Mucolytics for acute exacerbations of chronic obstructive pulmonary disease: a meta-analysis

Abstract

This meta-analysis explored the safety and effectiveness of mucolytics as an add-on treatment for chronic obstructive pulmonary disease (COPD) exacerbations. Based on a pre-registered protocol and following Cochrane methods, we systematically searched for relevant randomised or quasi-randomised controlled trials (RCTs). We used the Risk of Bias v2 tool for appraising the studies and performed random-effect meta-analyses when appropriate. We assessed certainty of evidence using GRADE. This meta-analysis included 24 RCTs involving 2192 patients with COPD exacerbations, entailing at least some concerns of methodological bias. We demonstrated with moderate certainty that mucolytics increase the rate of treatment success (relative risk 1.37, 95% CI 1.08–1.73, n=383), while they also exert benefits on overall symptom scores (standardised mean difference 0.86, 95% CI 0.63–1.09, n=316), presence of cough at follow-up (relative risk 1.93, 95% CI 1.15–3.23) and ease of expectoration (relative risk 2.94, 95% CI 1.68–5.12). Furthermore, low or very low certainty evidence suggests mucolytics may also reduce future risk of exacerbations and improve health-related quality of life, but do not impact on breathlessness, length of hospital stay, indication for higher level of care or serious adverse events. Overall, mucolytics could be considered for COPD exacerbation management. These findings should be validated in further, rigorous RCTs.

Short abstract

A meta-analysis of 24 RCTs (N=2192) demonstrated with moderate certainty that mucolytics can improve treatment success rate by 37% and significantly improve symptoms in COPD exacerbations #AECOPD. There were no safety signals. https://bit.ly/3RwTzlU

Introduction

Chronic obstructive pulmonary disease (COPD), the third leading cause of death globally, affects more than 300 million people worldwide [1, 2]. On average, COPD patients suffer from 0.5–3.5 exacerbations per year, that aggravate health status, quality of life and healthcare burden [3, 4]. Notably, COPD exacerbations, especially those with prolonged duration or incomplete recovery, precipitate unfavourable disease progression, which highlights the significance of their appropriate, rigorous management [5].

COPD exacerbations are characterised by substantial heterogeneity in their aetiology, clinical manifestations and underlying pathophysiological pathways, warranting a personalised approach to their management [6]. This is not yet fully realised in clinical practice, since their pharmacotherapy remains suboptimal, unchanged for decades, generic and mainly based on the administration of bronchodilators, antimicrobials and/or systemic corticosteroids [7]. The treatable traits theory inaugurates a patient-specific paradigm targeting clinically relevant, identifiable and modifiable features, and offers an effective, clinically practical methodology for personalising the management of airway diseases, including exacerbations [8–10]. For instance, decompensated type 2 respiratory failure represents the most established trait, which is consistently treated with noninvasive ventilation [11]. Bacterial infections and airway eosinophilic inflammation are increasingly targeted with biomarker-guided administration of antimicrobials or systemic corticosteroids, respectively [12–14].

Hypersecretion of mucus with increased viscosity represents another key treatable trait both during stable disease and acute exacerbations. It is burdensome to patients and associated with unfavourable outcomes [15] and disease progression [16]. Mucolytics regulate mucus viscoelastic properties, mostly by altering the degree of crosslinks and molecular interactions within mucin polymers, thereby potentiating mucociliary clearance and promoting sputum expectoration [17]. They are commonly prescribed in mucus hypersecretory conditions, such as respiratory tract infections or cystic fibrosis, and often for patients with COPD, in view of the mucus hypersecretion and ciliary dysfunction [18]. Furthermore, some possess antioxidant properties and exert anti-inflammatory or immunomodulatory effects, which could also benefit this patient group [17]. When administered in a stable disease state, various mucolytics, such as N-acetylcysteine, carbocysteine, erdosteine and ambroxol, have shown effectiveness in preventing COPD exacerbations and, perhaps, hospitalisation due to exacerbation [19]. Nevertheless, in the absence of conclusive evidence on their effectiveness in the management of exacerbations, there is no consensus around their use in clinical practice.

This systematic review sought to synthesise and appraise all available evidence from randomised controlled trials (RCTs) around the safety and clinical effectiveness of mucolytics as an add-on treatment to the standard care of COPD exacerbations.

Materials and methods

This systematic review and meta-analysis is based on a protocol that was prospectively registered with PROSPERO (ID: CRD42022314958) [20] and follows the methodology recommended by the Cochrane collaboration [21] and the Grading of Recommendation, Assessment, Development and Evaluation (GRADE) working group [22]. The present report adheres to the PRISMA statement [23].

Eligibility criteria

Randomised and quasi-randomised controlled trials assessing the clinical safety and effectiveness of mucolytics for COPD exacerbations were considered eligible. We included adult patients presenting with a clinical diagnosis of a moderate or severe COPD exacerbation. We considered eligible any pharmacological intervention primarily aiming to dissolve thick sputum that might have been administered via any route, in addition to usual care. In cases of inhaled mucolytics, normal saline was considered an acceptable placebo control. We excluded trials that recruited patients in the intensive care unit (ICU, those that focused on patients where COPD exacerbation was not the primary cause of their symptoms (e.g. decompensated heart failure and pneumonia) or assessed long-term administration of mucolytics during stable COPD.

Outcome measures

We assessed all the outcomes included in the European Respiratory Society (ERS) COPD exacerbations core outcome set [24] and additional secondary outcomes relevant to the intervention (details in supplementary table A1). The primary outcomes comprised: 1) patient-reported symptoms, preferentially assessed using standardised tools (e.g. the modified Medical Research Council dyspnoea scale), 2) treatment success [25], defined as a dichotomous measure of the overall outcome of the exacerbation that may be based on the judgement of a clinician and/or assessment of symptoms, signs and/or laboratory findings. Secondary outcomes comprised: mortality, indication for higher level of care (admission to hospital or ICU for the presenting exacerbation), hospitalisation duration, levels of oxygen and carbon dioxide in the arterial blood, health-related quality of life, activities of daily living, worsening of symptoms after initial treatment, disease progression assessed by pulmonary function tests, future exacerbations and future hospital admissions, serious adverse events, development of pneumonia, treatment adherence, overall microbiological outcome, and sputum characteristics (i.e. viscosity, volume and purulence).

Search strategy, study selection and data abstraction

Using a structured search strategy (see supplementary appendix), on 20 March 2022 we searched MEDLINE/PubMed, the Cochrane Central Register of Controlled Trials, the Cochrane Airways Trials Register, the World Health Organization International Clinical Trials Registry Platform and the conference proceedings of the ERS, American Thoracic Society and Asian Pacific Society of Respirology. The reference lists of all relevant studies were further perused. There were no language or time restrictions. Two authors independently assessed all identified studies for eligibility at title/abstract level, proceeding to full-text evaluation of all potentially eligible studies. Relevant data around the trial design, participant baseline characteristics, interventions and outcomes of interest were extracted in a structured and piloted Excel spreadsheet by two authors independently. Any disagreements were resolved through discussion or adjudication by a third reviewer.

Risk of bias and certainty of evidence assessment

The Cochrane Risk of Bias Tool v2 was used for assessing risk of bias at the study level [26, 27]. Moreover, certainty of the body of evidence per outcome was evaluated following GRADE methodology and taking into consideration the risk of bias, imprecision, inconsistency, indirectness and publication bias. We used a minimally contextualised approach for assessing GRADE. This approach focuses on assessing the presence rather than the magnitude of a potential treatment effect. We intended to use funnel plots for assessing publication bias in cases of meta-analyses involving at least 10 studies, but none of the meta-analyses fulfilled this criterion. The certainty was termed as “high”, “moderate”, “low” or “very low” according to GRADE [28].

Data synthesis

In anticipation of significant clinical and methodological heterogeneity, we performed random effect meta-analysis, using Review Manager 5.4.1 (Cochrane). Relative risks and mean differences (MDs), along with their 95% confidence intervals (95% CI), were calculated as effect estimates for dichotomous and continuous data, respectively. We used standardised mean differences (SMDs) to pool data from different scales/tools assessing the same outcome. SMD is used for conducting meta-analyses of various instruments assessing the same outcome (e.g. quality of life measured by different questionnaires). For calculating the SMD, the size of the intervention effect in each study is divided by the corresponding standard deviation. Heterogeneity was evaluated using the I² statistic and significant heterogeneity was explored via pre-specified subgroup analyses. We present narratively and in a tabulated format results of the included studies that were not reported in a format that could be included in the meta-analyses. All outcomes were assessed during treatment (we preferably captured measurements from the third to fifth treatment day), post-treatment (depending on the regimen) and long-term (>3 weeks) follow-up timepoints.

Sensitivity and subgroup analyses

For subgroup analyses, we intended to classify the trials according to exacerbation aetiology and severity, type or dosage of the mucolytics administered, and recruitment of patients with increased sputum viscosity or sputum volume.

Sensitivity analyses were performed using the fixed-effect model. We also intended to conduct a further sensitivity analysis only considering studies at a low risk of bias, but none of the included studies fulfilled this criterion.

Results

Study selection and baseline characteristics

The study selection process is depicted in the PRISMA flowchart (supplementary figure A1). 24 complete trials fulfilled the eligibility criteria totalling 2192 participants. Table 1 and supplementary table A2 present the characteristics of the included studies. Most trials were conducted in an inpatient clinical setting, with follow-up ranging from 7 days to 6 months. Patients with infective exacerbations were enrolled in seven trials [29–35], six of which only included patients with confirmed presence of bacteria in sputum cultures [30–35]. Another trial [36] recruited only patients receiving noninvasive ventilation. We did not identify significant baseline imbalances across the treatment arms of the included studies, although 10 studies only provided limited relevant information.

TABLE 1.

Baseline characteristics of the included randomised controlled trials (RCTs)

Study	Study design, clinical setting, follow-up	N	Withdrawal (M/C)	Mean (sd) age, years, males, %	Inclusion diagnosis	Sputum at presentation	Treatment regimen
Ansari et al. [ 42 ]	RCT NR, inpatient, ≥7 days	50	4/6	46.2 (10.7) 65.0%	Clinical AECOPD diagnosis with spirometric airflow obstruction	NR	NAC 1200 mg p.o. (twice daily), ≥7 days versus standard treatment
Aytemur et al. [ 53 ]	RCT DB, inpatient, 6 months	42	1/3	69 (8.8) 92.1%	AECOPD with previous spirometry confirmation	Increased volume (>50 mL·day−1)	NAC 600 mg p.o. (three times daily), 30 days versus placebo
Bisetti et al. [ 52 ]	RCT DB, inpatient, 7 days	28	0/1	62.3 (NR) 67.9%	Clinical AECOPD diagnosis Background: chronic bronchitis	Mucopurulent	Erdosteine 600 mg p.o. (twice daily), 7 days versus placebo
Black et al. [ 54 ]	RCT DB, inpatient, 7 days	50	0/0	73.3 (8) 60.0%	AECOPD with spirometric airflow obstruction admitted for ≤24 h	Mostly increased volume	NAC 600 mg p.o. (twice daily), ≤7 days versus placebo
Brocard et al. [ 29 ]	RCT DB, NR, 10 days	95	NR	NR	Acute infective exacerbation Background: chronic bronchitis	NR	NAC 600 mg p.o. (three times daily) 10 days versus placebo
El Hafiz et al. [ 55 ]	RCT DB, inpatient, >10 days	45	0/0	59 (7.5) 100%	AECOPD with airflow obstruction in spirometry GOLD criteria fulfilment	NR	NAC 600 mg p.o. (three times daily), NAC 1200 mg p.o. (three times daily), 10 days versus standard treatment
Galdi et al. [ 36 ]	RCT DB, inpatient, NR	41	2/4	75.8 (9.4) 34.1%	AECOPD requiring NIPPV Previous diagnosis of COPD	NR	Neb high molecular weight hyaluronan 0.3% 5 mL saline (twice daily) during NIPPV versus placebo
Jahnz-Rózyk et al. [ 37 ]	RCT DB, inpatient, NR	30	0/0	70.5 (6.9) 43.3%	Clinical AECOPD diagnosis Background: chronic bronchitis	NR	Neb ambroxol 30 mg (twice daily), up to clinical and spirometric improvement versus placebo
Langlands et al. [ 44 ]	RCT DB, inpatient, 20 days	31	3/1	NR 81.5%	AECOPD based on MRC criteria Background: chronic bronchitis	Mucoid	Bromhexine 24 mg p.o. (three times daily), 14 days versus placebo
Li et al. [ 40 ]	RCT NR, inpatient, 10 days	60	6/4	68.1 (8.8) 62%	Chinese medicine syndrome of retention of phlegm and heat in Fei with airflow obstruction in spirometry Onset ≤1 week	NR	Ambroxol 30 mg i.v. (twice daily), 10 days versus standard treatment
Maesen et al. [ 30 ]	RCT SB, inpatient, 17 days	22	NR	NR	Infective AECOPD Background: chronic bronchitis	Increased purulence	Bromhexine 72 mg (three times daily), 10 days versus placebo
Marchioni et al. [ 31 ]	RCT DB, NR, 11 days	237	6/5	64.1 (10.7) 76.4%	Infective, clinical AECOPD Background: chronic bronchitis	NR	Erdosteine 600 mg p.o. (twice daily), 7–10 days versus placebo
Mohanty et al. [ 50 ]	RCT DB, NR, NR	240	40 total	NR, NR	Clinical AECOPD diagnosis Background: chronic bronchitis	NR	Erdosteine 600 mg p.o., NR versus placebo
Moretti et al. [ 48, 49 ]	RCT SB, inpatient, 60 days	15	0/0	69.6 (5.6) NR	Clinical AECOPD diagnosis Hospitalisation ≤48 h from symptom onset	NR	Erdosteine 900 mg (three times daily), 10 days versus placebo
Moretti et al. [ 46, 56 ]	RCT SB, inpatient, 60 days	40	0/0	70.7 (5.8) 82.5%	Clinical AECOPD diagnosis with onset ≤24 h	Increased volume and purulence	Erdosteine 900 mg p.o. (three times daily), 10 days versus standard treatment
Paganin et al. [ 32 ]	RCT NR, NR, >10 days	24	4 (total)	61.5 (7.4) 79.2%	Infective, clinical AECOPD diagnosis with airflow limitation	Increased purulence	Ambroxol 90 mg p.o. (three times daily), 10 days versus standard treatment
Patel et al. [ 38 ]	RCT SB, inpatient, 30 days	70	9/0	61.2 (9.3) 49.2%	AECOPD	NR	Neb hypertonic saline 3.00% (every 6 h and as needed), 24 h versus neb normal saline 0.9%
Peralta et al. [ 33 ]	RCT DB, NR, 10 days	24	1 (total)	60 (8.4) NR	Infective, clinical AECOPD Background: chronic bronchitis	Increased purulence	Ambroxol 90 mg p.o. (three times daily), 10 days versus standard treatment
Reichenberger et al. [ 34 ]	RCT NR, NR, 21 days	24	NR	66 (10) 66.7%	Infective clinical AECOPD Background: chronic bronchitis	NR	NAC 1200 mg (twice daily), 21 days versus standard treatment
Ren et al. [ 39 ]	Q-RCT SB, inpatient, NR	78	0	80.8 (4.4) 62.8%	Clinical AECOPD diagnosis, according to the Respiratory Branch of the Chinese Medical Association	NR	Inh NAC 600 mg (twice daily), NR versus inh normal saline (6 mL)
Ricevuti et al. [ 35 ]	RCT DB, NR, 8 days	30	0	51.5 (NR) 46.7%	Infective, clinical AECOPD Background: chronic bronchitis	Increased viscosity and volume	Erythromycin-propionate-N-acetylcysteinate p.o. (three times daily), 7 days versus erythromycin stearate
Study 7171L01 [ 45 ]	RCT DB, NR, 30 days	714	18.21/17	NR NR	Anthonisen's criteria for AECOPD and BCSS ≥5 ATS/ERS criteria for COPD (documentation ≤1 year)	NR	N-Acetylcysteine 1200 mg, 600 mg p.o. (once), 10 days versus placebo
Zhang et al. [ 41 ]	RCT NR, NR, NR	80	NR	NR NR	AECOPD	NR	Ambroxol 60 mg i.v. (twice daily), NR versus standard treatment
Zuin et al. [ 51 ]	RCT DB, outpatient, 10 days	122	0.1/1	66.7 (12.4) 57.4%	Clinical AECOPD diagnosi	NR	NAC 600 mg p.o. (once and once placebo), 1200 mg p.o. (twice daily), 10 days versus placebo (twice daily)

AECOPD: acute exacerbation of chronic obstructive pulmonary disease; ATS: American Thoracic Society; BCSS: Breathlessness, Cough and Sputum Scale; C: control group; DB: double-blinded; ERS: European Respiratory Society; GOLD: Global Initiative for Chronic Obstructive Lung Disease; inh: inhaled; i.v.: intravenous; M: mucolytic group; MRC: Medical Research Council; N: number of patients; NAC: N-acetylcysteine; neb: nebulised; NIPPV: noninvasive positive pressure ventilation; NR: not reported; p.o.: per os; Q-RCT: quasi-randomised controlled trial; SB: single-blinded; sd: standard deviation.

The most commonly administered mucolytic was N-acetylcysteine (10 trials), followed by ambroxol (n=5), erdosteine (n=5) and bromhexine (n=2), whereas hypertonic saline and high molecular weight hyaluronan were assessed in one trial each. Mucolytics were usually administered orally, but inhaled [36, 37–39] and intravenous [40, 41] routes were also used. The median treatment duration was 10 days, generally ranging between 7 and 30 days across the trials, but was limited to only 24 h in one [38].

Most studies were double blinded (n=14), and some were single blinded (n=5), while five studies did not report on blinding. Adherence to treatment was not quantified in any trial, but patient withdrawal was significant, up to 24% for the control group in one study [42], up to 22.5% for the mucolytic group in another [43] and reaching more than a 10% difference between treatment arms in two trials, with higher occurrence in the mucolytics groups [43, 44]. Unfortunately, most of these trials did not describe the reasons for withdrawal.

Acceptable concurrent treatments included bronchodilators, inhaled corticosteroids, systemic corticosteroids, antimicrobials and/or methylxanthines, with or without oxygen supplementation, whereas one study only included patients receiving noninvasive positive-pressure ventilation [36]. In two studies where administration of the mucolytic was via inhalation, the control group received normal saline as an inhaled placebo [38, 39]. Concurrent chest physiotherapy to facilitate mucus expectoration was only reported in one trial [29].

Risk of bias assessment

All studies entailed some concerns or high risk of bias, mainly due to unclear allocation sequence, per-protocol analysis, utilisation of nonvalidated scores in outcome measurements and selection bias (figure 1 and supplementary figure A1).

FIGURE 1.

Risk of bias table. D1: bias arising from the randomization process; D2: bias due to deviations from the intended intervention; D3: bias due to missing outcome data; D4: bias in measurement of the outcome; D5: bias in selection of the reported result.

Meta-analyses

There was significant variability in the instruments used to assess the selected outcomes across the included studies. Moreover, information required for including results in the meta-analyses was often missing and, as a result, we were not able to pool all available data. Details of all relevant findings reported in the included trials are presented in supplementary table A3. Our meta-analyses are presented in figure 2 and supplementary figure A2. GRADE evidence profiles and detailed judgements for all clinically relevant outcomes are summarised in table 2.

FIGURE 2.

Forest plots depicting the overall effect estimates for the following outcomes upon treatment completion: a) treatment success; b) overall symptom scores; c) absence of breathlessness; d) absence of cough; e) ease of expectoration. M–H: Mantel–Haenszel method; SMD: standardised mean difference.

TABLE 2.

Grading of Recommendation, Assessment, Development and Evaluation (GRADE) evidence profile

Outcome	Measure	n (N)	Effect estimates	Certainty						Benefit
				Risk of bias	Inconsistency	Indirectness	Imprecision	Other	Overall certainty
Treatment success	N	4 (383)	Relative risk 1.37 (1.08–1.73), I2=63%	Serious#	Not serious	Not serious	Not serious	Not serious	Moderate	Yes
Overall symptom score	Post-treatment scores	3 (316)	SMD 0.86 (0.63–1.09), I2=0%	Serious#	Not serious	Not serious	Not serious	Not serious	Moderate	Yes
	Change from pre-treatment	1 (50)	SMD 0.80 (0.22–1.38)	Serious#	Not serious	Not serious	Not serious	Not serious
	N with improvement	1 (200)	Relative risk 1.18 (1.07–1.31)	Serious#	Not serious	Not serious	Not serious	Not serious
	Dyspnoea cough attacks	1 (30)	MD −1.40 (−3.51–0.71)	Serious#	Not serious	Not serious	Serious¶	Not serious
Breathlessness	N with absence	2 (276)	Relative risk 1.36 (0.55–3.38), I2=85%	Serious#	Serious+	Not serious	Serious¶	Not serious	Very low	No
	Change from pre-treatment in Borg's scale	1 (59)	MD 0.40 (−0.55–1.35)	Serious#	Not serious	Not serious	Serious¶	Not serious
	N with improvement	1 (27)	Relative risk 1.08 (0.69–1.68)	Serious#	Not serious	Not serious	Serious¶	Not serious
	N with dyspnoea at rest	1 (40)	Relative risk 0.18 (0.01–3.56)	Serious#	Not serious	Not serious	Serious¶	Not serious
Cough	N with absence	2 (276)	Relative risk 1.93 (1.15–3.23), I2=29%	Serious#	Not serious	Not serious	Not serious	Not serious	Moderate	Yes
	N with improvement in frequency/in intensity	1 (122)	Relative risk 1.40 (0.88–2.22)/Relative risk 2.06 (1.23–3.43)	Serious#	Not serious	Not serious	Serious¶	Not serious
Ease of expectoration	N with improvement	2 (149)	Relative risk 2.94 (1.68–5.12), I2=0%	Serious#	Not serious	Not serious	Not serious	Not serious	Moderate	Yes
	N with absence	1 (226)	Relative risk 3.71 (1.87–7.38)	Serious#	Not serious	Not serious	Not serious	Not serious
Health-related quality of life	Change from pre-treatment	1 (78)	MD 0.7 (0.6–0.8)	Serious#	Not serious	Not serious	Serious¶	Not serious	Low	Yes
Length of hospital stay	Mean duration in days	3 (114)	MD −1.04 (−4.66–2.58), I2=76%	Serious#	Not serious	Not serious	Serious¶	Not serious	Low	No
ICU admission and ventilation	N	2 (83)	Relative risk 0.36 (0.06–2.17), I2=0%	Serious#	Not serious	Not serious	Serious¶	Not serious	Low	No
Future exacerbations	N with re-exacerbation at day 30	2 (55)	Relative risk 0.08 (0.00–1.28)	Serious#	Not serious	Not serious	Serious¶	Not serious	Very Low	Yes
	HR, 2 months follow-up	1 (40)	HR 0.169 (0.033–0.875)	Serious#	Not serious	Not serious	Serious¶	Not serious
	Time to first (days), 6 months follow-up	1 (38)	MD 27.70 (−4.55–59.95)	Serious#	Not serious	Not serious	Serious¶	Not serious
Mortality	N	4 (173)	Relative risk 0.73 (0.11–4.90), I2=16%	Serious#	Not serious	Not serious	Serious¶	Not serious	Low	No
SAEs	N	2 (109)	No SAEs were observed	Serious#	Not serious	Not serious	Serious¶	Not serious	Low	No risk

^#All studies included in this systematic review were deemed to be of at least some concern of methodological bias. ^¶Broad confidence intervals and/or insufficient overall study population. ⁺Marchioni et al. [31] reported significant improvement in dyspnoea; however, significance was lost in a meta-analysis that included a second, smaller randomised controlled trial. n: Number of trials; N: number of patients; HR: hazard ratio; ICU: intensive care unit; MD: mean difference; SAE: serious adverse event; SMD: standardised mean difference.

Treatment success, defined as resolution or significant clinical, with or without laboratory, improvement, was assessed in four trials totalling 383 participants. Mucolytics significantly improved treatment success evaluated upon treatment completion (relative risk 1.37, 95% CI 1.08–1.73, I²=63%, moderate certainty). Two of these studies, totalling 253 of the 383 participants, included in this meta-analysis were double blinded [31, 52]. Marchioni et al. [31] defined treatment success as resolution of symptoms as assessed by a physician, while Li et al. [40] required resolution or significant improvement of the symptoms, signs and laboratory findings. The definition of treatment success was not adequately described by Bisetti et al. [52] and Zhang et al. [41]. Two of the studies also assessed treatment failure (relative risk 0.81, 95% CI 0.21–3.06, I²=0%, 287 participants, low certainty) [31, 40].

Patient-reported symptoms were addressed in 16 studies involving 1741 patients by means of composite scores, or by assessing treatment effects or specific symptoms. Of these, five tested composite symptom scores, evaluating breathlessness, cough and sputum expectoration, with or without fever [31, 40, 45–49]. Heterogeneous instruments were used for assessing symptom scores and for this reason we conducted meta-analyses using SMDs. Mucolytics were associated with better post-treatment overall symptom scores compared to controls (SMD 0.86, 95% CI 0.63–1.09, I²=0%, three trials with 316 participants). One trial that assessed change from pre-treatment in composite symptom scores revealed significant improvement with mucolytics compared to control (SMD 0.80, 95% CI 0.22–1.38, one trial with 50 participants) [40]. Finally, Mohanty et al. [50] reported a significantly higher percentage of participants with improved symptoms compared to pre-treatment in the mucolytics group but not in the control group (relative risk 1.18, 95% CI 1.07–1.31, 200 participants). Early treatment response at day 4–5 was explored in one trial that favoured the mucolytics group (MD 1.50, 95% CI 0.85–2.15, 226 participants) [31]. Overall, we found moderate-certainty evidence suggesting that mucolytics improve overall symptom scores.

Eight studies explored the impact of mucolytics on breathlessness, mostly using nonvalidated scores. An RCT including 226 participants suggested that mucolytics can improve breathlessness compared to control [31], but this finding was not corroborated in other studies, which, however, looked at smaller study populations. We did not identify a meaningful difference in our meta-analysis, which was based on two trials including the aforementioned one (relative risk 1.36, 95% CI 0.55–3.38, I²=85%, 276 participants, very low certainty) [31, 40].

Cough as an outcome was captured mostly using nonvalidated scores in eight trials, six of which demonstrated potential benefit with mucolytics. A higher proportion of participants receiving mucolytics experienced post-treatment absence of cough compared to the control group (relative risk 1.93, 95% CI 1.15–3.23, I²= 29%, 276 participants, moderate certainty) [31, 40]. In addition, post-treatment ease of expectoration improved significantly with mucolytics versus control (relative risk 2.94, 95% CI 1.68–5.12, I²=0%, 149 participants, moderate certainty). Sputum viscosity was assessed using nonvalidated scores in four trials, three of which showed significant improvement for the mucolytic group.

Health-related quality of life was addressed using the mean change from pre-treatment score in the clinical COPD questionnaire in one study with 78 participants [39], which suggested the superiority of mucolytics compared to control, although between-group difference did not exceed the minimal clinically important difference (MD 0.7, 95% CI 0.6–0.8, low certainty).

Mucolytics were not found to reduce length of hospital stay in five of the six trials that reported this outcome, with median lengths ranging from 6 to 10.5 days in the intervention and from 5.5 to 10.2 days in the control groups. Three trials reported adequate data for inclusion in a meta-analysis, which did not reveal between-group differences (MD −1.04, 95% CI −4.66–2.58, I²=76%, 114 participants, low certainty). Indication for a higher level of care was noted in two studies as admission to the ICU and intubation [36, 53] with no significant difference between groups (relative risk 0.36, 95% CI 0.06–2.17, I²=0%, 83 participants, low certainty).

Future exacerbations were captured in three trials. Two small studies assessed patients with re-exacerbation at 30 days and they did not reveal between-group differences (relative risk 0.08, 95% CI 0.0–1.3). Two studies assessed time-to-next exacerbation. One concluded that mucolytics delay the next exacerbation (hazard ratio 0.169, 95% CI 0.03–0.88, 40 participants, 2 months follow-up) [46], with the second showing a similar trend (MD 27.70 days, 95% CI −4.55–59.95, 38 participants, 6 months follow-up) [53]. The latter also suggested that mucolytics may delay the next hospitalisation (MD 21 days, 95% CI −14.5–56.5), but future hospitalisation rate was not affected (MD 0.30, 95% CI −0.34–0.94, 38 participants), as confirmed by a second study [38] (relative risk 1.03, 95% CI 0.23–4.71, 59 participants). Overall, mucolytics may delay future exacerbations and hospitalisation (very low certainty).

Safety data were only scarcely reported across the included studies. Mortality was reported in four trials, which did not reveal between-group differences (relative risk 0.73, 95% CI 0.11–4.90, I²=16%, 173 participants, very low certainty). Serious adverse events were monitored in two trials (109 participants); they did not capture any events in either treatment arm, supporting the safety of mucolytics (low certainty). Four studies reported data on adverse events that we were able to analyse. They did not reveal between-group differences (relative risk 0.74, 95% CI 0.23–2.32, I²=0%, 272 participants, very low certainty), while other trials narratively reported balanced side effects across the study groups.

Arterial blood gas parameters were evaluated in seven studies. We were able to analyse data on the partial pressure of oxygen in arterial blood (P_aO2 in mmHg) from four trials [39, 42, 53, 55] that revealed a slightly higher P_aO2 in the mucolytics group (MD 3.21, 95% CI 1.51–4.92 mmHg, I²=38%, 201 participants). We found similar results for oxygen saturation (S_aO2 in %, MD 0.96, 95% CI 0.27–1.66, I²=0%, 123 participants). On the other hand, mucolytics did not appear to have any impact on the partial pressure of carbon dioxide (P_aCO2 in mmHg, MD −1.09, 95% CI −2.40–0.23, I²=37%, 201 participants).

The ERS COPD exacerbations core outcome set recommended assessing the impact of exacerbations on disease progression as a change in forced expiratory volume in 1 s (FEV₁) from baseline [24]. While none of the included trials assessed this outcome, 13 reported on lung function after exacerbation. Post-treatment FEV₁ did not differ between treatment arms, both when assessing percentage of the predicted values (MD 2.74, 95% CI −1.24–6.72, I²=81%, four studies, 193 participants) and absolute values in millilitres (MD 120.29, 95% CI −103.45–344.03, I²=81%, 156 participants). Five out of eight trials that did not report adequate data for inclusion in our meta-analysis did not suggest between-group differences either. While we were not able to pool data for forced vital capacity due to poor reporting, most trials suggested a limited treatment effect on this parameter.

A microbiological outcome was assessed in five trials. Out of 74 patients with a positive sputum culture on admission, 45 achieved a negative post-treatment result, with similar proportions across treatment groups (relative risk 1.65, 95% CI 0.69–3.98, I²=68%).

Activities of daily living, worsening of symptoms after initial treatment, development of pneumonia and treatment adherence were not reported in any of the included studies.

Subgroup and sensitivity analyses

All planned subgroup analyses are presented in the online appendix. Moreover, table 3 summarises our findings for each mucolytic agent. While our subgroup analyses revealed some potential between-group differences, these analyses were based on very small study populations per subgroup, thus significantly limiting our confidence in the findings.

TABLE 3.

Summary of evidence per mucolytic, per outcome

Mucolytics/outcomes	N-Acetylcysteine	Ambroxol	Erdosteine	Bromhexine	Hypertonic saline	Hyaluronan
Treatment success	NA	N=50 (+) N with resolution or significant improvement of symptoms and laboratory findings [40] N=80 (+) N with clinical efficacy [41]	N=27 (+) N with clinical judgement of effectiveness [52] N=226 (+) N with resolution of symptoms [31]	NA	NA	NA
Patient reported	NA	30 (−) Number of dyspnoea and cough attacks [37]	226 (+) Global clinical assessment score [31] 200 (+) N with improvement [50]	NA	NA	NA
Overall symptom score	NA	50 (+) Chinese medical symptom score, mean change [40]	55 (+) Breathlessness–sputum–cough scale [46, 48]	NA	NA
Breathlessness	40 (−) N per severity [42] 88 (−) Likert score 0–7, nonvalidated [53, 54]	50 (−) Chinese medical symptom score [40]	226 (+) Nonvalidated score 0–3 [31] N with absence [31]	27 (−) N with improvement [44]	59 (−) Modified Borg scale [38]	NA
Cough	38 (−) Nonvalidated score 0–7 [53] 95 (+) Nonvalidated score 0–3 [29] 122 (+) N with improvement in intensity [51]	50 (+) N with severe cough affecting sleep, Chinese medical symptom score [40] 23 (−) Nonvalidated score 0–3 [33]	253 (+) Nonvalidated score 0–3 [31, 52] 226 (+) N with absence [31] (same study as above)	NA	NA	NA
Ease of expectoration	38 (−) Nonvalidated score 0–7 [53] 122 (+) N with improvement [51]	50 (−) Chinese medical symptom score [40] 23 (+) Nonvalidated score 0–3 [33]	253 (+) Nonvalidated score 0–3 [31, 52] 226 (+) N with absence [31] (same study as above)	27 (−) N with improvement [44]	NA	NA
Sputum viscosity	30 (+) Nonvalidated score 0–3 [35]	NA	27 (+) Centipoises, mean % change, compared to pre-treatment [52] 226 (+) Nonvalidated score 0–3 [31] N with fluid saliva [31]	27 (−) Centipoises and arbitrary units [44]	NA	NA
Health-related quality of life	78 (+) Clinical COPD questionnaire, mean change [39]
Hospitalisation duration	166 (−) [39, 53, 54]	NA	52 (−) [46, 47]	NA	NA	36 (+) [36]
Indication for higher level of care	42 (−) N with ICU admission and intubation [53]	NA	NA	NA	NA	41 (−) N needing invasive ventilation [36]
Future exacerbations	38 (−) Rate by 6 months [53] Time to first by 6 months [53]	NA	40 (+) N by day 30 [46] Hazard ratio by day 60 [46] 15 (−) Rate by day 30 [49] (+) Rate by day 60 [49]	NA	NA	NA
Mortality	42 (−) N [53]	NA	NA	31 (−) N [44]	59 (−) N up to 30 days [43]	41 (−) N [36]
Serious adverse events	50 (−) None [54]	NA	NA	NA	59 (−) None [43]	NA

Various measurement instruments were used for assessing some of the outcomes. Results are pooled per instrument and for each instrument we present the overall population in which it was tested and whether there was evidence (+) or no evidence (−) of a beneficial effect with mucolytics. ICU: intensive care unit; N: number of participants; NA: not assessed in any trial.

No study had a low risk of bias and it was therefore not feasible to perform any of the pre-specified sensitivity analyses. The fixed-effect model sensitivity analysis suggested a positive treatment impact of mucolytics on some additional outcomes, such as 1) post-treatment absence of breathlessness, 2) post-treatment P_aCO2, 3) post-treatment FEV₁ and 4) pathogen clearance in sputum culture. However, in view of the clinical and methodological heterogeneity identified, we believe our main analysis is more appropriate.

Discussion

This systematic review and meta-analysis, based on 24 RCTs looking at 2192 patients with a moderate or severe COPD exacerbation, demonstrated with moderate certainty that mucolytics increase the rate of treatment success by 37% and are associated with a clinically meaningful improvement in symptoms. Mucolytics also appear to reduce cough and ease sputum expectoration (moderate certainty). In addition, low or very low certainty evidence suggests that mucolytics may also reduce the risk of future exacerbations and improve health-related quality of life, but do not seem to have an impact on breathlessness, length of hospital stay, indication for higher level of care or serious adverse events. It should be highlighted that due to heterogeneity in the outcomes and the outcome measurement instruments used, as well as inadequate reporting of relevant outcome data by some trials, each of the meta-analyses presented in this report were informed by data from up to four trials.

Airway mucus hypersecretion presents a key treatable trait in COPD. In stable disease state, it is associated with lung function decline, impaired quality of life, risk of hospitalisation and mortality [57, 58]. During exacerbations, airway mucus hypersecretion aggravates expectoration difficulty, airway inflammation and obstruction, as well as bacterial adhesiveness, thus creating a vicious cycle [18, 59, 60]. By regulating the viscoelastic properties of mucus and facilitating sputum expectoration, mucolytics may resolve this vicious cycle, thus improving the outcomes of exacerbations.

Head-to-head comparisons of different mucolytics are only evaluated in a few small trials that do not allow us to draw confident conclusions [61–64]. Interestingly, in a double-blinded trial involving 426 patients with COPD exacerbation, Prabhu Shankar at al. [65] demonstrated the superiority of the combination of bromhexine and guaiphenesin, compared to the monocomponents, suggesting a potential role for combining mucolytics with different mechanisms of action.

Mucolytics for the management of stable COPD have been evaluated in 38 trials totalling 10 377 participants, which were synthesised in a Cochrane review by Poole et al. [19]. Firstly, consistent with our findings, their review did not reveal any meaningful difference in adverse events. Furthermore, moderate-certainty evidence suggested that maintenance treatment with mucolytics confers a modest benefit in preventing exacerbations, reducing the days of disability per month and, possibly, the rate of hospitalisation. Similar to our work, this meta-analysis was limited by the fact that most included RCTs did not specifically recruit patients suffering from the airway secretion of highly viscous or voluminous mucus. Therefore, these findings are also possibly weakened by the limited effect in patients lacking this trait.

A limitation of our work is that we were not able to access the full text of six older, potentially eligible studies. However, we were able to capture the main study characteristics and outcomes of three of these RCTs from their abstracts and we included them in our meta-analysis [34, 41, 50]. Overall, with 24 RCTs including almost 2200 patients with moderate or severe COPD exacerbations, this is a broad systematic review, aggregating the best currently available evidence, which could be used to drive clinical practice and future research. Indeed, previous systematic reviews considered up to 15 of the included studies [66, 67, 68]. In addition, we followed the rigorous methodologies recommended by Cochrane and the GRADE working group for synthesising the available evidence and assessing methodological quality at the level of the included studies and of the overall body of evidence.

Another potential limitation of our study is that we did not include any real-life evidence. However, considering the heterogeneity of COPD exacerbations and the established practice to administer mucolytics only to patients with expectoration difficulty or thick sputum, this data would be at a high risk of indication bias.

It has been postulated that mucolytics are particularly effective in patients with exacerbations characterised by hypersecretion of thick sputum that is difficult to expectorate. However, mucolytics have also been attributed other beneficial effects besides liquifying sputum, such as antioxidant or immunomodulatory effects [17]. Our meta-analysis was based on studies evaluating unselected patients with moderate or severe exacerbations; stronger effects might be anticipated among patients presenting with increased sputum volume and viscosity. We did not have adequate data to assess the impact of mucolytics on this group of patients. In fact, all our subgroup analyses were limited by the very small study populations per subgroup.

Our certainty of the overall evidence was moderate for the primary outcomes and moderate to very low for the remaining outcomes. Significant methodological limitations inherent to most of the included studies and considerable heterogeneity in the reported outcomes and outcome measurement instruments limited our availability to pool data from the different studies. In our review, we evaluated all critically important outcomes that have been prioritised in the recently published ERS COPD exacerbations core outcome set [24]. We note, however, that only a handful of these outcomes have been consistently adopted in the completed RCTs, with some studies not reporting on any at all. We would encourage the adoption of this core outcome set in future trials, with a view to increasing consistency in the selection of outcomes that matter most to patients and clinicians.

National and international clinical practice guidelines do not currently recommend mucolytics for the management of COPD exacerbations [69–72]. As a result, their uptake varies significantly between different countries, health systems and even clinicians working in the same department. As a result, observational studies report administration of mucolytics in 2.0–72.5% of all exacerbations [73, 74]. Our findings suggest that mucolytics should be considered for these patients.

Well-designed and adequately powered trials are warranted to further assess the impact of mucolytics on all COPD exacerbation outcomes that are critically important to patients and other stakeholders. Importantly, such trials will need to assess the effects of mucolytics in patients with exacerbations characterised by increased sputum volume or thickness. The MucAct is an important ongoing double-blind randomised controlled trial in the UK (EudraCT number: 2020–001949–39) that fulfils these characteristics and intends to assess hypertonic saline versus placebo in 860 patients with this trait. In addition, in view of the significant heterogeneity of COPD exacerbations that necessitates the introduction of precision medicine interventions, the DECODE-NET (DisEntangling Chronic Obstructive pulmonary Disease Exacerbations – an international clinical trials NETwork) aspires to launch a platform trial that will evaluate precision medicine interventions to address various treatable traits of COPD exacerbations [75].

In conclusion, our meta-analysis synthesised the best available evidence on the safety and clinical effectiveness of mucolytics for COPD exacerbations and concluded with moderate certainty that mucolytics appear to improve the treatment success rate and symptoms. Therefore, mucolytics could be considered for the management of patients with COPD exacerbations. Our work also revealed significant limitations of the available research evidence that warrant addressing in future adequately powered and well-conducted RCTs.

Questions for future research

Large and well-conducted RCTs are needed to further assess the clinical effectiveness of various mucolytics for unselected patients with COPD exacerbations, as well as those associated with hypersecretion of thick mucus. As a minimum, they should address the outcomes most important to patients and other stakeholders that are included in the ERS COPD exacerbations core outcome set.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material ERR-0141-2022.SUPPLEMENT ^{(2.5MB, pdf)}