Abstract
Background
The potential clinical implication of sleep talking in relation to stroke has not been explored to date. This study aimed to prospectively examine the association between sleep talking and the risk of developing stroke in a community‐based cohort.
METHODS AND RESULTS
Included were 8001 participants (mean age, 54 years) of the Kailuan Study, China. Sleep talking was measured by a questionnaire in 2012. Cases of incident stroke were confirmed by review of medical records. Cox proportional hazards models were used to explore the association between sleep talking and stroke, adjusting for several sleep parameters (ie, insomnia, daytime sleepiness, sleep duration, snoring, and use of hypnotics) and other potential confounders. During 8 years of follow‐up, 333 incident stroke cases were identified. Relative to participants without sleep talking at baseline, those with sleep talking had a higher risk of developing stroke (hazard ratio [HR], 1.30 [95% CI, 1.03–1.65]), adjusting for potential confounders. Compared with participants without probable rapid eye movement sleep behavior disorder and sleep talking, those with sleep talking and probable rapid eye movement sleep behavior disorder had a higher risk of stroke (adjusted HR, 1.93 [95% CI, 1.40–2.66]).
Conclusions
The presence of sleep talking was associated with a higher risk of developing stroke. Future studies with cases of clinically confirmed sleep talking and a longer follow‐up would be appropriate to further investigate this association.
Keywords: parasomnias, prospective cohort study, sleep talking, stroke
Subject Categories: Ischemic Stroke, Neurogenesis, Neurostimulation
Nonstandard Abbreviations and Acronyms
- pRBD
probable rapid eye movement sleep behavior disorder
- RBD
rapid eye movement sleep behavior disorder
Clinical Perspective.
What Is New?
The first community‐based, prospective population study regarding the association between sleep talking and risk of stroke.
Sleep talking was associated with a higher risk of developing stroke.
This association was independent of traditional risk factors of stroke and other sleep parameters, including sleep duration, daytime sleepiness, hypnotics use, insomnia, and snoring status.
What Are the Clinical Implications?
The results of this study suggest that sleep talking may be a novel risk factor for stroke and provide a basis for future biological mechanism research.
Future studies could benefit from including a questionnaire detailing the characteristics of sleep talking, such as frequency and duration.
Sleep talking, one of the most common parasomnias in the general population, is the utterance of speech during sleep, either spontaneous or provoked by external stimuli. 1 , 2 , 3 However, the potential clinical implication of sleep talking, to the best of our knowledge, has not been explored in population‐based studies to date.
Sleep talking has been reported to be associated with sleep fragmentation, poorer sleep quality 4 , 5 , 6 and other sleep disorders and nocturnal behaviors (eg rapid eye movement sleep behavior disorder [RBD], disorder of arousal and sleep moaning, and nonrapid eye movement parasomnia‐related phenomena). 6 , 7 Decreased sleep duration and increased sleep fragmentation are key triggers for intermediary mechanisms such as sympathetic activation and oxidative stress, which may increase the incidence and severity of cardiovascular diseases. 8 , 9 , 10 , 11 , 12 , 13 , 14 In addition, during sleep talking, the brain remains capable of engaging in complex and high‐level activities that potentially lead to elevated sympathetic output during sleep, which could heighten cardiovascular stress. 7 Given that these sleep factors have been suggested to be associated with altered stroke risk, 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 we hypothesized that individuals with sleep talking might have a higher stroke risk, relative to those without the symptoms.
To test this hypothesis, we conducted a prospective study including ≈8000 Chinese participants to examine the association between sleep talking and risk of developing stroke, during 8 years of follow‐up. The joint association of sleep talking and probable RBD (pRBD), another parasomnia that was shown to be associated with stroke risk in the same cohort, 17 with stroke risk was also examined.
METHODS
The data and code that support the findings of this study will be made available from the corresponding author upon reasonable request and approval.
Study Population
The Kailuan Study is an ongoing prospective cohort conducted in the Kailuan community in Tangshan City, China, and its study design and procedures were previously described. 18 , 19 , 20 Participants in the current study were a subset of the Kailuan Study, including 12 990 adults who were evaluated at the Kailuan hospital in 2012 and confirmed to be free of Parkinson disease and dementia, as previously detailed. 21 Figure S1 illustrates the inclusion and exclusion process for patients in this analysis. Individuals with a history of stroke who did not provide detailed information on sleep talking at baseline were excluded, leaving 8001 participants in the current analysis. The Kailuan General Hospital ethics committee approved this study. Patient consent for publication was not required.
Assessment of Incident Stroke
The outcome was defined as the first occurrence of stroke, including both hemorrhagic and ischemic types. The Municipal Social Insurance Institution database and Hospital Discharge Register for stroke were linked via identification number, which can cover all participants in the Kailuan Study. International Classification of Diseases, Tenth Revision (ICD‐10), codes were used to identify stroke cases: I61 for intracerebral hemorrhagic stroke and I63 for ischemic stroke. The diagnosis of stroke was confirmed through computed tomography, magnetic resonance imaging, and other diagnostic reports, as previously described. 17 , 18 Stroke cases were ascertained by 2 experienced cardiologists through reviewing medical records. In cases of disagreement, a third cardiologist was consulted for final adjudication.
Assessment of Sleep Talking and Other Sleep Parameters
In 2012, sleep talking was assessed using the following 2 questions of the RBD questionnaire–Hong Kong (RBDQ‐HK): (1) whether the participant had ever experienced sleep talking during their lifetime; and (2) whether they had experienced sleep talking once or more in the past year. Participants were considered to have experienced sleep talking if they exhibited ≥1 of the conditions mentioned above. The RBDQ‐HK is a diagnostic screening tool for RBD and was previously validated. 22 , 23 The questionnaire collected information in relation to RBD clinical symptoms. Participants were considered as pRBD with the total scale (score range 0–100) >18.
Insomnia status was evaluated using the validated Chinese version of the Athens Insomnia Scale, which has been validated as reliable in the Chinese population, and the cutoff to determine insomnia for the total scale was ≥6. 24 Daytime sleepiness was assessed using the Chinese version of the Epworth Sleepiness Scale, with a total score ≥10 indicating excessive daytime sleepiness. 25 Individuals were considered to have obstructive sleep apnea (OSA) if they scored positively on ≥3 of the 8 items on the STOP‐BANG questionnaire. 26 Sleep duration and snoring status were collected by questionnaire, and the participants were further categorized into 5 groups for sleep duration (<6, 6 to 7, 7 to 8, 8 to 9, and >9 hours per day) and 3 groups for snoring status (never/rare, occasionally, and frequently).
Covariates
Information on age, sex, marital status, education level, income level, occupation, physical activity, smoking, and drinking status was collected via questionnaires. Height and weight were measured in light clothing and without shoes in the 2012 interview by trained medical personnel. Body mass index was calculated by dividing weight in kilograms by height in meters squared. Two measurements of systolic and diastolic blood pressure were measured using a mercury sphygmomanometer with the participant in the seated position and the average of the 2 measurements used for the analysis. Fasting blood samples were collected, and further concentrations of glucose, triglycerides, low‐density lipoprotein cholesterol, and high‐density lipoprotein cholesterol were assessed using a Hitachi 747 autoanalyzer (Hitachi) at the central laboratory of the Kailuan General Hospital. 17 Hypertension was defined as ≥1 of the following: (1) systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg; or (2) taking antihypertensive medications. Diabetes was defined as ≥1 of the following: (1) fasting blood glucose ≥7.0 mmol/L; (2) taking oral hypoglycemic agents; or (3) active treatment with insulin.
Statistical Analysis
All statistical analyses were conducted using R software (R 4.1.1, The R Foundation for Statistical Computing). Characteristics of different groups were evaluated using Student t test or Wilcoxon rank sum test for continuous variables and χ2 test for categorical variables. Two‐tailed P values <0.05 were considered statistically significant.
The association between sleep talking and stroke was examined by a Cox proportional hazards model with adjustments for age, sex, marital status (married or single), education level (primary, middle, or college and higher), income level (<500, 500–1000, or >1000 renminbi per month), occupation (blue collar/white collar), physical activity (never, <4, or ≥4 times per week), smoking (never, past, or current smoker), alcohol status (yes/no), hypertension (yes/no), diabetes (yes/no), body mass index, and plasma concentrations of triglycerides, low‐density lipoprotein cholesterol, high‐density lipoprotein cholesterol, sleep duration (<6, 6–7, 7–8, 8–9, or >9 h/d), daytime sleepiness (yes/no), hypnotics use (yes/no), insomnia (yes/no), and snoring (never/rare, occasionally, or frequently). The proportional hazards assumption was checked using Schoenfeld residuals, 27 , 28 and no significant violations were found. Stroke subtypes (intracerebral hemorrhagic or ischemic stroke) were further explored.
To understand whether the potential association between sleep talking and stroke risk was attributable to inverse causality, ie, the preclinical stroke could lead to sleep talking, we examined the relationships between lifetime occurrence versus short‐time (recent/1‐year) occurrence of sleep talking and incident stroke.
To test the robustness of our observations, several sensitivity analyses were conducted by excluding participants with depressive disorder and other sleep disorders (including excessive daytime sleepiness, insomnia, and frequently snoring) or those using hypnotics or current alcohol drinkers at baseline, separately. We also conducted a sensitivity analysis by excluding participants with intermediate or high risk of OSA, based on the 2014 STOP‐BANG questionnaire.
The joint association of sleep talking and pRBD status with stroke was explored by a Cox proportional hazards model with adjustments for the aforementioned covariates. We examined potential interactions of presence of sleep talking (yes/no) with pRBD in relation to stroke risk, by including multiplicative terms in the Cox models, with adjustment for other potential confounders in the final model. To test the robustness of the joint association analysis, we further redefined pRBD by excluding the 2 sleep talking questions and used 18 as the cutoff point.
Traditional mediation analysis models 29 were constructed to examine whether the association between sleep talking and stroke was linked through sleep quality score or OSA. In the mediation models, the predictor variable was sleep talking, the mediator variables (M) were sleep quality score or OSA, and the outcome variable was incident stroke. Mediation analyses were performed using R package lavaan, adjusted for covariates in the final model.
RESULTS
Compared with participants without sleep talking, sleep talkers were more likely to be men, work blue collar jobs (compared with white collar jobs), smokers, alcohol drinkers, and be at lower education and physical activity levels (Table 1). Abnormal sleep parameters, such as daytime sleepiness, hypnotics use, and snoring, were also significantly associated with higher odds of experiencing sleep talking.
Table 1.
Baseline Characteristics by Sleep Talking
| Variables | Nonsleep talking | Sleep talking | P value |
|---|---|---|---|
| Total, n | 6028 | 1973 | |
| Men, n (%) | 4779 (79.3) | 1697 (86.0) | <0.001 |
| Age, y | 53.6±11.3 | 53.4±10.7 | 0.67 |
| BMI, kg/m2 | 25.0±3.4 | 25.2±3.2 | 0.02 |
| Marital status, n (%) | 0.78 | ||
| Married | 5893 (97.8) | 1926 (97.6) | |
| Single | 135 (2.2) | 47 (2.4) | |
| Education, n (%) | 0.02 | ||
| Primary | 253 (4.2) | 110 (5.6) | |
| Middle | 5222 (86.6) | 1699 (86.1) | |
| College | 553 (9.2) | 164 (8.3) | |
| Income, RMB per mo | 0.008 | ||
| <500 | 117±1.9 | 37±1.9 | |
| 500–1000 | 1192±19.8 | 330±16.7 | |
| >1000 | 4719±78.3 | 1606±81.4 | |
| Blue collar occupation, n (%) | 5374 (89.2) | 1800 (91.2) | 0.01 |
| Physical activity, n (%) | 0.02 | ||
| Never/rare | 2788 (46.3) | 961 (48.7) | |
| Occasional | 2378 (39.4) | 768 (38.9) | |
| Frequent | 862 (14.3) | 244 (12.4) | |
| Smoking, n (%) | <0.001 | ||
| Never | 3649 (60.5) | 992 (50.3) | |
| Past | 348 (5.8) | 118 (6.0) | |
| Current | 2031 (33.7) | 863 (43.7) | |
| Drinking, n (%) | 2296 (38.1) | 909 (46.1) | <0.001 |
| Hypertension, n (%) | 2056 (34.1) | 686 (34.8) | 0.61 |
| Diabetes, n (%) | 668 (11.1) | 240 (12.2) | 0.20 |
| Triglycerides, mmol/L | 1.2 (0.8–1.9) | 1.2 (0.9–1.9) | 0.40 |
| HDL‐C, mmol/L | 1.4 (1.2–1.6) | 1.3 (1.1–1.6) | <0.001 |
| LDL‐C, mmol/L | 1.9 (1.5–2.6) | 1.9 (1.5–2.5) | 0.38 |
| Follow years | 7.8 (7.5–8.2) | 7.9 (7.6–8.2) | 0.41 |
| Sleep parameters | |||
| Snoring, n (%) | <0.001 | ||
| Never/rare | 3922 (65.1) | 1049 (53.2) | |
| Occasional | 1371 (22.7) | 485 (24.6) | |
| Frequent | 735 (12.2) | 439 (22.3) | |
| Hypnotics use, n (%) | 697 (11.6) | 163 (8.3) | <0.001 |
| Daytime sleepiness, n (%) | 1317 (21.8) | 574 (29.1) | <0.001 |
| Sleep duration, h/d | 0.38 | ||
| 7.0–7.9 | 1044±17.3 | 319±16.2 | |
| <6.0 | 773±12.8 | 280±14.2 | |
| 6.0–6.9 | 1178±19.5 | 382±19.4 | |
| 8.0–8.9 | 2972±49.3 | 967±49.0 | |
| ≥9.0 | 61±1.0 | 25±1.3 | |
| Insomnia, n (%) | 136 (2.3) | 51 (2.6) | 0.45 |
| pRBD, n (%) | 58 (1.0) | 593 (30.1) | <0.001 |
| OSA, n (%) | 1746 (29.0) | 623 (31.6) | 0.03 |
Data are expressed as mean±SD or median (interquartile range) unless otherwise indicated. BMI indicates body mass index; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; OSA, obstructive sleep apnea; pRBD, probable rapid eye movement sleep behavior disorder; and RMB, renminbi.
During the follow‐up period, 333 incident stroke cases were documented. Compared with participants without sleep talking, those with sleep talking had a 30% higher risk (95% CI, 3%–65%) of developing stroke during the follow‐up, after adjusting for age, sex, marital status, education level, income level, occupation, physical activity, smoking status, drinking status, hypertension, diabetes, body mass index, plasma concentrations of lipids, and other sleep parameters (Table 2). A similar pattern was observed for 2 stroke subtypes: ischemic stroke (adjusted hazard ratio [HR], 1.30 [95% CI, 1.00–1.67]) and hemorrhagic stroke (adjusted HR, 1.36 [95% CI, 0.69–2.68]). Results remained consistent among all sensitivity analyses (Table 3).
Table 2.
Hazards Ratios and 95% CIs of Stroke According to Sleep Talking Status
| Nonsleep talking | Sleep talking | P value | |
|---|---|---|---|
| Total stroke | |||
| Cases/total, n (%) | 229/6028 (3.8) | 104/1973 (5.3) | |
| Model 1 | Reference | 1.39 (1.10–1.75) | 0.006 |
| Model 2 | Reference | 1.36 (1.08–1.72) | 0.01 |
| Model 3 | Reference | 1.30 (1.03–1.65) | 0.03 |
| Ischemic stroke | |||
| Cases/total, n (%) | 202/6028 (3.4) | 91/1973 (4.6) | |
| Model 1 | Reference | 1.38 (1.08–1.77) | 0.01 |
| Model 2 | Reference | 1.35 (1.05–1.74) | 0.02 |
| Model 3 | Reference | 1.30 (1.00–1.67) | 0.046 |
| Hemorrhagic stroke | |||
| Cases/total, n (%) | 27/6028 (0.4) | 13/1973 (0.7) | |
| Model 1 | Reference | 1.45 (0.75–2.81) | 0.28 |
| Model 2 | Reference | 1.44 (0.74–2.80) | 0.29 |
| Model 3 | Reference | 1.36 (0.69–2.68) | 0.38 |
Model 1: adjusted for age and sex. Model 2: further adjusted for marital status (married or single), education level (primary, middle, or college and higher), income level (<500, 500–1000, or >1000 renminbi [RMB] per month), occupation (blue collar/white collar), physical activity (never, <4, or >4 times per week), smoking (never, past, or current smoker), drinking status (yes/no), hypertension (yes/no), diabetes (yes/no), body mass index, and plasma concentrations of triglycerides, low‐density lipoprotein cholesterol, and high‐density lipoprotein cholesterol. Model 3: further adjusted for sleep duration (<6, 6–7, 7–8, 8–9, or >9 h/d), daytime sleepiness (yes/no), hypnotics use (yes/no), insomnia (yes/no), and snoring (never/rare, occasionally, or frequently snoring).
Table 3.
Sensitivity Analyses for Hazard Ratios (95% CIs) of Total Stroke According to Sleep Talking Status*
| Nonsleep talking | Sleep talking | P value | |
|---|---|---|---|
| Excluding 37 participants with depressive disorder | Reference | 1.30 (1.02–1.65) | 0.03 |
| Excluding 1891 participants with excessive daytime sleepiness† | Reference | 1.30 (0.98–1.71) | 0.07 |
| Excluding 1174 participants who reported frequent snoring | Reference | 1.35 (1.03–1.78) | 0.03 |
| Excluding 187 participants with insomnia‡ | Reference | 1.29 (1.02–1.65) | 0.04 |
| Excluding 860 participants who used hypnotics | Reference | 1.22 (0.95–1.57) | 0.13 |
| Excluding 3205 alcohol drinkers | Reference | 1.40 (0.97–2.00) | 0.07 |
| Excluding 911 participants with high risk of OSAδ | Reference | 1.36 (1.06–1.75) | 0.02 |
OSA indicates obstructive sleep apnea.
Adjusted for age, sex, marital status (married or single), education level (primary, middle, or college and higher), income level (<500, 500–1000, or >1000 renminbi per month), occupation (blue collar/white collar), physical activity (never, <4, or >4 times per week), smoking (never, past, or current smoker), drinking status (yes/no), hypertension (yes/no), diabetes (yes/no), body mass index, plasma concentrations of triglycerides, low‐density lipoprotein cholesterol, high‐density lipoprotein cholesterol, sleep duration (<6, 6–7, 7–8, 8–9, or >9 h/d), daytime sleepiness (yes/no), hypnotics use (yes/no), insomnia (yes/no), and snoring (never/rare, occasionally, or frequently snoring).
Based on 7 behavioral factors including sleep talking, shouting, limb movements, and sleep‐related injuries (score range 0–70, cutoff >7).
Athens Insomnia Scale >6.
STOP‐BANG score >3 in 2014.
Longer duration of sleep talking was associated with a higher risk of stroke (P trend=0.02). A significant association was only observed for the lifetime occurrence of sleep talking (adjusted HR, 1.41 [95% CI, 1.02–1.94]), but not for those occurring in the past year (adjusted HR, 1.23 [95% CI, 0.91–1.65]) (Figure 1).
Figure 1. Hazard ratios (95% CIs) of short‐/long‐term sleep talking status in relation to stroke.
The multivariable analysis was adjusted for age, sex, marital status (married or single), education level (primary, middle, or college and higher), income level (<500, 500–1000, or >1000 renminbi [RMB] per month), occupation (blue or white collar), physical activity (never, <4, or >4 times per week), smoking (never, past, or current smoker), drinking status (yes/no), hypertension (yes/no), diabetes (yes/no), body mass index, plasma concentrations of triglycerides, low‐density lipoprotein cholesterol, high‐density lipoprotein cholesterol, sleep duration (<6, 6–7, 7–8, 8–9, or >9 h/d), daytime sleepiness (yes/no), hypnotics use (yes/no), insomnia (yes/no), and snoring (never/rare, occasionally, or frequently snoring). Lifetime occurrence indicates a lifetime occurrence of sleep talking; short‐time occurrence, recent 1‐year occurrence of one or more sleep talking incidents.
Compared with participants without pRBD and sleep talking, those with sleep talking and pRBD had a higher risk of stroke (adjusted HR, 1.93 [95% CI, 1.40–2.66]) (Figure 2). There was no significant interaction in the association between sleep talking and pRBD (interaction P=0.65) (Figure S2). In the sensitivity analysis for the joint association (Figure S3), we found that compared with participants without pRBD and sleep talking, those with sleep talking still had a higher risk of stroke (adjusted HR, 1.30 [95% CI, 1.01–1.67]).
Figure 2. The joint association of sleep talking and probable rapid eye movement sleep behavior disorder (pRBD) status for stroke.
The multivariable analysis was adjusted for age, sex, marital status (married or single), education level (primary, middle, or college and higher), income level (<500, 500–1000, or >1000 renminbi [RMB] per month), occupation (blue or white collar), physical activity (never, <4, or >4 times per week), smoking (never, past, or current smoker), drinking status (yes/no), hypertension (yes/no), diabetes (yes/no), body mass index, plasma concentrations of triglycerides, low‐density lipoprotein cholesterol, high‐density lipoprotein cholesterol, sleep duration (<6, 6–7, 7–8, 8–9, or >9 h/d), daytime sleepiness (yes/no), hypnotics use (yes/no), insomnia (yes/no) and snoring (never/rare, occasionally, or frequently snoring).
We examined the mediation effects of sleep quality score and OSA on the association of sleep talking and incident stroke. Sleep quality score and OSA could only explain a small and insignificant proportion (estimated proportion <5% and P value >0.05 for both; Figure S4) on the sleep talking–stroke association.
DISCUSSION
In this community‐based prospective study, we found that sleep talking is associated with a 30% higher risk of developing stroke, independent of well‐known risk factors for stroke, such as age, sex, hypertension, diabetes, smoking, drinking status, and other sleep parameters. Longer duration of sleep talking was associated with higher stroke risk. Notably, the risk of incident stroke was 93% higher for those with both sleep talking and pRBD, compared with those with neither of the 2 parasomnias.
It has been reported that 20% to 25% of sleep talking occurs during rapid eye movement sleep and 75% to 80% occurs during nonrapid eye movement sleep. 30 A high frequency of sleep talking has been observed in patients with RBD, and it is considered a major symptom of RBD. 22 Based on the same prospective cohort—the Kailuan Study—our previous study demonstrated that participants with pRBD had a higher risk of stroke during 3 years of follow‐up, relative to those without pRBD. 17 In the current study, we expanded the prior analyses and found that, relative to participants without either of the 2 probable parasomnias (sleep talking and pRBD), the risk of incident stroke was 93% higher for those with the 2 parasomnias. Moreover, individuals exhibiting either sleep talking or pRBD alone did not show a significant risk of future stroke. This may indicate that the potential influence of pRBD on stroke risk may be through some underlying pathways related to sleep talking. In addition, we conducted a sensitivity analysis of the joint association of sleep talking and pRBD on stroke risk, excluding sleep talking as a defining item for pRBD. Interestingly, we observed that participants with sleep talking alone still faced a significantly higher risk of developing stroke (adjusted HR, 1.30 [95% CI, 1.01–1.67]), and this finding highlights the importance of sleep talking on stroke risk.
Several biological mechanisms may underlie the observed association between sleep talking and stroke, although the direct causal relationship has not been established. Sleep talkers had significantly worse self‐reported sleep quality, and sleep talking was positively correlated with sleep fragmentation. 5 Reductions in sleep duration and increased sleep fragmentation are primary triggers for intermediary mechanisms such as sympathetic activation, oxidative stress, and low‐grade inflammation. These factors could elevate the incidence and severity of cardiovascular and metabolic diseases. 8 , 9 , 10 , 11 , 12 , 13 , 14 Further, the brain during sleep talking remains to be capable of performing a complex and high‐level activity. 7 , 31 Previous studies have noted that stroke and transient ischemic attacks frequently occur at night or shortly after waking up (wake‐up strokes), particularly when accompanied by sleep disorders. 9 , 32 In our study, the significant association between sleep talking and stroke persisted even after adjustment for the cardiovascular factors, suggesting that sleep talking could be an independent risk factor for stroke beyond these conventional factors. Verbal sleep talking and language processing during wakefulness shared neural mechanisms, as assessed by electroencephalogram and polysomnography monitoring, 7 and sleep talking may lead to increased sympathetic output during sleep, which could elevate cardiovascular stress. Sleep talking is often related to other nocturnal behaviors, 6 , 7 making the study of its peculiarities complex. However, in our research, even after adjusting for other sleep parameters and controlling for related sleep disorders, the results remained similar.
The strengths of our study include the prospective design and detailed measurements of lifestyle factors and disease status. To the best of our knowledge, this is the first study to explore the relationship between sleep talking and a major chronic disease with a prospective design in a community‐based population. However, several limitations of the present study should be noted. The status of sleep talking was determined by self‐reported questionnaire, which could lead to misreporting and misclassification. In addition, responses to questions about parasomnias required a sleep partner, potentially introducing selection bias. To reduce the effects of potential misclassification of sleep talking status, we controlled other sleep parameters, such as daytime sleepiness, insomnia, and snoring. Further, detailed information on OSA was not collected until 2014, resulting in only 78% of participants with OSA data available, which might introduce residual bias. Moreover, OSA status was determined by the STOP‐BANG, rather than by polysomnography, potentially leading to misclassification. Like any observational study, our results may be limited by the possibility of residual confounding and measurement error. Further, no association was observed between sleep talking and hemorrhagic stroke because of fewer incident hemorrhagic stroke cases (n=40); thus, further study with longer follow‐up and a larger sample size are necessary. Another limitation is that we could not examine the sex‐specific association because of the small sample size of women. However, the interaction between sex and sleep talking in relation to stroke was not significant. Finally, because our study was conducted only in Tangshan City, an industrial city in northern China, the generalizability of our findings to other ethnic populations may be limited.
CONCLUSIONS
In this population‐based prospective study, we found that sleep talking was associated with a higher risk of developing stroke. These findings suggest that sleep talking may be a novel risk factor for stroke and provide a basis for future biological mechanism research. Future research in this field is warranted, including epidemiologic studies that objectively assess sleep and clinical studies investigating the biologic mechanisms underlying the behavior control of sleep‐related behaviors. Future studies could benefit from including a questionnaire detailing the characteristics of sleep talking, such as frequency and duration.
Sources of Funding
This study was supported by the Startup grant at Fudan University (JIF201015Y, JIF201017Y, and JIF201018Y) and the key projects in the 3‐year plan of the Shanghai Municipal Public Health System (2023–2025) (GWVI‐4).
Supporting information
Figures S1–S4
Acknowledgments
We appreciate the survey teams of the Kailuan Study group for their contribution and the study participants for their invaluable information for this project.
Author contributions: Y.L., Z.Y., X.G., M.P., and Y.L. generated the hypothesis and conceptualization of the study, data acquisition, data analysis, and interpretation. Y.L., L.S., J.Z., and S.C. contributed to data analysis and interpretation. S.C., Y.L., Z.L., and X.G. played a major role in the acquisition of data, study concept, and design. All authors reviewed and commented on the article and approved its final submission.
This article was sent to Michelle H. Leppert, MD, MBA, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.124.035813
For Sources of Funding, see page 7.
References
- 1. Arkin AM, Antrobus JS, Toth MF, Baker J, Jackler F. A comparison of the content of mentation reports elicited after nonrapid eye movement (NREM) associated sleep utterance and NREM “silent” sleep. J Nerv Ment Dis. 1972;155:427–435. doi: 10.1097/00005053-197212000-00006 [DOI] [PubMed] [Google Scholar]
- 2. Bjorvatn B, Grønli J, Pallesen S. Prevalence of different parasomnias in the general population. Sleep Med. 2010;11:1031–1034. doi: 10.1016/j.sleep.2010.07.011 [DOI] [PubMed] [Google Scholar]
- 3. Arkin AM. Sleep‐talking: a review. J Nerv Ment Dis. 1966;143:101–122. doi: 10.1097/00005053-196608000-00001 [DOI] [PubMed] [Google Scholar]
- 4. Honda K, Hashimoto M, Yatabe Y, Kaneda K, Yuki S, Ogawa Y, Matsuzaki S, Tsuyuguchi A, Tanaka H, Kashiwagi H, et al. The usefulness of monitoring sleep talking for the diagnosis of dementia with Lewy bodies. Int Psychogeriatr. 2013;25:851–858. doi: 10.1017/S1041610213000185 [DOI] [PubMed] [Google Scholar]
- 5. Camaioni M, Scarpelli S, Alfonsi V, Gorgoni M, De Bartolo M, Calzolari R, De Gennaro L. The influence of sleep talking on nocturnal sleep and sleep‐dependent cognitive processes. J Clin Med. 2022;11:6489. doi: 10.3390/jcm11216489 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Alfonsi V, D'Atri A, Scarpelli S, Mangiaruga A, De Gennaro L. Sleep talking: a viable access to mental processes during sleep. Sleep Med Rev. 2019;44:12–22. doi: 10.1016/j.smrv.2018.12.001 [DOI] [PubMed] [Google Scholar]
- 7. Mangiaruga A, D'Atri A, Scarpelli S, Alfonsi V, Camaioni M, Annarumma L, Gorgoni M, Pazzaglia M, De Gennaro L. Sleep talking versus sleep moaning: electrophysiological patterns preceding linguistic vocalizations during sleep. Sleep. 2022;45:zsab284. doi: 10.1093/sleep/zsab284 [DOI] [PubMed] [Google Scholar]
- 8. Wallace DM, Ramos AR, Rundek T. Sleep disorders and stroke. Int J Stroke. 2012;7:231–242. doi: 10.1111/j.1747-4949.2011.00760.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Pepin JL, Martinez‐Garcia MA. Sleep disorders and cerebrovascular disease: the long and winding road. Eur Respir J. 2020;55:1901977. doi: 10.1183/13993003.01977-2019 [DOI] [PubMed] [Google Scholar]
- 10. McDermott M, Brown DL, Chervin RD. Sleep disorders and the risk of stroke. Expert Rev Neurother. 2018;18:523–531. doi: 10.1080/14737175.2018.1489239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Fu X, Wan XJ, Liu JY, Sun Q, Shen Y, Li J, Mao CJ, Ma QH, Wang F, Liu CF. Effects of sleep fragmentation on white matter pathology in a rat model of cerebral small vessel disease. Sleep. 2023;47:zsad225. doi: 10.1093/sleep/zsad225 [DOI] [PubMed] [Google Scholar]
- 12. Cai H, Wang XP, Yang GY. Sleep disorders in stroke: an update on management. Aging Dis. 2021;12:570–585. doi: 10.14336/AD.2020.0707 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Calandra‐Buonaura G, Provini F, Guaraldi P, Plazzi G, Cortelli P. Cardiovascular autonomic dysfunctions and sleep disorders. Sleep Med Rev. 2016;26:43–56. doi: 10.1016/j.smrv.2015.05.005 [DOI] [PubMed] [Google Scholar]
- 14. Korostovtseva L, Bochkarev M, Sviryaev Y. Sleep and cardiovascular risk. Sleep Med Clin. 2021;16:485–497. doi: 10.1016/j.jsmc.2021.05.001 [DOI] [PubMed] [Google Scholar]
- 15. Bassetti CL. Sleep and stroke. Semin Neurol. 2005;25:19–32. doi: 10.1055/s-2005-867073 [DOI] [PubMed] [Google Scholar]
- 16. Khot SP, Morgenstern LB. Sleep and stroke. Stroke. 2019;50:1612–1617. doi: 10.1161/STROKEAHA.118.023553 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Ma C, Pavlova M, Liu Y, Liu Y, Huangfu C, Wu S, Gao X. Probable REM sleep behavior disorder and risk of stroke: a prospective study. Neurology. 2017;88:1849–1855. doi: 10.1212/WNL.0000000000003902 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wu Z, Jin C, Vaidya A, Jin W, Huang Z, Wu S, Gao X. Longitudinal patterns of blood pressure, incident cardiovascular events, and all‐cause mortality in normotensive diabetic people. Hypertension. 2016;68:71–77. doi: 10.1161/HYPERTENSIONAHA.116.07381 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Tu ZZ, Yuan Y, Xia PF, Lu Q, Chen SH, Wang GD, Zheng MY, Zhang YB, Chen JX, Zhou YF, et al. Trajectories of metabolic risk factors during the development of type 2 diabetes in Chinese adults. Diabetes Metab. 2022;48:101348. doi: 10.1016/j.diabet.2022.101348 [DOI] [PubMed] [Google Scholar]
- 20. Wu S, An S, Li W, Lichtenstein AH, Gao J, Kris‐Etherton PM, Wu Y, Jin C, Huang S, Hu FB, et al. Association of trajectory of cardiovascular health score and incident cardiovascular disease. JAMA Netw Open. 2019;2:e194758. doi: 10.1001/jamanetworkopen.2019.4758 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wong JC, Li J, Pavlova M, Chen S, Wu A, Wu S, Gao X. Risk factors for probable REM sleep behavior disorder: a community‐based study. Neurology. 2016;86:1306–1312. doi: 10.1212/WNL.0000000000002414 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Li SX, Wing YK, Lam SP, Zhang J, Yu MWM, Ho CKW, Tsoh J, Mok V. Validation of a new REM sleep behavior disorder questionnaire (RBDQ‐HK). Sleep Med. 2010;11:43–48. doi: 10.1016/j.sleep.2009.06.008 [DOI] [PubMed] [Google Scholar]
- 23. Shen SS, Shen Y, Xiong KP, Chen J, Mao CJ, Huang JY, Li J, Han F, Liu CF. Validation study of REM sleep behavior disorder questionnaire‐Hong Kong (RBDQ‐HK) in east China. Sleep Med. 2014;15:952–958. doi: 10.1016/j.sleep.2014.03.020 [DOI] [PubMed] [Google Scholar]
- 24. Soldatos CR, Dikeos DG, Paparrigopoulos TJ. Athens insomnia scale: validation of an instrument based on ICD‐10 criteria. J Psychosom Res. 2000;48:555–560. doi: 10.1016/s0022-3999(00)00095-7 [DOI] [PubMed] [Google Scholar]
- 25. Mw J. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14:545. doi: 10.1093/sleep/14.6.540 [DOI] [PubMed] [Google Scholar]
- 26. Chung F, Abdullah HR, Liao P. STOP‐Bang questionnaire: a practical approach to screen for obstructive sleep apnea. Chest. 2016;149:631–638. doi: 10.1378/chest.15-0903 [DOI] [PubMed] [Google Scholar]
- 27. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239–241. doi: 10.2307/2335876 [DOI] [Google Scholar]
- 28. Dey T, Mukherjee A, Chakraborty S. A practical overview and reporting strategies for statistical analysis of survival studies. Chest. 2020;158:S39–S48. doi: 10.1016/j.chest.2020.03.015 [DOI] [PubMed] [Google Scholar]
- 29. Baron RM, Kenny DA. The moderator‐mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51:1173–1182. doi: 10.1037//0022-3514.51.6.1173 [DOI] [PubMed] [Google Scholar]
- 30. Arkin AM, Toth MF, Baker J, Hastey JM. The frequency of sleep talking in the laboratory among chronic sleep talkers and good dream recallers. J Nerv Ment Dis. 1970;151:369–374. doi: 10.1097/00005053-197012000-00002 [DOI] [PubMed] [Google Scholar]
- 31. Arnulf I, Uguccioni G, Gay F, Baldayrou E, Golmard JL, Gayraud F, Devevey A. What does the sleeping brain say? Syntax and semantics of sleep talking in healthy subjects and in parasomnia patients. Sleep. 2017;40:40. doi: 10.1093/sleep/zsx159 [DOI] [PubMed] [Google Scholar]
- 32. Xiao Z, Xie M, You Y, Wu H, Zhou G, Li M. Wake‐up stroke and sleep‐disordered breathing: a meta‐analysis of current studies. J Neurol. 2018;265:1288–1294. doi: 10.1007/s00415-018-8810-2 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figures S1–S4