- Email: [email protected]
Findings before diagnoses of asthma among the elderly in a longitudinal study of a general population sample
Findings before diagnoses the elderly in a longitudinal population sample
of asthma among study of a general
Benjamin Burrows, MD, Michael D. Lebowitz, and Martha G. Cline, MS Tucson, Ariz.
PhD, Robert
A. Barbee, MD,
Forty elderly subjects who denied ever having asthma or emphysema on enrollment in a longitudinal epidemiologic study later reported consulting a doctor for asthma when they were older than 60 years of age. The average age at which the diagnosis was reported was 70.8 years, after a mean follow-up of 8.5 years. Findings on enrollment in the newly diagnosed subjects with asthma are compared with jindings in the 1145 subjects who provided follow-up information when they were older than age 60 years but had never developed asthma. At the time of enrollment, most subjects later diagnosed as having asthma already had wheezing symptoms, suggesting at least a mild asthmatic state, and many subjects had impaired ventilatory function, a positive allergy skin test (especially in association with rhinitis), and blood eosinophilia. Thirty-jive percent of the subjects recalled “respiratory trouble before age 16” despite denying prior asthma. The likelihood of a new asthma label was very closely related to the age-sex-standardized serum-IgE level before diagnosis. Newly diagnosed subjects with asthma demonstrated much greater rates of decline in FEV, than control subjects or than subjects who already had known asthma on enrollment. We conclude that (1) symptoms suggesting asthma are usually present for many years before the diagnosis of the disease in elderly subjects, (2) the serum-IgE level is closely related to the likelihood of a subsequent asthma diagnosis, even in this age group, and (3) a rapid fall in lung function often occurs around the time of initial diagnosis. (J ALLERGY CLIN IMMUNOL 1991;88:870-7.) Key words: Asthma, geriatrics,
allergy, lung function
There have been relatively few studies on asthma among the elderly, although, based on case series and cross-sectional studies, asthma does not appear to be a rare disorder.‘-’ However, we are not aware of previous prospective studies dealing specifically with findings before NDA in older adults. Such new diagnoses are occurring with reasonable frequency in our longitudinal study of a general population sample in Tucson, Ariz.* The purpose of this article is to describe findings in subjects who denied ever having asthma at the time of enrollment in our study but reported consulting a
From the Respiratory Sciences Center, University of Arizona College of Medicine, ‘lkson, Ariz. Supported by National Heart, Lung, and Blood Institute Specialized Center for Research Grant HL14136. Received for publication March 13, 1991. Revised July 1, 1991. Accepted for publication July 2, 1991. Reprint requests: Benjamin Burrows, MD, Respiratory Sciences Center, University of Arizona College of Medicine, Tucson, AZ 85724. l/1/32310
870
Abbreviations
NDA: COPD: STpos: SOBWZ: WHZ: EOS: OR: CI: AR:
used
Newly diagnosed asthma Chronic obstructive pulmonary disease Positive allergy skin tests Attacks of shortness of breath with wheezing Any wheezing complaint or SOBWZ Eosinophil Odds ratio Confidence interval Allergic rhinitis
physician for “asthma” for the first time when they were at least 60 years of age. Their findings before an asthma diagnosis (which was first reported at an average age of 71 years) and their courses during follow-up are contrasted with findings of other subjects in the population sample. MATERIAL AND METHODS Methods of selection and study of the general population sample have been reported previously.’
The study repre-
VOLUME NUMBER
a stratified random sample of white non-MexicanAmerican households residing in Tucson in 1971 to 1972. The present study is based on findings in the 1517 subjects who provided follow-up questionnaire data after the age of 60 years. OF these subjects, 176 stated on their initial questionnaire or on some questionnaire before the age of 60 years that they had had asthma at some time in their life and were excluded from the present study. An additional 156 subjects,reported physician-confirmed “emphysema” either on enrollment or before any asthma diagnosis. Based on our cross-sectional findings,’ we are reluctant to accept NDA in a subject who already has been diagnosed as having emphysema, and such subjects are also excluded from subsequent analyses. The remaining 1185 subjects form the basis of the present study. All subjects answered “No” to the direct question, “Have you ever had asthma?‘, on their initial questionnaire. They were not included if they answered “Yes,” even if subsequent questions indicated that the disease was no longer considered to be present or that the diagnosis was never confirmed by a physician. The study consisted of 725 women (61%) and 460 men (39%) whose average age at enrollment was 64.1 ? 10.3 years. Their average total questionnaire follow-up was 11.3 -t 5.3 years, and their average follow-up after reaching their sixtieth birthday was 8.7 ?: 5.3 years. Forty subjects first reported consulting a physician for asthma after the age of 60 years. If subjects answered affirmatively lo a specific question concerning asthma on at least one follow-up questionnaire, they were considered to have NDA. In most follow-up surveys, the question took the form of “During the past year, have you seen a doctor for asthma?’ In a few of the later surveys, the question was phrased as ton the initial questionnaire, but the diagnosis was accepted only if the subject not only answered “Yes” to “Have you ever had asthma?’ but also answered “Yes” to the additisonalquestion, “During the past year, have you seen a doctor for it?” In these elderly subjects, there is always doubt about a reported diagnosis of “asthma” and whether it distinguishes a group of subjects that is different from subjects with COPD. For ithis reason,we havealsoexaminedinitial findings amongthe subjectsin the study who never reported asthmaduring follow-up but whoselast spirometrictest indicateddefinite airway obstruction.Their last test was requiredto reveal a percentpredictedFEV, of <65% and an FEV,/FVC ratio of <70%. The 51 subjectswho met thesecriteria.areconsidered to havea nonasthmatic type of COPD.Their initial findingsarecomparedwith findingsin the 40 subje’cts with NDA. Methodsof spirometrictesting,allergyskintesting,measuringserum-IgElevel, andassessing bloodEOSshavebeen repOrted.V I2 Presentdata are incompletefor thesetests,but of the 40 NDA cases,39 subjects(98%) had satisfactory spirograms, .38(95%)hadallergyskintests,and38subjects (95%)hadserum-IgElevelsdetermined beforetheir asthma diagnosis.Only 30 subjects(75%) had blood EOS determinations.For the 1145other subjects,spirometrywas availablein 96%,allergyskintestsin 95%,serum-IgElevels in 87%, but percentbloodEOSsin only 69%. sented
Asthma in the elderlv
88 8
871
PercentpredictedFEV, wascalculatedfrom prediction equationsderived from this populationsample.9Subjects wereconsidered STposif theyhadanyreactionsto theprick testsappliedthat werelargerthanthe control reaction.Serum-IgE levels,expressedas internationalunits per milliliter, werelog transformed,and, to adjustfor ageandsex, the IgE levels are sometimesexpressedas age-sexstandardized “Z scores,”the numberof standarddeviations by which an individual’slog IgE differsfrom the meanfor their age-sex-specificgroupin the total population,as in previousarticles.‘. I3Sincethosesubjectsexcludedbecause of preexistingasthmatendedto have very high IgE levels, the meanZ scorefor the subjectsincludedin the present analysesis slightly
872
Burrows
et al.
TABLE I. Symptoms
J. ALLERGY
and smoking
habits on enrollment Subjects with NDA (Xl
Wheezing Any Apart from “colds” Attacksof SOBWZ Any More thanrarely Current“AR” Total reported ReportedplusSTpos Chroniccough chronic sputum Chroniccoughand/or sputum Exertionaldyspneat Smokinghabits Currentsmoker Exsmoker Nonsmoker Among ever smokers Smoked< 10 pack years Childhoodresptroublel( “Chronicbronchitis” Any wheezingand/or SOBWZ (WHZ)
in relation Other subjects I%)
50
19.9
25
10.4
37.5 20
11
65 34.2 20 27.5 27.5
CLIN. IMMUNOL. DECEMBER 1991
to NDA Relative risk*
95% Cl
3.8 2.7
(2.1-6.9) (1.4-5.5)
6
4.4 3.6
(1.7-7.5)
3.1 3.4 1.3 2.2 1.4 1.9
(1.7-5.9) (1.8-6.5)
50
36.2 12.5 15.6 14.5 20.5 33.9
30 32.5 37.5
24.7 26.2 49.1
1.65 1.65 0.69
(0.7-3.3) (0.8-3.3) (0.3-1.2)
36 35 20 62.5
17.2 15 7.4 23.3
2,6 2.9 2.9
(I .2-5.7) (1.5-5.4)
(1.4-6.2)
5.1
(2.7-9.5)
(2.4-8.2)
(0.6-2.9) (1.1-4.2) (0.7-2.9) (l.O-3.5)$
Resp, Respiratory. *Risk of later asthma in subjects with the symptom divided by the risk in subjects without the symptom, followed by 95% CI for this value. tBecome short of breath hying on level ground or watking up a stight hill. SLower limit of 95% CI is 1.04. $Relative risks for smoking based on current versus never, exsmoker versus never, and never versus ever smokers. None are statistically significant. (/“Respiratory trouble before age 16.” subjects with NDA and strongly predictive of a later
asthma diagnosis. There was also an increased rate of reported “AR” in subjects with NDA. This finding is especially striking if the diagnosis of AR is accepted only in subjects who are STpos. Cough, sputum, and exertional dyspnea tended to be more common in subjects developing asthma, but the differences in chronic cough and in chronic cough and/or sputum were not statistically significant. In contrast, a history of “respiratory trouble before age 16” was a highly significant predictor of a later asthma diagnosis. Twenty percent of the group with NDA reported that they had already consulted a physician for “chronic bronchitis” at the time of e~ollment. There was no significant difference between the
groups in the overall smoking status on enrollment, but of subjects who had ever smoked, a significantly higher prounion of the group with NDA had smoked ~10 pack years of cigarettes (p < 0.05). The spirometric findings, allergy skin test results, serum-IgE levels, and blood EOS percentagesin the
two groups are summarized in Table II. Subjects with NDA had significantly lower initial lung function, more allergy skin test reactivity, higher IgE levels, and more eosinophilia. The highly significant trend for NDA rates to increase with the age-sex-standardized IgE Z score is illustrated in Fig. 1. In a multiple logistic regression, attacks of SOBWZ was the best single predictor of NDA, but any wheezing added significantly to its prediction. The close relationship of any wheezing complaint and/or SOBWZ (WHZ) to later diagnosis is presented in the last line of Table I. Except for chronic sputum and
exertional dyspnea, all other symptoms and test variables in Tables I and II that had a signific~t overall relationship to a later asthma diagnosis (i.e., their 95% CIs did not include 1.0) added significantly to the prediction of NDA after controlling for WHZ. The rates of these findings in subjects with and without any WI-I% by the presence or absence of a later asthma diagnosis are presented in Table III. The combined ORs for NDA in relation to these findings
VOLUME NUMBER
Asthma
88 6
c-1.5
-.5<.5 (428)
-1.W.5
(75)
(271)
Age-Sex
Standardized
T’est
results
in relation
873
.54.5 (215)
IgE Z Score
FIG. 1. Rate of NDA in groups delineated by their age-sex-standardized Numbers of subjects at risk in each of the groups are presented proportions with increasing Z score is significant at p < 0.001.
TABLE II.
in the elderly
serum-IgE in parentheses.
Z scores. Trend in
to NDA % With finding or mean f SD Subjects with NDA
Other subjects
% Pred FEV, % With % FEN, <75%
82.5 zk 19.2 35.9%
95.7 -c 18.8 12.8%
3.6
FEV, /FVC ratio % With ratio <70%
75.5 2 13.0 30.8%
79.3 I!Z 8.5 13.2%
2.8
47.4%
26.4%
Mean serum IgE (log IU) (geometric mean) % With1 Z score >O
1.59 2 0.64 (38.9) 73.7%
1.25 + 0.72 (17.6) 45.4%
Mean % blood EOSs %With4+ %EOSs
2.80 k 2.67 36.7%
1.71 +- 1.78 13.0%
STpos
Relative risk*
-
2.4 3.2 3.6
p Value (or 95% Cl) p < 0.002
(1.9-6.7) 0.01 (1.4-5.4)
p <
(1.3-4.5) p < 0.005
(1.6-6.6) p < 0.002
(1.8-7.4)
% Pred, Percent predicted. *See legend fol Table I.
after stratifying by WHZ are also presented along with their 95% Cl[s. The likelihood of later asthma diagnosis is especially high in subjects with any WHZ plus either AR confirmed by STpos (l&2%), a percent predicted FEV, of <75% (17.5%), an FEV,/FVC ratio of <70% (19.2%), or eosinophilia (21.2%). The risk of NDA in the absence of WHZ or one of these
other findings is very low. For examaple, there is only a 1.5% risk of NDA among the 782 subjects with no WHZ and without AR confirmed by STpos. With the small number of subjects reporting NDA and the close interrelationships of most of the variables, further modeling of the predictors of NDA appears unjustified. Also, true ORs for the predictive value of serum
Burrows et al.
874
TABLE
III. Remaining
J. ALLERGY
relationships
to NDA after controlling
CLIN. IMMUNOL. DECEMBER 1991
for any WHZ and/or attacks of SOBWZ
% With finding Combined Without
(Total No. at risk) Without NDA With NDA Reported “AR’ Without NDA (%) With NDA (%) “AR” and STpos Without NDA (I) With NDA (%) ~hildh~ resp trouble Without NDA (%) With NDA (8) 8 Fred FEV, <75% Wi~out NDA (%) With NDA (%) FEV, / EVC ratio <70% Without NDA (“lo) With NDA (%) STpos Without NDA (%) With NDA (%) Serum IgE Z score >O Without NDA (%) With NDA (%) Blood EOSs 24% Without NDA (%) With NDA (%)
WHZ
With WHZ
OR* for later asthma
05% Cl)
(874) wl
(2671 (25)
32.6 46.7
47.9 76t
2.6 (1.3-5.3)
11.1 20
17.3 43.51
3.0 (1.5-6.3)
11.5 20
26.6 44
2.1 (l.O-4.2)P
11.1 28.6
18.4 40t
3.0 (1.5-6.1)
12.2 14.3
16.5 40‘1:
2.5 (1.2-5.4)
25.5 46.7
29.2 47.8
2.3 (1.2-4.3)
44.3 66.7
48.7 78.3i
3.2 (1.5-6.7)
12.4 28.6
14.4 43.8$
3.7 (1.7-8.2)
Resp, Respiratory; 96 Pred, Percent predicted. *Mantel-Haenszel OR for stratified data after stratification by the presence or absence of any WHZ. tp < 0.05 for difference within the group with any WHZ. $p< 0.01 for difference within the group with any WHZ. (All other intragroup differences are not significant.) ILower limit of 9.5% CI is 1.04.
IgE cannot be calculated because no asthma was diagnosed in subjects with the lowest IgE levels on enrollment,’ as illustrated in Fig. 1. As would be expected, the rates of wheezing complaints and exertional dyspnea increased by the time the subjects reported an asthma diagnosis. By that time, 62.5% reported wheezing apart from colds, 67.5% had SOBWZ, and 72.5% stated they became short of breath on hurrying on the level or walking up a slight hill. In contrast, rates of chronic cough and chronic sputum decreased slightly; only 22.5% of the subjects with NDA complained of chronic cough and/or sputum when they first reported their asthma diagnosis. Three of the 12 subjects with NDA who smoked on entry to the study had quit smoking by that time.
The course of lung function during follow-up is summarized in Table IV. Rates of decline in FEV, were almost twice as high in subjects with NDA as in other subjects. In these subjects with at least 4 years of spirometric follow-up, the average time from enrollment to the diagnosis of asthma was 9.3 years, whereas the average time of spirometric follow-up was 13.1 years; therefore, much of the calculated decline probably occurred before their new diagnosis. As presented in Table IV, the excessive rate of decline in EV, in subjects with NDA is especially striking in subjects with spirometry that was within the normal range on enrollment. In fact, the rate of decline in FEV, in the subjects with NDA was not only higher than in the control subjects but also was considerably higher than in subjects who were excluded because
VOLUME NUMBER
60 6
TABLE IV. FEV, decline
Asthma in the elderly
in subjects
with
875
4 or more years of follow-up Mean + SD Subjects with NDA
Total group (No. of subjects) Mean yearsof follow-up Meantestsper subject Meanchangein FEV, (ml/ yr) With initial % FEV, <75% (No. of subjects) Meanyearsof follow-up Meantestsper subject Meanchangein FEV, (ml/yr) With initial % FEV, 75%or more (No. of subjects) MeCan yearsof follow-up Meantestsper subject Meanchangein FEV, (ml/yr)
Other subjects
p Value
(31) 13.1 -e 4.0 7.4 It 2.5 -42.0 -c 43.4
(794) 12.2 2 3.9 6.8 + 2.4 -23.1 k 30.8
NS NS CO.001
(10) 13.1 2 4.3 7.5 f 2.2 -24.3 k 26.9
(76) 11.8 + 3.9 6.4 -c 2.5 - 15.9 t 32.3
NS NS NS
(21) 13.2 + 4.0 7.3 _’ 2.6 -50.5 AI 47.7
(718) 12.3 f 3.9 6.8 I 2.4 -23.8 -c 30.6
NS NS <0.0001
NS,Notsignificant. they reported ever having asthma at the time of enrollment, -42.0 +- 43.4 compared with -15.4 + 28.0 ml/yr (p < 0.0005).
Because the nature of our longitudinal study precluded our observing responses to therapy and because bronchial-challenge tests were not performed, it is reasonable to question whether or not our NDA might have been confounded by cases of nonasthmatic COPD. We therefore compared findings on entry of the subjects; with NDA with findings of subjects who had COPD on their last spirometric test, as defined under METHODS. In contrast to subjects with NDA, the s;ubjects with COPD had a higher rate in men than in women and more often were current smokers, although the differences between NDA and COPD cases were not statistically significant. However, the average pack years of smoking by subjects with COPDl (27.7 2 26.1) was significantly higher (p < 0.002) than in the remainder of the population (16.5 + 24.4), whereas there was no significant increase in pack years in the group with NDA (17.6 2 23.2). There were no significant differences between groups with NDA and COPD in initial complaints of cough, sputum production, or exertional dyspnea. However, initial WHZ was significantly more common in subjects with NDA than with COPD (62.5% versus 37.3%; p < 0.05), as was reported AR (65.0% versus 33.3%, p < 0.01) and reported AR plus a positive skin test (34.5% versus 8.0%, p < 0.005). There were significantly higher rates in
the group with NDA of STpos (47.4% versus 20.4%, p < 0.02), 4% or more blood EOSs (36.7% versus 8.3%,p < 0.02), and serum IgE Z scores >O (73.7% versus 47.7%, p < 0.05). Unlike the group with
NDA, there was virtually no trend for the frequency of COPD to increase with increasing levels of IgE Z score; the trend chi-square for the relationship of COPD to the levels of IgE Z score used in Fig. 1 was 0.056 (p > 0.8).
DISCUSSION Although the above-noted differences between subjects with NDA and subjects with apparently nonasthmatic COPD are in the direction that would be expected if many of our NDA cases were already manifesting evidence of asthma at the time of their enrollment, there is no way to exclude some confounding of the diagnoses. In the elderly, the distinction of persistent asthma from COPD can be difficult, even in a clinical setting. There are certainly no firm criteria for their differentiation based on epidemiologic criteria. However, any confounding of the diagnoses would tend to minimize rather than exaggerate most of the observed relationships. Whether the diagnoses of subjects’ physicians might have been influenced by the findings in our study might be questioned. However, results of responses to the questionnaire were not sent to the physicians, and the spirometric data, if these had any effect, were more likely to lead to a diagnosis of COPD rather than asthma, especially since the concept
876
Burrows
et al.
of chronic airway obstruction in asthma was not widely accepted in the early 1970s when our initial data were obtained. It is possible that reported allergy skin test reactivity could have created a diagnostic bias, but it appears doubtful that such a finding would have had much effect many years after the skin tests were done in subjects who were not already consulting a physician for their disease. Results of serum-IgE tests were never sent to subjects’ physicians, largely because we were uncertain of their clinical significance at the time these tests were performed. In any case, present data suggest that most NDA in subjects older than the age of 60 years occur in subjects with significant symptomatology more than 8 years before their diagnosis. Wheezing complaints were especially common. Also, many subjects already demonstrated some degree of obstructive ventilatory impairment. The overall impression is that most apparently new diagnoses were made in subjects who already had undiagnosed disease at the time of enrollment. A few subjects had already been told they had “chronic bronchitis”; the asthma diagnosis was simply added at a later date. In a disease such as asthma, there may be a great deal of difference between the onset of symptoms suggesting an asthmatic type of problem and the time when these symptoms become sufficiently severe to lead the subject to seek medical attention. Even in a long-term prospective study such as our study, with repeated questionnaires, it is impossible to be certain of the time of “onset” of the disease. The fact that 35% of these elderly subjects recalled “respiratory trouble before age 16” suggests that asthma-like symptoms may actually have begun during childhood but did not become sufficiently severe to lead to an asthma diagnosis until very late in adult life. Smoking did not appear to be a significant risk factor for a later diagnosis of asthma. However, among subjects who had ever smoked, a high proportion had smoked
J. ALLERGY
CLIN. IMMUNOL. DECEMBER 1991
in FEV, was especially rapid in subjects who had relatively well-preserved ventilatory function on enrollment. These data are compatible with the concept that many elderly subjects with asthma have relatively low ventilatory function throughout life but may have a period of increased symptomatology associated with a rapid loss of function that finally leads these subjects to seek medical care and receive a diagnosis. Once this period is over and the diagnosis is established, lung function appears to stabilize at a low level. Whether this is ascribable to the institution of antiasthma therapy or is simply part of the natural history of the disease cannot be determined from our data. Finally, even in these very elderly subjects, the likelihood of being diagnosed as having asthma is closely related to the age-sex-standardized serum-IgE level before diagnosis. This finding appears to be in conflict with the recent study by Braman et al6 in which elderly subjects with asthma are said to have no significant elevation of serum IgE compared to elderly, hospitalized, control subjects. The mean log IgE levels (mistakenly called “geometric means”) of their “early” and “late” onset of asthma in elderly subjects were 1.65 and 1.80, respectively. The overall mean log IgE for their elderly subjects with asthma can be calculated from the raw data presented to be 1.73 + 0.376, even higher than the value for subjects with later asthma in the present study and nearly as high as that previously reported in our elderly subjects with established disease.’ However, their control subjects had a mean log IgE of 1.49 (no standard deviation can be determined from the data provided), a higher value than in our control elderly population. If it is assumed that there is a similar standard deviation in the control subjects in the study of Braman et a1.,6 as in their subjects with asthma, which should be true if there are no differences between the groups, the IgE levels of their subjects with asthma would be significantly elevated (p < O.Ol), despite the fact that almost all the patients with asthma were taking oral corticosteroid therapy and that the control subjects, being hospitalized patients, might not be entirely appropriate. From our data, it is tempting to conclude that a high serum-IgE level is truly a risk factor for asthma and not simply an associated finding in the disease. However, it remains possible that the critical relationship of IgE is to a mild, and as yet, undiagnosed asthmatic state. In this sense, the present findings only help confirm our previous observations revealing the extremely close relationship of asthma prevalence to serum IgE.13 Final proof of cause and effect will probably be impossible to obtain from epidemiologic data. Present findings do indicate, however, that the serumIgE level helps identify those subjects who will later
VOLUME NUMBER
88 6
be diagnosed as having asthma, even in very elderly adults, and that the relationship of asthma to serum IgE does not depend on the presence of frank, clinical disease.
REFERENC:ES 1. Broder I, Barlow PP, Horton RJM. The epidemiology of asthma and hay fever in a total community, Tecumseh, Michigan. 1. Description of study and general findings. J ALLERGY 1962;33:513-23. 2. Lee Hl!, Stretton TB. Asthma in the elderly. Br Med J 1973;4:‘?3-5. 3. Lebowitz MD, Knudson RJ, Burrows B. Tucson epidemiologic study of obstructive lung diseases. I. Methodology and prevalence of disease. Am J Epidemiol 1975;102:137-52. 4. Burr ML, Charles TJ, Roy K, Seaton A. Asthma in the elderly: an epidemiologic survey. Br Med J 1979;1:1041-4. 5. Ford RM. Aetiology of asthma: a review of 11.55 1 cases(1958 to 1968). Med J Aust 1969;1:628-31. 6. Braman SS, Kaemmerlen JT, Davis SM. Asthma in the elderly: a comparison between patients with recently acquired and longstanding disease. Am Rev Respir Dis 1991;143:336-40. 7. Burrow:; B, Barbee RA, Cline MG, Knudson RJ, Lebowitz MD. Characteristics of asthma among elderly adults in a sample of the glsneral population. Chest 1991;100:935-42.
Asthma
in the elderly
877
8. Dodge RR, Cline MG, Burrows B. Comparisons of asthma, emphysema, and chronic bronchitis diagnoses in a general population sample. Am Rev Respir Dis 1986;133:981-6. 9. Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983;127:725-34. 10. Barbee RA, Lebowitz MD, Thompson HC, Burrows B. lmmediate skin-test reactivity in a general population sample. Ann Intern Med 1976;84: 129-33. 11. Barbee RA, Halonen M, Lebowitz MD, Burrows B. Distribution of IgE in a community population sample: correlations with age, sex, and allergen skin test reactivity. J ALLERGY CLIN IMMUNOL
1981;68:106-11.
12. Burrows B, Hasan FM, Barbee RA, Halonen M, Lebowitz MD. Epidemiologic observations on eosinophilia and its relation to respiratory disorders. Am Rev Respir Dis 1980;122:709-19. 13. Burrows B, Martinez FD, Halonen M, Barbee RA, Cline MG. Association of asthma with serum IgE levels and skin-test reactivity to allergens. N Engl J Med 1989;320:271-7. 14. Burrows B, Lebowitz MD, Camilli AE, Knudson RJ. Longitudinal changes in forced expiratory volume in one second in adults: methodologic considerations and findings in healthy nonsmokers. Am Rev Respir Dis 1986;133:974-80. 15. Burrows B, Bloom JW, Traver GA, Cline MG. The course and prognosis of different forms of chronic airways obstruction in a sample from the general population. N Engl J Med 1987;317:1309-14.
of asthma among study of a general
Benjamin Burrows, MD, Michael D. Lebowitz, and Martha G. Cline, MS Tucson, Ariz.
PhD, Robert
A. Barbee, MD,
Forty elderly subjects who denied ever having asthma or emphysema on enrollment in a longitudinal epidemiologic study later reported consulting a doctor for asthma when they were older than 60 years of age. The average age at which the diagnosis was reported was 70.8 years, after a mean follow-up of 8.5 years. Findings on enrollment in the newly diagnosed subjects with asthma are compared with jindings in the 1145 subjects who provided follow-up information when they were older than age 60 years but had never developed asthma. At the time of enrollment, most subjects later diagnosed as having asthma already had wheezing symptoms, suggesting at least a mild asthmatic state, and many subjects had impaired ventilatory function, a positive allergy skin test (especially in association with rhinitis), and blood eosinophilia. Thirty-jive percent of the subjects recalled “respiratory trouble before age 16” despite denying prior asthma. The likelihood of a new asthma label was very closely related to the age-sex-standardized serum-IgE level before diagnosis. Newly diagnosed subjects with asthma demonstrated much greater rates of decline in FEV, than control subjects or than subjects who already had known asthma on enrollment. We conclude that (1) symptoms suggesting asthma are usually present for many years before the diagnosis of the disease in elderly subjects, (2) the serum-IgE level is closely related to the likelihood of a subsequent asthma diagnosis, even in this age group, and (3) a rapid fall in lung function often occurs around the time of initial diagnosis. (J ALLERGY CLIN IMMUNOL 1991;88:870-7.) Key words: Asthma, geriatrics,
allergy, lung function
There have been relatively few studies on asthma among the elderly, although, based on case series and cross-sectional studies, asthma does not appear to be a rare disorder.‘-’ However, we are not aware of previous prospective studies dealing specifically with findings before NDA in older adults. Such new diagnoses are occurring with reasonable frequency in our longitudinal study of a general population sample in Tucson, Ariz.* The purpose of this article is to describe findings in subjects who denied ever having asthma at the time of enrollment in our study but reported consulting a
From the Respiratory Sciences Center, University of Arizona College of Medicine, ‘lkson, Ariz. Supported by National Heart, Lung, and Blood Institute Specialized Center for Research Grant HL14136. Received for publication March 13, 1991. Revised July 1, 1991. Accepted for publication July 2, 1991. Reprint requests: Benjamin Burrows, MD, Respiratory Sciences Center, University of Arizona College of Medicine, Tucson, AZ 85724. l/1/32310
870
Abbreviations
NDA: COPD: STpos: SOBWZ: WHZ: EOS: OR: CI: AR:
used
Newly diagnosed asthma Chronic obstructive pulmonary disease Positive allergy skin tests Attacks of shortness of breath with wheezing Any wheezing complaint or SOBWZ Eosinophil Odds ratio Confidence interval Allergic rhinitis
physician for “asthma” for the first time when they were at least 60 years of age. Their findings before an asthma diagnosis (which was first reported at an average age of 71 years) and their courses during follow-up are contrasted with findings of other subjects in the population sample. MATERIAL AND METHODS Methods of selection and study of the general population sample have been reported previously.’
The study repre-
VOLUME NUMBER
a stratified random sample of white non-MexicanAmerican households residing in Tucson in 1971 to 1972. The present study is based on findings in the 1517 subjects who provided follow-up questionnaire data after the age of 60 years. OF these subjects, 176 stated on their initial questionnaire or on some questionnaire before the age of 60 years that they had had asthma at some time in their life and were excluded from the present study. An additional 156 subjects,reported physician-confirmed “emphysema” either on enrollment or before any asthma diagnosis. Based on our cross-sectional findings,’ we are reluctant to accept NDA in a subject who already has been diagnosed as having emphysema, and such subjects are also excluded from subsequent analyses. The remaining 1185 subjects form the basis of the present study. All subjects answered “No” to the direct question, “Have you ever had asthma?‘, on their initial questionnaire. They were not included if they answered “Yes,” even if subsequent questions indicated that the disease was no longer considered to be present or that the diagnosis was never confirmed by a physician. The study consisted of 725 women (61%) and 460 men (39%) whose average age at enrollment was 64.1 ? 10.3 years. Their average total questionnaire follow-up was 11.3 -t 5.3 years, and their average follow-up after reaching their sixtieth birthday was 8.7 ?: 5.3 years. Forty subjects first reported consulting a physician for asthma after the age of 60 years. If subjects answered affirmatively lo a specific question concerning asthma on at least one follow-up questionnaire, they were considered to have NDA. In most follow-up surveys, the question took the form of “During the past year, have you seen a doctor for asthma?’ In a few of the later surveys, the question was phrased as ton the initial questionnaire, but the diagnosis was accepted only if the subject not only answered “Yes” to “Have you ever had asthma?’ but also answered “Yes” to the additisonalquestion, “During the past year, have you seen a doctor for it?” In these elderly subjects, there is always doubt about a reported diagnosis of “asthma” and whether it distinguishes a group of subjects that is different from subjects with COPD. For ithis reason,we havealsoexaminedinitial findings amongthe subjectsin the study who never reported asthmaduring follow-up but whoselast spirometrictest indicateddefinite airway obstruction.Their last test was requiredto reveal a percentpredictedFEV, of <65% and an FEV,/FVC ratio of <70%. The 51 subjectswho met thesecriteria.areconsidered to havea nonasthmatic type of COPD.Their initial findingsarecomparedwith findingsin the 40 subje’cts with NDA. Methodsof spirometrictesting,allergyskintesting,measuringserum-IgElevel, andassessing bloodEOSshavebeen repOrted.V I2 Presentdata are incompletefor thesetests,but of the 40 NDA cases,39 subjects(98%) had satisfactory spirograms, .38(95%)hadallergyskintests,and38subjects (95%)hadserum-IgElevelsdetermined beforetheir asthma diagnosis.Only 30 subjects(75%) had blood EOS determinations.For the 1145other subjects,spirometrywas availablein 96%,allergyskintestsin 95%,serum-IgElevels in 87%, but percentbloodEOSsin only 69%. sented
Asthma in the elderlv
88 8
871
PercentpredictedFEV, wascalculatedfrom prediction equationsderived from this populationsample.9Subjects wereconsidered STposif theyhadanyreactionsto theprick testsappliedthat werelargerthanthe control reaction.Serum-IgE levels,expressedas internationalunits per milliliter, werelog transformed,and, to adjustfor ageandsex, the IgE levels are sometimesexpressedas age-sexstandardized “Z scores,”the numberof standarddeviations by which an individual’slog IgE differsfrom the meanfor their age-sex-specificgroupin the total population,as in previousarticles.‘. I3Sincethosesubjectsexcludedbecause of preexistingasthmatendedto have very high IgE levels, the meanZ scorefor the subjectsincludedin the present analysesis slightly
872
Burrows
et al.
TABLE I. Symptoms
J. ALLERGY
and smoking
habits on enrollment Subjects with NDA (Xl
Wheezing Any Apart from “colds” Attacksof SOBWZ Any More thanrarely Current“AR” Total reported ReportedplusSTpos Chroniccough chronic sputum Chroniccoughand/or sputum Exertionaldyspneat Smokinghabits Currentsmoker Exsmoker Nonsmoker Among ever smokers Smoked< 10 pack years Childhoodresptroublel( “Chronicbronchitis” Any wheezingand/or SOBWZ (WHZ)
in relation Other subjects I%)
50
19.9
25
10.4
37.5 20
11
65 34.2 20 27.5 27.5
CLIN. IMMUNOL. DECEMBER 1991
to NDA Relative risk*
95% Cl
3.8 2.7
(2.1-6.9) (1.4-5.5)
6
4.4 3.6
(1.7-7.5)
3.1 3.4 1.3 2.2 1.4 1.9
(1.7-5.9) (1.8-6.5)
50
36.2 12.5 15.6 14.5 20.5 33.9
30 32.5 37.5
24.7 26.2 49.1
1.65 1.65 0.69
(0.7-3.3) (0.8-3.3) (0.3-1.2)
36 35 20 62.5
17.2 15 7.4 23.3
2,6 2.9 2.9
(I .2-5.7) (1.5-5.4)
(1.4-6.2)
5.1
(2.7-9.5)
(2.4-8.2)
(0.6-2.9) (1.1-4.2) (0.7-2.9) (l.O-3.5)$
Resp, Respiratory. *Risk of later asthma in subjects with the symptom divided by the risk in subjects without the symptom, followed by 95% CI for this value. tBecome short of breath hying on level ground or watking up a stight hill. SLower limit of 95% CI is 1.04. $Relative risks for smoking based on current versus never, exsmoker versus never, and never versus ever smokers. None are statistically significant. (/“Respiratory trouble before age 16.” subjects with NDA and strongly predictive of a later
asthma diagnosis. There was also an increased rate of reported “AR” in subjects with NDA. This finding is especially striking if the diagnosis of AR is accepted only in subjects who are STpos. Cough, sputum, and exertional dyspnea tended to be more common in subjects developing asthma, but the differences in chronic cough and in chronic cough and/or sputum were not statistically significant. In contrast, a history of “respiratory trouble before age 16” was a highly significant predictor of a later asthma diagnosis. Twenty percent of the group with NDA reported that they had already consulted a physician for “chronic bronchitis” at the time of e~ollment. There was no significant difference between the
groups in the overall smoking status on enrollment, but of subjects who had ever smoked, a significantly higher prounion of the group with NDA had smoked ~10 pack years of cigarettes (p < 0.05). The spirometric findings, allergy skin test results, serum-IgE levels, and blood EOS percentagesin the
two groups are summarized in Table II. Subjects with NDA had significantly lower initial lung function, more allergy skin test reactivity, higher IgE levels, and more eosinophilia. The highly significant trend for NDA rates to increase with the age-sex-standardized IgE Z score is illustrated in Fig. 1. In a multiple logistic regression, attacks of SOBWZ was the best single predictor of NDA, but any wheezing added significantly to its prediction. The close relationship of any wheezing complaint and/or SOBWZ (WHZ) to later diagnosis is presented in the last line of Table I. Except for chronic sputum and
exertional dyspnea, all other symptoms and test variables in Tables I and II that had a signific~t overall relationship to a later asthma diagnosis (i.e., their 95% CIs did not include 1.0) added significantly to the prediction of NDA after controlling for WHZ. The rates of these findings in subjects with and without any WI-I% by the presence or absence of a later asthma diagnosis are presented in Table III. The combined ORs for NDA in relation to these findings
VOLUME NUMBER
Asthma
88 6
c-1.5
-.5<.5 (428)
-1.W.5
(75)
(271)
Age-Sex
Standardized
T’est
results
in relation
873
.54.5 (215)
IgE Z Score
FIG. 1. Rate of NDA in groups delineated by their age-sex-standardized Numbers of subjects at risk in each of the groups are presented proportions with increasing Z score is significant at p < 0.001.
TABLE II.
in the elderly
serum-IgE in parentheses.
Z scores. Trend in
to NDA % With finding or mean f SD Subjects with NDA
Other subjects
% Pred FEV, % With % FEN, <75%
82.5 zk 19.2 35.9%
95.7 -c 18.8 12.8%
3.6
FEV, /FVC ratio % With ratio <70%
75.5 2 13.0 30.8%
79.3 I!Z 8.5 13.2%
2.8
47.4%
26.4%
Mean serum IgE (log IU) (geometric mean) % With1 Z score >O
1.59 2 0.64 (38.9) 73.7%
1.25 + 0.72 (17.6) 45.4%
Mean % blood EOSs %With4+ %EOSs
2.80 k 2.67 36.7%
1.71 +- 1.78 13.0%
STpos
Relative risk*
-
2.4 3.2 3.6
p Value (or 95% Cl) p < 0.002
(1.9-6.7) 0.01 (1.4-5.4)
p <
(1.3-4.5) p < 0.005
(1.6-6.6) p < 0.002
(1.8-7.4)
% Pred, Percent predicted. *See legend fol Table I.
after stratifying by WHZ are also presented along with their 95% Cl[s. The likelihood of later asthma diagnosis is especially high in subjects with any WHZ plus either AR confirmed by STpos (l&2%), a percent predicted FEV, of <75% (17.5%), an FEV,/FVC ratio of <70% (19.2%), or eosinophilia (21.2%). The risk of NDA in the absence of WHZ or one of these
other findings is very low. For examaple, there is only a 1.5% risk of NDA among the 782 subjects with no WHZ and without AR confirmed by STpos. With the small number of subjects reporting NDA and the close interrelationships of most of the variables, further modeling of the predictors of NDA appears unjustified. Also, true ORs for the predictive value of serum
Burrows et al.
874
TABLE
III. Remaining
J. ALLERGY
relationships
to NDA after controlling
CLIN. IMMUNOL. DECEMBER 1991
for any WHZ and/or attacks of SOBWZ
% With finding Combined Without
(Total No. at risk) Without NDA With NDA Reported “AR’ Without NDA (%) With NDA (%) “AR” and STpos Without NDA (I) With NDA (%) ~hildh~ resp trouble Without NDA (%) With NDA (8) 8 Fred FEV, <75% Wi~out NDA (%) With NDA (%) FEV, / EVC ratio <70% Without NDA (“lo) With NDA (%) STpos Without NDA (%) With NDA (%) Serum IgE Z score >O Without NDA (%) With NDA (%) Blood EOSs 24% Without NDA (%) With NDA (%)
WHZ
With WHZ
OR* for later asthma
05% Cl)
(874) wl
(2671 (25)
32.6 46.7
47.9 76t
2.6 (1.3-5.3)
11.1 20
17.3 43.51
3.0 (1.5-6.3)
11.5 20
26.6 44
2.1 (l.O-4.2)P
11.1 28.6
18.4 40t
3.0 (1.5-6.1)
12.2 14.3
16.5 40‘1:
2.5 (1.2-5.4)
25.5 46.7
29.2 47.8
2.3 (1.2-4.3)
44.3 66.7
48.7 78.3i
3.2 (1.5-6.7)
12.4 28.6
14.4 43.8$
3.7 (1.7-8.2)
Resp, Respiratory; 96 Pred, Percent predicted. *Mantel-Haenszel OR for stratified data after stratification by the presence or absence of any WHZ. tp < 0.05 for difference within the group with any WHZ. $p< 0.01 for difference within the group with any WHZ. (All other intragroup differences are not significant.) ILower limit of 9.5% CI is 1.04.
IgE cannot be calculated because no asthma was diagnosed in subjects with the lowest IgE levels on enrollment,’ as illustrated in Fig. 1. As would be expected, the rates of wheezing complaints and exertional dyspnea increased by the time the subjects reported an asthma diagnosis. By that time, 62.5% reported wheezing apart from colds, 67.5% had SOBWZ, and 72.5% stated they became short of breath on hurrying on the level or walking up a slight hill. In contrast, rates of chronic cough and chronic sputum decreased slightly; only 22.5% of the subjects with NDA complained of chronic cough and/or sputum when they first reported their asthma diagnosis. Three of the 12 subjects with NDA who smoked on entry to the study had quit smoking by that time.
The course of lung function during follow-up is summarized in Table IV. Rates of decline in FEV, were almost twice as high in subjects with NDA as in other subjects. In these subjects with at least 4 years of spirometric follow-up, the average time from enrollment to the diagnosis of asthma was 9.3 years, whereas the average time of spirometric follow-up was 13.1 years; therefore, much of the calculated decline probably occurred before their new diagnosis. As presented in Table IV, the excessive rate of decline in EV, in subjects with NDA is especially striking in subjects with spirometry that was within the normal range on enrollment. In fact, the rate of decline in FEV, in the subjects with NDA was not only higher than in the control subjects but also was considerably higher than in subjects who were excluded because
VOLUME NUMBER
60 6
TABLE IV. FEV, decline
Asthma in the elderly
in subjects
with
875
4 or more years of follow-up Mean + SD Subjects with NDA
Total group (No. of subjects) Mean yearsof follow-up Meantestsper subject Meanchangein FEV, (ml/ yr) With initial % FEV, <75% (No. of subjects) Meanyearsof follow-up Meantestsper subject Meanchangein FEV, (ml/yr) With initial % FEV, 75%or more (No. of subjects) MeCan yearsof follow-up Meantestsper subject Meanchangein FEV, (ml/yr)
Other subjects
p Value
(31) 13.1 -e 4.0 7.4 It 2.5 -42.0 -c 43.4
(794) 12.2 2 3.9 6.8 + 2.4 -23.1 k 30.8
NS NS CO.001
(10) 13.1 2 4.3 7.5 f 2.2 -24.3 k 26.9
(76) 11.8 + 3.9 6.4 -c 2.5 - 15.9 t 32.3
NS NS NS
(21) 13.2 + 4.0 7.3 _’ 2.6 -50.5 AI 47.7
(718) 12.3 f 3.9 6.8 I 2.4 -23.8 -c 30.6
NS NS <0.0001
NS,Notsignificant. they reported ever having asthma at the time of enrollment, -42.0 +- 43.4 compared with -15.4 + 28.0 ml/yr (p < 0.0005).
Because the nature of our longitudinal study precluded our observing responses to therapy and because bronchial-challenge tests were not performed, it is reasonable to question whether or not our NDA might have been confounded by cases of nonasthmatic COPD. We therefore compared findings on entry of the subjects; with NDA with findings of subjects who had COPD on their last spirometric test, as defined under METHODS. In contrast to subjects with NDA, the s;ubjects with COPD had a higher rate in men than in women and more often were current smokers, although the differences between NDA and COPD cases were not statistically significant. However, the average pack years of smoking by subjects with COPDl (27.7 2 26.1) was significantly higher (p < 0.002) than in the remainder of the population (16.5 + 24.4), whereas there was no significant increase in pack years in the group with NDA (17.6 2 23.2). There were no significant differences between groups with NDA and COPD in initial complaints of cough, sputum production, or exertional dyspnea. However, initial WHZ was significantly more common in subjects with NDA than with COPD (62.5% versus 37.3%; p < 0.05), as was reported AR (65.0% versus 33.3%, p < 0.01) and reported AR plus a positive skin test (34.5% versus 8.0%, p < 0.005). There were significantly higher rates in
the group with NDA of STpos (47.4% versus 20.4%, p < 0.02), 4% or more blood EOSs (36.7% versus 8.3%,p < 0.02), and serum IgE Z scores >O (73.7% versus 47.7%, p < 0.05). Unlike the group with
NDA, there was virtually no trend for the frequency of COPD to increase with increasing levels of IgE Z score; the trend chi-square for the relationship of COPD to the levels of IgE Z score used in Fig. 1 was 0.056 (p > 0.8).
DISCUSSION Although the above-noted differences between subjects with NDA and subjects with apparently nonasthmatic COPD are in the direction that would be expected if many of our NDA cases were already manifesting evidence of asthma at the time of their enrollment, there is no way to exclude some confounding of the diagnoses. In the elderly, the distinction of persistent asthma from COPD can be difficult, even in a clinical setting. There are certainly no firm criteria for their differentiation based on epidemiologic criteria. However, any confounding of the diagnoses would tend to minimize rather than exaggerate most of the observed relationships. Whether the diagnoses of subjects’ physicians might have been influenced by the findings in our study might be questioned. However, results of responses to the questionnaire were not sent to the physicians, and the spirometric data, if these had any effect, were more likely to lead to a diagnosis of COPD rather than asthma, especially since the concept
876
Burrows
et al.
of chronic airway obstruction in asthma was not widely accepted in the early 1970s when our initial data were obtained. It is possible that reported allergy skin test reactivity could have created a diagnostic bias, but it appears doubtful that such a finding would have had much effect many years after the skin tests were done in subjects who were not already consulting a physician for their disease. Results of serum-IgE tests were never sent to subjects’ physicians, largely because we were uncertain of their clinical significance at the time these tests were performed. In any case, present data suggest that most NDA in subjects older than the age of 60 years occur in subjects with significant symptomatology more than 8 years before their diagnosis. Wheezing complaints were especially common. Also, many subjects already demonstrated some degree of obstructive ventilatory impairment. The overall impression is that most apparently new diagnoses were made in subjects who already had undiagnosed disease at the time of enrollment. A few subjects had already been told they had “chronic bronchitis”; the asthma diagnosis was simply added at a later date. In a disease such as asthma, there may be a great deal of difference between the onset of symptoms suggesting an asthmatic type of problem and the time when these symptoms become sufficiently severe to lead the subject to seek medical attention. Even in a long-term prospective study such as our study, with repeated questionnaires, it is impossible to be certain of the time of “onset” of the disease. The fact that 35% of these elderly subjects recalled “respiratory trouble before age 16” suggests that asthma-like symptoms may actually have begun during childhood but did not become sufficiently severe to lead to an asthma diagnosis until very late in adult life. Smoking did not appear to be a significant risk factor for a later diagnosis of asthma. However, among subjects who had ever smoked, a high proportion had smoked
J. ALLERGY
CLIN. IMMUNOL. DECEMBER 1991
in FEV, was especially rapid in subjects who had relatively well-preserved ventilatory function on enrollment. These data are compatible with the concept that many elderly subjects with asthma have relatively low ventilatory function throughout life but may have a period of increased symptomatology associated with a rapid loss of function that finally leads these subjects to seek medical care and receive a diagnosis. Once this period is over and the diagnosis is established, lung function appears to stabilize at a low level. Whether this is ascribable to the institution of antiasthma therapy or is simply part of the natural history of the disease cannot be determined from our data. Finally, even in these very elderly subjects, the likelihood of being diagnosed as having asthma is closely related to the age-sex-standardized serum-IgE level before diagnosis. This finding appears to be in conflict with the recent study by Braman et al6 in which elderly subjects with asthma are said to have no significant elevation of serum IgE compared to elderly, hospitalized, control subjects. The mean log IgE levels (mistakenly called “geometric means”) of their “early” and “late” onset of asthma in elderly subjects were 1.65 and 1.80, respectively. The overall mean log IgE for their elderly subjects with asthma can be calculated from the raw data presented to be 1.73 + 0.376, even higher than the value for subjects with later asthma in the present study and nearly as high as that previously reported in our elderly subjects with established disease.’ However, their control subjects had a mean log IgE of 1.49 (no standard deviation can be determined from the data provided), a higher value than in our control elderly population. If it is assumed that there is a similar standard deviation in the control subjects in the study of Braman et a1.,6 as in their subjects with asthma, which should be true if there are no differences between the groups, the IgE levels of their subjects with asthma would be significantly elevated (p < O.Ol), despite the fact that almost all the patients with asthma were taking oral corticosteroid therapy and that the control subjects, being hospitalized patients, might not be entirely appropriate. From our data, it is tempting to conclude that a high serum-IgE level is truly a risk factor for asthma and not simply an associated finding in the disease. However, it remains possible that the critical relationship of IgE is to a mild, and as yet, undiagnosed asthmatic state. In this sense, the present findings only help confirm our previous observations revealing the extremely close relationship of asthma prevalence to serum IgE.13 Final proof of cause and effect will probably be impossible to obtain from epidemiologic data. Present findings do indicate, however, that the serumIgE level helps identify those subjects who will later
VOLUME NUMBER
88 6
be diagnosed as having asthma, even in very elderly adults, and that the relationship of asthma to serum IgE does not depend on the presence of frank, clinical disease.
REFERENC:ES 1. Broder I, Barlow PP, Horton RJM. The epidemiology of asthma and hay fever in a total community, Tecumseh, Michigan. 1. Description of study and general findings. J ALLERGY 1962;33:513-23. 2. Lee Hl!, Stretton TB. Asthma in the elderly. Br Med J 1973;4:‘?3-5. 3. Lebowitz MD, Knudson RJ, Burrows B. Tucson epidemiologic study of obstructive lung diseases. I. Methodology and prevalence of disease. Am J Epidemiol 1975;102:137-52. 4. Burr ML, Charles TJ, Roy K, Seaton A. Asthma in the elderly: an epidemiologic survey. Br Med J 1979;1:1041-4. 5. Ford RM. Aetiology of asthma: a review of 11.55 1 cases(1958 to 1968). Med J Aust 1969;1:628-31. 6. Braman SS, Kaemmerlen JT, Davis SM. Asthma in the elderly: a comparison between patients with recently acquired and longstanding disease. Am Rev Respir Dis 1991;143:336-40. 7. Burrow:; B, Barbee RA, Cline MG, Knudson RJ, Lebowitz MD. Characteristics of asthma among elderly adults in a sample of the glsneral population. Chest 1991;100:935-42.
Asthma
in the elderly
877
8. Dodge RR, Cline MG, Burrows B. Comparisons of asthma, emphysema, and chronic bronchitis diagnoses in a general population sample. Am Rev Respir Dis 1986;133:981-6. 9. Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983;127:725-34. 10. Barbee RA, Lebowitz MD, Thompson HC, Burrows B. lmmediate skin-test reactivity in a general population sample. Ann Intern Med 1976;84: 129-33. 11. Barbee RA, Halonen M, Lebowitz MD, Burrows B. Distribution of IgE in a community population sample: correlations with age, sex, and allergen skin test reactivity. J ALLERGY CLIN IMMUNOL
1981;68:106-11.
12. Burrows B, Hasan FM, Barbee RA, Halonen M, Lebowitz MD. Epidemiologic observations on eosinophilia and its relation to respiratory disorders. Am Rev Respir Dis 1980;122:709-19. 13. Burrows B, Martinez FD, Halonen M, Barbee RA, Cline MG. Association of asthma with serum IgE levels and skin-test reactivity to allergens. N Engl J Med 1989;320:271-7. 14. Burrows B, Lebowitz MD, Camilli AE, Knudson RJ. Longitudinal changes in forced expiratory volume in one second in adults: methodologic considerations and findings in healthy nonsmokers. Am Rev Respir Dis 1986;133:974-80. 15. Burrows B, Bloom JW, Traver GA, Cline MG. The course and prognosis of different forms of chronic airways obstruction in a sample from the general population. N Engl J Med 1987;317:1309-14.