INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Because the prevalence and mortality associated with asthma and chronic obstructive pulmonary disease (COPD) have increased worldwide (1), it is important to investigate the underlying genetic processes and pathophysiologic characteristics that are associated with host susceptibility and the development of these disorders. The investigation of genetic mechanisms should ultimately lead to a better understanding of the pathogenesis and pathophysiology of asthma and other obstructive airway diseases as well as to improved therapeutic and preventive measures (5, 6). However, it is difficult to compare and interpret the results of genetic studies without developing a standardized approach for the differential diagnosis of subjects with various forms of obstructive lung disease as well as a logical approach to classify individuals with some of the characteristics or early evidence of asthma.

Asthma and COPD are common diseases with a heterogeneous expression and may be difficult to differentiate from each other, especially in older persons (3, 7). Asthma produces intermittent or chronic symptoms caused by episodic airway obstruction. There appears to be a genetic or familial basis (10), but expression of the disease may be determined by environmental triggers that include allergic, infectious (viral), or other environmental exposures (17). The relative importance of each of these different modalities in the development and progression of asthma are not completely understood. In contrast, COPD is usually differentiated from asthma by a significant history of cigarette smoking and the presence of progressive, irreversible airway obstruction. However, some cigarette smokers have a clinical syndrome characterized by intermittent symptoms and bronchial hyperresponsiveness that has many similarities to asthma (25, 26). Some asthmatics, especially those in older age groups, have fixed or irreversible airway obstruction (27). Furthermore, only a minority of smokers develop severe symptomatic COPD, demonstrating the importance of individual susceptibility even in these subjects to the respiratory effects of cigarette smoke (4, 28). Thus, a possible genetic component that regulates individual susceptibility in both of these disorders may be an important risk factor for disease development and progression (32, 33). In order to study the epidemiology and familial aggregation of obstructive airway diseases, it is important to distinguish between the various forms of obstructive airway disease.

The overall purpose of this study was to develop a method to diagnose and differentiate patients with asthma from those with other obstructive airway diseases and healthy subjects. This approach is the initial step in the investigation of the role of a genetic component in the development of asthma since accurate phenotype assessment is necessary for genetic epidemiology and positional cloning (34). To accomplish this goal, we studied the members of 92 families ascertained through a proband, a parent with classic asthma first studied between 1962 and 1970 and reexamined in the present study. This allowed us to evaluate the effects of age and smoking on the diagnosis of asthma in the parent. In the current study, a comprehensive evaluation was performed on all family members that included pulmonary function testing, histamine bronchial provocation, allergen skin testing, and a standardized respiratory questionnaire. An algorithm (flow diagram) was developed based on clinical and physiologic parameters that are considered to characterize the asthma phenotype (27). The algorithm consisted of five classes that were used to differentiate asthma from other forms of obstructive airway disease and healthy subjects. Subsequently, all family members were then classified according to this algorithm. The presence of a prior physician diagnosis or therapy for asthma was compared with current clinical evidence of asthma. These data were examined to determine the usefulness of a prior physician's diagnosis to assess the presence of asthma and to assess whether there is familial clustering of asthma that could be consistent with a genetic component in this disorder.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

All probands in the present study were initially studied between 1962 and 1970 at Beatrixoord Hospital, Haren, The Netherlands, a regional asthma referral center. During that period of time, patients with symptomatic asthma without a current asthma exacerbation were referred to this hospital and admitted for 6 wk for a standardized, complete evaluation. At the time of initial testing, all probands had asthma symptoms, were hyperresponsive to histamine (PC₂₀ FEV₁ histamine  32 mg/ml) (22), and were younger than 45 yr of age. In the current study, 92 probands with their spouses, children, children's spouses, and grandchildren older than 7 yr of age were recruited and evaluated. This study was approved by the Medical Ethics Committee of the University Hospital Groningen, and all participants signed an informed written consent.

Questionnaire

At the first visit, 25 yr ago, all patients answered a validated Dutch version of the British Medical Society Respiratory Questionnaire (35). For current testing this questionnaire was again used with the addition of questions covering the following areas: a history of a previous physician's diagnosis or therapy of asthma, the presence of intermittent asthma symptoms associated with known asthma triggers, e.g., exposure to exercise, cold air, allergens, irritants, etc., as well as a more detailed smoking history. For children younger than 16 yr of age, living at home, the mother was asked to complete an extended respiratory symptom questionnaire for children with the addition of similar questions as outlined above.

Spriometry and Bronchodilator Testing

All pulmonary medication was withheld before testing, i.e., inhaled bronchodilators for at least 8 h, theophylline, and antihistamines at least 24 h, and inhaled corticosteroids at least 14 d. Participants had to be in a stable condition without a history of an exacerbation of asthma during the previous 6 wk. Spirometry was performed in a similar manner during initial and current testing. Two valid measurements of FEV₁ and IVC (inspiratory slow vital capacity) were obtained, using a water sealed spirometer (Lode Spirograph type DL; Lode b.v., Groningen, The Netherlands). The FEV₁ values had to be within 3% to be considered valid, and the highest value was recorded (36).

At the time of initial testing, reversibility was tested by repeating spirometry 30 min after an intramuscular injection of 25 mg Multergan (a potent anticholinergic drug with antihistaminic properties). During current testing reversibility was assessed 20 min after inhalation of 800 µg salbutamol (albuterol) administered with a spacehaler. This was performed after the histamine challenge test; FEV₁ had returned to within 5% of baseline in 99% of the participants at the time of bronchial challenge testing. Reversibility was defined as a difference of at least 9% in FEV₁% predicted (37).

Bronchial Responsiveness Testing

At the time of initial and recent testing, bronchial responsiveness was measured using the 30-s inhalation histamine challenge test using methods developed by De Vries and colleagues (22). Doubling concentrations of histamine phosphate were inhaled (0.03 to 32 mg/ml in children younger than 12 yr of age, and 0.5 to 32 mg/ml in adults). The test was stopped if there was a greater than or equal to 20% fall in FEV₁. The final concentration of 32 mg/ml histamine in this method is regularly used to distinguish between bronchial hyperresponsiveness (BHR) and unaffected (22, 38). When FEV₁ was below 1.20 L, histamine bronchial challenge testing was not performed. This occurred in 16 of the probands during current testing.

Allergy Testing

At the time of initial testing, intracutaneous allergy skin tests were performed with house dust, mixed molds, mixed grass pollens, mixed tree pollens, mixed spring pollens, mixed weeds, mixed animal dander and feathers, and hay dust. During current testing, all probands and family members (older than 12 yr of age) had intracutaneous tests with 16 common aeroallergens: mixed grass pollens*, two mixed tree pollens*, mixed weeds*, house dust mite* (Dermatophagoides pteronyssinus), storage mites (Lepidoglyphus destructor, Tyrophagus putrescentiae, Acarus siro), cat-*, dog*-, horse-*, rabbit/guinea-pig dander*, feather mix, and five molds (Aspergillus fumigatus*, Alernaria alternata*, Cladosporium herbarum, Penicillum notatum, Botrytis cineria). In children younger than 12 yr of age, a skin prick test was performed with the 10 allergens marked with an asterisk. A positive skin test control using histamine and a negative control using the diluent were always performed. Intradermal tests were considered as positive if the diameter of the weal was at least 5 mm (mean of the larger diameter and the perpendicular diameter), and prick tests were considered positive if the weal diameter was at least 2 mm. Subjects with at least one positive skin test were considered to be atopic. Total serum IgE (IU) was measured by solid-phase immunoassay (Pharmacia IgE EIA; Pharmacia Diagnostics AB, Uppsala, Sweden). The mean of duplicate tests was used. The values of the two measurements had to be within 5% to be considered valid. In two cases this level of reproducibility was not found and IgE measurements were repeated.

Algorithm

An algorithm to differentiate the asthmatic phenotype from other obstructive airway diseases and from normal unaffected family members was developed, on the basis of the following five clinical or laboratory findings: (1) BHR, (2) cigarette smoking, (3) asthma symptoms, (4) airway obstruction, and (5) reversibility to a bronchodilator (Figure 1). All participants were divided into the following five classes (Figure 2). Class 1: definite asthma; Class 2: probable asthma; Class 3: unclassified airway disease; Class 4: COPD; Class 5: unaffected, i.e., no evidence for obstructive airway disease.

View larger version (23K): [in this window] [in a new window]   Figure 1.   The sequence variables in the algorithm for the differential diagnosis of asthma and COPD.

View larger version (34K): [in this window] [in a new window]   Figure 2.   The asthma algorithm applied for classifying research subjects. BHR = bronchial hyperresponsiveness defined as PC20  32 mg/ ml histamine. PY = amount of pack-years; Symp = number of pulmonary symptoms reported; Obstr = airway obstruction defined as FEV1  95% CI; Rev = reversibility of airway obstruction defined as increase in FEV1  9% of FEV1 predicted after use of a bronchodilator. 1 = Class 1, asthma; 2 = Class 2, probable asthma; 3 = Class 3, unclassifiable airway disease; 4 = Class 4, COPD; 5 = Class 5, unaffected.

After reviewing the American Thoracic Society (ATS) guidelines for defining asthma (15), we chose the five parameters used in the algorithm (Figures 1 and 2). The ATS definition (1987) defines asthma as follows. "A clinical syndrome characterized by increasing responsiveness of the tracheobronchial tree to a variety of stimuli. Major symptoms are mild to severe and unremitting. The primary physiologic manifestation of hyperresponsiveness is variable airways obstruction, which reverses spontaneously or after treatment or increased obstruction caused by drugs or other stimuli. . . . evidence of mucosal edema of the bronchi, infiltration of the bronchial mucosa or submucosa with inflammatory cells, especially eosinophils; shedding of epithelium, obstruction of peripheral airways with mucus." Bronchial hyperresponsiveness, airway obstruction, and reversibility, which are components of this definition, can be objectively measured, whereas the symptoms of paroxysmal dyspnea, wheezing, and cough can be approximated by the use of a standardized questionnaire. Therefore, these parameters were used to develop this algorithm. In a family-based study such as this it was not feasible or reasonable to obtain bronchial alveolar lavage fluid or bronchial biopsies to directly assess inflammatory processes in the airways.

The following decision points were used in the development of this algorithm (Figures 1 and 2): The first step was BHR ( 32 mg/ml histamine) since BHR is widely accepted as a sensitive test for asthma (43) but is found in other obstructive diseases (25). Subjects with BHR were not considered to have normal airway function and, thus, were not classified as "unaffected" (Class 5). Subjects without BHR were not classified as "definite asthma" (Class 1).

Smoking was the second step as it is strongly associated with respiratory symptoms and the development of COPD (45). A subject with a history of more than 5 pack-years of cigarette smoking was thought to have a significant smoking history and was not classified as "definite asthma" (Class 1), the only exception was made in subjects with asthma symptoms and/or asthma attacks that preceded the onset of cigarette smoking. Thus, subjects with BHR and more than 5 pack-yr of cigarette smoking could fall into Classes 2 to 4 depending on the presence or absence of specific characteristics associated with asthma (BHR, symptoms, reversible airway obstruction). Likewise, nonsmokers were not classified as "COPD" (Class 4).

The third step was the assessment of symptoms by dividing the questionnaire into the following symptom groups: cough, dyspnea, wheeze, and nocturnal symptoms. The presence of one symptom group in subjects younger than 16 yr of age or two symptom groups in subjects 16 yr of age or older, were thought to be compatible with asthma if associated with the other criteria for asthma. A clear history of recurrent asthma attacks was considered strong evidence of asthma and was given the same significance as positive answers to two symptom groups.

Although airway obstruction (FEV₁  95% confidence interval of FEV₁% predicted) may not be present in all patients with asthma, it is an important component in the diagnosis of COPD. Thus, the absence of airway obstruction did not exclude asthma if an individual had BHR and asthma symptoms. The presence of airway obstruction is consistent with either one of these disorders. However, in asthma, airway obstruction should be reversible (increase in FEV₁  9% predicted) (37), which is the fifth or final step in the algorithm. In COPD, there may be no or only a small degree of reversibility after the administration of a bronchodilator (27, 37, 45).

Statistical Analyses

The data were analyzed using the computer program SPSS/PC+ (SPSS Inc., Chicago, IL). Results were expressed as mean ± standard deviation, or as the median with the range if the variable was not normally distributed. PC₂₀FEV₁ slope and IgE levels were log transformed and the results were presented as geometric means. Comparisons between groups were made using Student's t test, and p values < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Characteristics

We tested 534 subjects in 92 families: 92 probands, 92 spouses of probands, 265 first-degree offspring, 30 spouses of first- degree offspring, and 55 second-degree relatives (Table 1). Twenty-five family members, including 20 probands and five first-degree offspring were unable to stop oral or inhaled corticosteroids prior to testing. Twenty-three of these 25 subjects were hyperresponsive when tested. The majority of family members, including the probands, meeting criteria for definite asthma were found to be allergic.

Probands at Initial and Current Testing

There were more male than female probands. Although all probands were hyperresponsive to histamine initially, nine probands (10%) had no measurable bronchial hyperresponsiveness during current testing. All nine reported asthma symptoms, and two had airway obstruction. In 16 probands, baseline lung function was too low to safely permit histamine bronchial challenge testing (FEV₁ < 1.2 L), and in one, spirometry meeting ATS standards could not be performed (36). In two of these 16 subjects airway obstruction was not reversible with salbutamol.

Spouses

Fifteen of the 92 spouses (16%) had BHR (Table 1). Eight of these 15 reported asthma symptoms, and four had airway obstruction, in one of the spouses it was not reversible after the use of a bronchodilator. Three of these 15 spouses with BHR reported asthma attacks and a prior physician's diagnosis of asthma. In the spouses of offspring (n = 30), seven had BHR, two reported asthma attacks, and one had a prior diagnosis of asthma. Thus, the probands were the primary affected parent with asthma or an asthma-associated phenotype in these families.

Offspring

The first-degree offspring were of comparable age to their asthmatic parents at the time of initial testing (mean age, 25 and 27 yr, respectively) (Table 1). BHR was present in 93 (35%), 31 of these reported no pulmonary symptoms, whereas another 31 reported a prior physician's diagnosis of asthma. In the 265 first-degree offspring, 49 showed airway obstruction at the time of testing, 25 of them had BHR and 15 of them reported a physician's diagnosis of asthma. Eighteen of these subjects with airway obstruction did not report pulmonary symptoms and 30 did not have reversibility of their airway obstruction. In the second-degree offspring, 20 subjects showed BHR (36%), two reported a physician's diagnosis of asthma, and three had airway obstruction with asthma symptoms.

Allergy Status

Initially, 91% of the probands had at least one positive allergy skin test, whereas during current testing, approximately 25 yr later, only 76% showed at least one positive skin test (Table 1). In the first-degree offspring 54% were skin-test-positive, whereas in the second-degree offspring 29% were skin-test-positive, and 30 and 23% of the spouses of probands and of offspring, respectively, were skin-test-positive. Three of the eight probands who were initially skin-test-negative had a positive skin test when restudied. The geometric mean of the total serum IgE levels was higher in the probands (85.1 IU) and the first- and second-degree offspring (60.3 and 61.7 IU, respectively) than in the spouses of the probands and offspring (25.1 and 35.5 IU, respectively). At the time of initial testing, total serum IgE was not measured, so comparison with the current test results was not possible.

Algorithm

The 92 probands were classified twice with the algorithm using data from initial and current testing. The initial data were used to evaluate the confounding effect of increasing age and cigarette smoking. Probands with an FEV₁ too low to permit bronchial challenge testing during current testing were assumed to have BHR since they displayed BHR during initial testing. The results of the classification of the probands and spouses using the algorithm are reported in Table 2. Using the initial lung function and questionnaire data: 80 probands were classified as Class 1 (definite asthma), nine as Class 2 (probable asthma), two as Class 3 (uncertain), and one as Class 4 (COPD). Classification of the current data on the probands revealed an increased number of probands assigned to Class 3 (uncertain), which was due to three factors: loss of BHR, increase in number of pack-years of smoking, and/or the development of irreversible airway obstruction.

Although most spouses were nonasthmatic (Class 5), five spouses met criteria for Class 1 (definite asthma), and two of these reported a physician's diagnosis of asthma. Eight spouses met criteria for Class 2 (possible asthma). Six of these had BHR, five reported respiratory symptoms, and four had a prior physician's diagnosis of asthma. Of the four spouses with a physician's diagnosis, two with BHR reported asthma attacks, but they were smokers prior to the onset of asthma symptoms. The other two reported attacks but did not have BHR. One of the 81 spouses classified as unaffected (Class 5) reported a prior physician's diagnosis of asthma but did not have BHR, airway obstruction, or reported asthma attacks.

In the 320 first- and second-degree offspring, 113 had BHR. Most of these offspring were nonsmokers and were classified as Class 1 to 3 depending on symptoms (Figure 3). Five of the nine offspring with BHR had a history of smoking and were classified as Class 3 (uncertain), whereas the others were divided between Classes 1, 2, and 4 depending on onset of symptoms in relationship to smoking and airway obstruction. Seventy-six percent (n = 157) of the 207 offspring without BHR were classified as Class 5 (unaffected) (Figure 4). The others were divided between Class 3 or Class 4, except for nine offspring assigned to Class 2. 

View larger version (15K): [in this window] [in a new window]   Figure 3.   A diagram showing the classification of the 320 first- and second-degree relatives from 92 families after the decision steps outlined in the asthma algorithm.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Asthma classification of first- and second-degree offspring of 92 asthmatic probands. Class 1: definite asthma; Class 2: probable asthma; Class 3: unclassifiable airway disease; Class 4: COPD; Class 5: unaffected.

The relationship between a physician's diagnosis of asthma and the presence or absence of BHR in the 320 first- and second-degree relatives is shown in Figure 5. Of the 33 offspring with a physician's diagnosis of asthma and BHR, 32 were classified as "asthma" (Class 1) and one as "uncertain obstructive airway disease" (Class 3) because of a significant smoking history. The fifteen subjects without BHR but with a physician's diagnosis of asthma all reported symptoms. Four of these were classified as unaffected (Class 5) because they had no evidence of pulmonary disease, although three were smokers. All of the remaining 11 were classified as having some evidence of lung disease (three in Class 2, six in Class 3, and four in Class 4). There were 192 offspring without a physician's diagnosis of asthma or BHR (Figure 5). The majority of these were classified as unaffected (n = 153). It is of interest that there were 80 subjects with BHR and no physician's diagnosis of asthma. Twenty-one of these met criteria for Class 1 (definite asthma) and 20 met criteria for Class 2 (probable asthma). Two were classified as COPD (Class 4) and 37 as uncertain airway disease (Class 3).

View larger version (12K): [in this window] [in a new window]   Figure 5.   Comparison of a physician's diagnosis of asthma with a positive or negative histamine challenge test (BHR) in the 320 first- and second-degree relatives.

In the offspring classified as "definite asthma" (Class 1), the groups with and without a prior physician's diagnosis of asthma were compared (Table 3). The group with a prior physician's diagnosis of asthma showed a lower geometric mean of the PC₂₀ than did the group without a prior diagnosis. There were no other significant differences between the two groups. In the offspring with a prior physician's diagnosis of asthma, differences between the two groups with and without BHR were analyzed (Table 4). A larger percent of the group without BHR had a history of > 5 pack-yr of cigarette smoking (p < 0.01) and showed significantly less evidence for atopy (p < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Epidemiologic and genetic studies in asthma require accurate separation of subjects with asthma from those with other obstructive airway diseases such as COPD as well as from "healthy" or unaffected subjects. This approach is important since misclassification may underestimate or overestimate the prevalence of asthma. This study demonstrates that it is feasible to develop a logical method to diagnose and characterize patients with asthma, that is, based on both objective and subjective measures.

Asthma has previously been defined on the basis of clinical information obtained from standardized questionnaires, or by a prior physician's diagnosis of asthma. However, it is important to develop a more comprehensive approach based on all available clinical information, including subjective (questionnaire data) and objective measures (pulmonary function testing) to define the asthma phenotype. A true validation of the algorithm is not possible since there is no universally accepted standardized method for validation. Because physician's diagnosis of asthma was not used for classification, the algorithm was not biased by prior diagnostic labels. This approach also provided the opportunity to be able to compare the characteristics of subjects in the asthma group with and without a prior physician's diagnosis of asthma. There were a large number of offspring with clinical evidence of asthma who did not have a prior physician's diagnosis of asthma. Even though they had symptoms and/or bronchial hyperresponsiveness, they denied having asthma, and had not been seen by a physician for care. Some subjects compared their incidental dyspnea with the dyspnea of their affected parent (the proband) and thought their own symptoms insignificant. One even used his parent's asthma medication, but did not consider he had asthma. Comparison of the clinical data on the offspring with and without a prior physician's diagnosis of asthma in Class 1 (definite asthma group) demonstrates no major differences (except for an increase in the level of BHR) and considerable overlap in these two subgroups (Table 3). Furthermore, those offspring with a prior physician's diagnosis of asthma and no evidence of bronchial hyperresponsiveness had a greater frequency of cigarette smoking, had a lower total serum IgE level, and were less likely to have a positive allergy skin test (Table 4). It is possible that symptoms related to cigarette smoking formed the bases for their diagnosis of asthma. Thus, a physician's diagnosis of asthma seems to be a less valid method for epidemiologic and genetic purposes since it may result in underestimation or overestimation of the true prevalence of this disorder. The approaches outlined in this report to characterize obstructive airway disease are based on objective assessments that can be utilized in ongoing or future studies. These approaches should more accurately assess the prevalence and incidence of asthma.

This algorithm is based on the presence or absence of asthmatic symptoms, cigarette smoking, as well as objective laboratory testing, including spirometry with reversibility testing and bronchial responsiveness to histamine. In the algorithm, the parameters age, sex, and total serum IgE levels and positive skin tests were not included because they did not affect the ultimate classification. In addition, avoiding these parameters prevents a potential bias since patients tend to be labeled differently according to social class, age, sex, smoking, and allergic status (46). Physicians are more likely to label older men who are cigarette smokers with emphysema, women with asthma or bronchitis (49), and to favor a diagnosis of asthma when an allergic component is evident. However, one cannot deny or confirm asthma based on the allergic status of an individual patient, especially not in older patients. A skin-test-negative subject in this study meeting all criteria for asthma (Class 1) would still be classified as asthmatic. The latter is supported by the observation that 17 probands "lost" atopy based on skin test results after a period of 25 yr. Thus, these subjects (probands) would have been considered intrinsic or nonatopic asthmatics if their evaluation was based only on the results of the current evaluation.

Age was not used in this decision tree, but it was included by equating one symptom in a child younger than 16 yr of age with two symptoms in adults. The rationale behind the leniency in judging the symptoms in the group  16 yr of age is that although children may have bronchial hyperresponsiveness they often report only one symptom such as cough as a primary symptom of stoma (11, 27, 41). Nevertheless, age is important in the clinical expression of asthma. An advantage of this data set was the ability to evaluate some of the longitudinal effects of age on the clinical characteristics associated with asthma and atopy since the age of the probands at the time of their first evaluation was approximately the same as the age of the first-generation offspring when they were tested. If the initial data on the probands had not been available, in some of them asthma would not have been diagnosed during current testing. Of the initial 80 subjects in Class 1, 67 (84%) remained classified as definite asthma (Class 1), and 11 would not have been classified as asthmatic because of the confounding effect of cigarette smoking.

Bronchial hyperresponsiveness generally correlates with the presence and the severity of pulmonary symptoms in community-based populations (50). These data confirm observations by Rijcken and coworkers (39, 41) that many asymptomatic subjects may have BHR. In this study, 41 subjects with BHR (36%) were asymptomatic. Thus, the subjects without a history of asthma attacks or other pulmonary symptoms were not classified as "definite asthma" (Class 1), nor could they be considered unaffected (Class 5) since it has been shown that subjects with asymptomatic BHR may develop asthma in later life (51). Therefore, asymptomatic subjects with BHR, without reversibility after an inhaled bronchodilator were classified as "uncertain" (Class 3). Subjects who showed reversibility were classified as "probable asthma" (Class 2). Finally, subjects with irreversible airway obstruction were classified as "COPD" (Class 4) if they smoked cigarettes, and as "uncertain" (Class 3) if they did not smoke. The presence or absence of BHR did not affect this classification. Some subjects without BHR had airway obstruction with significant reversibility. It is difficult to classify these subjects as "definite asthma", although some of them may develop symptomatic asthma or COPD at a later stage of life (51). Therefore, it seems that longitudinal studies provide important information that will further improve the accuracy of current diagnostic approaches in asthma.

The increased frequency of bronchial hyperresponsiveness, high total serum IgE levels, and allergy skin tests in family members are consistent with previous observations that these phenotypes that are closely associated with asthma are more common in monozygotic twins than in dizygotic twins (52), and show familial aggregation (53). However, many of these investigators have relied primarily on questionnaires, a more subjective approach, to assess symptoms and the presence of asthma or atopy (56). The results reported in this study are based on objective testing that was performed on all of the family members. They demonstrate familial aggregation of asthma as well as of other phenotypes closely associated with asthma. Twenty-six percent of the offspring had a diagnosis of definite or probable asthma (Class 1 and Class 2), which is consistent with a genetic component of asthma in these families. There would have been underdiagnosis of this disorder if we had relied on a report of a prior physician's diagnosis of asthma and misclassification of the probands if longitudinal changes in disease expression were not taken into account.

In summary, all of the members of 92 families were classified using an algorithm that was developed to distinguish the presence or absence of obstructive airway disease. Only 49% of the offspring met the criteria for "unaffected" (Class 5), and had no evidence of obstructive or reactive airway disease. The large number of offspring with clinical evidence of asthma, but without a prior physician's diagnosis, suggests underdiagnosis of asthma in families of patients with asthma. In addition, young adults with asthma, when restudied after 25 yr, may develop changes in pulmonary status that makes them less easily classified as "typical" asthma since some of them no longer have asthmatic characteristics, or have a significant history of cigarette smoking (> 5 pack-yr), a confounding factor in the diagnosis of asthma. The high familial clustering of asthma and other obstructive airway diseases in these families suggests a strong genetic component. Therefore, additional genetic studies that include linkage and positional cloning are indicated, and they should provide more detailed information on the genetic basis of asthma.