Background
Cough both protects and clears the airway. Cough has three phases: breathing in (inspiration), closure of the glottis, and a forced expiratory effort. Chronic cough has a negative, far‐reaching impact on quality of life. Few effective medical treatments for individuals with unexplained (idiopathic/refractory) chronic cough (UCC) are known. For this group, current guidelines advocate the use of gabapentin. Speech and language therapy (SLT) has been considered as a non‐pharmacological option for managing UCC without the risks and side effects associated with pharmacological agents, and this review considers the evidence from randomised controlled trials (RCTs) evaluating the effectiveness of SLT in this context.

Objectives
To evaluate the effectiveness of speech and language therapy for treatment of people with unexplained (idiopathic/refractory) chronic cough.

Search methods
We searched the Cochrane Airways Trials Register, CENTRAL, MEDLINE, Embase, CINAHL, trials registries, and reference lists of included studies. Our most recent search was 8 February 2019.

Selection criteria
We included RCTs in which participants had a diagnosis of UCC having undergone a full diagnostic workup to exclude an underlying cause, as per published guidelines or local protocols, and where the intervention included speech and language therapy techniques for UCC.

Data collection and analysis
Two review authors independently screened the titles and abstracts of 94 records. Two clinical trials, represented in 10 study reports, met our predefined inclusion criteria. Two review authors independently assessed risk of bias for each study and extracted outcome data. We analysed dichotomous data as odds ratios (ORs), and continuous data as mean differences (MDs) or geometric mean differences. We used standard methods recommended by Cochrane. Our primary outcomes were health‐related quality of life (HRQoL) and serious adverse events (SAEs).

Main results
We found two studies involving 162 adults that met our inclusion criteria. Neither of the two studies included children. The duration of treatment and length of sessions varied between studies from four sessions delivered weekly, to four sessions over two months. Similarly, length of sessions varied slightly from one 60‐minute session and three 45‐minute sessions to four 30‐minute sessions. The control interventions were healthy lifestyle advice in both studies.

One study contributed HRQoL data, using the Leicester Cough Questionnaire (LCQ), and we judged the quality of the evidence to be low using the GRADE approach. Data were reported as between‐group difference from baseline to four weeks (MD 1.53, 95% confidence interval (CI) 0.21 to 2.85; participants = 71), revealing a statistically significant benefit for people receiving a physiotherapy and speech and language therapy intervention (PSALTI) versus control. However, the difference between PSALTI and control was not observed between week four and three months. The same study provided information on SAEs, and there were no SAEs in either the PSALTI or control arms. Using the GRADE approach we judged the quality of evidence for this outcome to be low.

Data were also available for our prespecified secondary outcomes. In each case data were provided by only one study, therefore there were no opportunities for aggregation; we judged the quality of this evidence to be low for each outcome. A significant difference favouring therapy was demonstrated for: objective cough counts (ratio for mean coughs per hour on treatment was 59% (95% CI 37% to 95%) relative to control; participants = 71); symptom score (MD 9.80, 95% CI 4.50 to 15.10; participants = 87); and clinical improvement as defined by trialists (OR 48.13, 95% CI 13.53 to 171.25; participants = 87). There was no significant difference between therapy and control regarding subjective measures of cough (MD on visual analogue scale of cough severity: −9.72, 95% CI −20.80 to 1.36; participants = 71) and cough reflex sensitivity (capsaicin concentration to induce five coughs: 1.11 (95% CI 0.80 to 1.54; participants = 49) times higher on treatment than on control). One study reported data on adverse events, and there were no adverse events reported in either the therapy or control arms of the study.

Authors' conclusions
The paucity of data in this review highlights the need for more controlled trial data examining the efficacy of SLT interventions in the management of UCC. Although a large number of studies were found in the initial search as per protocol, we could include only two studies in the review. In addition, this review highlights that endpoints vary between published studies.

The improvements in HRQoL (LCQ) and reduction in 24‐hour cough frequency seen with the PSALTI intervention were statistically significant but short‐lived, with the between‐group difference lasting up to four weeks only. Further studies are required to replicate these findings and to investigate the effects of SLT interventions over time. It is clear that SLT interventions vary between studies. Further research is needed to understand which aspects of SLT interventions are most effective in reducing cough (both objective cough frequency and subjective measures of cough) and improving HRQoL. We consider these endpoints to be clinically important. It is also important for future studies to report information on adverse events.

Because of the paucity of data, we can draw no robust conclusions regarding the efficacy of SLT interventions for improving outcomes in unexplained chronic cough. Our review identifies the need for further high‐quality research, with comparable endpoints to inform robust conclusions.