Find all editions of Cough Science News below and get access to the latest cough science developments, publications, and interviews with cough experts.
4 October 2026
Hello!
October cough research kept circling one question: how much of what a patient tells us about their cough shows up when we count it? Patient-reported and objective measures capture different things, and the studies below read best with that in mind.
Here's what caught my attention.
What the study found: In a prospective cohort of 149 adults at a Singapore tertiary multidisciplinary cough clinic, 111 had refractory chronic cough (RCC) and 38 had non-refractory chronic cough. RCC patients had lower Leicester Cough Questionnaire total scores (11.8 vs 13.7), more positive superior laryngeal nerve trigger points (36.4% vs 16.2%) and more abnormal videostroboscopy (58.7% vs 31.6%). After adjustment, only a lower laryngeal hypersensitivity (LHQ) irritation score remained independently associated with RCC (adjusted OR 0.71, 95% CI 0.52 to 0.97). Among RCC patients treated through a layered multidisciplinary pathway, 72.1% reached the LCQ minimal clinically important difference at some follow-up visit, at a median of 85 days.
Why it matters: Patients with refractory cough did report more throat irritation, and it was the only laryngeal measure that held up after adjustment, though not by a wide margin. The authors are careful not to oversell it, and rightly so: abnormal videostroboscopy turned up in about a third of the non-refractory patients too. I would also be cautious with the 72.1%. These patients got layered treatment with no control group and were followed with questionnaires only, so you can’t separate the treatment from placebo response or regression to the mean. I’d like to see these patients followed with cough counts too, to know whether the questionnaire gains show up in the cough itself.
What the study found: This meta-analysis pooled five randomized trials (719 participants) of aprepitant, orvepitant and serlopitant, covering refractory chronic cough, lung cancer cough and IPF cough. Against placebo, cough VAS fell by a mean 6.9 points more (95% CI 2.9 to 10.9) and cough-specific patient-reported scores improved (SMD 0.50, 95% CI 0.22 to 0.78). Awake objective cough frequency did not differ (4 studies, SMD 0.01, 95% CI -0.18 to 0.21). Non-severe adverse events were more common with treatment (RR 1.33) and severe events were not. Certainty of evidence was rated very low to low.
Why it matters: Patient-reported scores moved and the counts didn't. The authors offer two explanations: the drugs change how cough is perceived, or the counts were too noisy, and the placebo response too large, to show a real effect. They can't tell which, and pooling three different diseases means this isn't a clean class estimate either. To me it's one more argument for pairing patient-reported scores with continuous counts, which are better placed to settle the question.
What the study found: In 339 consecutive patients at a London specialist cough clinic (73% female, median age 57, median cough duration 4 years), 36 (11%) reported suicidal ideation on the PHQ-9, and only 11 of those 36 were being treated for depression. Moderate or worse depression was reported by 31% and anxiety by 20%. Suicidal ideation was associated with cough severity, cough-specific health status and depression and anxiety scores, but not with age, sex, cough duration or treatable traits. Cough frequency from a single 24-hour recording, available in 59 patients, was not associated with depression or anxiety scores. This was a single-center cross-sectional screen, and no suicide attempts were recorded over the following 12 months.
Why it matters: Mood scores went with how bad patients said their cough was, but not with how often they coughed. One possibility, which the authors raise, is that depression and anxiety change how a person experiences cough, so a questionnaire score is partly a mood score and counts are how you tell a real drop in cough from a drop in distress. The counts came from a single day in 59 patients, and our own data strongly suggest that one day can misrepresent a person. I'd like to see mood tracked alongside several days of counts.
Our October Cough Science Forum covered measurement bias in cough trials, with Joe Brew and Professor Giovanni Ferrara joining me. The recording is on demand at hyfe.com/cough-science-forum, where you can also register your interest in the next session. The forum now runs quarterly.
On hyfe.com, we work through the questions that come up repeatedly in conversations with sponsors and investigators. This month:
Why not just keep the audio and check the coughs ourselves?
Keeping the audio sounds rigorous. We explain why the rigor belongs in validation, and what retained audio costs in privacy and consent.
Can cough sit next to FVC in an IPF protocol?
IPF trials lean on FVC, and background antifibrotics shrink the decline it can detect. We look at what cough adds, from registry data to two positive trials that counted it objectively.
👉 See how the IPF case is built
Can cough tell you a COPD exacerbation is coming?
COPD exacerbation endpoints arrive late, often days after onset. We look at what cough does beforehand in published studies and our own interim data, and how far it can be trusted as a signal.
Is "mild, transient cough" enough for an inhaled program?
Inhaled programs expect cough but rarely describe it beyond a yes or no. We look at how much it varies with formulation and device, and what to measure before and after approval.
Until next month,
Peter Small
PS: If you find these monthly reviews useful, I'd appreciate you forwarding this to a colleague who might benefit. And if you'd like to continue the conversation, I'm always reachable at peter@hyfe.ai
