Induced sputum examination is a relatively non invasive method standardized by Pin et al[1] as an alternative to bronchoscopy procedure for collecting secretions and inflammatory cells from the airways of subjects with Chronic Obstructive Pulmonary Disease (COPD) and bronchial asthma. The method consists of the induction of airways secretions after the inhalation of progressively increased concentrations of saline aerosol. It is more feasible than bronchoscopy and is considered to provide repeatable and valid results [2].

However, the inhalation of normal or hypertonic saline, may also act as an indirect bronchoconstrictive stimulus in subjects with airflow limitation, leading to further lung function deterioration and worsening of symptoms such as dyspnea[2]. Concerns about the safety of the procedure have been raised especially after the report of a fatal asthma attack, precipitated by inhalation of distilled water[3]. In asthma, the method is considered safe when standard guidelines are applied[4].

In COPD lung function deterioration has been reported during the procedure [5–7]. European Respiratory Society (ERS) sputum induction Task group report[5] has underlined the lack of systematic studies addressing safety issues in patients with advanced COPD. In previous studies [6–10] the main objective was mainly the lung function changes and forced expiratory volume in one second (FEV₁) decline during SI but, little attention was paid in the worsening of dyspnea. However, dyspnea development during SI may affect the tolerability of the method in COPD[8, 10]. Thus, for research purposes or for clinical studies investigating cells or mediators measurable in sputum, it is important to realize the potential danger imminent in this procedure and to improve the tolerability of SI.

In the present study we performed SI by administrating normal and hypertonic saline in a group of COPD patients, as part of a longitudinal study of lung function decline. Our primary aim was 1) to assess dyspnea and adverse events during the procedure and, 2) to evaluate the relation between dyspnea and lung function change or oxygen desaturation that may occur during the procedure. A secondary endpoint in this study was to determine the diagnostic value of baseline parameters in distinguishing patients who will experience an adverse reaction during SI.

This prospective study represents the largest systematic report of adverse events and dyspnea evaluation during sputum induction in COPD. Our results suggest that COPD patients who undergo sputum induction, following a safety protocol, do not experience major adverse events. However, the patients may have an increase perception of dyspnea [overall increase in Borg score median[(IQR) 1.5(0–2)] and desaturation and may require short discontinuation of the procedure due to undesired symptoms. We found that dyspnea changes during sputum induction were significantly correlated with oxygen saturation, heart rate changes and mild adverse events. This adverse physiologic reaction was more frequent in subjects with advanced COPD. Notably, patients of stage IV in GOLD staging of severity, presented an increase risk to have unsuccessful sputum induction and to have minor adverse events during the procedure compared to patients with less severe disease (RR 2.9, 95%CI 1.4–6). In addition we found that the baseline values of post-bronchodilation FEV₁, of oxygen saturation and of 6MWT have a diagnostic value in distinguishing patients who develop an adverse reaction during SI. To the best of our knowledge this has not been reported.

In previous studies addressing SI safety in COPD, the measurement of FEV₁ was considered enough to diagnose acute lung responses[6, 7, 9]. However, an excessive fall in FEV₁ during SI is not always associated with clinical deterioration[9] or dyspnea development[6]. Thus, it is not unlikely that assessing only FEV₁ during the procedure, early signs of clinical deterioration may be undetected. Consequently, the development of adverse events may affect the tolerability of the procedure[8, 9]. In our study, special attention was paid to evaluate dyspnea, in addition to FEV₁ assessment. We found that patients undergoing SI may experience an adverse physiologic reaction, characterized by worsening of dyspnea, undesired symptoms and oxygen desaturation. This adverse reaction was not significantly related to FEV₁ decline but affected the tolerability of the procedure in a proportion of patients. This may have importance in research or clinical studies, especially in patients with advanced disease.

In the studied population, we included subjects with advanced disease since European Respiratory Society report has underlined the lack of systematic studies in this category of patients[5]. Normal and hypertonic saline was administered even in subjects with very severe disease. We found that COPD patients experienced significant Borg score and oxygen saturation changes, associated with disease severity. COPD subjects of stage IV had an increased risk of developing dyspnea, requiring subsequently discontinuation of the procedure. The degree of discomfort led eight out of 14 (57%) patients in this category to temporary discontinuation and two of them (14%) to early termination of the procedure. In addition, 4 out of 5 patients who did not provide sufficient sputum sample, had stage IV COPD. Hence, patients with advanced COPD may experience excessive dyspnea during sputum induction and they might be reluctant to repeat the procedure in the future. Therefore, sputum induction in this category of patients must be performed with great caution and in the ground of our findings careful monitoring of oxygen saturation and dyspnea is essential.

In the present study all subjects were premedicated with salbutamol. B₂-agonists and anticholinergic inhalers were withheld before SI in order to standardize further our assessment. It is known that inhaled salbutamol does not provide full protection from bronchoconstriction as it has been demonstrated by the adverse responses after saline inhalation[7, 9]. However, it remains unclear whether the magnitude of bronchoconstriction could be prevented by pretreatment with larger doses of inhaled salbutamol or with another type of bronchodilator or antinflammatory treatment[9, 10]. A prospective study is necessary to test this hypothesis.

In this study, 31% of the patients demonstrated an excessive fall (>20%) of FEV₁. According to previous reports, excessive FEV₁ decline ranges, between 11[6] and 50%[9], depending on the COPD population studied. Interestingly, we found that FEV₁ decline had almost reached the average decline 2 minutes after saline inhalation and that FEV₁ did not return to baseline during the procedure. In previous studies in COPD patients[8–10], there hasn't been any assessment before the 5^th minute following saline inhalation. In these studies, the greatest decline in FEV₁ seems to occur constantly at the beginning of the procedure, following similar time course patterns to our assessments. The time course pattern of FEV₁ could be explained by the underlying mechanism of the bronchoconstrictive response. The inhalation of normal or hypertonic saline may trigger mast cell and basophil degranulation, in response to an increase of airway osmolarity[17–20]. The release of bonchoconstrictive mediators from mast cells is rapid and essentially completed by five minutes[21]. In line with this early inflammatory response, it has been reported in a time course assessment study in asthmatic patients[18], that the maximal mean fall in FEV₁ occurs at 3 minutes post saline inhalation. In this ground, premedication of the patients with bronchodilators should be a standard safety measure of the procedure. Future studies of different design may identify which is the most effective bronchodilator to prevent this bronchoconstrictive response.

An interesting point in our study is that the development of dyspnea and the fall of the FEV₁ were not significantly correlated. One would expect patients with the most severe airway obstruction to be the most dyspnoeic. However, some patients with severe airway obstruction are minimally symptomatic, whereas others with little objective dysfunction appear to be very dyspnoeic[22]. Several studies have investigated the correlation between dyspnea and lung function[23]. Mahler et al reported that dyspnea and baseline pulmonary function are independent quantities in patients with COPD[24]. Subsequent studies employing newer techniques to quantify breathlessness found either no significant or weak correlations with FEV₁[25, 26]. Thus, an excessive fall in FEV₁ is not always correlated with symptoms development and with dyspnea scale scores in COPD[6, 9].

A reasonable explanation for this discrepancy between FEV₁ decline and clinical deterioration during SI, may be the subjectiveness of dyspnea perception[27]. Unlike asthmatic patients who experience episodic bronchoconstriction, those with COPD demonstrate chronic airflow limitation that might lead to desensitization. Ottanelli and colleagues have previously reported that a reduced perception of dyspnea during bronchoconstriction may be present in COPD patients[28]. A reduced perception of dyspnea might delay self referral or lead to underreport of discomfort during the procedure. In addition, it may be that some dyspnoeic patients in our study, did not develop their "potential maximal" drop in FEV₁, because they felt discomfortably and interrupted the procedure thus, demonstrating a submaximal effort in lung fuction testing. Furthermore, dyspnea during a bronchoconstrictive challenge is associated not only to airway obstruction but also to hyperinflation[29]. In fact, dyspnea perception may be better related with acute hyperinflation than with airflow obstruction sensation in patients with chronic airflow obstruction[28]. Patients with advanced COPD may develop dynamic hyperinflation in the setting of a bronchocontrictive stimulus[9, 29]. In the present study the most severely affected patients in terms of baseline disease severity presented the greatest perception of dyspnea during SI. Thus, it is likely that a proportion of patients may have experienced dyspnea during SI due to acute hyperinflation.

In this ground other lung function parameters may be considered in addition to FEV₁, when addressing safety in SI. It has been demonstrated that forced inspiratory rather than expiratory parameters were more sensitive in detecting SI related lung function deterioration and were better associated to dyspnea[9, 30]. Forced inspiratory volume in one second(FIV₁) is less affected by airway collapse than FEV₁, reflecting obstruction and hyperinflation[9]. In addition, acute inspiratory capacity changes (IC) account in part for the variability in the perception of dyspnoea after accounting for changes in FEV₁ during bronchoconstriction in patients with chronic airflow obstruction[28]. These data, along with the disassociation between dyspnea and FEV₁ in our study, suggest that possibly other parameters, like FIV₁ or IC should be brought forward to monitor lung function deterioration and adverse events development, during SI. In this ground a new insight for the reason of dyspnea during SI might also be provided. However, the present study was designed to assess dyspnea intensity and adverse events during SI based on ERS sputum induction task group report [5] and thus other lung function parameters were not assessed.

In the present investigation, ΔBorg was significantly correlated with ΔSpO₂. However the correlations between dyspnoea intensity and oxygen saturation changes were weak. A plausible explanation may be that the relationship between hypoxia – ventilatory response and breathlessness in patients with COPD is not linear. Thus, the level of breathlessness is related to hypoxaemia but not in all levels of desaturation[31]. In addition, dyspnea sensation may result from pathophysiological abnormalities that can be related to non respiratory mechanisms[32]. Therefore, the weak correlations between dyspnea and oxygen saturation changes could be attributable to other factors (emotional, cognitive) which have not been evaluated in the present study.

In the present study, we evaluated the diagnostic performance of baseline clinical characteristics of the patients undergoing SI, in order to distinguish those who will develop an adverse physiologic response during SI. This is important because clinical parameters which can be measured in a simple way, before performing the test, give useful information in advance. In addition, predictors of adverse events and lung function deterioration during SI are not yet widely known [6, 9, 27]. We found that the development of dyspnoeic events during SI, could be better predicted by the post-bronchodilation FEV₁(%pred), the MRC score, the oxygen saturation and the 6MWT. To the best of our knowledge, this has not been reported until now[5, 7, 10].

MRC score is a good predictor of exercise capacity. It has showed a consistent relationship with Borg rating and a significant correlation with breathlessness and dynamic hyperinflation measured during walking[33]. Baseline oxygen saturation is also reported to be associated with the hypoxemia during inhalation provocative tests[34]. In addition, our investigation showed that COPD patients with good performance status, by means of walking more than 155 meters during 6MWT, will be less prone to develop dyspnea during the procedure. This is likely to be due to dynamic hyperinflation. 6MWT performance is associated to the oxygen uptake, to the severity of chronic dyspnea in COPD patients and it may also be related with the dynamic hyperinflation which is developed in patients after certain stimuli[9, 35]. We believe that since subjects with moderate to severe COPD are characterized by hyperinflation and low performance, 6MWT is possibly a good predictor of developing dyspnea after a stimulus such as the inhalation of saline [9, 22].

In summary, we found that normal and hypertonic saline-induced sputum is a safe technique, when certain precautions are taken, in patients with COPD. It is safe even for patients in an advanced stage of the disease. However, excessive dyspnea is more likely to occur in these patients, leading in temporary or permanent discontinuation, affecting tolerance and success of the procedure. Therefore, sputum induction must be performed with great caution and careful monitoring of dyspnea and oxygen saturation in patients with very severe COPD. Post-bronchodilation FEV₁(%pred), oxygen saturation and 6MWT have a prognostic value for the development of dyspnea during SI and it would be useful to be evaluated in advance. We believe that this is important information and favors further the improvement of SI safety and tolerance especially in advanced COPD.