Prevalence and associated factors of cough syncope in patients with chronic cough: a multicentre prospective study

Abstract

Background

Cough syncope is a clinically significant complication of coughing, yet its epidemiology in patients with chronic cough remains poorly characterised. This study aimed to estimate the prevalence, and identify associated factors for cough syncope among patients with chronic cough.

Methods

Data were analysed from the Korean Chronic Cough Registry, a multicentre prospective observational study. Adults with newly referred chronic cough were followed for 1 year. Cough syncope was defined based on patient self-report at enrolment and the 1-year follow-up visit. We compared clinical characteristics between patients with and without cough syncope and performed multivariable logistic regression to identify independent associated factors.

Results

Among 682 eligible participants, cough syncope was identified in 13 (1.9%, 95% CI 1.0–3.2%). Patients with cough syncope were predominantly male (69.2% versus 31.1%; p=0.014) and had a higher median body mass index (30.5 versus 23.8 kg·m⁻²; p=0.001) compared to those without. They also exhibited a higher rate of refractory chronic cough (RCC) (84.6% versus 46.2%; p=0.006) and reported greater cough severity and hypersensitivity at 1 year, despite more frequent use of neuromodulators and opiates. Multivariable analysis identified obesity (adjusted (a)OR 4.65, 95% CI 1.24–17.47), RCC (aOR 4.08, 95% CI 1.09–15.23) and male sex (aOR 3.66, 95% CI 1.06–12.68) as independently associated with cough syncope.

Conclusion

Cough syncope affects ∼2% of patients with chronic cough and is associated with male sex, obesity and refractory chronic cough, defining a distinct high-risk phenotype that warrants targeted risk stratification and prevention.

Shareable abstract

Cough syncope may affect ∼2% of patients with chronic cough. Independent associated factors are male sex, obesity and refractory chronic cough. Recognising this high-risk phenotype is crucial for preventing severe complications. https://bit.ly/3QBKIUY

Introduction

Cough syncope is a clinical condition that involves a brief loss of consciousness triggered by paroxysmal coughing [1]. First described in 1876 as “laryngeal vertigo”, the condition is now recognised as a rare but clinically significant subtype of syncope, responsible for ∼5% of all reported cases [2], and regarded as a serious complication of coughing.

The clinical significance of cough syncope extends beyond the brief loss of consciousness. Episodes trigger serious secondary consequences, including motor vehicle accidents that occur when patients faint while driving [3]. Given that these events occur suddenly and unpredictably, effective preventive strategies remain poorly established. In several countries, current management primarily involves restricting driving privileges for patients with a confirmed history of cough syncope [4].

Despite its potential clinical influence, the epidemiology of cough syncope remains poorly characterised, with available data limited to case reports and retrospective studies [5]. Prior descriptions suggest a typical profile of a middle-aged, overweight male smoker with obstructive lung disease [1]. However, the prevalence and associated factors of cough syncope in patients with chronic cough remain unclear, leaving physicians without reliable evidence to support risk stratification, thereby preventing preventive counselling or therapeutic interventions. Therefore, this study aims to characterise patients with cough syncope, estimate its prevalence, and identify its associated factors within a prospective cohort of participants with chronic cough.

Material and methods

Study design and participants

This analysis was conducted as part of the Korean Chronic Cough Registry, a multicentre prospective, longitudinal observational study. The study followed previously established protocols [6] and received approval from the institutional review board (identifier 2019-0754). All participants provided written informed consent. Adults (aged ≥18 years) with newly referred chronic cough at a tertiary care hospital were enrolled. Demographics, comorbidities, cough characteristics and cough-related complications were evaluated at enrolment, followed by assessments of cough status and complications, including syncope, during scheduled follow-up visits. Exclusion criteria comprised loss to follow-up at 1 year, inability to determine the aetiology of cough, missing height or weight data, presence of red-flag signs, and refusal to participate.

Definition of cough syncope

The presence of cough syncope was determined using the self-report of the patient at each study visit. Information on cough syncope was collected as part of the assessment of cough-related complications and was recorded in a binary format (yes/no), based on patients’ responses to the question, “Have you experienced syncope associated with coughing?”. At the 1-year follow-up visit, the occurrence of newly developed cough syncope during the intervening period was assessed using the same approach. Patients who met these criteria at any time during enrolment and the 1-year follow-up period were classified into the cough syncope group, while all others were assigned to the control group.

Assessment

Obesity was defined using Korean criteria as a body mass index (BMI) ≥25 kg·m⁻², calculated as weight (kg) divided by height squared (m²) [7]. Data on height and weight were retrieved from medical records recorded at the enrolment visit.

Cough outcomes were evaluated at enrolment and at the 1-year follow-up visit, including subjective cough frequency, cough severity measured using the visual analogue scale (VAS), Cough Hypersensitivity Questionnaire (CHQ) scores, and cough-specific quality of life assessed using the Leicester Cough Questionnaire (LCQ) [6, 8–10]. Medications prescribed for chronic cough management were reviewed at the follow-up visit conducted 1 year after cohort registration [6].

The duration and subjective frequency of cough, alongside comorbid conditions such as hypertension, diabetes mellitus, heart failure, arrhythmia, allergic rhinitis, asthma, chronic rhinosinusitis, bronchiectasis, tuberculosis and gastro-oesophageal reflux disease (GORD), were evaluated based on the self-report of the patient. Following a protocol amendment in 2022, the presence of snoring was assessed by patient self-report at each study visit; as this item was added after study initiation, data were available for a subset of participants.

Refractory chronic cough (RCC) was defined as the absence of an identifiable underlying cause despite comprehensive clinical evaluation, or as a persistent cough without therapeutic response to appropriate evidence-based treatments for confirmed or suspected aetiologies [11]. The diagnosis of RCC was determined by attending physicians based on clinical assessment, diagnostic workup results, and treatment response evaluation during the 1-year observational period [6].

Statistical analysis

The overall prevalence of cough syncope in the cohort was calculated as the proportion of participants who experienced the event at baseline or at any time during follow-up. Exact (Clopper–Pearson) 95% confidence intervals were calculated for the prevalence estimate. Clinical characteristics of the cough syncope group were compared to those of participants with chronic cough who did not develop cough syncope as controls. Given the low incidence of cough syncope, continuous variables (including age, BMI, cough duration, cough severity VAS, CHQ and LCQ scores) were analysed using the Mann–Whitney U-test. Categorical variables, including sex, comorbidities and medication use, were assessed using the Chi-squared test. Continuous data are presented as median (interquartile range (IQR)), whereas categorical data are reported as counts with corresponding percentages.

To identify potential associated factors of cough syncope, multivariable logistic regression analyses were performed, with adjustment for variables retained through a backward elimination procedure. Results were presented as OR (95% CI). All statistical analyses were conducted using SPSS software, version 25.0 (IBM, Armonk, NY, USA).

Results

Prevalence of cough syncope

Among the 1004 patients enrolled in the registry across 19 institutions between July 2019 and July 2025, 682 met the eligibility criteria for the present analysis (figure 1). Cough syncope was identified in 13 patients (1.9%, 95% CI 1.0–3.2%): 12 reported at enrolment and one newly identified during follow-up.

FIGURE 1.

Study design and patient flow diagram. Of the 1004 patients enrolled in the chronic cough registry across 19 institutions, 682 met the inclusion criteria. Patients were categorised based on their history of cough syncope before enrolment and the occurrence of new episodes during the 1-year follow-up. The final analysis included 13 patients with cough syncope and 669 control participants.

Comparison of baseline clinical characteristics based on cough syncope

Baseline clinical characteristics differed between the cough syncope and control groups at enrolment (table 1). The cough syncope group included a higher proportion of men (69.2% versus 31.1%; p=0.014) and had a higher BMI (median 30.5 kg·m⁻², IQR 25.0–33.5 kg·m⁻² versus 23.8 kg·m⁻², 21.6–26.3 kg·m⁻²; p=0.001) than the controls. Age distribution showed no significant difference between groups (median 50 years, IQR 40.0–60.5 years versus 58 years, 42.0–66.0 years; p=0.158). Additionally, the two groups showed no differences in cough duration, subjective cough frequency, cough severity VAS, CHQ or LCQ scores. Comorbidity profiles were similarly comparable (table 1). Among participants with available data on snoring (n=495), snoring was more frequently reported in the cough syncope group than in the control group (42.9% (three out of seven) versus 21.1% (103 out of 488)); however, this difference did not reach statistical significance (p=0.164).

TABLE 1.

Comparison of baseline characteristics based on the presence of cough syncope

	Control	Cough syncope	p-value
Patients n	669	13
Age years	58 (42–66)	50 (40–60.5)	0.158
Male	31.1	69.2	0.014
BMI kg·m−2	23.8 (21.6–26.3)	30.5 (25–33.5)	0.001
Obesity (BMI ≥25 kg·m−2)	37.2	76.9	0.003
Snoring (n=495)	21.1 (n=488)	42.9 (n=7)	0.164
Smoking status			0.334
Never	75.3	53.8
Former	19.9	38.5
Current	4.8	7.7
Duration of cough months	72 (36–120)	60 (35–330)	0.516
Cough frequency			0.911
Daily	70.7	84.6
Weekly	22.6	15.4
Monthly	3.3	0
No cough	3.4	0
Cough severity VAS	60 (40–80)	60 (55–82.5)	0.452
CHQ score	8 (6–12)	10 (4–14)	0.595
LCQ score	10.8 (8.7–13.5)	7 (5.5–15.5)	0.155
Comorbidities
Hypertension	23.8	38.5	0.163
Diabetes	9.1	15.4	0.384
Arrhythmia	2.1	7.7	0.151
Heart failure	0.9	0	0.740
Allergic rhinitis	15.1	23.1	0.430
Asthma	7.2	7.7	0.890
Chronic rhinosinusitis	5.1	0	0.404
Tuberculosis	3.9	0	0.468
Bronchiectasis	0.3	0	0.843
GORD	19.4	15.4	0.790

Data are presented as median (interquartile range) or %, unless otherwise stated. BMI: body mass index; VAS: visual analogue scale; CHQ: Cough Hypersensitivity Questionnaire; LCQ: Leicester Cough Questionnaire; GORD: gastro-oesophageal reflux disease. Continuous variables such as age, BMI, duration of cough, cough severity VAS, CHQ score and LCQ score were analysed using the Mann–Whitney U-test. Categorical variables such as sex, obesity, smoking status, cough frequency and comorbidities were analysed using the Chi-squared test.

Association between cough syncope and refractory chronic cough

Changes in cough status were evaluated after 1 year of follow-up. At the 1-year follow-up, 320 (51.0%) patients were diagnosed with RCC. The proportion of RCC was higher in the cough syncope group than in the control group (84.6% versus 46.2%; p=0.006) (figure 2a). The cough syncope group was more likely to receive codeine (76.9% versus 32.9%; p=0.002), neuromodulators (38.5% versus 7.0%; p<0.001) and macrolides (38.5% versus 14.8%; p=0.019), whereas the use of other medications did not differ significantly between groups (figure 2b).

FIGURE 2.

Cough characteristics after 1 year of chronic cough treatment. Comparison of treatment outcomes between patients with cough syncope (n=13) and control participants (n=669) after 1 year of chronic cough management. a) Prevalence of refractory chronic cough was higher in the cough syncope group compared to the controls (84.6% versus 46.2%; p=0.006). b) Medication usage patterns showing higher use of codeine (76.9% versus 32.9%; p=0.002), neuromodulators (38.5% versus 7.0%; p<0.001), and macrolides (38.5% versus 14.8%; p=0.019) in the cough syncope group. ICS: inhaled corticosteroid; LABA: long-acting β₂-agonist; LAMA: long-acting muscarinic antagonist; PPI: proton pump inhibitor. *: p<0.05, **: p<0.01, ***: p<0.001 (Chi-squared).

Cough severity VAS scores were higher in the cough syncope group than in the control group at the 1-year follow-up visit (median 50, IQR 17.5–85.0 versus 20, 5.0–30.0; p=0.003) (figure 3a). The longitudinal change in cough severity VAS from enrolment to the 1-year follow-up was smaller in the cough syncope group (median −20, IQR −37.5– −0.0 versus −40, −60.0– −20.0; p=0.013) (figure 3b). In addition, the CHQ score at 1 year was higher in the cough syncope group (median 9, IQR 4.5–12.0 versus 5, 2.5–8.0; p=0.035) (figure 3c). Similarly, the longitudinal change in CHQ score from enrolment to the 1-year follow-up was smaller in the cough syncope group than in the control group (median −0.6, IQR −5.7–3.6 versus −3, −7.0– −1.0; p=0.047) (figure 3d). However, LCQ scores did not differ between groups at the 1-year follow-up (median 15.1, IQR 12.4–16.6 versus 16.6, 13.6–18.5; p=0.124) (figure 3e). The longitudinal change in LCQ score over 1 year was also comparable between groups (median 4.7, IQR −0.4–7.8 versus 5.1, 2.2–8.6; p=0.326) (figure 3f).

FIGURE 3.

Longitudinal changes in cough-related outcomes. a) Cough severity assessed using the visual analogue scale (VAS) at 1-year follow-up remained higher in the cough syncope group (median 50 versus 20; p=0.003). b) Longitudinal change (Δ) in cough severity showed smaller improvement in the cough syncope group (median change −20 versus −40; p=0.013). c) Cough hypersensitivity measured using the Cough Hypersensitivity Questionnaire (CHQ) at 1-year follow-up was higher in the cough syncope group (median 9 versus 5; p=0.035). d) Longitudinal change in cough hypersensitivity demonstrated smaller improvement in the cough syncope group (median change −0.6 versus −3; p=0.047). e,f) Quality of life assessed using the Leicester Cough Questionnaire (LCQ) (median 15.1 versus 16.6; p=0.124) and its longitudinal change (median change 4.7 versus 5.1; p=0.326) did not differ between groups. *: p<0.05, **: p<0.01 (Mann–Whitney U-test).

Factors associated with cough syncope in patients with chronic cough

In the univariable analysis, cough syncope was associated with RCC (OR 3.88, 95% CI 1.06–14.24; p=0.041), male sex (OR 4.99, 95% CI 1.52–16.38; p=0.008) and obesity (BMI ≥25 kg·m⁻²; OR 5.70, 95% CI 1.55–20.89; p=0.009). In the multivariable analysis, all three factors remained independently associated with cough syncope: RCC (aOR 4.08, 95% CI 1.09–15.23; p=0.036), male sex (aOR 3.66, 95% CI 1.06–12.68; p=0.041), and obesity (aOR 4.65, 95% CI 1.24–17.47; p=0.023) (table 2).

TABLE 2.

Factors associated with cough syncope (compared to no cough syncope)

	Univariable		Multivariable#
	OR (95% CI)	p-value	aOR (95% CI)	p-value
Age (years)	0.98 (0.94–1.01)	0.181
Sex
Female	1.00 (Reference)
Male	4.99 (1.52–16.38)	0.008	3.66 (1.06–12.68)	0.041
Classification
Non-RCC	1.00 (Reference)
RCC	3.88 (1.06–14.24)	0.041	4.08 (1.09–15.23)	0.036
Obesity (BMI ≥25 kg·m−2)
No	1.00 (Reference)
Yes	5.70 (1.55–20.89)	0.009	4.65 (1.24–17.47)	0.023

aOR: adjusted odds ratio; RCC: refractory chronic cough; BMI: body mass index. ^#: multivariable logistic regression analysis was performed using SPSS version 25.0 to identify independent associated factors, including age, sex, cough classification and obesity. A backward elimination procedure was applied in the binary logistic regression model.

Discussion

This study is the first to describe the epidemiology of cough syncope using a multicentre prospective registry of patients with chronic cough. We found that 1.9% of patients experienced at least one episode of cough syncope over a 1-year period. These patients exhibited persistent cough severity and hypersensitivity symptoms despite receiving more intense pharmacological treatments. Importantly, our analysis identified male sex, obesity and RCC as associated factors for cough syncope. Unlike previous reports largely limited to retrospective case series [1, 12, 13], our prospective cohort setting enabled the systematic characterisation of this phenotype, suggesting a specific high-risk subgroup within the chronic cough patient population.

The classic profile of cough syncope, a middle-aged, overweight, male smoker with obstructive lung disease is well-established [1]. Our findings align with this profile regarding male sex and obesity, but diverge regarding the underlying respiratory condition. Unlike earlier reports emphasising smoking and obstructive airway disease, we identified RCC as a major associated factor. This distinction probably reflects the nature of our cohort, which comprised patients referred for treatment-refractory or unexplained chronic cough rather than those with established respiratory diseases. Given that RCC represents a distinct clinical entity characterised by neuronal hypersensitivity [14–18], the significant association between RCC and cough syncope suggests that therapeutic strategies addressing neurogenic mechanisms may be warranted for this patient group. However, the temporal relationship between RCC and cough syncope warrants cautious interpretation. As most syncope cases were identified at enrolment, whereas RCC was determined after 1 year of follow-up, the observed association between them may reflect shared mechanisms, such as neuronal hypersensitivity, rather than a direct causal pathway from either condition. Longitudinal studies with incident cases would be needed to clarify the directionality of this relationship.

In the literature, data on the epidemiology of syncope in the Korean population are very limited. In our cohort, the overall prevalence of cough syncope was 1.9% (95% CI 1.0–3.2%). As the majority of cases were already present at enrolment, this figure primarily represents the prevalence among patients referred for chronic cough evaluation rather than a measure of new-onset incidence. For context, the Framingham Heart Study reported an incidence of general syncope of 6.2 cases per 1000 person-years (approximate annual risk of 0.6%) in the general population [19]. Although direct comparison between prevalence and incidence measures is inherently limited, the observed frequency in our chronic cough cohort appears notably higher, yet arguably lower than might be expected, given that chronic cough involves repetitive, forceful triggers for the event. This suggests that uncontrolled cough alone is insufficient to cause syncope in most patients; rather, it requires a specific host susceptibility, such as the “obese male” phenotype, to precipitate the event.

The fundamental mechanism of cough syncope involves vigorous, paroxysmal coughing that elevates intrathoracic pressure, thereby reducing venous return, cardiac output, and ultimately cerebral perfusion [20, 21]. While baroreceptor hypersensitivity may play a role [22], general population studies like the Framingham Heart Study in the United States of America did not identify obesity as a major associated factor for all-cause syncope [23], which is probably because cardiac or vasovagal causes predominate in the general population [24]. Thus, the associated factors we observed (obesity and male sex) appear to be specific to the pathophysiology of cough-induced situational syncope.

The susceptibility of obese patients to cough syncope can be attributed to three interrelated mechanisms: mechanical pressure, cough burden and autonomic dysregulation. First, obesity mechanistically elevates pressures. Studies show that obese individuals exhibit significantly higher oesophageal and gastric pressures at functional residual capacity compared to normal-weight controls, with pressure levels correlating strongly with BMI [25]. This suggests that abdominal adiposity acts as a mechanical compressor, maintaining chronically elevated intra-abdominal pressure [26]. Second, obesity is associated with a higher cough burden, partly due to increased prevalence of GORD, a common trigger of chronic cough [27]. Third, obesity may contribute to autonomic dysfunctions. Obese individuals often demonstrate impaired baroreflex sensitivity and sympathetic overactivity [28–30], which may hinder the compensatory cardiovascular responses required to maintain cerebral perfusion during paroxysmal coughing.

The male predominance in our study is consistent with historical data [1] and may be explained by anatomical and hormonal differences. Anatomically, men typically possess greater respiratory muscle mass and diaphragm thickness compared to women [31, 32]. This increased muscle mass probably generates more forceful expulsive efforts, leading to greater fluctuations in intrathoracic pressure. Hormonally, sex differences influence fat distribution; androgens promote the accumulation of visceral adipose tissue in the abdominal region in men, whereas oestrogens favour subcutaneous fat deposition in women [33–35]. Consequently, the combination of a stronger cough pump (due to muscle mass) and greater central resistance (due to visceral adipose tissue) in men likely creates a synergistic interaction that generates the critical intrathoracic pressure required to trigger syncope.

The phenotypic profile of cough syncope patients (male, obese and refractory to treatment) raises the possibility of undiagnosed sleep-disordered breathing, including obstructive sleep apnoea, as a contributing factor. In our cohort, snoring was reported more frequently in the cough syncope group (42.9% versus 21.1%), although this did not reach statistical significance, probably due to the limited sample size. As sleep-disordered breathing is associated with autonomic dysregulation and may exacerbate chronic cough through upper airway inflammation and vagal afferent sensitisation [36, 37], it could represent an underrecognised mediator linking obesity to cough syncope. Future studies should incorporate systematic screening for this condition in the chronic cough population.

Our study has limitations. First, cough syncope was defined solely on the basis of patient self-report without standardised clinical assessment. Therefore, we were unable to distinguish true syncope from pre-syncope, which may have resulted in outcome misclassification. If pre-syncope cases were included, the prevalence may have been overestimated. However, such misclassification is probably nondifferential with respect to the associated factors, which would bias the observed associations toward the null, suggesting that our estimates are conservative rather than inflated. Of note, a similar self-report approach was used in the European Respiratory Society NEuroCOUGH registry, which reported a prevalence of 14% [38], further suggesting that our estimate is not inflated and differences in population characteristics, such as obesity prevalence, may account for the between-study variation. Second, the relatively small number of syncope cases may limit the precision of the associated factor analysis. However, the findings are derived from a large, well-characterised prospective registry of chronic cough, and the low case number probably reflects the true epidemiology of the condition. Third, as the majority of cases with cough syncope were prevalent at enrolment, the temporal relationship between the identified factors and cough syncope cannot be established. The observed associations should therefore not be interpreted as causal. Fourth, the aetiological evaluation of syncope was limited by insufficient assessment of other potential contributing factors, including cardiac and neurological comorbidity.

In conclusion, this study estimates an overall prevalence of cough syncope of 1.9% (95% CI 1.0–3.2) in patients with chronic cough. The condition is associated with male sex, obesity and RCC. These findings suggest that cough syncope is not merely a consequence of cough severity, but results from specific physiological interactions in a susceptible phenotype. Clinicians should recognise the increased syncope risk in obese male patients with refractory cough and consider early interventions to mitigate this serious complication.