Sputum Induction in COPD—It's Safe, So Now What Do We Do?

Examination of the cellular and humoral constituents of induced sputum has, over the last decade, become a frequently-utilized assay of lower airway inflammation in diseases such as asthma and chronic obstructive pulmonary disease. Although the literature would seem to suggest that analysis of sputum cellularity, sputum supernatant cytokine concentrations and even gene expression profiles can each provide important insights into the pathobiology of chronic obstructive pulmonary disease (COPD), for reasons that are not well-characterized, this technique is less commonly applied to the study of COPD than asthma. In fact, a search of MEDLINE ([1]) crossing the term “induced sputum” with either “asthma” or “chronic obstructive pulmonary disease” indicated that published studies of sputum induction in asthma are approximately six times more common than those in COPD.

There is a paucity of data regarding the factors which might cause investigators to limit the use of sputum induction in COPD, but it has been proposed that unanswered questions regarding safety have thus far limited the widespread deployment of this investigative tool in COPD ([2]). Similar concerns existed when sputum induction was applied in the study of asthma, raised in part by reports that inhalation of hypertonic saline could cause bronchoconstriction ([3]) and oxygen desaturation ([4]), and possibly increase airway inflammation ([5]). Many of these concerns were allayed in 2001 when Fahy and colleagues in the Asthma Clinical Research Network reported ([6]) the results of a study which evaluated the safety and reproducibility of sputum induction with 3% hypertonic saline ([7]) in 79 subjects with moderate-to-severe persistent asthma with an FEV₁ range between 40% and > 80% of predicted. Although 14% of subjects experienced a fall in FEV₁ of > 20%, and although those with lower starting FEV₁ were at greater risk for a > 20% decline in FEV₁, all recovered to within 12% of the baseline FEV₁ after treatment with supplemental albuterol, and none experienced refractory bronchospasm. Furthermore, the reproducibility of measured inflammatory biomarkers such as sputum eosinophils percentage and tryptase and eosinophilic cationic protein concentrations were at least as reproducible as the methacholine PC₂₀ FEV₁ ([6]).

In this issue of COPD: Journal of Chronic Obstructive Pulmonary Disease, Wilson and colleagues ([8]) add to the literature evaluating the safety of sputum induction in COPD with a study notable for its size, range of severity of starting FEV₁ and incremental increase in saline concentration throughout the course of the procedure. The investigators evaluated 100 consecutive subjects with moderate to very severe COPD ([9]), beginning induction with either 0.9% or 3% saline and then increasing the concentration as allowed by FEV₁-based criteria. Five subjects experienced a fall in FEV₁ > 20%, and all recovered to within 10% of baseline after treatment with albuterol. They concluded, as have others ([10], [11], [12]), that sputum induction may be performed in patients across a wide range of COPD severity without causing significant or prolonged reductions in FEV₁. Wilson and colleagues did not measure oxygen saturation during their study, but complementary research has suggested that oxygen saturation also remains stable during sputum induction in moderate-to-severe COPD ([12]).

Understanding the safety characteristics of sputum induction is a critical first step toward increasing overall utilization of this investigative tool in COPD, and much work remains to be done to determine whether evaluation of induced sputum can yield one or more biomarkers that will improve our ability to diagnose and predict prognosis or response to therapy in COPD. Current diagnostic and therapeutic recommendations ([13]) focus on composites of symptoms and lung function as a guide to diagnosis, staging and treatment, but there is disconnect between these variables and overall prognosis, where additional factors such as body mass index and exercise performance have recently been shown to be important factors as well ([14]).

Important work in applying biomarkers to treatment has begun in asthma, with the publication in recent years of high-impact papers evaluating treatment strategies based on noninvasive biomarkers such as the degree of airway hyperresponsiveness ([15]), induced sputum eosinophil percentage ([16]) and exhaled nitric oxide ([17]). In the case of airway hyperresponsiveness, the work of Sont and colleagues demonstrated that titrating inhaled corticosteroid therapy to the degree of airway hyperresponsiveness resulted in a reduced cumulative incidence of first mild exacerbations at the cost of a higher overall utilization of inhaled corticosteroids ([15]), whereas Green and coworkers have demonstrated that titrating corticosteroid therapy to induced sputum eosinophils reduced exacerbations without increasing the need for oral or inhaled corticosteroids ([16]), and Smith and colleagues have reported that a protocol which altered controller therapy based on exhaled nitric oxide could achieve similar degrees of asthma control and exacerbation frequency as a symptom-based algorithm, with an approximately 40% reduction in the required dose of fluticasone ([17]). While the extent to which data like these can be integrated into clinical paradigms and applied to large groups of patients remains to be determined, this type of biomarker-based individualization of therapy should be as feasible in COPD as it is becoming in asthma.

Some initial steps in this direction have already been taken. In 2000, Brightling and colleagues published the results of a randomized controlled trial of prednisolone in moderate-to-severe COPD, demonstrating that eosinophilic airway inflammation with a mean eosinophil percentage of 2.35 ± 0.09 (SE) was associated with airflow limitation in subjects with COPD and that the short-term effect of prednisolone appeared to be modification of this component of airway inflammation ([18]). While this is one important example of utilizing sputum analysis to identify a biomarker of response to systemic corticosteroid therapy in COPD, it remains largely unknown whether sputum or other biomarkers predict response to the far more commonly-prescribed inhaled corticosteroids and bronchodilators or whether they correlate with important prognostic factors such as the rate of lung function decline or the BODE index ([14]).

Going forward, there is an urgent need for research which identifies biomarkers that are clearly associated with disease severity, are predictive of prognosis and therapeutic response, and which are themselves responsive to treatment such that therapeutic algorithms can be built around them. These are critical issues as the prevalence of COPD continues to increase worldwide, increasing the burden on both patients and clinicians, and increasing the need for a refined approach to this disorder.