Small airways disease (SAD) has long pathophysiologically linked to chronic obstructive pulmonary disease (COPD).1,2 In the seventies and eighties, early detection of SAD due to cigarette smoking was attempted by using single breath nitrogen (SBN2) test. Single breath nitrogen test abnormalities were shown to discriminate smokers from non-smokers in selected populations3 and symptomatic from asymptomatic subjects in general population samples.4 Conversely, other studies found abnormalities only in small proportions of those with a respiratory disorder not otherwise suspected or diagnosed;5 SBN2 test measurements were less sensitive than those from spirometry.6 Few longitudinal studies evaluated the relationships between SBN2 indexes and lung function decline. In US, an abnormal slope of phase 3 or alveolar plateau (N2-slope) failed to identify those with an accelerated forced expiratory volume in one second (FEV1) decline.7 In 50-60 year old men from a 7-year population-based Swedish study, an increased baseline N2-slope was associated with a higher FEV1 decline.8 A higher baseline closing capacity (CC) was associated with subsequent loss of FEV1 in adults followed-up for 4.7 years.9 In two different cohorts, CC/Total Lung Capacity ratio (CC/TLC) was significantly associated with FEV1 decline.10 Baseline N2-slope correlated with FEV1 and FEV1/Vital Capacity (VC) ratio at 13-year follow-up, in a group of 56 male smoking steelworkers.11

After 1980s, the prevailing opinion was that SBN2 test offered no or little advantage over monitoring individuals through spirometry. A renewed interest in SAD has occurred in the new millennium.12 Small airway pathological findings were related to COPD progression.13 Multiple breath washout test was shown to be abnormal after only 10 pack-years of smoking.14 N2-slope deterioration over 6 years was associated with exhaled nitric oxide measured in both smokers and non-smokers.15 Further studies were conducted in the last decade. Single breath nitrogen test was shown be useful in assessing the effects of inhaled medicines on small airways in asthma, by Scichilone et al.,16 and in COPD, by Pecchiari et al.17 Timmins et al.18 showed that the severities of COPD and airflow obstruction are independently predicted by the extent of both SAD and emphysema. Khurana et al.19 found that COPD patients with persistent sputum production had higher closing volume (CV). Boeck et al.20 demonstrated that N2-slope was increased in COPD patients compared with healthy subjects and associated with FEV1 predicted. Olofson et al. showed that the cumulative COPD event incidence rate over a 38-year follow-up increased with increasing of N2-slope,21 and it was shown to be higher among heavy smokers with the highest N2-slope and lowest FEV1.22

In this context, the availability of longitudinal data from a population-based study carried out in Italy in 1980-1991 prompted us to evaluate whether SBN2 test abnormalities are related to lung function decline and COPD incidence over an 8-year follow-up.

We have prospectively evaluated the association between impairment of small airways and lung function decline or incidence of COPD in a large longitudinal population-based study over 8-year follow-up. We observed that baseline N2-slope from SBN2 test is significantly associated with FEV1 decline and increased risk of incident airway obstruction as defined by GOLD or ATS-ERS criteria.

Our study is comparable to only few other similar studies conducted in the 1970-80s. Over a similar follow-up period, by using the SBN2 test technique in 460 men aged 50 and 60 years randomly sampled from the general population, Olofsson et al.8 found that 1 and 4% N2/L of baseline N2-slope produced a decline in FEV1 of 56 and 82 mL/year, respectively. Beaty et al.9 observed a significant negative relationship between baseline CC and ΔFEV1 in women, and between baseline CC or N2-slope and ΔFEV1 in men. Such study on 1912 adults had a shorter follow-up (average of 4.7 years between visits) and used a different SBN2 technique. By using our same SBN2 technique, Buist et al.10 analyzed data from two different cohorts, tested 3-5 times over a mean follow-up of 8.5 and 10.5 years. In both cohorts, ΔFEV1 rate was significantly associated with CV/VC (r = -0.16 in the Portland and -0.7 in the Screening center cohort) and with CC/TLC (r = -0.33 and -0.15), but not with N2-slope (r = -0.10 and -0.03). In a preliminary report on 389 subjects from the Tucson Epidemiological Study, Knudson et al.7 did not find an abnormal N2-slope to be a predictor of an accelerated ΔFEV1 after 4.2 years of follow-up. None of the above longitudinal studies evaluated SBN2 test indexes as predictors of incident airway obstruction or COPD. Stanescu et al. followed-up an original group of 105 male, long-term smoking, 45-55 years old steelworkers, after 6 (n=85)34 and 13 (n=56) years.11 Baseline N2-slope was significantly related to FEV1 decline over 6 years only in obstructed subjects with a baseline FEV1/VC < 66.6% (r = 0.59, p < 0.01),34 for whom baseline N2-slope was significantly related to FEV1 also at 13-year follow-up (r2 = 55%), but not to change in FEV1 over 13 years.11 No other SBN2 index was found related to FEV1 decline. Moreover, N2-slope was related to FEV1/VC (r = 0.61, p < 0.001) at baseline but not to change in FEV1/VC over 13-year follow-up.11 They concluded that middle-aged smokers are at no evident risk of functional deterioration if their FEV1/VC ratio is normal, even if results of small airways tests are abnormal.11 More recently, using data from a cohort surveyed in 1970s and followed-up for 38 years, Olofson et al. observed that N2-slope predicted the cumulative COPD event incidence (i.e. first hospital admission or death for COPD)21 and combining the N2-slope with FEV1 measurements and smoking history improved the prediction of future COPD events.22

Overall, previous longitudinal studies have observed links between impairment in small airways as detected by SBN2 test and subsequent loss in lung function. In our study, N2-slope appeared to predict lung function decline or airway obstruction. Our result is consistent with those from some previous longitudinal studies,8,21,22 but not with others.7,9–11 N2-slope reflects the distribution of ventilation in lung periphery while CV and CC denote the lung volume at which peripheral airway closure begins, and all these indexes may provide an indication of processes associated with parenchymal or small airways dysfunction.2 Different sampling procedures, SBN2 test techniques35,36 or statistical approaches of the quoted studies may account for the different findings with our study, along with a younger mean age of our population.

In our study there were 1228/3284 (37.4%) longitudinal subjects who had acceptable baseline SBN2 test and ΔFEV1 measures. Among these subjects only adults aged ≥ 20 yrs were selected (n=907), as lung function decline and incidence of COPD over a 8-year follow-up is not expected in subjects aged under 20 yrs. While all our subjects who completed the SBN2 test provided a measure of the N2-slope, only 83.9% and 85.5% provided a measure of CC and CV, respectively. This may be at least in part explained by the difficulty in performing SBN2 test maneuver among participants in this epidemiological survey.23 Subjects included in the analyses were significantly younger, slightly taller, and had better FEV1 with respect to those excluded. The regression analyses to estimate the risk for lung function decline associated with baseline SBN2 test indexes accounted for these confounders. A significant lower difference was found for ΔFEV1 between subjects included and excluded from analyses, which may be explained by the smaller number and the younger mean age of those providing acceptable spirometry measurements at both surveys 8-year apart. We observed that baseline N2-slope was associated with higher risks of lung function decline and airways obstruction incidence in a relatively young population; this result reinforces the importance of SAD in the natural history of COPD, as even higher risks may be expected in an older population possibly exposed to smoking for a longer period.

N2-slope has been repeatedly found to be the most sensitive SBN2 index in separating smokers from non-smokers,3,22,37 as we previously found.23 In the present longitudinal study, an association between N2-slope and decline in FEV1 was found adjusting for smoking habit, while current or former smoking were not significant determinants for incident airway obstruction. A shorter and lighter history of smoking in current young smokers and the inclusion of occasional among current smokers may explain these effects. Indeed, when only subjects aged over 30 years were analyzed, a statistically significant difference in ΔFEV1 of -7.4 mL/year was observed between current and never smokers, with no detectable difference between never and ex-smokers. The association between N2-slope and decline in FEV1 observed in this young study sample (mean age 37.4±11.1 years) from the general population highlights the relevance of the impairment of small airways in predicting lung function decline even independently from smoking.

Considering the scientific discussion on the spirometry definition of COPD (i.e., FEV1/FVC < 70% fixed ratio vs. FEV1/VC < 5th percentile of the predicted value),38 we evaluated SBN2 test as predictor of COPD incidence defined according to either GOLD32 or ATS-ERS33 criterion. We found N2-slope to be a significant predictor of COPD with both criteria. Among subjects without airway obstruction at baseline, approximately 10% vs. 5% developed COPD based on GOLD or ATS-ERS criteria, respectively. This difference may confirm that the use of a non-age-adjusted criterion to define airway obstruction may misclassify the older subjects as obstructed. Indeed, differently from the GOLD model with N2-slope predicting COPD incidence, in the ATS-ERS model age did not reach the statistical significance. This is consistent with the age-adjustment inherent in the ATS-ERS criterion.

We performed baseline spirometry without using bronchodilator, as at the time of our study there was no suggestion to perform a bronchodilator test when carrying out spirometry in general population samples. A Norwegian group, using a bronchodilator test in a community study, estimated post-bronchodilator airway obstruction prevalence (defined as FEV1/FVC < 0.70) to be 27% lower than the pre-bronchodilator values.39 However, the need of using a therapeutic agent in general population studies is still questionable, in view of the widespread lack of awareness of the disease in the general public who, in real life, is unlikely to seek medical advice at the early stage of the disease.40

Our study reinforces the usefulness of the SBN2 test in the early detection of airways obstruction. The reported difficulty in performing SBN2 test in population-based studies may not apply to the clinical setting. Indeed, high reproducibility for N2-slope has been shown in different populations of healthy subjects (correlation coefficient for duplicate measurements: r = 0.9635 or r = 0.9441). Other functional or biological tests for exploring small airways obstruction, possibly more patient friendly or easy to apply in clinical practice, are available: e.g., the impulse oscillometry (IOS) or exhaled nitric oxide measurements. Impulse oscillometry parameters have been shown to be reproducible in three measures in healthy children and children with cystic fibrosis (intraclass correlation coefficient = 0.9).42 Fractional exhaled nitric oxide (FeNO) measurements have been shown to be reproducible in healthy and asthmatic adults, as well as in children (intraclass correlation coefficient = 0.99).43 While no study has directly compared IOS with the SBN2 test, strong correlations have been found between the alveolar nitric oxide concentration (CAlv) and N2-slope in patients with severe asthma.44 The IOS parameter integrated area of low frequency X has been shown to predict the FEV1 decline in a group of COPD patients.45 There are no prospective data on the association between impairment of small airways, as assessed by IOS, and incidence of COPD. Alveolar nitric oxide concentration has been used to study the small airways in COPD, in addition to asthma,46 but longitudinal studies are lacking.

Thus, SBN2 test might be part of an extended evaluation of respiratory function, in addition to spirometry, in patients with clinically significant risk factors for COPD, e.g., tobacco smoking. Indeed, in the last years, the relevance of targeting treatment to small airways in asthma and COPD has been pointed out,47 and SBN2 test might be used to detect respiratory patients with functionally impaired small airways. In the clinical practice, highlighting abnormal results of the SBN2 test might also be used to motivate smoking patients to quit; indeed, it is known that diagnosis of airway obstruction motivates smokers to quit smoking48 and higher cessation rates are observed in those who are obstructed.49

N2-slope from SBN2 test is a significant predictor of lung function decline and COPD incidence defined according to either GOLD or ATS-ERS criteria. These results confirm the role of SAD in the natural history of COPD. Large prospective studies are needed to evaluate whether new proposed functional or imaging tests that measure small airways impairment may be useful in the early detection of COPD. Meanwhile, pulmonologists could re-discover an “old” test, which provide important information on their patients at risk for developing COPD.

This work was supported in part by: the National Research Council, Targeted Project "Prevention and Control Disease Factors - SP2 - Contract N 91.00171.PF41", the Italian Electric Power Authority (ENEL) - CNR Project "Interactions of energy system with human health and environment", Rome, Italy.