Sputum and Pulmonary Function in Asthma*

Transcription

1 Sputum and Pulmonary Function in Asthma* C. Kevin Connolly, MB; N. Krishna Murthy, MB; Susan M. Alcock, RGN; and Robin ]. Prescott, PhD Study objective: To assess the relevance of sputum production to pulmonary function, in particular, persistent obstruction in patients with a primary clinical diagnosis of asthma. Design: Cross-sectional study of all patients currently followed up in secondary care in a defined locality. Setting: National Health Service and private clinics in north-east England. Patients: All attenders, aged 18 years or older, with asthma, confirmed by reversibility of peak expiratory flow (PEF) by ~ 1 5 % and to ~ Umin. Interventions: Pro forma history. Pulmonary function at attendance. Assessment of best function according to protocol. Measurement of actual FEV 1, FVC, and PEF at attendance. Measurements and results: We studied 772 subjects; 387 (50%) were male; mean age was 55 years; atopic, 51%; current smokers, 11.5%; ex-smokers, 36%; and never smokers, 52.5%. Best pulmonary function was lower in chronic sputum producers (PEF, 83.2 vs 95.8; FVC, 67.9 vs 81.7% predicted). Chronic sputum production and its negative relationship with best function was strongly associated with smoking. There was little relationship between chronic sputum and persistent obstruction in nonsmokers. There were no univariate associations between sputum during attacks, or its color, and pulmonary function, but after allowing for demographic factors, including smoking, green sputum was associated with persistent obstruction. There was little relationship between sputum and actual/best function at attendance. Conclusions: Chronic sputum production is associated with persistent obstruction principally in those who have smoked, suggesting that the association reflects smoking rather than asthma. There is no interaction with atopy. After allowance for cigarette smoking, there is an association between green sputum production during exacerbations and persistent obstruction. Green sputum during relapse in asthma may indicate inflammation that is relevant to prognosis. (CHEST 1997; 112:994-99) Key words: actual/best pulmonary function; asthma; best pulmonary function; COPD; FEY 1 ; FVC; inflammation; peak expiratory flow; sputum Abbreviations: PEF=peak expiratory flow Asthma is increasingly being accepted as an inflammatory disease. 1-3 Sputum production might reflect the intensity of the inflammatory response, and indeed it has been reported as being associated with increased risk of hospitalization. 4 In one of the few studies of hypersecretion in asthma, there appeared to be little association *From the Department of Medicine (Dr. Murthy and Ms. Alcock), Darlington Memorial Hospital NHS Trust, and the Memorial Hospital, University of Newcastle upon Tyne (Dr. Connolly), and the Mecbcal Statistics Unit (Dr. Prescott), University of Edinburgh. Supported in part by a fellowship (Dr. Murthy) from Duncan Flockhart, now part of Glaxo Wellcome. Manuscript received March 27, 1996; revision accepted April 3, Reprint mquests: C. Kevin Connolly, MB, Memorial Hospital, Hollyhw-st Road, Darlington, Co, Durham DL3 6HX, UK between lability of peak flow measurements and sputum production, while persistent airway obstruction did appear to be most frequent in the groups characterized by no sputum at all or where large volumes were produced. 5 Sputum history was added to the database of the Darlington and Northallerton long-term follow-up study into asthma at the first quinquennial review. This article reports the prevalence of sputum production and examines its relationship to pulmonary function. Particular attention is paid to the relevance of sputum production to the coexistence of COPD as assessed by best after-bronchodilator function. 4 6 Any prognostic inferences arising from associations demonstrated in this cross-sectional analysis will be tested as the longitudinal data become available. 994 Clinical Investigations

2 Subjects MATERIALS AND METHODS All subjects entered into the Darlington and Northallerton asthma study were eligible. The entry criteria, which have been described in detail elsewhere, 6 are summarized below: (l) patients aged 18 years or older; (2) a clinical diagnosis of asthma; (3) reversibility of peak expiratory flow (PEF) spontaneously or after bronchodilator, by at least 15%; a level of peak flow of > 200 Umin must have been achieved on more than one occasion; and ( 4) referal for a specialist opinion and subsequent observation for at least 1 year. Cigarette smokers were accepted only if the clinical diagnosis was firm, usually supported by either blood or sputum eosinophilia. Subjects are encouraged to attend the clinic having taken their bronchodilators as normal. New entrants and those still attending were seen at routine visits, and those unavailable for follow-up were called specifically for this study. Patients were not entered during emergency visits and entry was deferred for a period of not less than 3 months before accepting the regimen as unstable. Subject to this, actual function was that recorded at the routine or recall visit. The present study includes the original cohort 6 8 and a second group entered at the first 5-year review. One of the authors (C.K.C. ) was the only respiratory physician in the districts concerned and private patients were included. Thus, all patients deemed by their general practitioner to require referral locally were included. The area concerned is a rural pati of the north-east of England approximately 8000 km 2, with Darlington, an industrial and commercial town, in the north-east corner. The prevailing wind is south-westerly. Data ColleLtion This has been described in detail previous!y. 6 7 The history was taken by a consultant physician, a research f ellow, or an experienced research nurse. The demographic and social variables recorded included the follmving: (1) atopic status-prick skin tests were performed to at least: cat, house dust mite (Dennatophagoides pteronyssoides ), grass pollen, and Aspergillus Jumigatus, and recorded as follows: positive if typical weal ~ mm 3 greater than control, negative if no reaction, and otherwise indeterminate; (2) smoking habit-(a) current smokers, tobacco smoked within the last 3 months; (b) ex-smokers, previous tobacco smokers \ \ ~ complete t h abstention for at least 3 months; (c) never smokers, lifetime amount smoked <365 cigarettes; (d) passive smokers, nonsmokers 'vvith current smoker in the household; amount smoked was expressed in lifetime pack-years; and (3) histmy of sputum production-this was entered for the first time at the first 5-year review. The following were recorded: (a) occurrence of chronic sputum (for at least 3 months in at least 2 consecutive years); (b) presence of or increase in sputum during most relapses; (c) color of chronic sputum; (d) color of sputum during relapse. Pulmonary Function PEF was measured by the mini-\vright meter (Clement Clarke; Essex, UK ), taking the best of three recordings. FEV l and FVC were measured using the rolling wedge spirometer (Vitalograph; Maidenhead, UK). Actual function was as recorded in the clinic without further use of bronchodilator (subjects were not asked to withhold bronchodilator treatment before clinic attendance). Best or maximum PEF and FVC were established between January 1988 and attendance. Function was expressed as percentage of predictedy The highest reading was accepted as reflecting best function, subject to the follmving conditions of recording: (1) >80%: after bronchodilator; (2) 70 to 80%: after bronchodilator on normal suppressive medication (sodium cromoglycate, nedocromil, inhaled steroids, maintenance oral steroids) with 4-hourly recordings of peak flow for not less than 5 days; and (3) <70%: after bronchodilator and receiving prednisolone, 30 mgld for not less than 5 days, peak flow readings to be recorded 4-hourly, and to be consistent for the last 48 h. vvhere these conditions were not satisfied, best function was immediately established for the purpose of this study according to the above protocol. Actual function was that recorded at the review appointment, whether or not control was deemed as satisfactory by the physician at that time. It was expressed as percentage of predicted. Control of potentially reversible airway obstruction at attendance was assessed by actual/best function, expressed as a percentage. Analysis and Statistical Methods For the purposes of analysis, the subjects were divided into four groups by history of sputum production: (1) chronic sputum producers 'vvith increases during relapse; (2) chronic sputum producers only; (3) subjects who produced sputum during relapses only; and (4) subjects who never had sputum. Conventional statistical methods of analysis were employed. Association between categorical variables was by the x 2 test. Comparison of continuous variables between categories of patients was by one-way and two-way analysis of variance, as appropriate. Comparison between groups with allowance for demographic vajiab!es was by analysis of vmiance. All significance tests reported are based on two-tailed tests of significance. RESULTS Seven hundred seventy-two subjects (387 male, 385 female) were recruited. Mean age of the sexes (52 years; range, 19 to 86 years) was similar, as was the incidence of atopy (52% male; 50% female). Fifty-four (14%) men and 35 (9%) women were current smokers, and 151 (39%) men and 254 (66%) women had never smoked. Nearly all (97%) were receiving maintenance corticosteroid therapy, which was taken orally by 16%. The characteristics of the subjects in each of the sputum categories are shown in Table l. There is a positive association between chronic sputum production and age, particularly in men (men, 62 vs 50 years; women, 59 vs 53 years). Sputum production during relapse is associated with relative youth in both sexes. As expected, chronic sputum production was generally related to smoking. Nevertheless, 29% of 151 nonsmoking men and 34% of 254 nonsmoking women did have chronic sputum. Furthermore, the distribution of cigarette smoking in the 54 women with chronic only sputum was similar to that of the population as a whole (expected nonsmokers 66%, actual nonsmokers 69%), and the sputum was more likely to be purulent than in men (women, 39% vs men, 21 %; p<0.01). Although the CHEST I 112 I 4 I OCTOBER,

3 Table 1-The Characteristics of the Patients With the Different Histories of Sputum Production, Showing the Mean Age, the Proportion Atopic, and the Distribution of Smoking Habit Within Each Group Male, Mean Age, Atopic, Groups No. No.(%) yr No.(%) Chronic sputum increased (55) (45) in relapse Chronic sputum only (58) (40) Relapse sputum only (39) (62) No sputum (52) (52) p value* < *Overall significance of difference between sputum groups. Never Smoked, Current Smoker, Ex-Smoker, No. (%) Purulent, No. (%) No. (%) (Includes Passive) No. (%) 32 (18) 75 (43) 68 (39) 135 (77) 17 (13) 59 (46) 52 (41 ) 35 (27) 21 (10) 57 (27) 134 (63) 134 (63) 18 (7) 85 (33) 154 (60) < < number of male (134) and female (123) subjects with no sputum was approximately equal, chronic sputum production was more frequent in men (170 vs 133), and relapse only sputum more frequent in women (83 vs 129) (sputum category by gender l= 15.2, p=0.002). The proportion of atopic subjects was highest in those with relapse only sputum (vs no sputum x 2 =4.5, p=0.03), and lowest in those who produced sputum chronically, which was not exacerbated during relapse. There was no interaction with gender. The relationships between pulmonary function and sputum production are shown in Table 2. There were no interactions with gender or atopy. Poorer actual pulmonary function was associated with chronic sputum production, whether worse in relapse or not, but not when sputum was seen only in relapse. There was a small difference in actuavbest function between chronic sputum producers and the rest (82% vs 85%). This was significant (p=0.04) and consistent throughout the smoking categories (smoking categories, p=0.99; interaction of smoking/sputum, p=0.50). However, when allowance was made for the other social and demographic variables by multiple regression analysis, actuavbest function was no longer significantly less in chronic sputum producers. Since control, as assessed by actuavbest function, Table 2-The Actual Percent Predicted PEF and FEV 1 Stratified by History of Sputum Production* Sputum Production PEF FEV 1 Chronic sputum increased 68.3 (170) 58.4 (168) in relapse, o/o pred (No.) Chronic sputum only, 70.8 (126) 63.2 (124) o/o pred (No.) Relapse sputum only, 85.0 (207) 75.9 (205) o/o pred (No.) No sputum, o/o pred (No.) 81.9 (242) 74.7 (237) Analysis of variance: F 22.7 (p< ) 23.9 (p < ) *% pred=percent predicted. 996 was responsible for only a small proportion of the relationship between pulmonary function and sputum production, differences in pulmona1y function were almost entirely associated with the persistent component as assessed by best function. Best function was consistently worse in subjects who produced chronic sputum. There was a highly significant difference between the best function of those who produced sputum chronically and those who did not (PEF, 83.2 vs 95.8%; FEYv 67.9 vs 81.7%; FVC, 78.9 vs 89.4% predicted; p< in all cases). This applied particularly when sputum increased or became more purulent during relapse (Table 3). The differences remained highly significant after the multivariate analysis allowing for social and demographic variables, but there was an interaction with smoking habit. The greater association between poor best function and sputum production in smokers and ex-smokers than in nonsmokers is demonstrated in Table 3. However, in passive but otherwise nonsmokers, the differences in best function behveen chronic sputum producers and the rest are similar to those seen in ex-smokers (best PEF, 84% vs 97%, p<0.01; best FEVv 66% vs 83%, p<o.ol). Stratification by atopic status demonstrated no interactions. Although best FEV 1 was higher in atopies than nonatopics (80.3% vs 73.5%, p=0.007), this difference disappeared after correction for age. Multiple tests for interaction between the interrelationships of age, gender, sputum production, sputum color, and pulmonary function with atopy were all negative. There was an inverse relationship between increasingly intense therapy (dose of inhaled corticosteroid and then dose of oral corticosteroid) and sputum production (p=0.0001) and all measurements of pulmonary function (p=0.001). However, multiple tests for interaction of the interrelationships among age, gender, sputum production, atopy, and pulmonary function with therapeutic step were all negative. After allowance for these factors, there was a consistent relationship between sputum production and pulmonary function across the therapeutic steps. A Clinical Investigations

4 Table 3-The Best Pulmonary Function of Subjects With Differing Histories of Sputum Production, Showing the Values for Each Smoking Category Sputum Production Current Smoker (No. ) Ex-Smoker (No.) Never Smoked (No. ) All (No. ) Best PEF, % predicted* Chronic sputum increased in relapse 74.9 (32) 75.4 (74) 92.1 (65) 81.6 (171) Chronic sputum only 79.6 (17) 80.1 (59) 93.0 (51) 85.2 (127) Relapse sputum only 88.9 (20) 96.4 (57) (130) 98.1 (207) No sputum 95.7 (16) 92.5 (80) 94.3 (143) 93.8 (240) Best FEY, % predicted 1 Chronic sputum increased in relapse 52.2 (32) 63.0 (73) 77.1 (63) 66.2 (188) Chronic sputum only 67.7 (17) 66.1 (57) 75.8 (49) 70.2 (123) Relapse sputum only 75.8 (19) 80.8 (56) 85.7 (130) 83.5 (205) No sputum 81.8 (16) 78.6 (79) 84.2 (142) 82.1 (237) *Significance levels from two-way analysis of vatiance: smoking categories, p< O.OOOl; sputum categories, p< ; interaction of smoking! sputum, p= Significance levels from two-way analysis of variance: smoking categories, p< ; sputum categories, p< ; interaction of smoking! sputum, p=0.09. significant interaction at the 5% level with social class suggests that the relationship between poor pulmonary function and sputum production may be enhanced in the socially deprived, irrespective of therapeutic step, but this result should be treated with caution in view of multiple testing. On univariate analysis, there was no association between sputum color dming attacks and pulmonary function. After allowing for social and demographic variables, there was a significant trend to decreasing PEF with increasing discoloration of sputum from white to yellow to green (Table 4). Similarly, after correction, there was a significant decline in both best FEV 1 and best PEF but no significant difference in best FVC (not shovm in the table). The differences in actual!best PEF were only a little less than those demonstrated for best PEF, but they did not quite reach statistical significance. DISCUSSION This article presents the cross-sectional analysis from a parallel cohort study of adult asthmatics. Asthma and COPD are regarded as different entities. The hypotheses to be tested include, first, that asthma is a risk factor for COPD and second, that COPD is a risk factor for mortality in asthmatics. We have regarded maximal best function on optimal therapy as the physiologic measurement reflecting the chronic or persistent component, and the proportion of that achieved (actual/best) as reflecting the uncontrolled reversible component at the time of observation. Although actual/best function at any one time does not exclude the possibility of poor control at other times, in a large group of subjects, mean actual/best function is likely to be a crude reflection of the control of the group as a whole. In contrast to best function, it has been shown to be independent of therapeutic steps in asthmatics treated by a group of interested physicians.jo We accept that the use of a single measure of actual/best function reduces the chances of showing the associations between sputum production and control of asthma. However, and for similar reasons, the use of best rather than actual function (the traditional value in epidemiologic studies) enhances Table 4-Pulmonary Function in Subjects Producing Different Colors of Sputum During Attacks* Pulmonary Function White (No. ) Yellow (No. ) Raw values Actual PEF 76.6 (199) 76.6 (148) Best PEF 90.5 (200) 89.2 (149) Actual/best PEF 83.4 (200) 83.5 (150) Best FEV (197) 74.7 (146) After correction Actual PEF 78.6 (196) 77.0 (146) Best PEF 91.9 (197) 89.5 (147) Actual/best PEF 84.2 (197) 83.5 (148) Best FEV, 77.2 (194) 75.5 (144) *The results are given before and after correction for demographic variables. 1 Analysis of variance. 1 Analysis of covariance. Green (No.) p Value 73.6 (147) (146) (145) (144) (143) (144) (143) (142) CHEST I 112 I 4 I OCTOBER,

5 the changes of demonstrating relationships. Furthermore, the hypothesis that COPD as reflected by poor best function is a risk factor for mortality has been confirmed in the analysis of mortality at 10 and 5 years, respectively, in the 1983 and 1988 cohorts We did confirm that a substantial minority of asthmatics produced sputum despite being lifelong nonsmokers. 5 In view of the importance of decline in best function, we were patticularly interested whether this sputum production was a risk factor for COPD in smokers as well as nonsmokers. It was fortunate for the power of the study that the proportion of smokers and nonsmokers was approximately equal. Most were ex-smokers. The reliability of the claim to be an ex-smoker was not checked at this review, but it was on entry of the 1983 cohort, when >85% of claimed ex-smokers had a carbon monoxide level of <0.5%. 7 In view of this high proportion, self-reported ex-smoking has been accepted as valid on subsequent occasions in this study. As anticipated, there was a close relationship between tobacco smoking and chronic sputum production. There was no interaction between this relationship and atopy, although atopic asthmatics are more likely than nonatopics to produce sputum dming relapse. This finding is in conflict with that of Openshaw and Turner-Warwick, 5 who found that sputum volumes were greater in those with negative prick tests. This was a clinical study of asthmatics and not an epidemiologic study. The relationship between atopy and bronchial hyperresponsiveness and other abnormalities of pulmonary function is likely to be closer in a group chosen for their asthmatic symptoms than in the general public. The results, therefore, cannot be extrapolated from diagnosed asthmatics to the population as a whole and do not exclude the possibility of an interaction with atopy in the relationships between chronic bronchitis and chronic airway obstruction in the general population. We did not show any interaction with atopy in the association between chronic sputum production and pulmonary function, suggesting that atopic status is not important in determining the risk of accelerated decline of pulmonary function. However, the interaction demonstrated with cigarette smoking does confirm that there is an increased tendency to develop persistent obstruction in those asthmatics who are smokers or ex-smokers who produce chronic sputum,l' particularly when this is increased in relapse. In population studies, tobacco smoking appears to be the major determinant of the development of airway obstruction, principally in male subjects.l8 However, in this study, the negative test results for interaction suggest that it is as relevant in women as in men. Virtually all subjects were receiving inhaled or oral corticosteroids as preventive therapy, which might suppress sputum production due to inflammation, particularly that due to the asthmatic process. Nevertheless, there were two findings which suggested that sputum production during exacerbation may indeed reflect relevant inflammation. First, there was a weak association between poor actuallbest function and chronic sputum production independent of smoking habit. Although control of wheeze at attendance is a crude measure of asthma control, we have shown that in clinical practice, actuallbest function may be titrated successfully against therapy in the stepwise management of asthma. 10 We do not discourage patients from taking bronchodilators before attending the clinic, for two reasons: first, it is difficult to be sure how long treatment with bronchodilators should be withheld to ensure no bronchodilator activity; and secondly, if treatment with bronchodilators is not withheld, then the actuallbest function demonstrated at the clinic is a measure of the control that the patient regards as acceptable. It is recognized that many subjects may have apparently well-controlled conditions at attendance, witl1 actuallbest function at or near 100%, but nevertheless they are symptomatic. These patients may include those whose reversible component is controlled at the time by appropriate therapy and others who have wheeze that is currently irreversible by [3-agonist due to mucus retention or the development of secondarily irreversible obstruction. 19 However, if actuallbest function is actually low at attendance, then control is clearly unsatisfactory, certainly at that time and probably at other times. Correction by multivariate analysis, including social variables, apparently denies the relevance of the association between poor control and sputum production. However, poor control of asthma has already been shown to be associated with age and lower social class, 7 where sputum production tends to be higher. This suggests that there may be a common association of poor control and chronic sputum production in the elderly socially deprived. It is accepted that some of the inflammation in some of these subjects may not have been allergic in nature. Irritation by the very inhaled steroids intended to reduce inflammation might contribute. Furthermore, the dose of inhaled steroids may have been suboptimal, but actuallbest function for the distribution of corticosteroid dose was similar to that in adjacent districts, 10 and there was no interaction with therapeutic step. If treatment had been suboptimal in some subjects, then one would expect greater sputum production in the lower therapeutic steps. It is also accepted that it is difficult to be certain of the extent to which "sputum" production reflects upper airway secretions. Green 998 Clinical Investigations

6 sputum may occur with the degradation of either neutrophils or eosinophils It may reflect stasis as well as severe inflammation. Although simple chronic bronchitis in cigarette smokers may not be associated with the development of persistent obstruction, 22 COPD is more likely to become more prevalent in smokers with increasing degrees of purulence, reflecting inflammatory response to bacteria and cigarette smoke. The study suggests that, with certain exceptions, sputum production is not a good measure of inflammation in asthma. The results presented herein strongly suggest that chronic sputum production in asthma is mainly relevant where it reflects respiratory damage due to tobacco smoke, with a possible small further contribution from adverse factors associated with social deprivation. Green sputum during exacerbations may be more relevant to deterioration associated with asthma itself. The green color of sputum might reflect a stasis following the development of COPD, but the multivariate analysis enhances the inverse relationship between green sputum and best function. Any relationship between green sputum and persistent obstruction depending on stasis consequent upon deterioration should not be affected by this exercise. As smoking, the principal determinant of chronic sputum production, was included in the model, this finding is compatible with the hypothesis that in nonsmoking asthmatics, green sputum during times of respiratory distress may indeed r eflect a more severe inflammatory response, which is associated with development of persistent obstruction. All the conclusions of this analysis are necessarily restricted b y the cross-sectional design of the study, but there will be an opportunity to test them as the longitudinal data become available. REFERENCES 1 H olgate ST. Asthma: past, present and future. Eur Respir J 1993; 6: Barnes PJ. Frontiers in mechcine: new aspects of asthma. J Intern Med 1992; 231: Wenzel SE. Asthma as an in flammatory disease. Ann Allergy 1994; 72: Connolly CK. Management of asthma in out-patients. J R Coli Physicians Lond 1983; 17: Openshaw PJ, Turner-Warwick M. ObseJvations on sputum production in patients with variable airflow obstruction: implications for the chagnosis of asthma and chronic bronchitis. Respir Med 1989; 83: Connolly CK, Chan NS, Prescott RJ. The relationship between age and duration of asthma and the presence of persistent obstruction in asthma. Postgrad Med J 1988; 64: Connolly CK, Chan NS, Prescott RJ. The influence of social factors on the control of asthma. Postgrad Med J 1989; 65: Connolly CK, Prescott RJ. Pulmonary function and drug regimens in asthmatics. Br J Clin Pract 1990; 44: Cotes JE. Lung function: assessment and application in medicine. 4th ed. Oxford: Blackwell, 1979 lo Connolly CK, Prescott RJ, Alcock SM, et al. Actual over best function as an outcome measure for asthma. Respir Med 1994; 88: ll Connolly CK, Roy PM, Alcock SM, et al. The Darlington and Northallerton asthma study: mortality at 10 years. Eur Respir J 1995; 8:316S 12 Connolly CK, Alcock SM, Roy PM, et al. Darlington and Northallerton asthma study: mortality 1988/93. Am J Respir Crit Care Med 1996; 153:A Altounyan REC. Changes in histamine and atropine responsiveness as a guide to chagnosis and evaluation of therapy in obstructive airways chsease. In: Pepys J, Frankland AW, eds. Disochum cromoglycate in allergic airways disease. London: Butterw01th, 1970; Cockroft DW, Ruffin RE, Dolovitch J, et al. Allergen-induced increase in non-allergic bronchial reactivity. Clin Allergy 1977; 7: Sears MR, Herbison GP, Holdaway MD, et al. The relative risks of sensitivity of sensitivity to grass pollen, house dust mite and cat dander in the development of childhood asthma. Clin Exp Allergy 1989; 84: Bessot JC, de Blay F, Pauli G. From allergen sources to reduction of allergen exposure. Eur Respir J 1994; 7: Krzyzanowski M, Camilli AE, Lebowitz MD. Relationships between pulmonary function and changes in chronic respiratory symptoms: compmison of Tucson and Cracow longituchnal studies. Chest 1990; 98: Sherrill DL, Lebowitz MD, Knudson RJ, et al. Smoking and symptom effects on the curves of long function growth and decline. Am Rev Respir Dis 1991; 144: Alcock SM, Prescott RJ, Connolly CK. The inter-relationships of actual/best and best function, regimen status and symptoms in chronic asthma. Thorax 1995; 50(suppl 2):A53 20 Von Hoesslin, H. Das sputum. Berlin: Verlag von Julius Springer, 1926; Hirsh JG. Neutrophils and eosinophil leukocytes. In: Zweifach BW, Grant L, McCluskey RT, eds. The inflammatory process. New York: Academic Press, 1965; Fletcher C, Peto R. The natural history of chronic airflow obstruction. BMJ 1977; 1: CHEST I 112 I 4 I OCTOBER,