Smoking is by far the most common cause of chronic obstructive pulmonary disease (COPD) in the world, and is responsible for more than 80%–90% of new cases.1 COPD has conventionally been attributed to chronic inflammation triggered by tobacco smoke, which releases proteolytic substances and free radicals, causing airway damage and breakdown of the pulmonary parenchyma, leading to airflow limitation (AFL) and significant symptoms.2 Nevertheless, early involvement of the small airways has been known for several decades to play an important role in the pathogenesis of this disease.3,4

Various studies have used computed tomography (CT) to investigate smokers who have not yet developed AFL.5,6 However, this technique does not directly evaluate the small airways, and instead analyzes the walls of the larger bronchioles and the probable impact of small airway dysfunction (SAD) on attenuation of the lung parenchyma.7 The forced oscillation technique (FOT) or impulse oscillometry system (IOS), on the other hand, provides a specific, systemized, functional assessment of the small airways that shows a pattern of involvement characteristic of SAD.8,9

In recent years, smokers without AFL have been known to constitute a heterogeneous group with various symptomatic manifestations that impact on their health-related quality of life (HRQOL).10,11 However, scant information is available on the functional or morphological parameters that might differentiate these individuals and, in some way, contribute to their symptoms. Given its potential pathogenic importance, it is interesting to speculate that SAD might be an early step in the development of AFL12 and could, as such, demonstrate incipient effects on quality of life.

We hypothesize, therefore, that smokers without AFL who have developed SAD have a poorer quality of life. Our main objective was to compare the quality of life of smokers without AFL but with evidence of SAD on IOS, with that of non-smokers, smokers without SAD, and smokers with AFL consistent with COPD. As secondary objectives, we also aimed to determine the frequency of SAD in smokers without AFL and to compare lung attenuation densities between the 4 groups of subjects.

According to the established criteria, we recruited 20 non-smokers, 53 active smokers without AFL, and 20 active smokers with AFL consistent with COPD (Table S1, supplementary material). The three groups were homogeneous in terms of sex, age, height, weight, and body mass index (Table S2, supplementary material).

In the group of smokers without AFL, 19 subjects had SAD, i.e., a rate of 35.8% (95% CI: 22.9–48.81%). None of the non-smokers showed SAD, so they were all considered control subjects. Table 1 shows the anthropometric characteristics and tobacco consumption of the 4 final groups. The anthropometric parameters were similar, and no differences were observed in the characteristics of tobacco consumption, nicotine dependence, or psychological, social or gestural dependence among smokers.

Apart from obvious differences in the forced oscillation variables used to define the groups, the only spirometric parameters that were lower in smokers with SAD compared with smokers without SAD were FEF25% and FEF25–75% (Table 2).

The morphological involvement of lung parenchyma showed a lower attenuation density in smokers with AFL compared to the other 3 groups (Table 3). In both inspiratory and expiratory phases, COPD patients showed a lower mean lung density, higher percentages of low attenuation volume, and attenuation volumes of between −1000 to −951HU and between −950 to −901HU, and a lower P15. However, no differences were detected in attenuation densities between smokers with or without SAD. Furthermore, no correlation was found between the measurements of lung attenuation and forced oscillation variables between the 2 groups of smokers without AFL.

Table 4 shows a comparison of HRQOL between the 4 study groups. In addition to the predictable differences between smokers with AFL and the other groups, it is interesting to note that smokers with SAD scored worse in the physical and mental components of the SF-36 questionnaire and on the EQ-5D visual analog scale (VAS), compared to both non-smoker controls and smokers without SAD, evidence of an intermediate situation of deterioration in quality of life between these subjects and smokers with COPD (Fig. 2). The analysis of the different domains of the SF-36 questionnaire shows a similar behavior, mainly affecting the domains related to the physical component (Table 4 and Fig. 3).

Fig. 2.

Comparison of global scores for the physical component of the SF-36 questionnaire and the EQ-5D scale in the 4 study groups. Error bars correspond to the standard mean error. AFL: airflow limitation; EQ-5D: European Quality of Life-5 Dimensions; SAD: small airway dysfunction; SF-36: Short Form 36 Health Survey.

(0.49MB).

Fig. 3.

Comparison of the scores for the different SF-36 domains among the study groups. Error bars correspond to the standard mean error. AFL: airflow limitation; SAD: small airway dysfunction; SF-36: Short Form 36 Health Survey.

(0.46MB).

In smokers without AFL, both with and without SAD, no correlation was found between HRQOL and anthropometric characteristics, age of onset, intensity of tobacco use, years of cigarette smoking, or Fagerström or Glover-Nilsson scores.

However, the quality of life of these subjects was associated with various functional and morphological parameters (Table 5). Thus, the SF-36 physical component score showed a moderate inversely proportional relationship with R5, R20, X5 and AX and a weak directly proportional relationship with FVC. FEV1, and FEF25–75%, and various measurements of lung attenuation (MLD and P15) in inspiration. The score obtained on the EQ-5D scale was proportionally inversely associated with R5 and X5 (Table 5). Finally, the multiple linear regression models only retained R5 and X5 as independent predictors of the SF-36 questionnaire physical component and the EQ-5D scale, respectively (Fig. 4 and Table 6).

Fig. 4.

Relationship between the oscillometric parameters and health-related quality of life in smokers without airflow limitation. The upper panel (A) shows the relationship between respiratory resistance at 5Hz (R5) and the physical component of the Short Form 36 Health Survey (SF-36) questionnaire, while the lower panel (B) shows the relationship between capacitive reactance at 5Hz (X5), and European Quality of Life-5 Dimensions (EQ-5D) questionnaire.

(0.09MB).

The principal finding of our study is that SAD affects over one third of smokers without AFL. Although these changes do not affect spirometry or cause changes in lung attenuation, they are associated with an incipient decline in HRQOL, to the extent that an independent relationship could be identified between HRQOL and some forced oscillation parameters.

SAD was detected in our smokers at a higher rate than in previous studies, which reported that 5%–13% of smokers with normal spirometry might have altered IOS.22,23 However, these studies did not specify a cut-off point for the detection of AFL and did not distinguish between active smokers and former smokers with different times of abstinence. In contrast, when predetermined cut-off points for R5–R20 were used, the presence of SAD was reported in 33%–42% of subjects with asthma and normal spirometry.24,25 Although the information available in smokers without AFL is scarce, Jetmalani et al.26 found at least one abnormality in up to 51% of such individuals, mainly in respiratory resistance, while Berger et al.27 reported that R5 is high in 56% of asymptomatic smokers with normal spirometry. In any case, the percentage of SAD identified in smokers without AFL is in line with figures determined by multiple breath nitrogen washout, which identified conductive airway involvement in 24% of subjects and acinar airway involvement in 38%.28

A preliminary analysis of a subsample of the ECLIPSE study revealed the potential usefulness of IOS in the preliminary stages of smoking-induced lung damage.22 Although that study found no differences in the absolute values of the IOS parameters between patients with COPD, smokers without AFL, and non-smokers, probably due to the notable differences in anthropometric characteristics and smoking among the groups, it did show a discreet, clinically acceptable intra-individual variability in respiratory system impedance in the short and medium term.22 But the most interesting finding was the identification of an “apparently healthy” group of smokers who had a high AX, leading the authors to speculate that the presence of small airway changes could provide early evidence of AFL that would otherwise be undetectable with conventional spirometric parameters.22 Along the same lines, Frantz et al.29 compared patients with normal spirometry and a self-reported diagnosis of chronic bronchitis or emphysema with subjects without this diagnosis, and found that the first group had higher values of R5, R5–R20, X5, and AX, in addition to greater use of respiratory medication.

It is interesting to note that no correlation between oscillometric parameters and lung attenuation measurements in inspiratory or expiratory phases was observed in our smokers with no evidence of AFL. A similar finding was reported in the ECLIPSE study with regard to CT measurements carried out in inspiration.22 Three types of pulmonary involvement detectable on CT have been proposed to explain this discrepancy: airway wall thickening, increased lung attenuation (emphysema), and mixed.30 Changes in R5, R5–R20, fres and X5 may correspond mainly to airway involvement and mixed involvement, so their relationship with attenuation changes is not predictable.30 Another explanation may lie in the fact that IOS and CT are sensitive to changes in pulmonary mechanics that occur at various stages in the progression of airway damage. Resistance is affected by many factors, including airway caliber, particularly when narrowing is heterogeneous and widespread, and involves the central and peripheral airways.14 Reactance is also sensitive to the heterogeneous and widespread narrowing of the bronchial and bronchiolar lumen, but it also depends on airway closure,9 which may also be compromised in the case of air trapping. However, air trapping in smokers without AFL is much lower than in asthmatics,31 so this change is probably too insignificant to induce important changes in lung attenuation and generate an additional increase in X5. In short, we might suggest that CT can be used to evaluate distal parenchymal involvement, whereas respiratory impedance, particularly respiratory resistances, reflect the caliber of the small airways to an even better extent than bronchial wall thickness, since this latter measurement is performed on higher bronchial generations.31

In our opinion, the most singular finding of this study is the loss of HRQOL in active smokers with SAD who have not yet developed AFL. A greater frequency of wheezing was previously reported in subjects with high R5.28 The decline in HRQOL among smokers with SAD, which lies midway between non-smokers or smokers without SAD and smokers with AFL, could explain the differences in perceived health among smokers. Smokers are also known to have worse HRQOL than non-smokers32 and a persistent smoking habit is associated with a progressive deterioration of HRQOL,11 while smoking cessation leads to an improvement in quality of life.33 As in our study, HRQOL involvement was detected by changes in EQ-5D scores and in all domains of the SF-12 and SF-36 questionnaires,33–35 although those authors reported a slightly higher degree of involvement than found in our patients, probably because their patients were older, predominantly men, with an unknown cumulative smoking rate, who were resident in other geographic areas.

However, the most notable finding regarding HRQOL in smokers with normal spirometry was the heterogeneity of its intensity, which to date could not be adequately explained merely by a more intense habit34 or greater nicotine dependence.35 The finding that smokers with SAD are those who really experience a deterioration in their quality of life helps identify this functional alteration as a potential risk, and is consistent with information from patients with COPD, in whom SAD evaluated by IOS is associated with health status and dyspnea level.36

It is reasonable to assume that the presence of SAD in smokers without AFL is caused by oxidative stress and the direct damage caused by local inflammation in the peripheral airways.37 In fact, in smokers with normal spirometry, individuals with SAD have shown higher lymphocyte and neutrophil counts in their bronchoalveolar lavage and higher cytokines IL-8, eotaxin, and chemokine CD3Cl1 levels.27 The latter is of particular importance due to its involvement in the differentiation of Th17 cells38 and, as such, the progression of chronic respiratory diseases. In turn, the existence of a directly proportional relationship between R5 and bronchoalveolar lavage levels of IL-8, eotaxin, CD3CL1, IL-12p70, and TGFalfa supports its association with the degree of inflammation of the distal airways.27

Evidence on the prognostic significance of SAD in smokers without AFL is still emerging, although in the COPDGene Study changes in the small airways in smokers and ex-smokers without COPD were found to correlate with the rate of annual FEV1 decline, which may reach the same levels as those of patients with moderate COPD, yet could be reversible if distal airway inflammation is controlled.39 These results seem consistent with those obtained in a small group of patients with moderate COPD followed over 5 years, in whom X5, AX and fres were also associated with the annual rate of FEV1 decline.40

Our study has some limitations. It was a small, single-center study, although the sample size was sufficient to detect differences in HRQOL among the different groups of subjects. This makes it difficult to extrapolate results to the general population of smokers. Unlike other analyses, we included a large number of women and middle-aged individuals in order to obtain a more representative sample of the smoking population, but this means that comparison with other studies must be approached with caution. Lung attenuation was evaluated using software for clinical use which does not measure airway caliber thickness. Moreover, we do not have measurements of the diffusing capacity of the lung, although the possibility of finding changes in this parameter among smokers with or without SAD seems unlikely, as attenuation densities did not differ between these groups. Finally, the cross-sectional design of our study prevented us from establishing prognostic factors or demonstrating any kind of association or causation. Similarly, our results do not allow us to make any recommendations on the management or treatment of active smokers with SAD.

In conclusion, SAD in active smokers who have not yet developed AFL or changes in lung attenuation density affects their symptomatic perception of health, since smokers with SAD have a lower HRQOL, which lies midway between non-smokers or smokers without SAD and smokers who have developed COPD. Forced oscillation is a simple, fast, safe, and economic procedure that can identify SAD in more than a third of active smokers.

Research project partially funded by Chiesi. The sponsor has not participated in the study design, data collection, data analysis and interpretation, manuscript preparation, or the decision to submit the article for publication.

The authors state that they have no conflict of interests.