Phenotypes of Obstructive Lung Diseases in Elderly: A Real-Life Study in a Non-Clinical Setting

Background: This study aimed to characterize the clinical, functional, inammatory and immunological features of the main obstructive lung diseases (OLD) in elderly. Methods: We included subjects aged ≥ 65 years. In Phase I, participants answered standardized questionnaires, performed spirometry, fractional exhaled nitric oxide, an atopy assessment and blood analysis. In Phase II, patients with COPD or asthma criteria and a control group were invited to undergo a more detailed evaluation that included the determination of IL-1-β, IL-5, IL-6, IL-13, TNF-α and periostin. In Phase III was assessed the frequency of non-scheduled medical visits due to respiratory exacerbations, hospitalization and mortality. Results: Phase I included 286 participants. The median age was 85.8 (P27-P75: 81.0 - 90.2) years and 69% were females: 27.3% had chronic obstructive pulmonary disease (COPD) without asthma (Group 1), 2.8% asthma without COPD (Group 2) and 5.2% presented asthma and COPD traits (Group 3). The remaining participants (64.7%), without asthma or COPD, were classied as Group 4. At this stage, Group 3 was more atopic and presented a lower post-bronchodilator FEV 1 and FEV 1 /FVC. In Phase II, that included 82 participants, Group 3 presented the higher median values of TNF-α (p-value: 0.072) and higher lung volumes (p=0.001). No statistically differences were found for other cytokines among groups. In Phase III, Group 3 patients needed more non-scheduled visits. Conclusions: In our study patients with asthma and COPD traits presented more severe airow limitation, higher values of TNF-α, higher lung volumes and more non-scheduled medical visits.


Introduction
Asthma and COPD prevalence vary widely according to survey methods and diagnostic criteria. COPD is more frequent in smokers, in those older than 40 years and in men. Different studies assessed the prevalence of COPD [1] , [2] , [3]. According to a systematic review, in 2010 the global COPD prevalence for the population aged 30 years or older was 11.7% (95% con dence interval: 8.4%-15.0%) [4]. In a recent meta-analysis [5], the global mean prevalence of COPD in Europe was 12.38% without signi cant differences among the four regions with available data. According to the Global Burden of Disease Study, COPD was one of the ve leading causes of disability-adjusted life years (DALYs) in 2017 [6].
Asthma is usually considered a disease more frequent in younger people and for this reason it may be under or misdiagnosed in the elderly. Nonetheless different authors showed that asthma is also present in older people and may overlap with COPD [7]. The coexistence of asthma and COPD was until recently named as asthma-COPD overlap (ACO) [8], although the GOLD report of 2020 considered that we should no longer refer to ACO, as they constitute two different diseases that may share some common traits and clinical features.
Ageing of the population is a global problem particularly in industrialized countries. Old age is frequently associated with respiratory symptoms [9]. COPD prevalence increases with age [2] , [10]. In elderly, asthma is more frequent in women [11]. Furthermore, COPD could coexist with asthma and mask the latter [12] contributing to undertreatment. In epidemiological studies conducted in United States and United Kingdom, 17-19% of patients with obstructive airway disease reported having asthma and COPD simultaneously [13], comprising more than 50% of the patients with obstructive airway disease whose age was over 50 years [13]. Therefore, patients with concomitant diagnosis of asthma and COPD are usually older. A fully characterization of older persons with obstructive lung diseases [14] (OLD) selected from the general population is lacking in the literature, as most studies were conducted in speci c samples of people with asthma or COPD followed at Health Care Departments.
The present manuscript reports results from the Obstructive Lung Disease in Elders (OLDER) study. The primary objective was to characterize in a non-clinical setting the clinical, functional, in ammatory and immunological features of older persons with the most frequent obstructive lung diseases.

Study design, setting and participants
The OLDER study took place in Lisbon, Portugal. It was an observational study, divided in three phases.
In Phase I, which took place from April to December 2016, residents of 15 Lisbon's elderly care centers (ECC) from different civil parishes were invited to participate.
In Phase I, besides spirometry, fraction of exhaled nitric oxide (FENO), an atopy assessment and blood analysis, participants answered standardized questionnaires administered by a trained interviewer and collected a blood sample.
In Phase II (April 2016 to January 2017), every patient with COPD or asthma criteria was invited to go in a twoweek period after Phase I to the Nova Medical School Lung Function Laboratory, in order to perform a more detailed assessment. A subsample of participants without asthma, COPD, respiratory symptoms and with normal lung function was used as control group. The evaluation included a body plethysmography, a carbon monoxide diffusing capacity (DLCO) and additional in ammatory blood biomarkers.
Phase III assessed the frequency of non-scheduled medical visits due to respiratory exacerbations (de ned by deterioration of breathing symptoms that affected usual daily activities), hospitalisation and mortality in the 12 months after the inclusion in Phase I.
To be eligible to the study, participants should be ≥ 65 years, present cognitive and collaboration capabilities su cient to perform a spirometry and should not have any contraindication for lung function tests. Sample size was calculated in order to estimate the frequency of participants with a forced expiratory volume in 1st second /forced vital capacity (FEV 1 /FVC) < 0.70. According to published data for the Portuguese population [10], a prevalence of 30% was considered for this age group. For a con dence level of 95% and a 4.5% margin of error, we would need to include 293 participants. According to our experience in a previous study [15], we considered that about 70% of the ECC residents would not ful l the inclusion criteria. In order to achieve this number, we planned to screen 1000 residents.
The procedures followed were in accordance with those of the Code of Ethics of the World Medical Association Depression Scale (GDS-15). In Phase II the COPD assessment test (CAT) and the Control of Allergic Rhinitis and Asthma Test (CARAT) were also administered. Authorizations from the authors of these questionnaires were obtained when needed. SGRQ was administered only to those with COPD, asthma or ACO criteria and also to every participant who reported to have any respiratory symptom (cough, phlegm, wheezing or dyspnoea). Additionally, the Charlson comorbidity index (CCI) was calculated for each participant.

Lung function tests and fractional exhaled nitric oxide (FENO)
In Phase I a spirometry with bronchodilatation and a FENO measurement were performed according to recommendations[16] , [17].

Atopy assessment
Atopy was assessed in Phase I by skin prick tests to common airborne allergens (Leti®, Barcelona, Spain) or, if anergy or refusal, by an inhalant panel analysis (Phadiatop®, Thermo-Fisher Scienti c, Uppsala, Sweden).

In ammatory blood biomarkers
A complete blood count (Advia120, Siemens, Munich, Germany) and a high sensitivity c-reactive protein (hs-CRP) determination (Dimension EXL200, Siemens, Munich, Germany) were carried out in Phase I.

Health outcome de nitions
Obstructive lung diseases (OLD) were de ned as illnesses associated with airways obstruction [14]. For the present study only COPD and asthma were considered. Asthma was de ned as a previous medical diagnosis OR reported wheezing in the last 12 months plus reversibility (considered if there was a post-bronchodilator improvement in FEV 1 or FVC of at least 12% and 200 mL) [22]. A participant was classi ed as having COPD if had a post-bronchodilator FEV 1/ FVC < 0.70, in accordance with the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [8]. Global Initiative for Chronic Obstructive Lung Disease (GOLD) severity classi cation of air ow obstruction was also used.
Presence of allergic rhinitis was established on the basis of a previous medical diagnosis.
Smoking was assessed as more than 20 packs of cigarettes or 12 oz of tobacco in a lifetime or at least one cigarette a day for one year. A subject who smoked during the last 30 days was considered a current smoker.
Patients with asthma and COPD traits (previously named as ACO) were de ned according to the de nitions used in previous large epidemiological surveys, as the PLATINO and PUMA studies [22][23]: a postbronchodilator FEV 1 /FVC < 0.70 AND the presence of a previous medical diagnosis of asthma or wheezing in the last 12 months plus reversibility.
Participants were classi ed in groups according to these de nitions [23]: Group 1 -COPD without asthma, Group 2 -Asthma without COPD, Group 3 -patients with asthma and COPD traits / clinical features. Participants without COPD or asthma were classi ed as Group 4.

Statistical analysis
An exploratory analysis of the variables of interest was carried out in all the sample. To compare differences between groups, Kruskal-Wallis was used for continuous variables and Chi-square or Fisher Exact tests for categorical variables.
The frequency of female gender, smoking, depression (according to GDS-15), cognitive impairment (according to MMSE), atopy, wheezing in the previous 12 months, reported asthma diagnosis, reported allergic rhinitis, frequency of FEV 1 /FVC < 0.7 and airways reversibility were estimated in the sample.
The level of signi cance used was 0.05, although p values greater than 0.05 and lower than 0.1 were still considered as indicating an evidence. Data analysis was performed using STATA (Stata Statistical Software: Release 12; StataCorp LP, Lakeway, TX, USA).

Results
In Phase I, 305 out of 1034 screened residents ful lled the inclusion criteria and 286 were able to complete all the assessments planned for this phase. Non-participation was mostly related with lack of cognitive and collaboration capabilities. Descriptive analysis of these 286 participants is presented in Table 1 and 2. The median age was 85.8 (P27-P75: 81.0 -90.2) years and 69% were females. The ow diagram is presented in Figure 1.
Eight participants (2.8%) had wheezing within the last 12 months plus reversibility and 5.6% (n=16) reported lifetime diagnosis of asthma. The combined frequency of asthma based in reported lifetime diagnosis (medical diagnosis) and wheezing within the last 12 months plus reversibility was 8%. Fifteen (65%) of these asthma patients had also COPD.
The remaining 185 participants (64.7%) did not present any criteria of COPD or asthma (Group 4), although 62 (21.7%) of these reported at least one respiratory symptom, namely cough and wheezing. The overlap of these groups is presented in Figure 2.
Considering only the 101 participants with COPD and / or asthma, 38% were under medication for CRD. Cytokines, lung volumes and DLCO comparisons between Groups 1, 2, 3 and 4 are presented in Table 3.
Patients with both asthma and COPD (Group 3) presented the higher median values of TNF-a (differences among groups: p-value: 0.073) and higher lung volumes (p-value: 0.001). No statistically differences were found for other cytokines among groups.

Non-scheduled visits, hospitalization and mortality -Phase III
Information about one-year non-scheduled medical visits due to respiratory exacerbations, hospitalization and mortality was available for 264 out of 286 participants from Phase I.
During the study, 27% (n=77) of the older persons needed to schedule an appointment with a doctor due to respiratory symptoms, 4.9% (n=14) were hospitalized and 4.2% (n=12) died. Four out of 12 people died as consequence of a respiratory illness. Patients with asthma and COPD traits / clinical features patients needed more non-scheduled visits to a doctor due to respiratory complaints (differences among groups: p-value = 0.046). Results are presented in Table 4.
Non-scheduled medical visits were associated with lower FEV 1 (p-value: 0.005) and a lower FEV 1 /FVC post bronchodilator (p-value: 0.003). Hospitalization was associated with a lower FEV 1 post bronchodilator (pvalues: 0.021). These results are presented in Table 5.
No associations were found between Th cytokine pattern and medical non-scheduled visits, hospitalization or mortality.

Discussion
In this study we aimed to characterize the clinical, lung function, in ammatory and immunological characteristics of older persons with the main OLD, namely asthma and COPD. These diseases were de ned as reported by other authors and participants were allocated in different groups in order to assess differences between them. Additionally, we determined different cytokines related with Th-1 and Th-2 pathways.
Asthma prevalence, de ned as the combined frequency of asthma diagnosis reported in lifetime and wheezing within the last 12 months plus reversibility was 8%, which is according to the Portuguese National Asthma Survey [24] data for this age group.
In our sample, the frequency of COPD according to GOLD de nition was 32.5%. Our results are similar with data from the Portuguese BOLD study that estimated COPD prevalence in people older than 70 years of 30.8% [10], using also the xed FEV 1 /FVC ratio criteria. The xed ratio criteria may overdiagnoses COPD in elderly, although we already showed in a previous study that in this age group, spirometry interpretation is highly in uenced by the reference equations used and meaningful differences may be found [15] when using the lower limit of normality. This is in agreement with the Global Initiative for Chronic Obstructive Lung Disease (GOLD) report 2020[8].
COPD represented the most frequent OLD, which is not surprising considering the physiologic lung function decline associated with age [25] that compromises mostly the FEV 1 . This might have contributed also for the high frequency of patients with both asthma and COPD traits, previously classi ed as ACO (5.2%) as compared to other authors that have used the same de nition as we, namely in PLATINO [22] and PUMA [23] studies.
There were different ACO de nitions available in the literature and consensual one was been not established[26] [27]. This has contributed to the wide prevalence range found in population based studies and in asthma and COPD patient surveys [28]. Most de nitions were based in phenotypic features and a biomarker is still lacking. In our study, higher levels of TNF-α were associated with the presence of asthma and COPD traits in the same patient. Higher levels of TNF-α have been related with the severity of airway obstruction in COPD patients [29] which may explain our ndings as patients with both diseases presented more severe airways obstruction.
To the moment, no single endotype of asthma or COPD has been fully characterized in elderly. Despite the lack of strong evidence, the presence of Th-2 in ammation allows to classify asthma as Th-2 "high" or Th-2"low" endotypes, as proposed by Wenzel S. Th-2 "high" is associated with a better response to inhaled corticosteroids [30].
More recently, it was considered that type 2 in ammation may constitute also a distinct endotype in COPD, similar to the asthma Th-2 "high" endotype. Other existent COPD endotypes seem to be related with Th-1 in ammation and type 17 helper T cells [31].
Previous studies also found diverse in ammatory patterns in asthma, COPD and may overlap in certain patients, leading to a mixture of Th-2 and non-Th-2 cytokine expression. Patients with asthma and COPD may show therefore diverse in ammatory processes that overlap the characteristics of both conditions. For this reason the previously ACO entity cannot be considered a speci c phenotype of asthma or COPD but a blend of both components [32] , [33].
Our ndings suggest that at least in this age group COPD and asthma are heterogenous in terms of immunological endotype and Th-1 and Th-2 cytokines are common independently of the considered disease, even in COPD.
Patients with both asthma and COPD traits / clinical features patients have been considered to present an increased risk of exacerbation and hospitalization [22] as they usually present more severe disease. We found lower lung function parameters and more frequent non-scheduled visits to a doctor in this group, indicating that this phenotypic presentation should raise awareness. According to our results this is particularly relevant as patients with features of both diseases presented a future risk of exacerbations. This trend for increased risk of healthcare utilizations has been previously reported [34].
Surprisingly, COPD patients reported the best SGRQ results. This might be related with the characteristics of our sample, as older patients tend to have substantial comorbidities and to be less active, which may have contributed for a wrong perception of the real burden of COPD on the quality of life.
According to the literature, comorbidities are increased in patients with asthma and COPD features compared to other obstructive airway diseases [26], mainly in smokers. In our study we could not nd this association probably due to the high "baseline" degree of comorbidities among this age group.
The major strengths of our study are the inclusion of a carefully selected sample of older persons with the main OLD recruited in a non-clinical setting, the detailed characterization in terms quality of life, disease control, comorbidities, lung function, airways in ammation and cytokine assessment. Additionally, we were able to assess the future risk in these patients. To our knowledge, this is the rst study providing such information in older persons.
We were not able to include a high number of patients with asthma or asthma and COPD traits and this constitute a limitation. Additionally, we do not know to which extent these results could be extrapolated.
However, we found that the presence of asthma and COPD traits in the same patient is associated with worst lung function, more non-scheduled visits to a doctor and with a higher level of TNF-α.

Conclusions
Patients with asthma and COPD traits tend to have more severe airways obstruction and a higher risk to exacerbate. In older people, levels of Th-1 and Th-2 related cytokines may be similar independently of the OLD.