Annual decline in forced expiratory volume and airway inflammatory cells and mediators in a general population-based sample

Chronic obstructive pulmonary disease (COPD) is characterized by reduced forced expiratory volume in 1 s (FEV₁) and is associated with an enhanced inflammatory response in the airways and the lungs [1]. Environmental factors can contribute to airway inflammation, including tobacco smoke, occupational agents, urban air pollutants, and fuel wood [2]. There is evidence that exposure for air pollution is associated with impairment of lung growth and increased risk of reduced adult lung function. This is shown by a study from Los Angeles region that investigated the percentage of 18-year-old participants with attained FEV₁ below 80% of the predicted value. This percentage turned out to be much higher in the more polluted zones as compared to the less polluted zones [3].

Several types of cells are involved in COPD inflammation, including macrophages, epithelial cells, dendritic cells, neutrophils, eosinophils, and T and B lymphocytes, and higher levels of these cells in the lungs are associated with higher levels of inflammation and with disease progression. There are studies describing normal sputum composition. Induced sputum from healthy subjects is rich in macrophages and neutrophils and poor in the eosinophils, lymphocytes and epithelial cells and number of macrophages is higher than neutrophils. The relative number of neutrophils compared with macrophages increases with COPD developing and aging process [4‐7].

The eosinophilic phenotype of COPD shares some features with asthma, such as the reversibility of airway obstruction and a greater response to corticosteroids than other types of COPD [8]. Inflammatory cells release inflammatory mediators that contribute to the pathophysiology of the disease, including cytokines, chemokines, lipid mediators, and growth factors [10]. Cytokines reflect both T helper 1 (IFN-γ), T helper 2 (IL-6), and T helper 17 (IL-17a, IL-17f, IL-23) responses and the activation of pro-inflammatory M1 macrophage phenotypes (CCL2, CCL3, CCL4, CXCL8). Notably, these cytokines play central roles in the pathogenesis of COPD. Activated murine M1 macrophages contribute to the pathology of the disease by releasing tissue-destructive enzymes that can cause lung destruction and airway damage. T helper 1 and T helper 2 cells are involved in hyper mucous secretion [9, 10].

Airway inflammation can be examined using a non-invasive procedure that involves sputum collection after the inhalation of a hypertonic salt solution. This test has been standardized and is used as a diagnostic tool to research airway diseases such as asthma and COPD [11]. In a longitudinal study of healthy individuals, Fletcher and co-workers showed that subjects who developed COPD had accelerated decline in FEV1 [12] compared to individuals who did not develop COPD. This hallmark study also found that dFEV₁ values were higher in smokers than in non-smokers and that smoking cessation normalized dFEV₁ [13]. Newer studies suggested that only 50% of adult smokers developed reduced lung function because of accelerated fall in FEV1 another part never archive the normal level of lung function in early adulthood [14].

Some retrospective studies have examined bronchial inflammation by collecting induced sputum from healthy individuals and investigating the associations between bronchial inflammation and systemic inflammation and dFEV₁. These studies found that neutrophils are elevated in the sputum of smelter workers and healthy smokers and that rapid dFEV₁ is associated with an increased level of sputum neutrophilia [15, 16]. However, only a few studies have prospectively investigated the association between dFEV₁ and indices of airway inflammation [17, 18]. As sputum induction is time consuming, these studies only included a few participants, and more studies are needed.

The current prospective study collected sputum data from 120 randomly selected healthy subjects living in the Akershus University Hospital catchment area in Norway who were followed-up for 10 years. Of the 120 subjects, 62 were available for re-examination. We hypothesized that in this population, inflammatory markers in sputum would predict increased decline in FEV1 or the development of COPD. The main aim of this study was to investigate the associations between airway inflammatory markers in induced sputum and subsequent dFEV₁. A second aim was to investigate the relationship between inflammation at baseline and chronic airflow limitation at the end of the 10-year follow-up period.

The demographic data collected at baseline showed no significant differences between the follow-up participants, and those who did not participate in follow-up with regards to sex, age, or lung function (Table 1). The subjects in the follow-up group smoked more (9.9 pack-years), and there were fewer current smokers in the follow-up group than in the group that did not participate in follow-up. No one started to smoke during the observation period.

The mean dFEV₁ was 32.9 (26.3) ml/year. Table 2 shows the distribution of dFEV₁ values in tertiles according to demographic data, spirometry results, peripheral white blood cell count, and induced sputum cells using t-test.

We found no associations between dFEV₁ and any of the covariates in Table 2. Surprisingly, the group with the largest dFEV₁, which was > 38 ml/year, had more never smokers then the other groups, although there was no difference in age and sex distribution. Notably, 8 of the 10 participants who were current smokers at baseline quit smoking during the follow-up period.

We also investigated cytokine levels in the sputum. Table 3 shows the analysis of associations between dFEV₁ and sputum cytokines and chemokines. We found no significant associations between sputum cytokines or chemokine levels and dFEV₁ but correlation between log-transformed IL8 and log-transformed number of sputum neutrophils were highly significant (p-value < 0.001).

We investigated the associations between dFEV1 as well as FEV1 at baseline as outcome variables, and IL-6, IL-8, IL-23, IL-17f, IL-17a, IFN-γ, TNF- α, CXCL8, CCL2, CCL3, CCL4 and sputum neutrophils as covariates. None of these relationships reached significant levels. It is likely that the variation in the outcome as well as explanatory variables were too wide, as the residuals of these regressions were normally distributed and in the normal range for FEV1 as well as dFEV1. During the follow-up period, 7 participants developed irreversible AFL. We compared the baseline data of these subjects with the data of participants who had no post-BD AFL (n = 49 subjects) (GOLD criterion). The groups showed no difference in sex distribution and only a negligible difference in dFEV₁. The mean subject age was higher in the group with irreversible post-BD AFL than in those in the non-AFL group.

Table 4 shows that seven participants had post-BD AFL. We identified two distinct groups among the participants who had post-BD AFL: the first group of participants who had positive reversibility test (2 subjects) and the second group of participants, who did not respond to salbutamol (5 subjects). Participants who had irreversible post-BD AFL at the end of follow-up had significantly lower predicted FEV₁%, lower DLCO and KCO, higher peripheral white blood cell count (limited to neutrophils), and higher tobacco consumption at baseline compared with participants who had normal lung function at the end of follow-up (Table 4). The dFEV₁ in first group was steeper than in the second group, but the FEV₁/FVC ratio at baseline was similar in the two groups. The diffusion capacity of CO in the group with positive reversibility test was higher at baseline than in group without positive reversibility test and was equal to that of the healthy group. The baseline number of sputum eosinophils in subjects in the first group was higher than in subjects in the second group.

The primary aim of this study was to test the hypothesis that airway inflammation in presumably healthy subjects is a determinant of accelerated dFEV₁. We found no association between dFEV1 and inflammation at baseline. There was, however, a positive relationship between irreversible airflow limitation at the 10-year follow-up and sputum eosinophils at baseline, foremost in participants with the combination of airflow limitation and positive reversibility test.

Induced sputum testing is a safe, non-invasive procedure that can provide valuable information about the immune response in the central airways and in airway diseases [22]. However, it may be uncomfortable for the person being tested and is time-consuming. This may explain why few longitudinal studies have investigated the relationship of sputum and clinical outcomes. As early as 1996, Stanescu found that an accelerated annual decline in FEV₁ correlated with the percentage of neutrophils and inversely correlated with the percentage of macrophages [16]. However, dFEV₁ was analysed using variables obtained at follow-up rather than at baseline. Hence, the association between dFEV₁ and sputum inflammation could be the result of, rather than a predictor of, the accelerated decline. More recent prospective studies are more in line with our study [17, 18]. Neither of these studies found any associations between inflammatory variables at baseline and subsequent dFEV.

Induced sputum has mostly been compared between healthy subjects (who are frequently young students) and patients with established clinical lung diseases, such as asthma or COPD. In this study, none of the subjects had diagnosed COPD at baseline. The participants lived in area without big industry and low air pollution. Accordingly, the study could only detect associations between dFEV₁ and inflammatory indices in a healthy population. The number of sputum cells was in the same range as found in other sputum studies in general populations [7]. As the variability in cell counts was as high as 20%, sputum studies of dFEV₁ in the general population should be considerably larger than this study.

There are several possible explanations for our negative findings. dFEV₁ variability could be too small to measure. Notably, the mean dFEV₁, as well as dFEV₁ variation, was in the same range as found in other longitudinal studies of lung function in general populations [23, 24]. Moreover, some smokers may quit smoking during the follow-up, as in this study.

There were several interesting findings related to the second aim of the study. In particular, participants who had both post-BD AFL and a positive reversibility test at follow-up had significantly more sputum eosinophils at baseline than the other participants. Moreover, they had normal diffusion of CO at baseline and steeper dFEV₁ than the other participants, although the latter relationship was not significant. Both of these observations are in line with other studies [25, 26]. Participants with post-BD AFL and positive reversibility test had a lower FEV₁/FVC-ratio at baseline than those who had no post BD AFL at follow-up, but they were not older than the other participants. This was not in agreement with other studies [25].

The participants who had BD AFL and showed no reversibility with salbutamol at the follow-up had significantly lower diffusion of CO, higher peripheral white blood cell counts (neutrophils), higher cumulative consumption of tobacco, and they were older than the healthy group at baseline. All these findings are in agreement with other studies [25, 27] and indicate that these subjects had emphysema. However, they had roughly the same dFEV₁ as those who had no post-BD AFL. This finding may be surprising, but recent studies have shown that COPD patients with low initial FEV₁ may have dFEV₁ that is within the normal range [28, 29].

The main limitation of this study was its small sample size, and therefore sensitive to clustering of the participants. However, the age and sex distribution of the sample was close to the target population. Hence, we do not believe that the sample was biased. However, as a sputum study this sample size should be sufficient to detect a clinically relevant relationship between airway inflammation and the decline in lung function. Sputum samples with more than 20% squamous cells (32%) were excluded insofar as sputum cells were included as explanatory variables. Several of the associations are interesting and merit further investigation in a larger cohort. In accordance with our results, such a prospective cohort include should more subjects with an increased risk of developing post-BD AFL or accelerated dFEV₁. Regrettably, we were not able to identify participants with these risk profiles. Second, the total inhalation time varied between participants which are likely to have increased the variability of sputum neutrophil counts. The first portions produced are generally rich in neutrophils and later portions richer in macrophages. It could be one of the reasons why we did not find association between sputum cells and dFEV1. However, the total inhalation time was not recorded. Therefore, the lack of information on inhalation time should be considered as a limitation of the study.

Notably, DTT could influence the detection level of cytokines but it is difficult to say to what degree, and this fact could have impacted our results.

In this cohort study of healthy subjects, we did not identify any inflammatory markers in sputum that predicted dFEV₁ 10 years after the initial examination. However, we identified two distinct groups with post-BD AFL at the end of follow-up: one group with a positive reversibility test and another group with an irreversible BD test. The latter group had reduced DLCO and a history of much higher tobacco consumption than the other participants, a habit that suggests possible emphysema. The group with a positive reversibility test was characterized by having sputum eosinophilia at baseline. This study was relatively small, so this should be confirmed in a larger population.