Qual Life Res (2007) 16:1555–1565 DOI 10.1007/s11136-007-9267-4
Impact of smoking on asthma symptoms, healthcare resource use, and quality of life outcomes in adults with persistent asthma Oren Shavit Æ Arlene Swern Æ Qian Dong Æ Kathleen Newcomb Æ Vasilisa Sazonov Kocevar Æ Stephanie D. Taylor
Accepted: 11 September 2007 / Published online: 5 October 2007 Springer Science+Business Media B.V. 2007
Abstract Background Smoking habits of asthmatics are similar to those of the general population. However, little attention has been paid to the associations between smoking and asthma-related outcomes. Objective To evaluate relationships between cigarette smoking, asthma symptoms, and asthma-related resource utilization in subjects with persistent asthma. Method A stratified, random sample of adults from France, Germany, and the UK with persistent asthma were surveyed in 2001 through 2004. Statistical analyses compared asthma symptoms and healthcare resource utilization for cigarette smokers compared with those for nonsmokers. Results Analyses included 1109 subjects with persistent asthma symptoms in 2001–2003 and 852 subjects with persistent asthma symptoms in 2004. Using multivariate analysis of data from eligible subjects in 2004 that did not have concomitant chronic obstructive pulmonary disease and adjusting for subjects’ baseline and demographic characteristics, cigarette smokers were more likely to experience nighttime symptoms (OR 1.46, 95% CI 1.07, 1.97 P = 0.015) and were more likely to use healthcare
O. Shavit School of Pharmacy, Temple University, Philadelphia, PA, USA A. Swern Q. Dong K. Newcomb Merck Research Laboratories, Merck & Co. Inc, Rahway, NJ, USA V. Sazonov Kocevar S. D. Taylor (&) Outcomes Research – WS2E65 Global Human Health Marketing, Merck & Co. Inc., P.O. Box 100, Whitehouse Station, NJ 08889-0100, USA e-mail:
[email protected]
resources than were non-smokers (P £ 0.004). Findings were similar in a secondary analysis of subjects £55 years of age. Conclusions Cigarette smoking appears to be associated with more asthma symptoms and more ED visits and hospitalizations in adults with persistent asthma. Keywords
Asthma Health survey Smoking
Abbreviations AIDS Autoimmune deficiency syndrome BMI Body mass index CI Confidence interval COPD Chronic obstructive pulmonary disorder ED Emergency department GERD Gastroenteric reflux disease GP General practitioner HIV Human immunodeficiency virus ICS Inhaled corticosteroid OR Odds ratio OTC Over-the-counter QoL Quality of life SABA Short-acting b-agonist
Introduction How ironic that tobacco was recommended as a treatment for asthma for centuries following its introduction by Sir Walter Raleigh in the 16th century [1]. The harmful influence of smoking tobacco on the development of asthma and worsening of its conditions really was not recognized until well into the 20th century. The World Health Organization has estimated that approximately 25%
123
1556
of adults in developed countries smoke cigarettes, with rates as high as 38% in some European countries [1, 2]. The prevalence of cigarette smoking in the United Kingdom (UK) and the United States (USA) declined during the 1970s and early 1980s, reaching a low of 26% of women and 28% of men in the UK in 2001 and approximately 19% of women and 23% of men in the USA in 2004 [2–4]. Although it would be a logical conclusion that asthma patients would have a lower prevalence of smokers because of the additive effects on respiratory function, smoking rates among asthmatic patients are, in fact, relatively similar to those of the general population [5–7]. Former smokers also account for a large proportion of adult asthmatic patients. There is conflicting evidence regarding the risk of active smoking in the development of asthma; however, a common finding is that cigarette smoking exacerbates asthma symptoms [1, 5, 6]. The effect of cigarette smoking on pulmonary function has not been extensively studied because smokers are often excluded from population samples in asthma studies; however, in studies that have included smokers, smoking has been associated with several adverse effects on clinical, physiologic, radiologic, and inflammatory features of asthma [4, 5, 8, 9]. A limited number of studies have shown smokers have more severe and more frequent symptoms; an accelerated decline in lung function; reduced efficacy of certain drugs, including asthma medications; increased hospital rates; and increased mortality rates compared with these events in non-smokers [4, 8, 10]. The purpose of these analyses was to evaluate the relationship between smoking, asthma symptoms, asthmarelated resource utilization, and quality of life (QoL) scores in a sample of survey participants from Germany, France, and the UK.
Methods Data for the analyses were obtained during 2001–2004 from the National Health and Wellness Survey (NHWS), a cross-sectional, self-administered randomized annual survey of non-institutionalized adults (‡18 years old) in Europe and the USA [11]. The survey is conducted annually from a population of approximately 42,000 adult consumers in the USA and 37,000 in the UK, France, Germany, Spain, and Italy, covering more than 75 therapy areas. Participants are recruited by invitation. The survey collects an abundance of information about the overall health of participants, including histories of illnesses, smoking status, resource use, illness symptom severity, medications and other treatments they use, their attitudes towards their health, and subsequent behaviors. In addition,
123
Qual Life Res (2007) 16:1555–1565
the survey has several questions relating to demographics. Data are then verified against the national health statistics of each country. The NHWS sampling frame is stratified based on knowledge of previous response rates by different groups. For instance, if young, minority male subjects are less likely to respond than the general population, invitations would be sent to a greater number of young minority males, to provide an adequate representation within the total sample. That said, the sample itself is found to be acceptable for analysis and it does in fact approximate the demographic characteristics of the national population. For the present analyses, consumer healthcare data were limited to a nationally representative, stratified, random sample of survey participants from Germany, France, and the UK. Participants in 2001–2003 received and submitted the questionnaire by postal mailing. They completed the questionnaire at home, with instructions to complete the survey for only themselves, not for any others in the household. In 1999, pilot questionnaires, identical to those used in 2001–2003, were mailed to a sample population in each country (approximately 10% of the target population) to determine that the questionnaire was clear and comprehensive for the data to be evaluated and that the demographics were representative of the population-atlarge. In 2004, participants were drawn from an Internet panel maintained by Harris Interactive, and questionnaires were completed and submitted via the Internet, with a similar warning as for the mailed surveys for people to assess only themselves when responding to the questions. Before launching the online survey to the target population, a pilot release of the questionnaire, which had been modified from the mailed survey used in 2001–2003, was completed by a sample population that represented approximately 10% of the target population. Before release, responses from the online sample were evaluated side-by-side with the mailed surveys that had been previously submitted for assurance that the responses received via the Internet also were representative of the target population. Subjects in the 2004 survey received at least one email reminder to complete and submit the survey. The questionnaire in 2004 asked direct questions regarding specific medications used by the subjects for each disease condition, whereas in 2001–2003, respondents were simply asked to list all prescription and over-thecounter (OTC) medications, without specifying the associated disease or disorder. Since several questions specific to asthma were asked in the 2004 survey when the respondent indicated they had experienced asthma within the past 12 months, data from that survey are presented here as the primary analysis. Data collected in 2001–2003 have been collapsed into a single data set for the analysis since the same questionnaire was used for all 3 years.
Qual Life Res (2007) 16:1555–1565
Subjects The original data set in 2004 comprised 1895 subjects with physician-diagnosed asthma. Of these, 950 were excluded for indicating they had intermittent asthma, and 93 were excluded for reporting they had been diagnosed with chronic obstructive pulmonary disease (COPD), leaving 852 subjects who responded to questions regarding their smoking status and who indicated they had physiciandiagnosed persistent asthma with no diagnosed COPD (Fig. 1) [12]. In 2004, the survey asked: ‘‘Have you ever smoked cigarettes?’’ ‘‘Do you currently smoke cigarettes?’’; and ‘‘How many cigarettes do you smoke per day?’’ Of the 3920 subjects in the 2001–2003 survey, 1946 reported having been persistent asthma that was diagnosed by a physician and provided a response to their smoking status (Fig. 1). In 2001–2003, data regarding smoking status was based on two questions: ‘‘Do you smoke cigarettes?’’ and ‘‘If yes, how many cigarettes do you smoke per day?’’ The 2001–2003 survey did not ask specific questions regarding the presence of COPD, so 837 subjects over 55 years of age were excluded from analysis in an attempt to eliminate as many subjects as possible that had respiratory symptoms that might have been caused by COPD rather than (or in addition to) asthma, leaving 1109 subjects for inclusion in the secondary analysis. This was justified
1557
by the 2004 survey. In 2004, 18% of subjects over 55 reported having COPD compared with approximately 7% of subjects 55 and under that reported having COPD in addition to their asthma. Reported baseline characteristics included age, body mass index (BMI), education (high school or less, some college, college degree or higher), marital status, current employment status, health insurance status, years of having asthma, asthma severity, asthma exacerbation triggers, asthma medications, compliance, concomitant nasal allergies or hay fever, and the presence of co-morbidities. The range of co-morbidities and focus were limited to selfreported clinical problems that were the most common or are potentially associated with asthma, including gastroenteric reflux disorder (GERD), angina pectoris, arrhythmia, atrial fibrillation, cancer, congestive heart failure, diabetes, myocardial infarction, HIV/AIDS, stroke, epilepsy, non-cardiac chest pain, and pain.
Outcomes Outcomes included asthma symptoms and related resource use. Symptom-related outcomes included frequency of use of asthma medications, such as short-acting b-agonists (SABAs) in the past month, and subjective evaluation regarding control of one’s health. Resource use outcomes included visits to a general practitioner (GP) or specialist (pulmonologist, cardiologist, allergist, otolaryngologist, oncologist, or internist) in the past 6 months and the frequency of those visits, number of visits to the emergency department (ED) or hospitalizations in past 6 months. In the 2004 survey, outcomes also included identification of subjects having daytime and nighttime asthma symptoms, the frequency of these symptoms, and results of the SF-8TM Health Survey [13]. Translations for the SF-8 Health Survey were developed for these countries following standard translation and validation procedures that were first developed for the SF-36 Health Survey during 1991–1993 [13, 14]. These outcomes data from the SF-8 Health Survey were not collected in 2001–2003.
Analyses
Fig. 1 Schematic representation of data samples in analyses
The primary analysis in 2004 compared outcomes data between smokers and non-smokers with persistent asthma and no COPD (Fig. 1). A secondary analysis of mutual outcomes between smokers and non-smokers included subjects 55 years of age and younger from the original data sets in both surveys (Fig. 1). Although questionnaires for 2001–2003 and 2004 were not identical, the characteristics of the subjects were similar enough to provide an
123
1558
opportunity to independently examine several outcomes from the populations within each survey. Within each survey, Student’s t test was used to evaluate continuous demographic and outcomes variables, and chisquare tests were used for categorical demographic and outcomes variables. Logistic regression models were used to calculate the odds of reporting daytime or nighttime asthma symptoms during the previous month; use of SABAs or other asthma medications during the previous month; visits to a GP, specialist, alternative health provider, or ED during the previous 6 months; hospitalization during the previous 6 months; or not feeling in control of one’s health, based on smoking status. Subjects reporting daytime and nighttime symptoms were categorized by frequency (i.e., nighttime symptoms more than 2 nights per month vs. 2 or fewer nights per month and daytime symptoms more than 2 times per week vs. 2 or fewer times per week). All baseline characteristics were then included in multivariate logistic regression models using both stepwise forward and backward procedures with significant level of entry into the model set at 0.1 and significant level to stay of 0.1. All covariates found to be significant (P \ 0.10) in any of the models were included in the final multivariate logistic regression models to control for their potential confounding effect. Covariates considered in the multivariate models are age (continuous by year); gender; asthma severity, education level; employment status; health insurance status; years since asthma diagnosis; taking controller medication; GERD; nasal allergy, selfreported compliance; having any of the selected co-morbidities of interest; and whether asthma exacerbations were caused by allergy, stress, exercise, weather, illness, or pollutants. BMI was compared between overweight and obese subjects versus ‘‘normal’’ subjects. Among the covariate factors, age, asthma severity, education level, employment status, health insurance status, and causes of asthma were significant.
Results Primary analysis 2004—sample characteristics The distribution of the 852 subjects included in the primary analysis of data in the 2004 survey is presented by various characteristics of interest in Table 1. Approximately 32% of the subjects reported they currently smoked cigarettes and 32% indicated they had never smoked. Although 37% reported they were former smokers, there were no data provided for duration of smoking or duration of smoking cessation; therefore, for this analysis, subjects who never smoked or were former smokers are considered non-smokers.
123
Qual Life Res (2007) 16:1555–1565
Non-smokers were approximately 4 years older (P \ 0.001), had a higher education level (P \ 0.001), and were more likely to be overweight or obese (65%) than were smokers (55%) (P = 0.024). Other socioeconomic indicators (marital status, employment status, and insurance status) were similar for smokers and non-smokers (data not shown). On average, non-smokers had asthma nearly 3 years longer than smokers had (P = 0.004). There was no significant difference (P = 0.130) between smokers and nonsmokers in self-reported severity of asthma (moderate or severe asthma versus mild asthma) (Table 1). Stress was the most frequent asthma trigger for smokers (48% vs. 41% for non-smokers; P = 0.039), and pollutants was the most frequent trigger for non-smokers (48% vs. 38% for smokers; P = 0.010) (Fig. 2). Among all subjects that reported using a controller medication, 61% reported they were using inhaled corticosteroids (ICS) alone. Use of controller medications and ICS was similar between smokers and non-smokers. There was no significant difference in compliance between nonsmokers (75%) and smokers (69%) (Table 1). One or more selected co-morbidities of interest were reported by 59% of all subjects, with no significant differences between smokers and non-smokers. Other comorbidities were reported by 97% of the population and were significantly associated with cigarette smoking (99% for smokers vs. 96% for non-smokers; P = 0.019). There was no significant difference in the reporting of nasal allergies and/or hay fever by non-smokers (63%) than by smokers (56%) (Table 1).
Primary analysis 2004—outcomes Outcomes data for smokers and non-smokers in the primary analysis of data from the 2004 survey are shown in Table 2. A small percentage (11%) of all subjects reported using OTC medications during the previous month for control of asthma symptoms. SABAs were used to relieve asthma symptoms by 75% of the subjects, irrespective of smoking status. Use of SABAs or OTC medications and the number of days SABAs were used for asthma symptom relief were not associated with smoking status (Table 2). Daytime symptoms were reported by 84% of all subjects. Of these, 67% had daytime symptoms more than twice a week (considered mild asthma or worse), and 29% experienced daytime symptoms at least once daily (moderate asthma or worse) [15]. The odds of experiencing daytime symptoms did not differ significantly between smokers and non-smokers (Table 2).
Qual Life Res (2007) 16:1555–1565 Table 1 Characteristics of subjects with persistent asthma and without COPD in 2004 survey
1559
Variable
Smokers (n = 271)
Non-smokers (n = 581)a
Female gender, n (%)
183 (68)
364 (63)
Age, yr (mean ± SD)
41.20 ± 15.08
45.36 ± 16.54
Cigarettes per day (mean ± SD)
17 ± 11
–
High school graduate or less
123 (45)
178 (31)
Some college
71 (26)
153 (26)
College graduate/Advanced
77 (28)
250 (43)
\25 kg/m2 (normal)
118 (45)
198 (36)
25–\30 kg/m2 (overweight)
80 (31)
187 (34)
‡30 kg/m (obese)
62 (24)
169 (31)
Asthma duration, yr (mean ± SD)
15.36 ± 11.95
18.28 ± 14.03
Body mass index, n (%)
0.024
Asthma severity, n (%)
a
Includes subjects that reported they were former smokers COPD = chronic obstructive pulmonary disease, GERD = gastroenteric reflux disease; ICS = inhaled corticosteroid; SD = standard deviation
0.167 \0.001 \0.001
Education, n (%)
2
P value
0.004 0.130
Mild
105 (39)
266 (47)
Moderate Severe
138 (52) 23 (9)
255 (45) 49 (9)
Use of controller medications, n (%)
179 (66)
406 (70)
0.262
Use of ICS controller only, n (%)
74 (27)
153 (26)
0.765
Self-reported compliance, n (%)
188 (69)
436 (75)
0.082
GERD, n (%)
20 (7)
61 (10)
0.148
Nasal allergy/hay fever, n (%)
153 (56)
364 (63)
0.085
Selected co-morbidities, n (%)
165 (61)
334 (57)
0.348
All other co-morbidities, n (%)
268 (99)
557 (96)
0.019
Fig. 2 Self-reported triggers of asthma exacerbations in smokers and non-smokers surveyed in 2004
Nighttime symptoms were reported by 61% of all subjects, and 85% of these subjects reported having nighttime symptoms at least twice a week (mild or worse asthma). More than half of those reporting nighttime symptoms (55%) experienced these symptoms at least one night a week (moderate asthma or worse). The odds of experiencing nighttime symptoms were 46% higher for smokers then for non-smokers (OR = 1.46; 95% CI = 1.07–1.97; P = 0.015) (Table 2). Multivariate logistic regression results show that the OR for nighttime symptoms did not change appreciably after adjusting for covariates. Irrespective of smoking status, approximately half of the subjects (52%) reported they did not feel in control of their
health, with no significant difference (P = 0.250) detected between smokers and non-smokers. The mental component score from the SF-8 Health Survey was significantly higher (i.e., better) (P \ 0.001) for non-smokers (45.60 ± 11.63) than the score for smokers (42.13 ± 12.04) (Fig. 3). The mean physical score was similar between smokers (42.96 ± 11.43) and non-smokers (43.44 ± 11.17), however. Visits to a GP for any reason in the previous month were reported by 89% of the subjects; 39% reported a visit to a specialist, and 29% had at least one visit to an alternate healthcare provider. Smoking status was not a significant factor for requiring these visits; however, visits to the ED for any reason (OR 1.86; 95% CI 1.30–2.67; P \ 0.001) as well as visits to the ED because of an asthma exacerbation (OR 1.78; 95% CI 1.17–2.72; P = 0.007) were more frequent for smokers than for non-smokers. The odds of being hospitalized were 80% higher for smokers than for nonsmokers (OR 1.80; 95% CI 1.20–2.68; P = 0.004), with no appreciable difference in the odds after adjusting for covariates (Table 2).
Secondary analysis—sample characteristics Distribution of subjects with persistent asthma and under 55 years of age included in the secondary analysis is shown
123
1560
Qual Life Res (2007) 16:1555–1565
Table 2 Results of logistic regression for primary analysis of association of smoking and asthma-related symptoms, medications, and well-being and asthma-related resource use in 2004 survey Primary analysis (persistent asthma, no COPD) Smokers (n = 271)
Non-smokers (n = 581)
Odds ratio (95% CI)
230 (86%)
475 (83%)
1.28 (0.84, 1.93)
128 (27%)
P = 0.247 1.26 (0.89, 1.78)
Asthma-related symptoms and medications Has daytime asthma symptoms Has daytime symptoms at least once daily (moderate or severe)
73 (32%)
P = 0.187 Has daytime symptoms ‡2 times/wk (mild or worse)
161 (70%)
310 (65%)
1.24 (0.88, 1.74) P = 0.211
Has nighttime asthma symptoms
177 (67%)
329 (58%)
1.46 (1.07, 1.97) P = 0.015
Has nighttime symptoms at least once per wk (moderate or worse)
105 (59%)
174 (53%)
1.30 (0.90, 1.88) P = 0.166
Has nighttime symptoms ‡2 times/mo (mild or worse)a
156 (88%)
276 (84%)
1.43 (0.83, 2.45) P = 0.199
Uses SABA reliever
189 (76%)
395 (75%)
1.01 (0.71, 1.44) P = 0.947
Uses OTC meds to control asthma
31 (12%)
65 (11%)
1.02 (0.65, 1.62)
150 (55%)
297 (51%)
1.19 (0.89, 1.58) P = 0.250
Subjects with visit to GP
246 (91%)
512 (88%)
1.33 (0.82, 2.15)
Subjects with visit to specialist
108 (40%)
228 (39%)
P = 0.915 Does not feel in control of own health Asthma-related resource use P = 0.251 1.03 (0.76, 1.38) P = 0.865 Subjects with visit to alternative provider
82 (30%)
161 (28%)
1.13 (0.82, 1.55) P = 0.443
Subjects with visit to ED for any reason
67 (25%)
87 (15%)
Subjects with visit to ED due to asthma exacerbation
44 (16%)
57 (10%)
1.86 (1.30, 2.67) P \ 0.001 1.78 (1.17, 2.72) P = 0.007
Subjects with hospitalizations
50 (18%)
65 (11%)
1.80 (1.20, 2.68) P = 0.004
a
Symptom occurrences are only for patients who reported having these symptoms
Variables with significant differences between smokers and non-smokers are shown in bold type for emphasis CI = confidence interval; COPD = chronic obstructive pulmonary disease; ED = emergency department; GP = general practitioner; OTC = over-the-counter; SABA = short-acting b agonist
by various characteristics in Table 3. In 2001–2003, 32% of the subjects were smokers compared with 36% of the subjects in 2004.
Secondary analysis—outcomes Outcomes data for the secondary analysis are shown in Table 4. Visits to a GP, specialist, or an alternative health
123
care provider during the previous 6 months did not differ significantly between smokers and non-smokers in either survey. Smokers in both surveys, however, reported significantly more visits to the ED for any reason (P £ 0.009). In 2001–2003, a significantly greater percentage of smokers were hospitalized during the previous 6 months (14%) than were non-smokers (10%; P = 0.027). In 2004, hospitalizations between smokers (15%) and non-smokers (10%) did not reach significance (P = 0.066).
Qual Life Res (2007) 16:1555–1565
1561
Fig. 3 Association of quality of life SF-8 health survey scores and smoking status for subjects with persistent asthma and no reported COPD surveyed in 2004 (higher scores indicate a more favorable status)
Discussion The prevalence of smoking in the population sample in the primary analysis was 32%. In the WHO European Database on Tobacco Control [2], smoking prevalence (defined as current daily cigarette smoker) was estimated at
approximately 25–34% for France, Germany, and the UK in 2002 through 2005, and the population-weighted estimate of smoking prevalence in the WHO European Region was roughly 29% in 2002 and 2005 [2]. Although the surveys were designed to be nationally representative when they were conducted, the primary analysis may have underrepresented persons older than 55 years because nearly 18% of subjects older than 55 years were excluded for having COPD, whereas only 7% of respondents less than 55 years of age were excluded for this reason. In the secondary analysis, all subjects over 55 years of age were excluded from the analyses. Furthermore, the surveys in 2001–2003 were sent by postal mailing, whereas the survey in 2004 was conducted via the Internet, which may account for the observation that subjects in 2004 were an average of 2 years younger and had a higher education level than those in 2001–2003. Surveys in the USA and UK indicate that elderly persons are not as likely as younger people to access the Internet. These surveys also indicate that Internet users tend to be better educated and may be of higher socioeconomic status than persons who do not have access to the Internet [16, 17].
Table 3 Characteristics of subjects with persistent asthma £55 years of age in 2001–2003 and 2004 surveys Variable
2001–2003 Survey
2004 Survey P value
Smokers (n = 245)
Non-smokers (n = 428)a
490 (65)
0.297
172 (70)
293 (68)
0.637
39.58 ± 10.06
0.008
36.88 ± 10.26
36.91 ± 10.78
0.958
375 (51)
104 (42)
130 (30)
79 (23) 54 (16)
147 (20) 220 (30)
65 (27) 76 (31)
96 (22) 202 (47)
196 (56)
358 (47)
108 (46)
160 (40)
Smokers (n = 351)
Non-smokers (n = 758)a
Female gender, n (%)
239 (68)
Age, yr (mean ± SD)
37.60 ± 10.57
Cigarettes per day (mean ± SD)
17.5 ± 11.7
–
High school graduate or less
213 (62)
Some college College graduate/Advanced
\0.001
Education, n (%)
Body mass index, n (%) \25 kg/m2 (normal)
\0.001
0.005
0.153
25– \ 30 kg/m2 (overweight)
96 (27)
212 (28)
71 (30)
124 (31)
‡30 kg/m2 (obese)
59 (17)
188 (25)
55 (24)
121 (30)
Asthma duration, yr (mean ± SD)
13.87 ± 11.05
16.02 ± 12.42
15.58 ± 11.69
17.21 ± 12.10
Asthma severity, n (%)
0.008
P value
0.354
0.094 0.044
Mild
154 (47)
367 (51)
96 (40)
Moderate
145 (44)
283 (40)
124 (52)
200 (48) 175 (42)
Severe
31 (9)
64 (9)
20 (8)
45 (11)
Use of controller medications, n (%)b
87 (25)
249 (33)
0.007
166 (68)
302 (71)
0.447
Use of ICS controller only, n (%)b
35 (10)
113 (15)
0.025
64 (26)
107 (25)
0.748
\0.001
0.545
Self-reported compliance, n (%)
254 (73)
612 (82)
168 (69)
303 (71)
GERD, n (%)
44 (13)
95 (13)
0.974
21 (9)
42 (10)
0.595
Nasal allergy/hay fever, n (%)
191 (56)
485 (65)
0.003
151 (62)
294 (69)
0.063
a
Includes subjects who reported they were former smokers
b
These questions were asked in different ways in the 2001–2003 and 2004 surveys
GERD = gastroenteric reflux disease; ICS = inhaled corticosteroid; SD = standard deviation
123
1562
Qual Life Res (2007) 16:1555–1565
Table 4 Results of logistic regression in analysis of association of smoking and asthma-related resource use in the previous 6 months in subjects with persistent asthma and £55 years of age in surveys conducted in 2001–2003 and 2004 Variable
Subjects with visit to GP
2001–2003 Survey Smokers (n = 351)
Non-smokers (n = 758)
278 (85%)
622 (88%)
2004 Survey Odds ratio (95% CI)
Smokers (n = 245)
Non-smokers (n = 428)
0.78 (0.53, 1.13)
223 (91%)
371 (87%)
P = 0.189 Subjects with visit to specialist
50 (14%)
97 (13%)
Subjects with visit to alternative provider
17 (5%)
39 (5%)
Subjects with visit to ED for any reason
59 (17%)
Subjects with visit to ED due to asthma exacerbation
ND
Subjects with hospitalizations
47 (14%)
1.13 (0.78, 1.63)
95 (39%)
155 (36%)
71 (29%)
121 (28%)
1.61 (1.12, 2.31) –
1.04 (0.73, 1.46) P = 0.844
61 (25%)
68 (16%)
P = 0.009 ND
1.12 (0.81, 1.54) P = 0.508
P = 0.831) 85 (11%)
1.58 (0.93, 2.62) P = 0.094
P = 0.508) 0.94 (0.52, 1.68)
Odds Ratio (95% CI)
1.76 (1.19, 2.59) P = 0.005
33 (13%)
45 (11%)
1.32 (0.82, 2.14) P = 0.250
70 (10%)
1.56 (1.05, 2.32)
37 (15%)
44 (10%)
P = 0.027
1.55 (0.97, 2.48) P = 0.066
Variables with significant differences between smokers and non-smokers are shown in bold type for emphasis CI = confidence interval; ND = not done; ED = emergency department; GP = general practitioner
No significant differences in gender distribution were detected between smokers and non-smokers. In the primary analysis, age was a significant factor in that non-smokers were approximately 4 years older than smokers were. This age difference and the observation that non-smokers had asthma nearly 3 years longer than smokers could be an indication that some of the non-smokers were actually former smokers and had quit smoking due to negative health effects of concomitant asthma and several years of tobacco consumption; however, this cannot be determined from the data collected in this survey. There is an abundance of data that indicates smoking increases asthma symptoms, but it is less clear whether smoking increases the severity of asthma [1, 3, 5, 18–24]. Data are scarce regarding airway pathology in asthmatic patients who smoke, and results from community studies provide inconsistent results [25]. Studies have shown that the combination of smoking and asthma negatively affect lung functionality and inflammation to a greater degree than either factor alone [1, 5, 19, 24]. No difference was found in subject-reported asthma severity between smokers and non-smokers in the primary analysis. It is possible that smoking status was not a significant factor for asthma severity because smoking may not affect chronic symptoms of asthma to the same extent it precipitates acute reactions that lead to ED visits and/or hospitalizations [19]. It is also possible that smokers with more severe asthma had to stop smoking or that a subject’s perception of asthma severity was inaccurate, particularly since no guidelines for rating asthma severity were provided in the questionnaires.
123
Daytime asthma symptoms evaluated in 2004 did not occur with increased frequency in smokers compared with non-smokers. Using a patient self-administered questionnaire in a cross-sectional study of outpatients in Japan, Suzuki et al. [9] reported daytime respiratory symptoms (i.e., sputum, coughing, wheezing, shortness of breath) were significantly higher in current smokers relative to non-smokers or ex-smokers, despite a high rate of ICS use among patients in the sample. Nighttime symptoms and sleep disturbances, however, were reported by significantly more smokers than by nonsmokers in 2004. Other studies have found similar results in asthmatics who smoke, and levels of nighttime symptoms in ex-smokers were similar to those for non-smokers [9, 20]. One study indicated that many patients who experience nighttime symptoms do not report these findings to their physician, so these symptoms may also be underreported in questionnaires, including the ones reported here [26]. The present surveys attempted to exclude as many subjects as possible with COPD by direct query regarding its presence in 2004 and by excluding all subjects over 55 years of age in the secondary analysis. Most cases of COPD are found in adults 45 years and older. The estimated prevalence of COPD in the 45- to 54-year age group in the USA was 59.2 per 1000 population in 2002, and the highest prevalence rates were found in the age groups 55–64, 65–74, and [ 75 (estimated prevalence of 79.5–106 per 1000 population) [27]. Selection of 55 years as the criterion for exclusion from the secondary analysis left in 45 subjects that had been excluded in the primary analysis
Qual Life Res (2007) 16:1555–1565
and may have included other subjects with undiagnosed COPD. Although it is likely that some outcomes reported here were actually related to early COPD rather than asthma (or at least in addition to asthma since subjects reported their asthma was diagnosed by a physician), based on recommendations from clinicians participating in a clinical trial for the sponsor of this present analysis, exclusion of patients over 55 years of age would provide a reasonable compromise for evaluation of factors associated with smoking. Although subjects reported their daily cigarette use, evaluations of the effects of different quantities of cigarettes smoked per day on outcomes of interest were not performed because categories based on quantity of cigarettes smoked per day had insufficient statistical power to provide meaningful conclusions. Surveys did not request information regarding duration of smoking or duration of smoking cessation among subjects who were former smokers, which could be essential factors for determination of the effects of smoking cessation on outcomes in this study. Previous studies have found former smokers align with current smokers for some variables (e.g., asthmaspecific QoL and several SF-36 domains, including physical functioning, social functioning, mental health, vitality, bodily pain, and general health) [28]. However, ex-smokers aligned with never-smokers in the longitudinal analysis of hospital-based asthma care. The SF-8 Health Survey evaluates responses using ‘‘norm-based’’ comparisons. Specific information is used from various subgroups formed by age, gender, and the presence of diseases, so that scores and interpretations from the surveyed sample can be made relative to the given global subgroup. Therefore, QoL responses as applied to the data in the present study are believed to represent a reliable interpretation of health status [13]. It has been suggested that QoL questionnaires and Asthma Control Questionnaires may not accurately report asthma severity, morbidity, respiratory symptoms, work or school absenteeism, and health resource use as they relate to smoking status [5]. Possible reasons for reporting inconsistencies mentioned by Boulet et al. [5] include failure of the patient to recognize inadequate disease control and that patients develop a tolerance for the disease and its limiting factors [26]. Many asthma patients who smoke do not perceive (or admit) their smoking is a detrimental factor in the worsening of their asthma [19, 29]. In a previous survey of adults presenting to the ED with uncontrolled asthma symptoms, 50% recognized that smoking was responsible for worsening of their symptoms; however, only 4% felt that smoking was responsible for the asthma exacerbation they were experiencing [19]. Only 69% of all subjects with persistent asthma used preventive treatment with asthma controller medication,
1563
with no difference in their frequency of use between smokers and non-smokers. ICS were the predominant controller treatment used by subjects in this survey. There is evidence from several clinical studies that smoking alters responses to asthma treatment, [1, 4, 8–10, 20, 25, 26, 29– 32] and some studies indicate that the reduced effect of these medications can be at least partially restored by smoking cessation [8]. The majority of subjects indicated they used a SABA for relief of acute asthma symptoms, with virtually no difference between smokers and non-smokers in the frequency of using these medications. Little is known regarding effects of smoking on the effects of SABAs [4]. Laforest et al. [33] found that obese patients are more apt to have uncontrolled asthma symptoms (irrespective of smoking). In a post hoc analysis, lower responses to placebo with increasing BMI in non-smoking patients with moderate asthma suggest BMI may influence the progression of the patient’s asthma [34]. It was an unexpected finding that significantly more non-smokers were overweight or obese than were smokers in the current survey. The present analyses suggest that in addition to aggravating asthma symptoms, smoking increased utilization of healthcare resources. It is recognized that subjects may not accurately recall some events for a 6-month period, and previous studies suggest that data reported in surveys may be both over- and under-reported [35]. However, one poll conducted in the USA indicated somewhere around 42% of 2242 adults surveyed keep personal and family health records on file that could be accessed for reference [36]. It also has been reported that long-term recall was best for anthropometric information, smoking and alcohol consumption, whereas recall for dietary information and physical activity was less accurate [37]. A meta-analysis of data from several studies indicated that salience of an event and task difficulty were more critical factors affecting recall than time; recall declined rapidly in the days and weeks immediately following an event, but did not decline substantially after that time [38]. One study [39] found that older age, cognitive impairment, and male gender all contributed to reduced reproducibility in the elderly population that was studied. In the present study, older age was not felt to be a significant factor that would affect recall; average age in the primary analysis was approximately 44 years, and subjects over 55 were excluded from the secondary analysis. Overall, we believe the responses provide reasonably accurate and meaningful recall regarding asthma-related activities and healthcare issues over a 6-month period prior to completion of the questionnaire. In the primary analysis, smoking was significantly associated with increased visits to the ED for any reason as well as for asthma-related reasons. Having health insurance
123
1564
was not significantly different between smokers and nonsmokers. Therefore, one might infer that preference for use of the ED over a visit to a GP or specialist is motivated by something unrelated to the incremental cost of providing the services. One possible scenario is that asthmatic patients who smoke allow their condition to deteriorate markedly before seeking medical attention, in an attempt to avoid confrontation by their GP or specialist regarding their smoking habits. However, the surveys did not solicit information regarding reasons for visits to a GP, so many of the reported visits could have been unrelated to effects of smoking or asthma. The current surveys were designed to be a comprehensive evaluation of the health status of respondents using a selfreported questionnaire and were not designed to capture mortality data or lifetime tobacco consumption. However, in a sample of 1075 asthmatic adult patients, relative risk of death from asthma was significantly increased with increasing pack-years of smoking [40]. Other studies provide conflicting evidence regarding the association of smoking with fatal or near-fatal asthma [18, 41–43]. A review of various lifestyle-related endpoints in recent asthma surveys that included nearly 35,000 patients suggest that asthma, by itself and without regard for its severity, negatively affects lifestyle and QoL [26]. Several studies provide evidence that asthma symptoms are more frequent for smokers and are more severe for heavy smokers than for moderate smokers [18]. Results of several studies indicate that many clinical aspects of asthma for exsmokers are statistically similar to those for never smokers and that the increased use of health care resources by smokers reaches levels similar to non-smokers within 4 years of quitting [28].
Conclusions The limitations of the data in the present surveys are similar to those with any database, namely availability and quality of data and interpretation of the questions asked. Nevertheless, the data are useful for assessment of the impact of smoking on asthma symptoms, healthcare resource use, and QoL outcomes in adults with persistent asthma. In the present surveys, cigarette smoking was associated with increased frequency of nighttime symptoms and healthcare utilization in patients with persistent asthma. Findings suggest that asthmatic patients who do not smoke will have improved QoL outcomes and reduced medical resource use. Due to the prevalence of smoking among asthmatics and the implication that smoking negatively affects clinical responses to asthma treatments, smokers should be included (and identified as smokers) in clinical trials that evaluate treatment strategies and other
123
Qual Life Res (2007) 16:1555–1565
outcomes in patients with asthma. Further study is warranted on the effect of smoking on asthma symptoms and on optimal treatment strategies for asthma in patients who smoke. Acknowledgement Authors Swern, Dong, Sazonov Kocevar, Newcomb, and Taylor are all paid employees of Merck & Co., Inc. Author Shavit was supported by a post-doctoral research fellowship from Merck & Co., Inc. This study was funded by a research grant from Merck Research Laboratories, Merck & Co., Inc., Rahway, NJ USA.
References 1. Thomson, N. C., Chaudhuri, R., & Livingston, E. (2004). Asthma and cigarette smoking. European Respiratory Journal, 24, 822–833. 2. WHO European Database on Tobacco Control. Available at: http://www.data.euro.who.int/tobacco/?TabID=2404. Accessed June 26, 2007. 3. US Department of Health and Human Services, Centers for Disease Control and Prevention. Health, United States, 2006 with Chartbook on Trends in the Health of Americans. Table 63. Retrieved April 5, 2007 from http://www.cdc.gov/nchs/data/ hus/hus06.pdf. 4. Livingston, E., Thomson, N. C., & Chalmers, G. W. (2005). Impact of smoking on asthma therapy. A critical review of clinical evidence. Drugs, 65, 1521–1536. 5. Boulet, L. P., Lemiere, C., & Archambault, F., et al. (2006). Smoking and asthma: Clinical and radiologic features, lung function, and airway inflammation. Chest, 129, 661–668. 6. Piipari, R., Jaakkola, J. J. K., Jaakkola, N., & Jaakkola, M. S. (2004). Smoking and asthma in adults. European Respiratory Journal, 24, 734–739. 7. Troisi, R. J., Speizer, F. E., Rosner, B., Trichopoulos, D., & Willett, W. C. (1995). Cigarette smoking and incidence of chronic bronchitis and asthma in women. Chest, 108, 1557–1561. 8. Chaudhuri, R., Livingston, E., McMahon, A. D., Thomson, L., Borland, W., & Thomson, N. C. (2003). Cigarette smoking impairs the therapeutic response to oral corticosteroids in chronic asthma. American Journal of Respiratory and Critical Care Medicine, 168, 1308–1311. 9. Suzuki, K., Tanaka, H., Kaneko, S., et al. (2003). Respiratory symptoms and cigarette smoking in 3,197 pulmonologist-based asthmatic patients with a highly prevalent use of inhaled corticosteroids. Journal of Asthma, 40, 243–250. 10. Tomlinson, J. E. M., McMahon, A. D., Chaudhuri, R., Thompson, J. M., Wood, S. F., & Thomson, N. C. (2005). Efficacy of low and high dose inhaled corticosteroid in smokers versus non-smokers with mild asthma. Thorax, 60, 282–287. 11. NHWS. National Health and Wellness Survey. Retrieved December 15, 2006 from http://www.nhwsurvey.com. 12. National Institutes of Health (1997). Expert panel report 2: Guidelines for the diagnosis and management of asthma. Clinical Practice Guidelines (NIH Publication No. 97–4051). Bethesda, MD: US Department of Health and Human Services. 13. Ware, J. E., Kosinski, M., Dewey, J. E., & Gandek B. (2001). How to score and interpret single-item health status measures: A manual for users of the SF-8 health survey. Lincoln RI: QualityMetric Incorporated. 14. The IQOLA Project. Available at: http://www.iqola.org/project. aspx. Accessed June 26, 2007. 15. GINA Assembly (2006). Pocket guide for asthma management and prevention. A pocket guide for physicians and nurses.
Qual Life Res (2007) 16:1555–1565 16. 50+ Health. Available at: http://www.50plushealth.co.uk/ index.cfm?articleID=1502. Accessed June 26, 2007. 17. US Department of Commerce. Economics and Statistics Administration, National Telecommunications and Information Administration. A Nation Online: How Americans are expanding their use of the Internet. Available at http://www.ntia.doc.gov/ ntiahome/dn/anationonline2.pdf. Accessed June 26, 2007. 18. Siroux, V., Pin, I., & Oryszczyn, M. P., et al. (2000). Relationships of active smoking to asthma and asthma severity in the EGEA study. Epidemiological study on the genetics and environment of asthma. European Respiratory Journal, 15, 470–477. 19. Silverman, R. A., Boudreaux, E. D., Woodruff, P. G., et al. (2003). Cigarette smoking among asthmatic adults presenting to 64 emergency departments. Chest, 123, 1472–1479. 20. Althuis, M. D., Sexton, M., & Prybylski, D. (1999). Cigarette smoking and asthma symptom severity among adult asthmatics. Journal of Asthma, 36, 257–264. 21. Zimlichman, E., Mandel, D., Mimouni, F. B., et al. (2004). Smoking habits in adolescents with mild to moderate asthma. Pediatric Pulmonology, 38, 193–197. 22. Prechtm, D. H., Keiding, L., & Madsen, M. (2003). Smoking patterns among adolescents with asthma attending upper secondary schools: A community-based study. Pediatrics, 111, (5 Pt 1), e562–e568. 23. Eisner, M. D., Yelin, E. H., Trupin, L., et al. (2001). Asthma and smoking status in a population-based study of California adults. Public Health Reports, 116, 148–157. 24. Lange, P., Parner, J., Vestbo, J., et al. (1998). A 15-year followup study of ventilatory function in adults with asthma. New England Journal of Medicine, 339, 1194–1200. 25. Frank, P., Morris, J., Hazell, M., Linehan, M., & Frank, T. (2006). Smoking, respiratory symptoms, and likely asthma in young people: Evidence from postal questionnaire surveys in the Wythenshawe Community Asthma Project (WYCAP). BMC Pulmonary Medicine, 6, 10. 26. Holgate, S. T., Price, D., & Valovirta, E. (2006). Asthma out of control? A structured review of recent patient surveys. BMC Pulmonary Medicine, 6, (Suppl 1), S2. 27. Mannino, D. M., Homa, D. M., Akinbami, L. J., Ford, E. S., & Redd, S. C. (2002). Chronic obstructive pulmonary disease surveillance– United States, 1971–2000. MMWR Surveillance Summaries, 51, 1–16. 28. Sippel, J. M., Pedula, K. L., Volmer, W. M., Buist, A. S., & Osborne, M. L. (1999). Associations of smoking with hospitalbased care and quality of life in patients with obstructive airway disease. Chest, 115, 691–696. 29. Wakefield, M., Ruffin, R., Campbell, D., et al. (1995). Smokingrelated beliefs and behaviour among adults with asthma in a representative population sample. Australia and New Zealand Journal of Medicine, 25, 12–17.
1565 30. Dijkstra, A., Vonk, J. M., Jongepier, H., et al. (2006). Lung function decline in asthma: Association with inhaled corticosteroids, smoking and sex. Thorax, 61, 105–110. 31. Chalmers, G. W., MacLeod, K. J., Little, S. A., et al. (2002). Influence of cigarette smoking on inhaled corticosteroid treatment in mild asthma. Thorax, 57, 226–230. 32. Livingston, E., Chaudhuri, R., McMahon, A. D., Fraser, I., McSharry, C. P., & Thomson, N. C. (2007). Systemic sensitivity to corticosteroids in smokers with asthma. European Respiratory Journal, 29, 64–70. 33. Laforest, L., Van Ganse, E., Devouassoux, G., et al. (2006). Influence of patients’ characteristics and disease management on asthma control. Journal of Allergy and Clinical Immunology, 117, 1404–1410. 34. Peters-Golden, M., Swern, A., Bird, S. S., Hustad, C. M., Grant, E., & Edelman, J. M. (2006). Influence of body mass index on the response to asthma controller agents. European Respiratory Journal, 27, 495–503. 35. Tsui, E. L. H., Leung, G. M., Woo, P. S., Choi, S., & Lo, S-V (2005). Under-reporting of inpatient services utilization in household surveys – a population-based study in Hong Kong. BMC Health Services Research, 5, 31. 36. Taylor, H. (2004). Two in five adults keep personal or family health records and almost everybody thinks this is a good idea. Health Care News, 4, 1–5. 37. Rothman, K. J. (1994). Improving long-term recall in epidemiologic studies. Epidemiology, 5, 1–4. 38. Mathiowetz, N. A. (2000). The effect of length of recall on the quality of survey data. Paper presented at 4th Conference on Methodologic Issues in Official Statistics, Stockholm, Sweden [Invited paper]. 39. Cumming, R. G., & Klineberg, R. J. (1994). A study of the reproducibility of long-term recall in the elderly. Epidemiology, 5, 116–119. 40. Ulrik, C. S., & Frederiksen J. (1995). Mortality and markers of risk of asthma death among 1,075 outpatients with asthma. Chest, 108, 10–15. 41. Turner, M.O., Noertjojo, K., Vedal, S., et al. (1998). Risk factors for near-fatal asthma. A case-control study in hospitalized patients with asthma. American Journal of Respiratory and Critical Care Medicine, 157(6 Pt 1), 1804–1809. 42. Ryan, G., Musk, A. W., Perera, D. M., et al. (1991). Risk factors for death in patients admitted to hospital with asthma: A followup study. Australia and New Zealand Journal of Medicine, 21, 681–685. 43. Leson, S., & Gershwin, M. E. (1996). Risk factors for asthmatic patients requiring intubation. III. Observations in young adults. Journal of Asthma, 33, 27–35.
123
