Introduction

Interstitial lung disease (ILD) encompasses a heterogenous group of respiratory conditions characterised by inflammation and/or fibrosis of the lung parenchyma [1, 2]. Idiopathic pulmonary fibrosis (IPF), the most common and well-studied type of ILD, is associated with progressive lung function decline and poor prognosis [3]. Several other types of ILD, including hypersensitivity pneumonitis (HP), sarcoidosis and connective tissue disease-associated ILD (CTD-ILD) such as systemic sclerosis associated-ILD (SSc-ILD) and rheumatoid arthritis-associated ILD (RA-ILD), carry a risk of developing a similar progressive phenotype, referred to as ‘progressive pulmonary fibrosis’ or ‘progressive-fibrosing ILD’ (PPF/PF-ILD) [2, 4].

Among patients with ILD, cough is a prevalent symptom and can sometimes manifest as the initial sign of the disease [5,6,7]. Cough tends to persist over time in ILD populations, with chronic cough reported in 50─90% of patients with IPF [8, 9]. The pathophysiology of cough in ILD is considered multifactorial, involving mechanisms such as mechanical distortion, heightened cough sensitivity reflex, increased mucus production, the presence of inflammatory mediators and the influence of comorbidities such as gastroesophageal reflux disease, asthma, non-asthmatic eosinophilic bronchitis and obstructive sleep apnoea [6, 10,11,12,13,14]. It has been suggested that cough may contribute to a profibrotic feedback loop that drives disease progression in ILD [6, 9]. While there is some evidence that cough is an independent predictor of prognosis in IPF [15], findings have been mixed [16].

Although advances have been made in the treatment of IPF and other ILDs [17, 18], the lack of specific treatments for cough remains a significant concern for patients, with up to one-third of patients with ILD ranking cough as their worst symptom [7]. To inform future treatment and care models, it is vital to understand the breadth and magnitude of the burden of cough in these populations.

While the wide-ranging and multifaceted impact of cough has been explored in various populations [19,20,21,22], there has been no previous systematic synthesis of literature focussed on the impact in IPF and other ILDs. The objective of this review is therefore to bridge this knowledge gap by providing the first systematic evidence synthesis of the full spectrum of humanistic and economic impact related to cough in IPF and other ILDs. Through this synthesis, we aim to shed light on the multidimensional impact of cough, identify any gaps in existing literature, discuss considerations for future research and ultimately help pave the way for the development of effective treatment and support strategies for cough in IPF/ILD.

Methods

Search strategy

The protocol for this systematic review was prospectively registered with PROSPERO (CRD42022369379). Search strategies were developed following the guidance provided by the Centre for Reviews and Dissemination (CRD) and the Cochrane Handbook. Electronic databases, including Embase (OVID), MEDLINE (OVID), PubMed, Europe PMC, Cochrane Database of Systematic Reviews (CDSR), Cochrane Central Register of Controlled Trials (CENTRAL) and NHS Economic Evaluation Database (NHS EED), were initially searched on 31 August 2022, to identify studies on the burden of cough in IPF and other ILDs. Update searches were conducted on 3–4 October 2023.

Searches included combinations of relevant indexing and free-text terms. Terms were adapted to meet the syntax requirements of each database, and Boolean operators were employed to combine concepts effectively. Searches were limited to articles published between January 2010 and October 2023. The full search strategy, including the search terms and syntax used for each database, is provided in the Supplementary Methods within Additional file 1.

Database searches were supplemented with hand searching and berry-picking techniques, including free-text searches using Google Scholar, and a review of the reference lists of included studies.

Study selection

Studies in patients with IPF and other ILDs reporting cough characteristics and relevant humanistic or economic outcomes were eligible for inclusion. Non-human studies, case studies (based on study design), letters, editorials, commentaries and studies that did not report relevant outcomes were excluded. Two independent reviewers used Rayyan software to screen the retrieved articles based on the predefined inclusion and exclusion criteria shown in Table 1. Initially, titles and abstracts were reviewed within the software to assess their eligibility for inclusion. Subsequently, the full texts of potentially eligible articles were assessed for final inclusion. Discrepancies between reviewers were resolved through discussion with a third reviewer until consensus was reached. For articles reviewed in full, reasons for exclusion were recorded.

Table 1 Inclusion/exclusion criteria
Full size table

Data extraction and synthesis

For included articles, data extraction was performed by a primary reviewer using a standardised data-extraction template. The extracted data included article details, study design, participant demographics, cough characteristics, intervention details, control/comparison groups, outcome measures and relevant findings. A narrative synthesis approach was employed to summarise and interpret the findings from the included studies. This involved the use of text and tables to provide a comprehensive summary of the data and identify key trends.

Included studies were stratified by the types of ILD they examined and their design. Additionally, to allow for an analysis of heterogeneity and comparison of outcomes across different study designs and populations, studies were also grouped by whether chronic cough was specified, the prevalence of cough and whether cough was the major focus of the study:

  • Chronic cough: Studies in patients with chronic cough (definition not standard across studies).

  • Majority cough population: Studies where the majority of patients reported cough (not further specified).

  • Minority cough population: Studies where < 50% of patients reported cough (not further specified) or the prevalence of cough was not reported.

  • Broader includes: Studies where the prevalence of cough was not reported and cough was not the focus, but cough outcome measures were collected.

The risk of bias was assessed by two independent reviewers using design-specific appraisal tools. These included the appraisal tool for cross-sectional studies (AXIS) [23], the Cochrane risk-of-bias tool 2 for randomised controlled trials (RCTs) [24], the Critical Appraisal Skills Programme for qualitative studies [25] and the consensus-based standards for the selection of health measurement instruments for studies validating patient-reported outcomes [26].

Results

The searches yielded 7,439 articles, with 7,364 identified through electronic database searches and 75 through free-text and hand-searching. After screening, 261 unique articles remained and were subject to full-text assessment, with 65 eligible articles, based on 61 unique studies, ultimately included in the synthesis. The study selection process is illustrated in Fig. 1.

Fig. 1
figure 1

Flow of articles through the different phases of the systematic synthesis

Full size image

Among the unique studies, there were 33 focussed on patients with IPF, 18 on patients with mixed ILDs (various types of ILD, including CTD-ILDs in some cases), six on patients with CTD-ILD, and four on patients with sarcoidosis. The number of unique studies by disease group and study design is illustrated in Fig. 2. For each disease group, a comprehensive overview of the study designs, patient characteristics and outcomes for the included studies is available in the Supplementary Results within Additional file 2. Key findings are summarised in the subsections below.

Fig. 2
figure 2

Overview of included studies by disease group and study design*

*Totals were calculated based on the number of unique studies rather than articles. For IPF, two articles were based on the same trial (NCT00600028), counted as a single trial and two articles were derived from the same cohort/registry study (PROOF), counted as one observational study. Similarly, for CTD-ILD, three articles were based on the same trial, counted as a single trial (NCT00883129).

CTD-ILD, connective tissue disease-associated ILD; ILD, interstitial lung disease; IPF, idiopathic pulmonary fibrosis; PS, pulmonary sarcoidosis; RCT, randomised controlled trial

Full size image

Based on design-specific appraisal tools, 29 studies were rated as having a low risk of bias and 28 as having a medium risk. Four studies were not assessed for bias due to the lack of suitable tools to assess their design. An overview of the bias assessment for each study is provided in the Supplementary Results within Additional file 2.

While treatment effects are not the focus of this review and we do not discuss impact of treatments on cough outcomes, a proportion of the studies were investigating pharmacological interventions including antibiotics, antifibrotics, corticosteroids, opioids, anti-acid medication, immunosuppressants, neuromodulators, sodium cromoglycate and thalidomide.

Studies in IPF

IPF: study characteristics and measures

Among the 33 studies in patients with IPF, there were nine interventional trials [27,28,29,30,31,32,33,34,35], 15 observational studies [8, 36,37,38,39,40,41,42,43,44,45,46,47,48,49], two validation studies [50, 51], three mixed-methods studies [52,53,54] and four qualitative studies [55,56,57,58]. An overview of the quantitative and mixed-methods studies is provided in Table 2.

Table 2 Quantitative and mixed-methods studies in IPF
Full size table

Eight of the studies were classed as chronic cough studies, comprising seven trials [27,28,29,30,31,32,33] and one observational study [36]. These studies used various definitions for chronic cough, including history of cough (with or without exertional dyspnoea) [28], self-reported chronic cough [30] and cough for > 8 weeks [27, 29, 30, 59]. Some of the trials required further criteria, such as stable cough frequency for > 4 weeks [30], cough affecting daily life or quality of life [29, 33, 59], refractory cough [27, 33], 24-hour cough count of > 10/15 coughs per/hour [27, 31], and/or visual analogue scale (VAS) for cough severity > 40 mm [27, 30, 31].

Across the studies in IPF, a wide range of cough, health-related quality of life (HRQoL) and other impact measures were used (Table 2). The most frequently used cough measures were cough severity VAS (N studies = 11 [27, 30, 31, 33, 34, 37,38,39,40, 45, 59]; N patients = 1543; reported as mean score range 30.5–73.7, reported as median score range 32–40.3; possible score range 0–100 [no cough–worst cough]) and objective cough counts (N studies = 5 [27, 28, 30, 31, 40]; N patients = 186; mean count/hr range 8.9–48.0). Apart from one observational study in a majority cough population [40], the only studies that used objective cough counts were interventional trials in patients with chronic cough. Outside of interventional trials, cough was often assessed using cough items or domain scores from other HRQoL or impact measures. The St. George’s Respiratory Questionnaire (SGRQ) was the most frequently used HRQoL measure (N studies = 14 [32, 33, 35, 37, 39, 42, 44,45,46,47,48, 51, 52, 59], N patients = 3988; mean score range 29.09–57.4; possible score range 0–100 [lower–greater impact on HRQoL]), followed by the Leicester Cough Questionnaire (LCQ; N studies = 14 [8, 27, 28, 31,32,33,34,35,36, 38, 40, 42,43,44], N patients = 1532; reported as mean score range 11.0–16.16, reported as median score range 14.8–18.2; possible score range 3.0–21.0 [lower scores indicate worse cough specific quality of life]).

IPF: comparisons with other populations

Six studies compared cough severity, HRQoL and/or other impact measures in patients with IPF with other populations ( [27, 36, 38, 40, 48, 54], Table 2; Fig. 3). A trial investigating the use of PA101 (sodium cromoglicate) found that patients with IPF and chronic cough had a higher baseline objective cough frequency than patients with chronic idiopathic cough (CIC), but better VAS and LCQ scores [27]. PA101 reduced objective daytime cough counts in patients with IPF but not in patients with CIC, indicating a potentially different mechanism of cough [27]. However, an observational study comparing patients with IPF and chronic cough and patients with CIC found no between-group differences in LCQ scores or other features of cough, including cough responses to paint or fumes, cough-related sleep disturbance or self-reported cough frequency [36]. An observational study in a majority-cough population similarly found that objective cough rates in patients with IPF were comparable with other patients with chronic cough and higher than previously published rates in healthy controls and patients with asthma [40]. Based on worse VAS, LCQ and/or Cough and Sputum Assessment Questionnaire (CASA-Q) scores, other studies found that patients with IPF had more severe cough compared with healthy volunteers, patients with combined pulmonary fibrosis and emphysema, and previously published values in patients with COPD and chronic bronchitis [38, 48, 54].

Fig. 3
figure 3

Overview of studies assessing cough severity and/or HRQoL measures in (a) IPF and (b) ILD in comparison with other populations

Full size image

IPF: association of cough with impact/HRQoL and economic measures

Nine studies investigated concurrent/baseline associations between cough, HRQoL and/or other impact measures in patients with IPF, with all of these reporting at least one association ( [27, 33, 37, 38, 40,41,42, 53, 59], Table 2). These included two interventional trials in patients with IPF and chronic cough where cough severity VAS was correlated with Cough Quality of Life Questionnaire (CQLQ) (N patients = 23, R = 0.63), LCQ (N patients = 20, R range = -0.42) or SGRQ scores (N patients = 20, R = 0.42) [33, 59]. While one trial found an association between cough severity VAS and objective cough counts (N patients = 24, R = 0.683), another found no association between cough severity VAS and objective polygraphy-derived indices of cough [27, 33]. In observational studies in majority cough populations, cough severity VAS was similarly associated with LCQ scores (N patients = 27, Rho = − 0.72) and objective cough counts (N patients = 19, R = 0.80) and found to be an independent predictor of SGRQ scores (N patients = 516, β = 0.20) [37, 38, 40]. Correlations were also found between ER-S cough domain scores and EuroQol EQ-5D indices of pain/discomfort and anxiety/depression (N patients = 168, R range = 0.23–0.26) and between LCQ scores and Beck’s Depression Inventory (BDI) scores (N patients = 98, R = − 0.57) [42, 53].

Three studies reported longitudinal associations between cough, HRQoL and/or other impact measures: the trial of PA101 in patients with IPF and chronic cough, which found associations between change in cough severity VAS and change in daytime cough frequency (N patients = 23, R = 0.415) [27]; a post hoc trial analysis that found correlations between change in CASA-Q and SGRQ domain scores (N patients = 1061, R range = − 0.29 to − 0.45) [51]; and an observational study in the ‘broader includes’ category which found no association between change in LCQ scores and physical activity decline over 12 months (N patients = 54) [44].

IPF: economic burden of cough

No studies on the economic burden of cough in IPF were identified.

IPF: experiences of cough

Four qualitative studies in IPF used focus groups or interviews to gain patient and/or caregiver perspectives on symptoms and their burden [55,56,57,58]. Across these four studies (N patients = 83), cough emerged as one of the most troubling symptoms, with wide-ranging impact. Patients described relentless coughing throughout the day, leading to feelings of ‘exhaustion’ [56,57,58]. Although some patients felt that oxygen therapy helped alleviate their cough, coughing was described as ‘incredibly bothersome’ during the morning and evening, often occurred during exertion, and caused sleep disruption and incontinence [53, 54, 56,57,58]. Caregivers described feelings of distress, being perpetually vigilant and witnessing ‘draining coughing fits’ with a sense of helplessness [56, 58].

Studies in patients with various ILDs and ILDs other than IPF

ILD: study characteristics and measures

Among the 18 studies in mixed ILD populations, there were two interventional trials [60, 61], seven observational studies [7, 62,63,64,65,66,67], four validation studies [68,69,70,71], four qualitative studies [72,73,74,75]. and one mixed-methods study [76]. Of the six studies in CTD-ILDs, there were three interventional trials in SSc-ILD [77,78,79], one observational study in a mixed CTD-ILD population [80], one validation study in SSc-ILD [81], and one mixed-methods study in a mixed CTD-ILD population [82]. Of the four studies for sarcoidosis, there was one interventional trial [83] and three observational studies [84,85,86]. Tables 3, 4 and 5 respectively provide an overview of the quantitative and mixed-methods studies in mixed ILDs, CTD-ILDs and sarcoidosis.

Table 3 Quantitative and mixed-methods studies in mixed ILD
Full size table
Table 4 Quantitative and mixed-methods studies in CTD-ILD
Full size table
Table 5 Quantitative and mixed-methods studies in sarcoidosis
Full size table

Of the studies in mixed ILDs, four were classed as chronic cough studies, compared with none in CTD-ILDs and one in sarcoidosis. Across these studies, definitions for chronic cough included self-reported chronic cough [62, 83] cough for > 8 weeks [7, 72], and cough intensity and frequency VAS > 10 mm [61].

Similar to the studies in IPF, studies in ILD used a range of measures to assess cough (Tables 3, 4 and 5). While cough severity VAS was the most frequently used cough measure in mixed ILDs (N studies = 9 [7, 60,61,62,63, 65, 67, 70, 72], N patients = 473; reported as mean score range 25–52.3, reported as median score range 18–63) and sarcoidosis (N studies = 3 [83,84,85], N patients = 393; mean score range 25–53), VAS was not used to assess cough in CTD-ILDs; instead, most CTD-ILD studies relied on cough items or domain scores from other measures, including the SGRQ and the LCQ. Objective cough counts were absent in studies in mixed ILDs and CTD-ILDs and were only employed in two studies in sarcoidosis (N patients = 93 [83, 84], mean count/hr range 6–244). The most frequently used HRQoL measure was the SGRQ (mixed ILD: N studies = 7 [62,63,64, 68, 69, 71, 76], N patients = 693, mean score range 32.9–78.65; CTD-ILD: N studies = 4 [78,79,80,81], N patients = 829, reported as mean score range 27.1–43.9, reported as median score range 27.1–79.4; sarcoidosis: 0 studies) followed by the LCQ (mixed ILD: N studies = 5 [7, 61, 64, 65, 68], N patients = 528, reported as mean score range 14.9–16.7, reported as median score range 18.7–19.6; CTD-ILD: N studies = 3 [78, 80, 81], N patients = 254, reported as mean score range 16.7–17.5, reported as median score 18.5; sarcoidosis: N studies = 4 [83,84,85,86], N patients = 668, mean score range 14–17.5).

ILD: comparisons with other populations

Only one study comparing cough in various types of ILD with other populations was identified; it found cough severity to be higher in patients with ILD than in those with COPD, as measured by Edmonton symptom assessment scores [66]. Seven studies in mixed ILD populations compared cough and/or impact/HRQoL measures among various ILD types ( [7, 62,63,64,65, 69, 70], Tables 3, 4 and 5). Four of these studies found the prevalence and/or severity of cough to be higher in patients with IPF or idiopathic interstitial pneumonias (IIPs) [7, 62, 63, 65], two reported worse HRQoL in IPF (particularly in cough-related domains) [63, 69], and one observed greater increase in cough severity over time in IPF [63]. Conversely, one study observed no differences in cough severity between IPF and non-IPF groups [70] and in another HRQoL was lower in CTD-ILD than IIPs (including IPF) [64]. Additionally, two studies comparing patients with IIPs – including IPF – with patients with CTD-ILDs, found no between-group differences in LCQ scores [64, 65]. However, one of these studies reported higher VAS scores for cough intensity in patients with IIPs [65].

In CTDs specifically, one observational study compared cough measures in patients with RA and other CTDs with and without associated ILD [80]. The study found SGRQ scores to be worse in patients with RA-ILD compared with other CTD-ILDs, although there were no differences in LCQ scores. Further, while LCQ scores did not differ between patients with and without associated ILD, SGRQ outcomes were worse in those without associated ILD. In sarcoidosis, one observational study found objective cough counts and cough reflex sensitivity to be higher in patients than healthy controls [84].

ILD: association of cough with impact/HRQoL measures

In mixed-ILD populations, six studies investigated concurrent/baseline associations between cough, HRQoL and/or other impact measures [62,63,64,65, 68, 70], with five of these reporting at least one significant association ([62,63,64,65, 68], Tables 3, 4 and 5). These included an observational study in patients with chronic cough, which found cough severity VAS to be an independent predictor of SGRQ total and/or domain scores in patients with IPF (N patients = 77, R = 0.33–0.55) and SSc-ILD (N patients = 67, R = 0.34–0.51) but not in chronic HP (N patients = 32, R = -0.10–0.03) [62]. Conversely, while an observational study in a majority cough population with fibrotic ILDs observed a significant association between cough severity VAS and King’s Brief Interstitial Lung Disease (KBILD) scores (N patients = 35, R = − 0.54), no such association was observed with SGRQ scores [63]. In mixed ILD studies in the ‘broader includes’ category, one observational study and one validation study found LCQ scores to be correlated with total SGRQ scores (N patients = 139, β=–3.55; N patients = 55, R= − 0.70, respectively) [64, 68], with the observational study also finding a correlation with Short Form 36 (SF-36) physical component scores (N patients = 139, β = 0.55–1.34) [64].

In CTD-ILD, a trial in a majority cough population with SSc-ILD similarly observed a correlation between total LCQ and SF-36 physical component scores (N patients = 87, R = 0.258) [88]. Three trials in SSc-ILD also reported correlations between the presence of cough (frequency) and cough impacts (as measured by LCQ scores) and measures of disease severity (extent of fibrosis, cutaneous sclerosis and/or dyspnoea) [77, 79, 88]. However, a study validating the Patient-Reported Outcomes Measurement Information System 29 tool (PROMIS-29) in SSc-ILD found no correlations between PROMIS domains and LCQ social or psychological domain scores, although there was a correlation with LCQ physical domain scores (N patients = 73, R = 0.36) [81]. In sarcoidosis, three studies investigated and observed correlations between total LCQ scores and cough severity VAS scores (N patients = 355, R = − 0.83), objective cough counts (N patients = 32, R = − 0.61) and/or 15-dimension HRQoL scores (N patients = 275, β = 0.24) [84,85,86].

Only two mixed ILD studies investigated longitudinal associations between cough, HRQoL and/or other impact measures. These included an observational study where correlations were observed between changes in LCQ scores and SGRQ scores at 6 and 12 months (N patients = 147, R = − 0.56 to − 0.58) [64] and a validation study where changes in cough VAS scores at 3–6 months were found to correlate with KBILD total scores (N patients = 64, R = − 0.36), but not KBILD cough domain scores or measures of disease severity [70]. Further, the validation study found that cough VAS was unable to detect changes in cough symptoms over time, as assessed by KBILD cough domain scores [70]. In CTD-ILD, the only trial investigating longitudinal associations between cough and impact/HRQoL measures found correlations between 24-month changes in LCQ scores and SF-36 physical scores (N patients = 142, R = 0.54) and mental component scores (R = 0.45) as well as Health Assessment Questionnaire Disability Index scores (R = − 0.03) [88]. In sarcoidosis, a trial investigating longitudinal associations observed correlations between 3-month changes in objective cough counts and LCQ scores (N = 19, R = − 0.64) and King’s Sarcoidosis Questionnaire general health scores (R = − 0.59) [83], while an observational study reported a significant correlation in individual patients between changes in cough VAS scores and the LCQ between clinic visits (N = 891, R = not reported, p < 0.0001) [85]. Overall, this heterogeneity of results demonstrates that while cough may impact a patient’s HRQoL, available assessment tools are sensitive to different changes and do not always correlate. This means a broad range of assessments may be necessary to capture the multifaceted dimensions of cough until a standardised and disease-specific cough assessment tool can be developed and validated.

ILD: economic burden of cough

No studies specifically investigating the economic burden of cough were identified in ILD populations. However, one observational study in patients with fibrotic ILDs, including IPF, IIP and unclassifiable ILD, found an association between cough severity and workplace productivity loss [67]. Specifically, the study showed that the odds of productivity loss (N = 148) increased by 3% for every 1 mm increase in cough severity VAS, with an estimated annual cost of CAN$11,610 per employee.

ILD: experiences of cough

Four qualitative studies in mixed ILD populations (N patients = 63) explored symptoms and their impact on patients through interviews or focus groups [72,73,74,75]. Across these studies, patients identified cough as a significant symptom, describing it as ‘debilitating’, ‘hard to control’ and ‘embarrassing’ [72]. They reported that coughing not only led to fatigue but also hampered their daily activities and caused social discomfort and frustration/irritability [72,73,74,75]. In one of the studies, cough was also described as being especially disruptive to caregivers [73]. These experiences were mirrored in a mixed-methods study in patients with CTD-ILD, where cough was noted to negatively affect physical function, social participation, daily activities and sleep quality [82]. In sarcoidosis, no studies on patient experiences were identified.

Discussion

Our study is the first to systematically synthesise literature on the humanistic and economic impact related to cough in IPF and other ILDs. The findings confirm that cough is a pervasive and persistent symptom in many patients with IPF and other ILDs.

Quantitative studies in patients with IPF consistently demonstrated the detrimental effect of cough on HRQoL, while qualitative studies in this population highlighted the significant cough-related burden experienced by patients and their caregivers, including disruption of daily activities, sleep deprivation, fatigue, incontinence, social embarrassment and psychological distress.

Indications of similar impact were reported in patients with CTD-ILDs, sarcoidosis, and other ILDs, but data were more limited. While no dedicated studies on the economic burden or healthcare resource use associated with cough were found, one study in ILD indicated that cough severity significantly predicted workplace productivity loss [67]. Studies in other populations with chronic cough suggest substantial economic burden related to increased healthcare utilisation, challenges in the workplace, and cough-related comorbidities [20], underscoring the need for more research in this area in IPF and ILD.

Across disease groups and study designs (with the exception of CTD-ILD), cough severity VAS was the most frequently used measure of cough, despite its lack of validation in patients with IPF and ILD. Objective cough monitoring devices were rarely used outside of trials in patients with IPF and chronic cough. While score ranges were wide, studies enriched for patients with IPF and chronic cough reported the highest mean VAS scores (56.0–73.7 mm). Conversely, lower mean/median scores were generally observed in studies in patients with IPF where cough chronicity was not specifically reported (reported as mean 30.5 mm; reported as median 32.0–40.3 mm) and mixed ILD (reported as mean 25.0–52.3 mm; reported as median 18.0–31.0 mm) populations.

Beyond cough characteristics, differences in patient populations, disease severity and treatments may have contributed to the variability observed in cough severity and duration.

Notably, cough severity VAS scores were variable even within the different chronic cough cohorts, which may relate to between-study differences in definitions of chronic cough as well as variability in study designs, populations and the tools used to assess cough. While the 8-week minimum duration was the most frequently used criterion, the use of additional thresholds based on subjective and/or objective criteria varied considerably, mirroring challenges in other populations with chronic cough and highlighting the need for consensus-driven criteria to ensure consistency. Future studies should aim to establish uniform inclusion criteria and standardised assessment tools for chronic cough to enhance the comparability of results.

Irrespective of the presence of chronic cough, in most studies that directly compared patients with IPF with other ILDs, the prevalence and severity of cough tended to be greater in patients with IPF. Outside of IPF, there were few studies specifically focussed on cough burden and impact. In most cases, cough-related data were collected as secondary measures in larger studies, and there was substantial heterogeneity in the tools used. Various cough-related HRQoL tools, such as the LCQ, CQLQ and CASA-Q were used alongside broader ILD-specific HRQoL tools such as the KBILD, Living with Pulmonary Fibrosis (L-PF) questionnaire and the A Tool to Assess Quality of Life in Idiopathic Pulmonary Fibrosis (ATAQ-IPF) questionnaire. Studies also used more general HRQoL tools, such as the SGRQ, Functional Assessment of Chronic Illness Therapy (FACIT) and EQ-5D as well as measures focussed on psychological distress, including the BDI. Even when the same tool was used, there was variability in the versions and scales used as well as in the reported statistics, exacerbating discrepancies in the synthesised data. While the LCQ was the most frequently used impact measure for cough both in IPF and ILD, it was not specifically designed for these populations and may not capture the full impact of cough in these contexts.

The divergence in approaches may be related to the absence of a widely accepted and validated cough-specific measure for IPF and ILD. Development of a standardised and disease-specific cough assessment tool, alongside validation of existing tools in these populations, could facilitate more consistent and reliable between-study comparisons. In addition, establishing such tools could pave the way for future advancements in personalised medicine, with treatments guided by the severity of cough in IPF and ILD.

A notable strength of our study is its comprehensive search strategy, which allowed for in-depth examination of the full spectrum of impact related to cough not only in IPF but also other ILDs. The high level of heterogeneity in the included studies and relative low number of studies in ILDs are significant limitations, which posed challenges for bias assessment and direct comparisons. Differences in study designs, patient populations, disease severity and assessment tools also limit the generalisability of the results. Other limitations include the exclusion of letters, graphical abstracts, case series and articles published before 2010, which may have resulted in certain aspects of cough burden and impact being missed. Despite these challenges, our study offers valuable insights into the current state of research on the burden of cough in IPF and other ILDs, emphasising the importance of standardisation to advance knowledge in this area.

The lack of specific treatments and management for cough in these patients remains a significant unmet clinical and patient need. Evidence for impact of different therapeutic approaches on cough is limited, with many of the trials negative or requiring confirmation in larger studies [7, 89]. We hope by highlighting the impact and breadth and magnitude of the burden of cough in patients with IPF and ILD we can encourage further research.

Conclusions

In conclusion, studies consistently confirm a negative effect of cough on HRQoL in IPF, with indications of a similar impact in other ILDs, though less well studied. However, differences in definitions and assessment methods across studies hinder meaningful comparisons and there is a notable lack of research on the economic burden of cough in both IPF and other ILDs. Establishing standard measures for cough assessment in IPF and other ILDs is vital to enhance understanding of cough and inform future research, treatment and care models.