doses of HS by doubling the duration of nebulization until either a 15% fall in FEV1 was observed or the maximum inhalation time of 8 min (representing a cumulative inhalation time of 15.5 min) had been achieved. FEV1 was measured in duplicate 60 s after each step of the challenge. The nebulizer canister and tubing were weighed prior to the challenge, and after the final inhalation step, on an electronic balance (Mettler BB-3000, Switzerland) to assess the total amount of aerosol nebulized and the nebulizer output/min. A subject was defined as responsive to HS if he or she experienced a fall in FEV1 of 15% or more from their baseline value during the challenge. Measurement of atopy. Atopy was assessed from skin prick test reactions to eight common environmental allergens (house dust mite, cat hair dander, dog hair dander, cockroach, mould mix, rye grass, Bermuda grass, and grass pollen mix), to a positive control of histamine (always expected to produce a reaction), and to a negative control of allergen diluent (never expected to produce a reaction). Allergens were applied to the forearm, pricked, and blotted. After 15 min, any resultant weal was measured to determine the average weal size, defined as the mean of its long axis and perpendicular. If there was no response to the positive control, the test was considered incomplete and no classification of atopy was made. If any of the allergens produced a mean weal size of 3 mm or greater than the weal produced by the negative control, the subject was classified as atopic.

Zygosity assessment. Opposite sex twin pairs were defined as DZ. Zygosity testing was performed for pairs of sex-concordant twins. DNA was extracted from Guthrie cards using the Chelex method (12) and genotyped for up to eight highly polymorphic loci, all from different chromosomes: D9S942, TYRP2B, FGF3, D12S356, D6S89, TGFA, D4S192, and D1S214. Fluorescently labeled polymerase chain reaction (PCR) products were then run on an acrylamide gel and visualized on an ABI 373 DNA sequencer. Based on the population frequencies of the polymorphisms, the estimated probability of incorrectly defining a pair as MZ was less than 0.005. For all sex concordant pairs, zygosity was then determined using the following criteria, where available, in a hierarchical system: DNA test (if all identical: MZ; if not identical: DZ; if not done: next criterion), difference in hair/eye color judged by the respiratory physician (if different: DZ; otherwise next criterion), parents’ response to zygosity question on the screening questionnaire (either MZ or DZ; if not done: next criterion), and zygosity as recorded on the Australian Twin Registry, based on parents’ opinion according to the zygosity question asked at the time of registering. Protocols were approved by the Royal Children’s Hospital Ethics in Human Research Committee, and written consent was obtained from the subjects or their parents or guardians.

Statistical Methods The method of analysis of measuring associations between the binary traits (asthma in the last 12 mo, atopy defined by positive skin prick test, and BHR to hypertonic saline) within an individual and between twins of a pair follows that of Hopper and coworkers to which the reader is referred for technical details (8). Association is expressed as an odds ratio (OR). Initial analyses considered each trait separately, and estimated the prevalence and the twin pair association. The difference in association between MZ and DZ pairs was tested, and interpreted according to the classic twin method. For each pair of binary traits, we first considered the association between traits in the same individual. If there was an association, we then considered the correlation between the first trait in one twin and the second trait in the cotwin, and tested if this was greater for MZ pairs. Under the classic twin method, this suggests the existence of genetic factors that influence both traits, and explain at least part of the association between the two traits. The recruitment involved two stages, and our analysis allowed for the different response rates at the second stage of sampling according to their zygosity and wheeze status. Based on the responses by parents to questions on wheeze in the past 12 mo, within each zygosity response rates were calculated separately for three groups: both positive for wheeze, one positive for wheeze, and neither positive for wheeze. A weighted analysis was performed using weights inversely proportional to the response rates in each subgroup. Estimation was performed by generalized linear modeling techniques using STATA (13), and robust estimates of standard errors that took into account the weighting above were calculated using the information-sandwich method (14).

RESULTS Phase 1—The Screening Questionnaire

The screening questionnaire was mailed to the parents of 1262 twin pairs. Questionnaires were returned for 1060 pairs (84% response rate). The breakdown, according to the zygosity report given by parents, was MZ female–female, 200 pairs (19%); MZ male–male, 182 pairs (17%); DZ female–female, 179 pairs (17%); DZ male–male, 165 (16%); DZ female–male, 248 pairs (23%); and unknown zygosity sex concordant, 86 pairs (8%). Figure 1 shows the current wheeze status of the 1049 twin pairs for whom their parents reported current wheeze status. Phase 2—The Clinical Study

The clinical study was attempted on 399 pairs from Phase 1. For the analysis we used data from the 381 twin pairs that completed all tests, and for whom zygosity was determined.

2190

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 162

2000

Figure 1. Flow diagram of responses at the various phases of the study. Zygosity is best estimate based on blood test, physical difference, parents’ perception and twin registry.

As shown in Figure 1, in the Phase 2 clinical study those pairs in which both twins were reported to have current wheeze were overrepresented, and pairs in which neither twin had current wheeze were underrepresented. These differences were accounted for by the method of statistical analysis. According to the final zygosity determination, there were 183 MZ pairs (96 female–female, 87 male–male) and 198 DZ pairs (57 female–female, 54 male–male, 87 female–male). Of the MZ pairs, 13 (7%) were perceived by their parents to be DZ, and for 32 (17%) the parents were unsure. Of the sexconcordant DZ pairs, 12 (11%) were perceived by their parents to be MZ, and for 9 (8%) the parents were unsure. The prevalence of current asthma, atopy and BHR, by sex, is shown in Table 1. BHR was approximately 50% more prevalent (p ⫽ 0.003), and atopy approximately 25% more prevalent (p ⫽ 0.0008), in males compared with females. Overall, approximately one in four twins had current asthma, one in five had BHR, and one in two had atopy. Associations between traits within an individual. Table 2 shows that for each of the three conditions (asthma, atopy, BHR), the presence of any other condition increased the probability of the individual having that condition. The magnitude of the association between pairs of conditions is given by an odds ratio. All odds ratios were independent of zygosity and greater than unity (p ⬍ 0.001). The associations were highest between BHR and both asthma and atopy, and least between asthma and atopy. Associations between traits across a twin pair. Table 3 shows, separately for MZ and DZ pairs, the estimated probability of an individual having a condition (asthma, atopy, or BHR) given that his or her twin has, or has not, another condition. For all conditions and pairs of conditions, the odds ratios

within MZ and DZ pairs were greater than unity (p ⬍ 0.05), and were greater for MZ pairs than for DZ pairs (p ⬍ 0.05). Consider the probability a twin has asthma, conditional on another condition. For example, if he or she has BHR, it is 0.71 (Table 2). If the cotwin has BHR, it is a little less, at 0.64 for MZ pairs, but about half that, at 0.34 for DZ pairs (Table 3). Conditioning instead on atopy, these probabilities are 0.37, 0.37, and 0.29, respectively. The same conditional probabilities of a twin having BHR, conditional on asthma, are 0.59, 0.53, and 0.29, and conditional on atopy are 0.36, 0.35, and 0.27. For atopy, they are 0.84, 0.78, and 0.67 conditional on asthma and 0.94, 0.89, and 0.76 conditional on BHR. Figure 2 illustrates the conditional probabilities within an individual and within pairs. For any two traits, probabilities of each trait were highest when conditional on the presence of the other trait in the same individual. They were next highest conditional on the other trait in the MZ cotwin, which in turn was greater than being conditional on the other trait in the DZ cotwin. All conditional probabilities were higher than the prevalence of the condition in the sample indicating that the

TABLE 2 ESTIMATED PROBABILITY OF AN INDIVIDUAL HAVING TRAIT 2, GIVEN THE PRESENCE OR ABSENCE OF TRAIT 1 FOR MZ AND DZ PAIRS COMBINED* Probability of Being Affected Trait 2 Trait 1

Affection Status

Asthma

Yes No

TABLE 1 PREVALENCE (STANDARD ERROR IN PARENTHESES) OF ASTHMA ATOPY AND BHR Trait Asthma Atopy BHR

Asthma

Males (n ⫽ 375)

Females (n ⫽ 387)

Combined (n ⫽ 762)

0.26 (0.02) 0.65 (0.03) 0.26 (0.03)

0.25 (0.02) 0.51 (0.03) 0.17 (0.02)

0.26 (0.02) 0.58 (0.02) 0.22 (0.02)

Definition of abbreviation: BHR ⫽ bronchial hyperresponsiveness. * Asthma is defined as asthma in the past 12 mo diagnosed by respiratory physician at interview, atopy is defined by a positive skin prick test to any of eight common allergens, and BHR to hypertonic saline, from twins aged 8 to 18 yr.

BHR

Atopy

0.59

0.84

0.09

0.49

1 BHR

Atopy

Yes

0.71

0.94

No

0.13

Yes

0.37

0.36

No

0.10

0.03

2

0.48 3

Definition of abbreviation: BHR ⫽ bronchial hyperresponsiveness. * (1). Asthma and BHR association: OR ⫽ 15.6 (1.00–24.3); (2) asthma and atopy association: OR ⫽ 5.4 (3.6–8.2); (3) BHR and atopy association: OR ⫽ 18.5 (9.0–38.0).

2191

Clarke, Jenkins, Hopper, et al.: Twin Study of Asthma, Atopy, and BHR TABLE 3 ESTIMATED PROBABILITY OF A TWIN HAVING A TRAIT GIVEN THE PRESENCE OR ABSENCE OF THEIR COTWIN HAVING A TRAIT, AND THE ASSOCIATION BETWEEN TRAITS WITHIN EACH TWIN PAIR AS AN ODDS RATIO (95% CONFIDENCE INTERVALS), BY ZYGOSITY* Probabillity of Being Affected Status of Cotwin Trait Asthma

BHR

Atopy

Present or Absent Present Absent OR (95% CI) Relative OR (95% CI) Present Absent OR (95% CI) Relative OR (95% CI) Present Absent OR (95% CI) Relative OR (95% CI)

Asthma

BHR

Atopy

MZ

DZ

MZ

DZ

0.74 0.10 25.6 (11.3–57.8)

0.35 0.22 1.9 (1.0–3.5)

0.53 0.11 9.1 (4.9–16.9)

0.29 0.19 1.7 (1.0–3.0)

13.4 (4.8–37.3) 0.64 0.34 0.16 0.23 9.1 1.7 (4.9–16.9) (1.0–3.0) 5.2 (2.3–12.0) 0.37 0.29 0.14 0.20 3.6 1.7 (2.0–6.5) (1.1–2.6) 2.2 (1.0–4.6)

5.2 (2.3–12.0) 0.63 0.43 0.11 0.15 14.1 4.2 (6.4–31.0) (2.1–8.6) 3.3 (1.2–9.7) 0.35 0.27 0.06 0.12 8.5 2.7 (4.0–18.1) (1.5–4.9) 3.1 (1.2–8.1)

MZ 0.78 0.49 3.6 (2.0–6.5)

DZ 0.67 0.55 1.7 (1.1–2.6)

2.2 (1.0–4.6) 0.89 0.76 0.48 0.54 8.5 2.7 (4.0–18.1) (1.5–4.9) 3.1 (1.2–8.1) 0.82 0.67 0.24 0.46 14.6 2.5 (7.1–30.1) (1.4–4.5) 5.9 2.3–15.1

Definition of abbreviations: BHR ⫽ bronchial hyperresponsiveness; CI ⫽ confidence interval; DZ ⫽ dizygous; MZ ⫽ monozygous; OR ⫽ odd ratio. * Relative OR is the ratio of the MZ odds ratio and the DZ odds ratio.

existence of one trait consistently increased the probability of the other trait in the same individual or the other twin of a pair. The probability of asthma is higher given the presence of BHR compared with the presence of atopy, indicating asthma may be more closely linked to BHR. Likewise the probability of BHR is higher given the presence of asthma compared with the presence of atopy.

DISCUSSION These results are consistent with the hypothesis that genetic factors are involved in the etiology of asthma, atopy, and BHR, and that these three traits may be due to an overlap between the genetic factors involved in each of these three traits. This conclusion is derived from the following observations: (i) for each of asthma, atopy, and BHR, the association between genetically identical twins (MZ) is significantly greater than

between DZ twins, who on average share half their genes from their parents; (ii) for each pair of conditions, the association between the conditions within an individual is strong and independent of the zygosity of the individual; and (iii) for each pair of conditions, the association across twins within an MZ pair is significantly greater than for twins in a DZ pair. These results support the findings of an earlier study of adult Australian twins, based on self-completed questionnaires, which suggested that genetic factors are implicated in both asthma and hay fever (a disease due to an atopic reaction to airborne allergens), and that at least some of these genetic factors are common to both traits (7, 8). Further, they demonstrate that a third trait, BHR, may also have a genetic etiology in common with asthma and atopy. Because asthma is a disease defined by symptoms that appear in episodes and vary in severity, and a natural history where the majority of children with asthma no longer have

Figure 2. Probabilities of each trait given the presence of the other traits in the same individual (Self), in the MZ cotwin (MZ), or in the DZ cotwin (DZ). Also the prevalence of the trait in the twins (Prev).

2192

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

symptoms when adults (15), it is difficult to determine who has asthma and who does not. Rather than relying on a parent report of asthma in the child, we have used a diagnosis by a respiratory physician following a consultation for all children as our “gold standard” for defining current asthma. A Finnish study of 2,483 twin families defined twins as asthmatic if, according to their parents, asthma had been previously diagnosed (16). It is therefore possible that, compared with our study, the Finnish study was studying a more severe form of asthma. This, and possible geographic differences, may have resulted in the differing prevalence rates: 4.7% and 3.1% in boys and girls, respectively, in the Finnish study compared with 26% and 25%, respectively, in our study. Such differences in prevalence can influence measures of association. From Table 1 of the Finnish study, we have calculated the odds ratios (and 95% confidence intervals) for the association between twins of MZ and DZ pairs to be 33.4 (11.6–96.6) for MZ pairs and 7.5 (4.1–13.9) for DZ pairs. The ratio of these two odds ratios, 4.5 (1.3–15.2), is substantially less than the corresponding ratio of 13.4 (4.8–37.3) in our study (see Table 3), indicating that the role of genes on the susceptibility to clinically defined current asthma in the Australian population may be greater than that of parent-reported asthma in the Finnish population. We defined atopy as a positive response to any one of eight common allergens. Therefore we have measured the presence of the allergy following a specific exposure to a trigger (the resultant hypersensitivity in an atopic individual following chronic exposure to the allergens) rather than the broad condition of atopy, defined as the potential to become allergic. However, as the allergens we have used are those most commonly shown to elicit an allergic response in atopic individuals, our measure of allergy is likely to be sensitive to the existence of atopy. Our definition of BHR is based on airways response to hypertonic saline. This has been shown to have short-term repeatability (11), is relatively safe compared with methacholine and histamine, the pharmacological agents more commonly used in previous studies of BHR, and is readily available and inexpensive. Therefore it is likely to be of relevance to further studies of BHR. Twin studies of binary traits have traditionally calculated tetrachoric correlations to estimate, under the classic twin model, the proportion of the variance of an unmeasured underlying normally distributed liability of the trait that is attributable to genetic factors. This has sometimes erroneously been referred to as the heritability of the condition itself (17). We have instead used simple odds ratios to represent trait associations within an individual and between twins of the same pair. We have expressed differences between MZ and DZ associations using a ratio of odds ratios (Table 3). All these ratios were statistically significant (95% confidence intervals not including 1) suggesting that for each pair of traits, MZ twins were more similar to their cotwin than were DZ twins. Under the classic twin model this finding is consistent with asthma, atopy and BHR each having a genetic etiology, and that part of this genetic etiology is common to all three traits. In terms of tetrachoric correlations within an individual, these associations would be asthma–atopy 0.56, asthma–BHR 0.79, atopyBHR 0.78. Across a pair these associations would be asthma– asthma MZ 0.86, DZ 0.24; BHR–BHR MZ 0.77, DZ 0.50; atopy–atopy MZ 0.79, DZ 0.33; asthma–BHR MZ 0.70, DZ 0.20; asthma-atopy MZ 0.45, DZ 0.19; BHR-atopy 0.65, DZ 0.35. We have not tried to interpret these as explaining genetic or environmental components of variation of “liability” because we cannot test the assumptions underlying the classic twin model for binary traits.

VOL 162

2000

Attempts at discrimination between genetic models can be made without making the assumptions of the classic twin model by using the risk ratio method described by Risch (18). For any pair of traits, i and j, let ␭R ⫽ ␭R(i,j) ⫽ P(Yi2 ⫽ 1 | Yj1 ⫽ 1)/P(Yj1 ⫽ 1) within a pair of related individuals where Yi2 is the value of trait i in twin 2 and Yj1 is the value of trait j in twin 1. Under a single locus model with no dominance variance, the value ␭R ⫺ 1 would be expected to decline by half for each degree of relatedness. That is, ⌳(i,j), the ratio of ␭R ⫺ 1 for DZ twin pairs to ␭R ⫺ 1 for MZ twin pairs, should be 0.5. Under more complicated multilocus models ⌳(i,j) is expected to differ from 0.5. For example, under a two-locus multiplicative model in which both loci are required for disease, ⌳(i,j) is expected to be approximately 0.33. Applying this method to this data for the same traits within a pair we calculated ⌳(i,j) to be 0.19 for asthma, 0.51 for BHR, and 0.37 for atopy. Therefore the within-pair associations are consistent with a single locus additive model for BHR but more complicated multilocus models for asthma and atopy. If this can be extended to crosstrait associations, then ⌳(i,j) was 0.21 for asthma and BHR, 0.23 for BHR and asthma, 0.36 for asthma and atopy, 0.45 for atopy and asthma, 0.38 for BHR and atopy, and 0.51 for atopy and BHR, indicating that except for atopy and BHR, the cross-trait associations are consistent with multilocus models. There has been a strong recent interest in searching for susceptibility genes for asthma, atopy, and BHR, using either genomic searches to identify chromosomal regions linked with the conditions or candidate gene approaches, which screen specific chromosomal regions that contain mapped genes that may be involved in the disease, based on their known function. A major difficulty for both these approaches is the difference in definition of the phenotypes used by different studies. For example, using a genomic search approach, Cookson and coworkers reported linkage of the chromosome 11 marker D11S97 to atopy (defined as positive skin test or RAST or elevated total serum immunoglobulin E [IgE]) (19), which led to the identification of the locus encoding the ␤ chain of the highaffinity IgE receptor, Fc␧RI␤ (20). However, an Australian study found linkage of the same gene with BHR to methacholine, even in the absence of atopy (21). Studies of candidate genes on chromosome 5q31–5q33, including those that code for the TH2 cytokines, have observed linkage with total serum IgE, but not with specific IgE (22). The majority of linkage studies have used the phenotype atopy, and only a few have considered asthma and BHR. Linkages have been established for the allergen-specific IgE with the region 6p21-3; total serum IgE and BHR with the region 5q31-33; atopy, BHR, total serum IgE, and asthma with the region 11q13; atopy with the marker D13S153 on chromosome 13; atopy and asthma with the marker D16S289 on chromosome 16; total serum IgE and atopy with the marker D4S426 on chromosome 4; and specific IgE and total IgE with markers on chromosome 14 (23). If, as our data show, there may be some genes that have etiological roles in asthma and atopy, asthma and BHR, and BHR and atopy, or even asthma, BHR, and atopy, then a locus cannot be ruled out as being etiologically important unless all three conditions are tested for linkage at that locus. In a recent review of the genetic epidemiology of asthma, it was concluded that there was sufficient evidence for the existence a gene on 11q, exerting influence on atopy, asthma, and BHR (5). As asthma and allergy are likely to be genetically heterogeneous diseases, it is likely that other genes, such as those on 5q31-33, may also influence the etiology of all three traits. There has been much discussion about the definition of appropriate phenotypes in segregation and linkage studies for

2193

Clarke, Jenkins, Hopper, et al.: Twin Study of Asthma, Atopy, and BHR

asthma and related diseases (24). If the definition is broad, such as “nonspecific allergy,” then segregation and linkage studies will have greater power to detect genes that when mutated are causes for a variety of conditions that comprise the broad definition. There will be less power, however, to detect genes specific to those conditions that comprise a minor constituent. If instead, a narrow phenotype definition is used, such as “allergy to cat hair,” the power to detect a gene involved only in allergy to cats is increased if the same number of “affecteds” are studied. However, as only a small proportion of atopic individuals is allergic to cats, these families may be hard to find. Our study has shown that a substantial proportion of the susceptibility to asthma, atopy, and BHR is likely to be attributable to a group of genes that when mutated increase risk on all three conditions. Therefore restricting phenotype definition to one of these traits will lead to reduced power to detect genes that affect susceptibility to the three conditions. In conclusion, under the assumptions of the classic twin model, these data suggest that (i) genetic factors play a role in determining susceptibility to asthma, atopy, and BHR, and (ii) most of the cross-sectional association between these three traits could be due to an overlap between the genetic factors involved in each of these traits. References 1. Robertson CF, Heycock E, Bishop J, Nolan T, Olinsky A, Phelan PD. Prevalence of asthma in Melbourne school children: changes over 26 years. Br Med J 1991;302:1116–1118. 2. Dalton MF, Powell CVE. Ordonez G, Robertson CF. Bronchial hyperresponsiveness to 4.5% NaCl is related to the clinical expression of asthma [abstract]. Am J Respir Crit Care Med 1998;157:A543. 3. Shaw RA, Crane J, O’Donnell TV. Asthma symptoms, bronchial hyperresponsiveness and atopy in a Maori and European adolescent population. NZ Med J 1991;104:175–191. 4. Clifford RD, Howell JB, Radford M, Holgate ST. Associations between respiratory symptoms, bronchial response to methacholine, and atopy in two age groups of school children. Arch Dis Child 1989;64:1133–1139. 5. Duffy D. The genetic epidemiology of asthma. Epidemiol Rev 1997;19: 129–143. 6. Edfors-Lubbs ML. Allergy in 7000 twin pairs. Acta Allergol 1971;26:249– 285. 7. Duffy DL, Martin NG, Battistutta D, Hopper JL, Mathews JD. Genetics

8.

9. 10. 11.

12.

13. 14. 15.

16.

17. 18. 19.

20.

21.

22.

23. 24.

of asthma and hay fever in Australian twins. Am Rev Respir Dis 1990; 142:1351–1358. Hopper JL, Hannah MC, Macaskill GT, Mathews JD. Twin concordance for a binary trait: III. A bivariate analysis of hay fever and asthma. Genet Epidemiol 1990;7:277–289. Hopper JL. Commentary: genes for osteoarthritis: interpreting twin data. Br Med J 1996;312:943–944. Smith CM, Anderson SD. Hyperosmolarity as the stimulus to asthma induced by hyperventilation? J Allergy Clin Immunol 1986;77:729–736. Riedler J, Reade T, Robertson CF. Repeatability to the response of 4.5% saline challenge in children with mild to severe asthma. Pediatr Pulmonol 1994;18:330–336. Walsh PS, Metzger DA, Higuchi R. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques 1991;10:506–513. StataCorp. Stata Statistical Software: Release 5.0. College Station, TX: Stata Corporation; 1997. Royall RM. Model robust inference using maximum likelihood estimators. Int Statist Rev 1986;54:221–226. Jenkins MA, Hopper JL, Bowes G, Carlin JB, Flander LB, Giles GG. Factors in childhood as predictors of asthma in adult life. Br Med J 1994;309:90–93. Laitinen T, Räsänen M, Kaprio J, Koskenvuo M, Laitinen L. Importance of genetic factors in adolescent asthma: a population-based twin-family study. Am J Respir Crit Care Med 1998;157:1073–1078. Hopper JL. Heritability. In: Armitage P, Colton T, editors. Encyclopedia of biostatistics. New York: Wiley; 1998. p. 1905–1906. Risch N. Linkage strategies for genetically complex traits: I. Multilocus models. Am J Hum Genet 1990;46:222–228. Cookson WOCM, Sharp PA, Faux JA, Hopkin JM. Linkage between immunoglobulin E responses underlying asthma and rhinitis and chromosome 11q. Lancet 1989;1:1292–1294. Sandford AJ, Shirakawa T, Moffatt MF, Daniels SE, Ra C, Faux JA, Young RP, Nakamura Y, Lathrop GM, Cookson WOCM, Hopkin JM. Localisation of atopy and beta subunit of high-affinity IgE receptor on chromosome 11q. Lancet 1993;341:332–334. van Herwerden L, Harrap SB, Wong ZY, Abramson MJ, Kutin JJ, Forbes AB, Raven J, Lanigan A, Walters EH. Linkage of high affinity IgE receptor gene with bronchial hyperreactivity even in the absence of atopy. Lancet 1995;346:1262–1268. Marsh DG, Neely JD, Breazeale DR, Ghosh B, Freidhoff LR, EhrlichKautzky E, Schou C, Krishnaswamy G, Beaty TH. Linkage analysis of IL-4 and other chromosome 5q31.1 markers and total serum IgE concentration. Science 1994;264:1152–1156. Ruffilli A, Bonini S. Susceptibility genes for allergy and asthma. Allergy 1997;52:256–273. Sandford AJ, Weir TD, Pare PD. Phenotypic heterogeneity in asthma and allergy. Clin Exp Allergy 1998;28(Suppl 1):26–31.