Age‐Dependent Role of Cough Hypersensitivity in Transition to Refractory Chronic Cough

Summary

• During a one‐year follow‐up, approximately 20% of patients with aetiology‐explained chronic cough transitioned to refractory cough.

• Cough hypersensitivity showed an age‐dependent association with transition to refractory chronic cough.

To the Editor,

The natural history of chronic cough remains incompletely understood. Chronic cough initially attributable to aetiologies—such as asthma, upper airway cough syndrome (UACS), or gastroesophageal reflux disease (GERD)—sometimes evolves into refractory chronic cough (RCC). However, factors associated with such progression remain poorly defined. Therefore, we sought to describe the transition to RCC and explore factors associated with subsequent transition.

We analysed patients from the Korean Chronic Cough Registry who were newly referred with chronic cough (> 8 weeks). The registry protocol has been previously reported [1]. At the first follow‐up visit, patients were classified as having cough attributed to specific aetiologies based on clinical response to diagnostic evaluation and cause‐specific therapy, in accordance with clinical guidelines [2, 3]. RCC was defined as a cough refractory to aetiology‐targeted management or a cough without an identifiable aetiology [3]. Participants were assessed at scheduled visits (1, 6, 12, 24, 36, 48, and 60 months), with diagnosis re‐evaluated at each visit [1]. Transition was defined as physician‐determined reclassification from aetiology‐attributed chronic cough to RCC at any subsequent visit.

Logistic regression models evaluated associations with transition to RCC. Quadratic age terms were used to assess non‐linear age‐related patterns, and age‐by‐patient‐reported outcome interactions were included to assess effect modification by age. All participants provided written informed consent, and the study was approved by institutional review boards at participating centres.

Among 1,097 registry participants, 845 were newly referred with chronic cough, and 501 completed at least two follow‐up visits. Of these, 245 had aetiology‐attributed chronic cough at the first follow‐up visit and comprised the analytic cohort. They were predominantly female (66.1%), with a median age of 54 years (interquartile range [IQR], 41–65), and a median follow‐up duration of 357 days (IQR, 202–524). Of these, 53 (21.6%) transitioned to RCC over time. Among transitioners, the median time to transition was 262 days (IQR, 176–371), with continued follow‐up for a median of 380 days (IQR, 241–680). Patients who did not transition to RCC (n = 192) were followed for a median of 379 days (IQR, 212–680) until their last available visit (right‐censored).

At baseline, patients who transitioned to RCC were generally similar to those who did not, including cough aetiology, with asthmatic cough being the most common aetiology in both groups (58.3% vs. 56.6%), followed by GERD (17.2% vs. 17.0%) and UACS (13.5% vs. 9.4%). Baseline Cough Hypersensitivity Questionnaire (CHQ) scores were also identical (median 8.0 [IQR, 6.0–12.0] vs. 8.0 [6.0–12.0]; p = 0.865). However, patients who transitioned had significantly lower baseline Leicester Cough Questionnaire (LCQ) scores (9.8 [IQR, 7.9–12.4] vs. 11.6 [9.2–14.3]; p = 0.012) and higher cough severity visual analogue scale (VAS) (70.0 [IQR, 50.0–80.0] vs. 60.0 [40.0–70.0]; p = 0.060). Transitioners also received more codeine‐based antitussives than non‐transitioners. Detailed baseline characteristics, including prescribed medications, are provided in Table S1 at the Open Science Framework repository (https://doi.org/10.17605/OSF.IO/N5WUM).

RCC transition rates showed an inverted U‐shaped pattern across age groups (Figure 1A). Age was not linearly associated with transition (odds ratio [OR] 1.01/year, 95% confidence intervals [95% CI] 0.99–1.04; p = 0.218), but a quadratic age term improved model fit (likelihood ratio test; p = 0.028), with transition probability peaking around 55 years (28.3%).

FIGURE 1.

Age‐specific transition patterns and longitudinal trajectories of patient‐reported outcomes in chronic cough. (A) Age‐specific transition rates to refractory chronic cough. Red circles indicate observed rates within 10‐year age groups, with circle size proportional to group size. The solid blue line shows predicted probabilities from a quadratic logistic regression model [logit(P) = β₀ + β₁·Age + β₂·Age²; β₁ = 0.201, β₂ = −0.00182], and the shaded area indicates the 95% confidence interval. (B) Age‐specific odds ratios for cough‐related patient‐reported outcomes in relation to transition to refractory chronic cough. Odds ratios per 1‐point increase are shown for the Cough Hypersensitivity Questionnaire, Leicester Cough Questionnaire, and cough severity visual analogue scale. Estimates were derived from logistic regression models including each score, age, and their interaction term, with centred variables. Shaded areas indicate 95% confidence intervals. (C) Mean CHQ scores (±95% confidence intervals) are shown at three time points: Baseline (enrollment), the first follow‐up visit at which cough was classified as explainable by a specific aetiology, and the endpoint visit (transition to RCC or last available follow‐up for non‐transitioners). Of the 245 patients in the analytic cohort, 164 (66.9%) had CHQ data available at all three time points and are included in this analysis. CHQ, Cough Hypersensitivity Questionnaire; LCQ, Leicester Cough Questionnaire; OR, odds ratio; RCC, refractory chronic cough; VAS, visual analogue scale.

We hypothesised that cough hypersensitivity may contribute to this age‐related RCC transition because cough hypersensitivity is the key feature of RCC [3, 4] and has been linked to the middle‐age predominance in chronic cough [5]. We used the CHQ score as a proxy of cough hypersensitivity [6].

Although baseline CHQ score alone showed no association with RCC transition, the CHQ × age interaction was significant in the unadjusted model (OR 1.066 per 1‐point CHQ × 10‐year age increase, 95% CI 1.004–1.132; p = 0.035), with a crossover age of 50 years and higher CHQ‐associated risk at older ages (Figure 1B). The effect size remained similar after adjustment for sex and baseline LCQ, although statistical significance was attenuated (OR 1.056, 95% CI 0.991–1.124; p = 0.091). No significant age interaction was observed for LCQ or VAS (p = 0.817 and 0.721). Lower LCQ predicted transition (OR 0.90 per 1‐point increase; p = 0.022), whereas VAS showed only a borderline association (OR 1.013 per 1‐point increase; p = 0.053), with no age‐dependent variation.

The relatively short median time to transition (262 days), combined with the age‐dependent association between CHQ and RCC transition, suggests a possible “unmasking” phenomenon. In some older patients, cough hypersensitivity less responsive to aetiology‐targeted treatment may already coexist at presentation and become more apparent only after the aetiology has been controlled. Longitudinal CHQ trajectories support this interpretation. CHQ scores were similar between groups at baseline and at the aetiology‐explained cough visit (p = 0.758 and 0.965, respectively) but subsequently diverged. Non‐transitioners showed a further decline in CHQ score (Δ = −0.9, p = 0.018), whereas patients who later transitioned to RCC showed a significant re‐increase (Δ = +1.5, p = 0.017), with a significant difference over time (p = 0.002) (Figure 1C).

Several limitations should be noted. First, the small number of transition events limited multivariable analysis. Second, diagnoses of cough aetiologies and their transition to RCC relied on physician judgement. Given the possibility of spontaneous improvement or placebo effects [7], aetiological diagnoses may have been overestimated [7]. However, these assessments were made in accordance with clinical guidelines [2, 3]. Accordingly, our findings should be interpreted as real‐world observations. Third, as the registry did not enforce fixed visit intervals, the exact timing of transition carries possible imprecision inherent to interval‐censored observation. Finally, the CHQ is a self‐reported questionnaire rather than an objective measure of cough reflex sensitivity [8]. These findings should therefore be regarded as hypothesis‐generating and require confirmation.

In conclusion, among patients initially diagnosed with aetiology‐explained chronic cough, approximately 20% transitioned to RCC. CHQ scores showed an age‐dependent association with RCC transition, a pattern not observed for cough severity. Further studies are warranted to determine whether treatment targeting both cough aetiologies and cough hypersensitivity could lead to better outcomes in older patients.

Author Contributions

W.‐J.S. conceived and supervised the study. S.‐Y.K. and W.‐J.S. analysed and interpreted the data. S.‐Y.K. and W.‐J.S. drafted the manuscript. All authors contributed to patient recruitment, data acquisition, or data interpretation, critically revised the manuscript, approved the final version, and agree to be accountable for all aspects of the work.

Funding

The study was supported in part by a research grant from the Investigator‐Initiated Studies Program of Merck Sharp & Dohme Corp. The grantor had no role in the study design, data collection, analysis, result interpretation, publication decision, or manuscript preparation. The opinions expressed in this paper are entirely those of the authors.

Ethics Statement

The study was approved by the local Ethics Committees of participating centres.

Consent

All participants provided signed informed consent to participate in the study.

Conflicts of Interest

W.J.S. declares grants from Merck Sharp & Dohme Corp. and AstraZeneca, consulting fees from Merck, Bellus, AstraZeneca, Shionogi, and G.S.K., and lecture fees from Immunotek/Thermo Fisher, Merck, AstraZeneca, G.S.K., Sanofi, and Novartis. Other authors have none to declare.

Supporting information

Table S1: Baseline Characteristics of Study Participants.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.