Clustering Patients With Chronic Cough Using Reported Sensations and Triggers: Results from the Triggers and Sensations Provoking Coughing Questionnaire.

BACKGROUND

Chronic cough (CC) is one of the most common symptoms reported to primary care and to respiratory outpatient clinics. The Triggers and Sensations Provoking Coughing (TOPIC) questionnaire is a 15-item questionnaire designed to capture sensations and triggers associated with CC in a fashion that discriminates between refractory CC (RCC) and other causes of CC.

RESEARCH QUESTIONS: Does a relationship exist between TOPIC questionnaire scores and other measures of cough in patients referred to 3 specialist cough clinics? Can patients with CC be clustered based on reported sensations and triggers as captured by the TOPIC questionnaire?

STUDY DESIGN AND

METHODS

This was a multicenter observational study. Data were collected including the Leicester Cough Questionnaire (LCQ) score, verbal cough severity, day and night cough visual analog scale score, 24-hour cough frequency, and TOPIC questionnaire results. Patterns of TOPIC questionnaire responses were explored using cluster analysis.

RESULTS

Baseline data were collected from 101 participants. Demographics of study participants were typical of patients with CC; 65% were female with a mean (SD) age of 59 (12.9) years. Baseline TOPIC questionnaire score correlated moderately negatively with baseline LCQ score (r = -0.59; P < .001), but not with other subjective or objective measures of cough. Hierarchical and K mean cluster analysis were used to group study participants into 4 distinct clusters based on triggers and sensations of cough as captured by the TOPIC questionnaire: (1) high sensations burden, (2) vocal triggers, (3) eating triggers, and (4) need to clear throat. These groups showed statistically significant differences in demographics and subjective measures of cough (P < .01).

INTERPRETATION: Patients with CC reported unpleasant somatic sensations and cough triggers captured by the TOPIC questionnaire, associated with lower LCQ scores and worse quality of life. Our results suggest that TOPIC questionnaire responses can be used to cluster patients into clinical phenotypes based on reported sensations and triggers.