Abstract
Study Objectives:
Our cohort study aimed to study the association between sleep duration and risk of mortality due to respiratory diseases.
Methods:
We included 498,200 participants from UK Biobank (2006–2021). We classified sleep duration as short sleep duration (< 7 hours), long sleep duration (> 9 hours for adults, > 8 hours for older adults), and midrange sleep duration (7–9 hours). We used the Cox proportional hazards model and restricted cubic spline analysis to explore the association between sleep duration and respiratory diseases mortality.
Results:
During a median follow-up of 12.49 years, 2,477 deaths due to respiratory diseases were recorded, of which 1,099 were deaths due to chronic lower respiratory diseases. Cox models with penalized splines showed U-shaped associations of sleep duration with mortality due to total respiratory diseases and chronic lower respiratory diseases. Compared with midrange sleep duration, short sleep duration was associated with 14% higher risk of total respiratory diseases mortality (hazard ratio = 1.14; 95% confidence interval: 1.04, 1.25), and long sleep duration was associated with 35% higher risk of total respiratory diseases mortality (hazard ratio = 1.35; 95% confidence interval: 1.19, 1.55), after adjustment of baseline characteristics, health status, and lifestyle habits. Similarly, the hazard ratios for chronic lower respiratory diseases mortality were 1.20 (95% confidence interval: 1.04, 1.38) and 1.44 (95% confidence interval: 1.19, 1.74), respectively.
Conclusions:
There was a U-shaped association between sleep duration and respiratory diseases mortality. Appropriate sleep duration may improve the progress of respiratory diseases.
Citation:
Du M, Liu M, Liu J. U-shaped association between sleep duration and the risk of respiratory diseases mortality: a large prospective cohort study from UK Biobank. J Clin Sleep Med. 2023;19(11):1923–1932.
Keywords: sleep duration, respiratory diseases, mortality, cohort, U-shaped
BRIEF SUMMARY
Current Knowledge/Study Rationale: Although the sleep quality of patients with respiratory diseases has been well demonstrated, a comprehensive evaluation of the impact of sleep duration on respiratory disease mortality has not been undertaken.
Study Impact: To our knowledge, this is the first study to examine how sleep duration associates with mortality due to eight specific types of respiratory diseases. Our study of a large cohort with high-quality mortality data revealed U-shaped associations between sleep duration and mortality due to total respiratory diseases and chronic lower respiratory diseases. This study suggests that sleep duration may be a pragmatic target for interventions to improve the progress of respiratory diseases.
INTRODUCTION
Respiratory diseases are some of the most fatal diseases and consistently cause a large burden across health systems.1 In 2017, lower respiratory infections and tuberculosis resulted in 2.5 million and 1.2 million deaths, respectively.1 Mortality due to respiratory disease, especially obstructive, interstitial, and infectious respiratory disease, was high in the United Kingdom between 1985 and 2015.2 The World health organization reported that over 3 million people died each year because of chronic obstructive pulmonary disease (COPD), which accounted for nearly 6% of all deaths worldwide.3 Chronic respiratory diseases were the third leading cause of death in 2017, behind cardiovascular diseases and neoplasms.4 Deaths due to chronic respiratory diseases numbered 3. 9 million in 2017, an increase of 18.0% since 1990.4 Respiratory diseases including lung cancer, COPD, and lower respiratory tract infections are consistently becoming the leading causes of death across different health systems.1–3
Lifestyle factors are commonly viewed as important influencing factors for respiratory systems. Healthy lifestyles might alleviate the health burden of respiratory diseases.5,6 Numerous studies have examined the contribution of an individual lifestyle factor or several lifestyle factors to respiratory disease mortality, including consumption of tea, smoking, alcohol consumption, physical activity, and diet.5,7 Recently, sleep health has been considered as a key factor influencing all-cause, cardiovascular, and cancer-specific mortality.8–10 However, limited studies focus on the association between sleep and respiratory diseases. A number of studies demonstrated the effects of respiratory diseases on sleep.11 Sleep is an active and regulated process with restorative functions for physical and mental conditions.12 In fact, it modulates autonomous nervous system functions including respiration, and thus poor sleep quality may also predict more rapid disease progression and even mortality.11,13 Using a cohort study to explore the association of sleep and respiratory diseases mortality is critical to understand the possible mutual relationship.
However, up to now there has been a lack of cohort studies to evaluate the relationship of sleep with respiratory disease mortality among a general population. In order to help clarify the association, we used large-sample-size cohort data from the UK Biobank to evaluate the effect of sleep duration on different types of respiratory disease mortality including acute upper respiratory infections, chronic lower respiratory diseases, and lung diseases due to external agents, after controlling for baseline characteristics, health status, and lifestyle habits.
METHODS
Study population
We used data from the UK Biobank (application 79114), a large prospective cohort study of 502,414 participants aged 37–73 years. All participants were registered with the UK National Health Service (NHS) and resided within 40 km of 1 of 22 assessment centers across England, Wales, and Scotland between 2006 and 2021. Sociodemographic, lifestyle, and health-related information was assessed by using a touchscreen at assessment centers. Participants also received physical measurements and provided biological samples. Details have been described elsewhere.14,15 We excluded 4,214 (0.84%) participants with missing or incomplete information on sleep duration, yielding an analytic cohort of 498,200 participants. The UK Biobank study was approved by the Northwest Multi-Center Research Ethics Committee (reference no. 21/NW/0157). All participants provided written informed consent to participate. Information about ethics oversight can be found at https://www.ukbiobank.ac.uk/ethics/.
Assessment of sleep duration
Sleep duration was assessed by the question, “About how many hours sleep do you get in every 24 hours (please include naps)?” The answer could only contain integer values. Participants answered either the number of average hours of sleep, “Do not know,” or “Prefer not to answer.” If participants answered < 1 or > 23 hours for their sleep duration, they were excluded from further analysis. If participants answered > 12 hours for their sleep duration, they were asked to confirm their response. If participants activated the help button, they were shown the following message: “If the time you spend sleeping varies a lot, give the average time for a 24-hour day in the last 4 weeks.” Sleep duration was categorized into 3 groups: short sleep duration, midrange sleep duration, and long sleep duration based on the National Sleep Foundation’s sleep time duration recommendations for adults and older adults.16 The recommended sleep duration was 7 to 9 hours for adults and 7 to 8 hours for older adults.
Assessment of respiratory diseases mortality
The primary outcomes including vital status, date of death, and underlying primary cause of death were provided by the NHS Information Centre (England and Wales) and the NHS Central Register (Scotland) to June 30, 2020.17 Specific causes of death were defined using the following codes from the International Classification of Diseases, 10th Revision (ICD-10): J09–J18 influenza and pneumonia; J20–J22 other acute lower respiratory infections (acute bronchitis, acute bronchiolitis, and unspecified acute lower respiratory infection); J40–J47 chronic lower respiratory diseases; J60–J70 lung diseases due to external agents; J80–J84 other respiratory diseases principally affecting the interstitium (adult respiratory distress syndrome, pulmonary edema, pulmonary eosinophilia, not elsewhere classified, other interstitial pulmonary diseases); J85–J86 suppurative and necrotic conditions of lower respiratory tract; J90–J94, J95–J99, and J30–J39 other diseases of the respiratory system.
Assessment of covariates
We used the baseline questionnaire to assess the potential confounders including baseline characteristics, health status, and lifestyle habits. Baseline characteristics included age, sex, race and ethnicity, educational level, and Townsend deprivation index. The Townsend deprivation index was used as an indicator of socioeconomic status, and negative Townsend deprivation index values indicated relative affluence.18 Health status included body mass index, waist circumference, general health, cancer, diabetes, cardiovascular disease, poor psychological status, and family history of diseases (stroke, high blood pressure, chronic bronchitis/emphysema, Alzheimer’s disease/dementia, type 2 diabetes, Parkinson’s disease, severe depression, lung cancer, bowel cancer, and breast cancer). The body mass index was constructed from standing height and weight measured during the initial assessment center visit. Participants reported perceived general health and comorbid conditions. Poor psychological status was investigated by self-reported responses to questions including “Have you seen a doctor for nerves, anxiety, tension or depression?” and “Have you seen a psychiatrist for nerves, anxiety, tension or depression?” If the response was yes for one of these questions, poor psychological status was defined as yes. Lifestyle habits included morning chronotype; getting up in the morning; napping during the day; daytime sleepiness/insomnia; snoring; daytime dozing/sleeping (narcolepsy); smoking status; alcohol consumption; physical activity; consumption of fruits, vegetables, red meat, processed meat, oily fish (eg, sardines, salmon, mackerel, and herring), coffee, and tea; and temperature of hot drinks consumed. Except for physical activity and diet, other lifestyle factors were assessed by self-reported questionnaire. Physical activity was assessed by the international physical activity questionnaire. We categorized participants into low, moderate, and high activity level groups based on categorical criteria.19 Participants self-reported their usual intake of 17 preselected foods and beverages at baseline using a touchscreen food frequency questionnaire.20 Processed meat, red meat (unprocessed beef, lamb/mutton, or pork), fruit intake (fresh, dried), vegetable intake (cooked, raw), and oily fish consumption were categorized from lowest to highest quartile (Q1, Q2, Q3, and Q4).20
Statistical analysis
Baseline characteristics were presented as mean ± standard deviation for continuous variables and number (percentage) for categorical variables. Differences among groups were tested by analysis of variance for continuous variables and χ2 test for categorical variables. Dose–response relationships were examined using restricted cubic spline analysis with knots at 5, 6, 7, 8, 9, and 10 hours between sleep duration and mortality.
Cox proportional hazards regression was used to estimate hazard ratios and 95% confidence intervals (CIs) for the prospective association of sleep duration with mortality. The proportional hazards assumption was tested using Schoenfeld residuals. We adjusted for baseline age (continuous), sex (male, female), Townsend deprivation index (continuous), educational level (college or university degree, advanced [A] levels/advanced subsidiary [AS] levels or equivalent, general certification of education ordinary [O]/general certificate of secondary education [GCSE] or equivalent, certificate of secondary education [CSE] or equivalent, national vocational qualification [NVQ]/higher national diploma [HND]/higher national certificate [HNC] equivalent, other professional qualifications, no degree), race and ethnicity (White, Black, Asian, mixed, other), body mass index (continuous), waist circumference (continuous), general health (excellent, good, fair, poor), cancer (yes, no), diabetes (yes, no), cardiovascular disease (yes, no), poor psychological status (yes, no) and family history (yes, no, unknown), morning chronotype (definitely a “morning” person, more a “morning” than “evening” person, more an “evening” than a “morning” person, definitely an “evening” person), getting up in the morning (not at all easy, not very easy, fairly easy, very easy), napping during the day (never/rarely, sometimes, usually), daytime sleepiness (never/rarely, sometimes, usually), snoring (yes, no), daytime dozing (never/rarely, sometimes, often/all of the time), smoking status (never, former, current), alcohol consumption (never, former, current), physical activity (low, moderate, high), fruit intake, vegetables intake, red meat intake, processed meat intake, oily fish intake, hot drink temperature (Q1, Q2, Q3, Q4), coffee intake (yes, no), and consumption of tea (continuous). Detailed information on the missing covariates is presented in Table S1 (211.9KB, pdf) in the supplemental material, and we used multiple imputation by chained equations to impute any missing covariate values.21 All covariates were included in the multiple imputation model.
In order to explore the effect of other sleep characteristics including morning chronotype, getting up in the morning, napping during the day, daytime sleepiness, snoring, and daytime dozing on the relationship between sleep duration and mortality, we also did subgroup analysis. We performed the following sensitivity analyses to assess the robustness of the results. We excluded participants who had an outcome event during the first 5 years of follow-up. Moreover, we excluded participants with missing covariates in the sensitivity analyses.
All analyses were done using STATA software, version 17.0 for Windows (StataCorp LLC, College Station, Texas). Two-sided P values less than .05 were considered statistically significant.
RESULTS
Population characteristics
Table 1 shows baseline characteristics of participants from UK Biobank. Among 498,200 participants (56.53 ± 8.09 years of age, 45.65% men), 357,451 (71.75%), 123,230 (24.74%), and 17,519 (3.52%) reported midrange, short, and long sleep duration, respectively.
Table 1.
Baseline characteristics of participants.
| Characteristics | n | Midrange Sleep Duration, n (%) | Short Sleep Duration, n (%) | Long Sleep Duration, n (%) |
|---|---|---|---|---|
| Total | 498,200 | 357,451 (71.75) | 123,230 (24.74) | 17,519 (3.52) |
| Mean age (SD), years | 56.53 (8.09) | 56.30 (8.10) | 56.36 (7.86) | 62.25 (7.50) |
| Sex | ||||
| Male | 227,433 | 161,305 (45.13) | 57,407 (46.59) | 8,721 (49.78) |
| Female | 196,146 | 196,146 (54.87) | 65,823 (53.41) | 8,798 (50.22) |
| Race and ethnicity | ||||
| White | 471,532 | 341,374 (95.50) | 113,576 (92.17) | 16,582 (94.65) |
| Black | 7,983 | 4,010 (1.12) | 3,682 (2.99) | 291 (1.66) |
| Asian | 11,300 | 7,466 (2.09) | 3,433 (2.79) | 401 (2.29) |
| Mixed | 2,920 | 1,856 (0.52) | 974 (0.79) | 90 (0.51) |
| Other | 4,465 | 2,745 (0.77) | 1,565 (1.27) | 155 (0.88) |
| Educational level | ||||
| College or university degree | 160,744 | 122,688 (34.32) | 12,811 (10.40) | 1,400 (7.99) |
| A/AS levels or equivalent | 55,136 | 40,925 (11.45) | 12,811 (10.40) | 1,400 (7.99) |
| O/GCSEs or equivalent | 104,679 | 75,297 (21.06) | 26,193 (21.26) | 3,189 (18.20) |
| CSEs or equivalent | 26,705 | 18,554 (5.19) | 7,493 (6.08) | 658 (3.76) |
| NVQ/HND/HNC equivalent | 32,491 | 22,448 (6.28) | 8,657 (7.03) | 1,386 (7.91) |
| Other professional qualifications | 114,697 | 74,855 (20.94) | 32,475 (26.35) | 7,367 (42.05) |
| No degree | 3,748 | 2,684 (0.75) | 942 (0.76) | 122 (0.70) |
| Mean Townsend deprivation index (SD) | −1.31 (3.08) | −1.49 (2.98) | −0.87 (3.28) | −0.88 (3.28) |
| Mean BMI (SD), kg/m2 | 27.43 (4.80) | 27.18 (4.64) | 27.98 (5.11) | 28.60 (5.21) |
| Mean waist circumference (SD), cm | 90.29 (13.48) | 89.64 (13.21) | 91.56 (13.95) | 94.64 (14.00) |
| General health | ||||
| Excellent | 81,887 | 64,616 (18.08) | 15,470 (12.55) | 1,801 (10.28) |
| Good | 289,844 | 215,037 (60.16) | 66,554 (54.01) | 8,253 (47.11) |
| Fair | 104,305 | 66,900 (18.72) | 32,367 (26.27) | 5,038 (28.76) |
| Poor | 22,164 | 10,898 (3.05) | 8,839 (7.17) | 2,427 (13.85) |
| Cancer | ||||
| No | 459,731 | 330,535 (92.47) | 113,748 (92.31) | 15,448 (88.18) |
| Yes | 38,469 | 26,916 (7.53) | 9,482 (7.69) | 2,071 (11.82) |
| Diabetes | ||||
| No | 471,990 | 340,774 (95.33) | 115,808 (93.98) | 15,408 (87.95) |
| Yes | 26,210 | 16,677 (4.67) | 7,422 (6.02) | 2,111 (12.05) |
| Poor psychological status | ||||
| No | 326,123 | 240,173 (67.19) | 75,965 (61.64) | 9,985 (57.00) |
| Yes | 172,077 | 117,278 (32.81) | 47,265 (38.36) | 7,534 (43.00) |
| Cardiovascular disease | ||||
| No | 349,867 | 257,666 (72.08) | 82,840 (67.22) | 9,361 (53.43) |
| Yes | 148,333 | 99,785 (27.92) | 40,390 (32.78) | 8,158 (46.57) |
| Family history | ||||
| No | 38,551 | 28,570 (7.99) | 8,822 (7.16) | 1,159 (6.62) |
| Yes | 446,939 | 320,548 (89.68) | 110,632 (89.78) | 15,759 (89.95) |
| Unknown | 12,710 | 8,333 (2.33) | 3,776 (3.06) | 601 (3.43) |
| Morning chronotype | ||||
| Definitely “morning” person | 134,195 | 91,921 (25.72) | 37,780 (30.66) | 4,494 (25.65) |
| More “morning” than “evening” person | 198,506 | 146,350 (40.94) | 45,889 (37.24) | 6,267 (35.77) |
| More “evening” than “morning” person | 125,700 | 92,718 (25.94) | 28,158 (22.85) | 4,824 (27.54) |
| Definitely “evening” person | 39,799 | 26,462 (7.40) | 11,403 (9.25) | 1,934 (11.04) |
| Getting up in morning | ||||
| Not at all easy | 19,467 | 11,261 (3.15) | 6,364 (5.16) | 1,842 (10.51) |
| Not very easy | 73,542 | 51,037 (14.28) | 19,393 (15.74) | 3,112 (17.76) |
| Fairly easy | 245,497 | 183,003 (51.20) | 55,222 (44.81) | 7,272 (41.51) |
| Very easy | 159,694 | 112,150 (31.37) | 42,251 (34.29) | 5,293 (30.21) |
| Nap during day | ||||
| Never/rarely | 280,153 | 203,682 (56.98) | 71,705 (58.19) | 4,766 (27.20) |
| Sometimes | 191,493 | 136,738 (38.25) | 46,166 (37.46) | 8,589 (49.03) |
| Usually | 26,554 | 17,031 (4.76) | 5,359 (4.35) | 4,164 (23.77) |
| Daytime sleepiness | ||||
| Never/rarely | 120,593 | 97,606 (27.31) | 18,175 (14.75) | 4,812 (27.47) |
| Sometimes | 238,020 | 186,173 (52.08) | 43,325 (35.16) | 8,522 (48.64) |
| Usually | 139,587 | 73,672 (20.61) | 61,730 (50.09) | 4,185 (23.89) |
| Snoring | ||||
| No | 323,549 | 231,116 (64.66) | 81,643 (66.25) | 10,790 (61.59) |
| Yes | 174,651 | 126,335 (35.34) | 41,587 (33.75) | 6,729 (38.41) |
| Daytime dozing/sleeping (narcolepsy) | ||||
| Never/rarely | 378,465 | 279,701 (78.25) | 87,807 (71.25) | 10,957 (62.54) |
| Sometimes | 105,925 | 70,305 (19.67) | 30,532 (24.78) | 5,088 (29.04) |
| Often/all of the time | 13,810 | 7,445 (2.08) | 4,891 (3.97) | 1,474 (8.41) |
| Smoking status | ||||
| Never | 272,688 | 199,606 (55.84) | 64,903 (52.67) | 8,179 (46.69) |
| Former | 173,048 | 123,291 (34.49) | 42,631 (34.59) | 7,126 (40.68) |
| Current | 52,464 | 34,554 (9.67) | 15,696 (12.74) | 2,214 (12.64) |
| Alcohol use status | ||||
| Never | 22,087 | 14,281 (4.00) | 6,616 (5.37) | 1,190 (6.79) |
| Former | 18,317 | 11,507 (3.22) | 5,658 (4.59) | 1,152 (6.58) |
| Current | 457,796 | 331,663 (92.79) | 110,956 (90.04) | 15,177 (86.63) |
| Physical activity level | ||||
| Low | 56,682 | 39,379 (11.02) | 14,932 (12.12) | 2,371 (13.53) |
| Moderate | 164,655 | 121,216 (33.91) | 38,161 (30.97) | 5,278 (30.13) |
| High | 276,863 | 196,856 (55.07) | 70,137 (56.92) | 9,870 (56.34) |
| Vegetable intake, servings/day | ||||
| Q1 | 172,183 | 122,218 (34.19) | 43,879 (35.61) | 6,086 (34.74) |
| Q2 | 101,543 | 74,012 (20.71) | 24,116 (19.57) | 3,415 (19.49) |
| Q3 | 97,996 | 69,114 (19.34) | 25,227 (20.47) | 3,655 (20.86) |
| Q4 | 126,478 | 92,107 (25.77) | 30,008 (24.35) | 4,363 (24.90) |
| Fruit intake, servings/day | ||||
| Q1 | 136,638 | 95,747 (26.79) | 35,603 (28.89) | 5,288 (30.18) |
| Q2 | 205,078 | 149,523 (41.83) | 48,746 (39.56) | 6,809 (38.87) |
| Q3 | 97,049 | 68,736 (19.23) | 24,796 (20.12) | 3,517 (20.08) |
| Q4 | 59,435 | 43,445 (12.15) | 14,085 (11.43) | 1,905 (10.87) |
| Oily fish intake, times per week | ||||
| Never | 54,732 | 36,671 (10.26) | 15,929 (12.93) | 2,132 (12.17) |
| 1 | 353,840 | 257,363 (72.00) | 84,664 (68.70) | 11,813 (67.43) |
| 2–6 | 88,438 | 62,656 (17.53) | 22,278 (18.08) | 3,504 (20.00) |
| 7 | 1,190 | 761 (0.21) | 359 (0.29) | 70 (0.40) |
| Processed meat intake, times per week | ||||
| Never | 46,580 | 32,651 (9.13) | 12,350 (10.02) | 1,579 (9.01) |
| 1 | 297,496 | 215,912 (60.40) | 71,555 (58.07) | 10,029 (57.25) |
| 2–6 | 150,074 | 106,275 (29.73) | 38,072 (30.90) | 5,727 (32.69) |
| 7 | 4,050 | 2,613 (0.73) | 1,253 (1.02) | 184 (1.05) |
| Red meat intake, times per week | ||||
| Never | 33,590 | 23,735 (6.64) | 8,842 (7.18) | 1,013 (5.78) |
| 1 | 389,870 | 281,258 (78.68) | 95,379 (77.40) | 13,233 (75.54) |
| 2–6 | 74,046 | 52,075 (14.57) | 18,743 (15.21) | 3,228 (18.43) |
| 7 | 694 | 383 (0.11) | 266 (0.22) | 45 (0.26) |
| Hot drink temperature | ||||
| Does not drink | 5,294 | 3,529 (0.99) | 1,574 (1.28) | 191 (1.09) |
| Very hot | 84,831 | 58,729 (16.43) | 23,209 (18.83) | 2,893 (16.51) |
| Hot | 330,508 | 240,301 (67.23) | 78,866 (64.00) | 11,341 (64.74) |
| Warm | 77,567 | 54,892 (15.36) | 19,581 (15.89) | 3,094 (17.66) |
| Coffee intake | ||||
| No | 114,093 | 79,426 (22.22) | 30,144 (24.46) | 4,523 (25.82) |
| Yes | 384,107 | 278,025 (77.78) | 93,086 (75.54) | 12,996 (74.18) |
| Mean consumption of tea (SD) cups/d | 3.41 (2.90) | 3.38 (2.80) | 3.44 (3.11) | 3.67 (3.28) |
All P values < .0001. BMI = body mass index, Q = quartile, SD = standard deviation.
Adults with short sleep duration were more likely to identify as non-White and be less educated, from a lower social class, to have high physical activity, to definitely be a “morning” person, to get up in the morning very easily, never/rarely to nap during day, usually to have daytime sleepiness, less snoring, and more daytime dozing, and to be current smokers.
Adults of long sleep duration were more likely to be older, male, and less educated, to have higher body mass index and waist circumference and poorer general health, to report cancer, diabetes, poor psychological status, cardiovascular disease and family history of diseases, and to have low physical activity, be more of an evening person, not get up in the morning easily, nap during the day, have less daytime sleepiness, snore, doze in the daytime, smoke less, and drink less alcohol and coffee and more tea.
Sleep duration and respiratory diseases mortality
During a median follow-up of 12.49 years (interquartile range, 11.57–13.30 years; total person-years, 5,619,335), 2,477 deaths due to respiratory diseases were recorded (including 582 influenza and pneumonia, 63 other acute lower respiratory infections, 1,099 chronic lower respiratory diseases, 69 lung diseases due to external agents, 580 other respiratory diseases principally affecting the interstitium, 11 suppurative and necrotic conditions of lower respiratory tract, and 43 other diseases of the respiratory system).
Cox models with penalized splines showed statistically significant U-shaped associations of sleep duration with mortality due to total respiratory diseases (P < .0001) and chronic lower respiratory diseases (P < .001). However, the U-shaped associations were not statistically significant for other type of respiratory diseases mortality (Figure 1 and Figure S1 (211.9KB, pdf) ). When sleep duration was categorized into 3 groups, compared with the midrange sleep duration group the short sleep duration and long sleep duration groups had higher risk of total respiratory diseases mortality after adjustment for all covariates, with hazard ratios of 1.14 (1.04, 1.25) and 1.35 (1.19, 1.55), respectively. Similarly, the respective estimates for chronic lower respiratory diseases mortality were 1.20 (1.04, 1.38) and 1.44 (1.19, 1.74) (Table 2). Table S2 (211.9KB, pdf) presents results from the sensitivity analysis, which showed relatively consistent results when we excluded participants with missing covariates. Results were slightly attenuated when we excluded participants who had an outcome event during the first 5 years of follow-up.
Figure 1. Multivariable Cox regression model with penalized splines on association of sleep duration and total respiratory diseases mortality in the UK Biobank.
All models were adjusted for age, sex, Townsend deprivation score, educational level, race and ethnicity, body mass index, waist circumference, perceived general health, cancer, diabetes, cardiovascular disease, poor psychological status and family history of diseases, smoking status, alcohol status, physical activity, fruit intake, vegetables intake, red meat intake, processed meat intake, oily fish intake, coffee intake, consumption of tea, and hot drink temperature. The solid line represents hazard ratio; the dotted line represents 95% confidence interval.
Table 2.
Associations of sleep duration with respiratory diseases mortality.
| Outcomes | Midrange Sleep Duration | Short Sleep Duration | Long Sleep Duration |
|---|---|---|---|
| HR (95% CI) | HR (95% CI) | ||
| Total respiratory diseases | |||
| Number | 357,451 | 123,230 | 17,519 |
| Death, n | 1,402 | 753 | 292 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.42 (1.30, 1.55) | 1.99 (1.75, 2.27) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.14 (1.04, 1.25) | 1.49 (1.31, 1.70) |
| Multivariable-adjusted* | 1 (Ref) | 1.14 (1.04, 1.25) | 1.35 (1.19, 1.55) |
| Influenza and pneumonia | |||
| Death, n | 345 | 173 | 64 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.35 (1.12, 1.62) | 1.95 (1.49, 2.56) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.14 (0.94, 1.37) | 1.49 (1.13, 1.96) |
| Multivariable-adjusted* | 1 (Ref) | 1.12 (0.92, 1.36) | 1.31 (0.99, 1.74) |
| Other acute lower respiratory infections | |||
| Death, n | 37 | 17 | 9 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.28 (0.72, 2.28) | 2.41 (1.14, 5.10) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.03 (0.57, 1.86) | 1.72 (0.81, 3.65) |
| Multivariable-adjusted* | 1 (Ref) | 1.03 (0.56, 1.90) | 1.57 (0.72, 3.42) |
| Chronic lower respiratory diseases | |||
| Death, n | 583 | 367 | 149 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.58 (1.39, 1.81) | 2.37 (1.97, 2.85) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.17 (1.02, 1.33) | 1.63 (1.35, 1.96) |
| Multivariable-adjusted* | 1 (Ref) | 1.20 (1.04, 1.38) | 1.44 (1.19, 1.74) |
| Lung diseases due to external agents | |||
| Death, n | 41 | 23 | 5 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.51 (0.90, 2.53) | 1.23 (0.48, 3.16) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.32 (0.78, 2.22) | 0.96 (0.37, 2.47) |
| Multivariable-adjusted* | 1 (Ref) | 1.20 (0.69, 2.07) | 0.86 (0.33, 2.26) |
| Other respiratory diseases principally affecting the interstitium | |||
| Death, n | 358 | 161 | 61 |
| Baseline characteristics-adjusted | 1 (Ref) | 1.28 (1.06, 1.55) | 1.52 (1.15, 2.00) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 1.11 (0.92, 1.35) | 1.26 (0.95, 1.66) |
| Multivariable-adjusted* | 1 (Ref) | 1.09 (0.89, 1.32) | 1.24 (0.94, 1.65) |
| Suppurative and necrotic conditions of lower respiratory tract | |||
| Death, n | 8 | 2 | 1 |
| Baseline characteristics-adjusted | 1 (Ref) | 0.62 (0.13, 2.94) | 1.65 (0.20, 13.87) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 0.53 (0.11, 2.58) | 1.16 (0.14, 9.87) |
| Multivariable-adjusted* | 1 (Ref) | 0.54 (0.11, 2.74) | 1.09 (0.12, 9.96) |
| Other diseases of the respiratory system | |||
| Death, n | 30 | 10 | 3 |
| Baseline characteristics-adjusted | 1 (Ref) | 0.89 (0.43, 1.83) | 1.19 (0.35, 3.96) |
| Baseline characteristics and health status-adjusted | 1 (Ref) | 0.80 (0.39, 1.66) | 0.97 (0.29, 3.25) |
| Multivariable-adjusted* | 1 (Ref) | 0.68 (0.32, 1.45) | 0.99 (0.29, 3.38) |
Models were adjusted for age, sex, Townsend deprivation score, educational level, race and ethnicity, body mass index, waist circumference, perceived general health, cancer, diabetes, cardiovascular disease, poor psychological status and family history of diseases, smoking status, alcohol status, physical activity, fruit intake, vegetables intake, red meat intake, processed meat intake, oily fish intake, coffee intake, consumption of tea, and hot drink temperature. HR = hazard ratio, CI = confidence interval, Ref = reference.
The effect of other sleep characteristics on associations of sleep duration and respiratory diseases
Table S3 (211.9KB, pdf) and Table S4 (211.9KB, pdf) present results from subgroup analyses for associations of sleep duration with mortality due to total respiratory diseases and chronic lower respiratory diseases. The estimates of the long sleep duration group on total respiratory diseases mortality were slightly enhanced for participants who were more an evening person (hazard ratio = 1.63; 95% CI: 1.28, 2.08) and did not snore (hazard ratio = 1.41; 95% CI: 1.20, 1.65); the estimates of the short sleep duration group on total respiratory diseases mortality were slightly enhanced for participants who did not snore (hazard ratio = 1.21; 95% CI: 1.08, 1.35) and sometimes dozed in the daytime (hazard ratio = 1.32; 95% CI: 1.11, 1.57) (Figure 2).
Figure 2. Multivariable Cox regression model on association of sleep duration and total respiratory diseases mortality by subgroups analysis of sleep characteristics in the UK Biobank.
All models were adjusted for age, sex, Townsend deprivation score, educational level, race and ethnicity, body mass index, waist circumference, perceived general health, cancer, diabetes, cardiovascular disease, poor psychological status and family history of diseases, smoking status, alcohol status, physical activity, fruit intake, vegetables intake, red meat intake, processed meat intake, oily fish intake, coffee intake, consumption of tea and hot drink temperature. Red represents short sleep duration; blue represents long sleep duration; point represents HR; line represents 95% CI. CI = confidence interval, HR = hazard ratio.
DISCUSSION
This prospective analysis found that sleep duration was associated with total respiratory disease mortality and chronic lower respiratory disease mortality. Both associations were U-shaped and persisted after adjustment for baseline confounders.
Recently, cross-sectional studies found that poor sleep quality or short sleep duration were common among respiratory disease patients.22–24 Patients with respiratory diseases, especially COPD patients, often have poor sleep quality.11,25–27 However, sleep disorders may indicate more rapid disease progression, even mortality.13 Previous studies reported that healthy lifestyles including healthy diet and sufficient exercise were associated with lower risk of respiratory diseases mortality.6,28 Locke et al proposed that more robust studies were needed to truly understand the relationship between sleep and respiratory diseases.29 However, few studies have analyzed the effect of sleep on respiratory disease mortality. Our results were critical to understand the possible mutual relationship. We found a U-shaped association of sleep duration with mortality due to total respiratory diseases and chronic lower respiratory diseases after controlling for covariates. Short sleep duration and long sleep duration both were associated with higher risk of mortality due to total respiratory diseases and chronic lower respiratory diseases. Our results indicated that maintaining midrange sleep duration may decrease the risk of deaths due to respiratory diseases, especially chronic lower respiratory diseases.
However, our results should be interpreted cautiously. We should pay attention to the influence of the bidirectional association between sleep and respiratory diseases on our results. Sleep duration may be influenced by respiratory illness itself.23,24 Studies reported that daytime and nocturnal symptoms including cough and dyspnea caused poor sleep quality and extreme sleep duration among COPD patients.25,26 Although our study used the cohort design to provide more effective evidence on their casual associations than a cross-sectional study, severity of respiratory illness may influence the changeable degree of sleep quality and sleep duration, which could lead to a misconstrued association between sleep duration and mortality. Up to now, specific potential mechanisms of extreme sleep duration–related respiratory diseases death are still unclear. Inflammation may be the most important biological explanation. Research showed that shorter and longer sleep duration were associated with higher levels of inflammatory cytokine, which may worsen respiratory system injury.29–31 Additionally, sleep fragmentation and deterioration can compound the fatigue and physical exhaustion often experienced by chronic lung disease patients, further reduce quality of life and exacerbate disease.29,32 What’s more, due to lower baseline oxygen saturations at sleep onset and severe episodic desaturation caused by an altered breathing pattern during sleep, COPD patients face sleep hypoxemia during more than 30% of sleep time, especially patients with chronic bronchitis.11 In addition, a reduction in muscle tone and recumbent position also have profound effects on ventilation, even causing ventilation–perfusion mismatch during sleep.11 Furthermore, we found that for total respiratory diseases mortality the risk estimates of long sleep duration were slightly enhanced for participants who were more an evening person and of short sleep duration were slightly enhanced for participants who sometimes dozed in the daytime. The above results may reveal that other sleep indexes also should be assessed. Collop proposed that COPD patients were likely to be asked about daytime drowsiness but were rarely asked about insomnia and other sleep disturbances.11 When considering treatment for sleep disturbances, comprehensive assessment of sleep may benefit underlying COPD as well.11
In our study, adults who had high physical activity had short sleep duration, whereas those who had low physical activity had long sleep duration. Exercise generally increases total sleep duration, because sleep serves to conserve energy and restore the body and thermoregulatory functions.33 However, our results may hint that higher the physical activity was, shorter the sleep duration was. More cohort study considering exercise timepoint was needed to explore their relationship. In addition, current smokers had a higher proportion of short sleep duration, whereas others had a higher proportion of long sleep duration. Patients who smoked had shorter sleep duration due to nicotine in cigarettes, which induces cell proliferation, edema, epithelial thickening, and/or ciliary dysfunction, inducing chronic inflammation of the upper respiratory tract and leading to sleep breathing disorders.34
The major strength of this study was the large sample size, which allowed our results to have sufficient statistical power. Also, UK Biobank used death data from the NHS Information Centre and the NHS Central Register, which helps ensure the accuracy of the data. However, this study still had several limitations. First, information on sleep duration was self-reported, and thus measurement error was inevitable owing to recall error. In addition, we could not capture baseline exposure changes over time. Future studies with repeated measurements would be preferred. Second, although our main analysis adjusted for comorbidities at baseline, considering the short follow-up duration and those who died due to serious diseases at baseline the possibility of residual confounding cannot be fully eliminated. In addition, unmeasured confounding also had an effect on results.
In conclusion, based on a large UK cohort, this prospective analysis found U-shaped associations of sleep duration with mortality due to total respiratory diseases and chronic lower respiratory diseases after adjustment of baseline characteristics, health status, and lifestyle habits. Therefore, promoting healthy sleep duration to improve the progress of respiratory diseases could be considered. Our novel findings may provide evidence for clinical and public health related to respiratory disease management. In addition, objective measures of sleep duration to address this same question are also needed in the future.
DISCLOSURE STATEMENT
All authors have seen and approved the manuscript. This work was partly supported by the National Natural Science Foundation of China (72122001,71934002). The authors report no conflicts of interest.
ACKNOWLEDGMENTS
This study was conducted using the UK Biobank resource (application no. 79114). The UK Biobank was established by the Wellcome Trust, the Medical Research Council, the UK Department of Health, and the Scottish Government. The UK Biobank has also received funding from the Welsh Assembly Government, the British Heart Foundation, and Diabetes United Kingdom. This work used the computational resources of the NIH High-Performance Computing Biowulf cluster.
Data sharing statement: The UK Biobank datasets are openly available by submitting a data request proposal from https://www.ukbiobank.ac.uk/. We are authorized to access the database through the Access Management System (AMS) (application no. 79114).
Author contributions: Min Du: conceptualization, data curation, formal analysis, methodology, writing original draft, writing review and editing. Min Liu: conceptualization, writing review and editing. Jue Liu: conceptualization, writing-review and editing. All authors contributed significantly to the interpretation of the findings, the writing of the manuscript, and approval of the final version.
ABBREVIATIONS
- CI
confidence interval
- COPD
chronic obstructive pulmonary disease
- NHS
National Health Service
REFERENCES
- 1. GBD 2017 Causes of Death Collaborators . Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017 . Lancet. 2018. ; 392 ( 10159 ): 1736 – 1788 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Salciccioli JD , Marshall DC , Shalhoub J , Maruthappu M , De Carlo G , Chung KF . Respiratory disease mortality in the United Kingdom compared with EU15+ countries in 1985-2015: observational study . BMJ. 2018. ; 363 : k4680 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. WHO . Chronic Respiratory Diseases. https://www.who.int/health-topics/chronic-respiratory-diseases#tab=tab_2 . Accessed October 22, 2022. .
- 4. GBD Chronic Respiratory Disease Collaborators . Prevalence and attributable health burden of chronic respiratory diseases, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017 . Lancet Respir Med. 2020. ; 8 ( 6 ): 585 – 596 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Zhu N , Yu C , Guo Y , et al. China Kadoorie Biobank Collaborative Group . Adherence to a healthy lifestyle and all-cause and cause-specific mortality in Chinese adults: a 10-year prospective study of 0.5 million people . Int J Behav Nutr Phys Act. 2019. ; 16 ( 1 ): 98 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wu E , Ni JT , Zhu ZH , Xu HQ , Tao L , Xie T . Association of a healthy lifestyle with all-cause, cause-specific mortality and incident cancer among individuals with metabolic syndrome: a prospective cohort study in UK Biobank . Int J Environ Res Public Health. 2022. ; 19 ( 16 ): 9936 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Inoue-Choi M , Ramirez Y , Cornelis MC , Berrington de González A , Freedman ND , Loftfield E . Tea consumption and all-cause and cause-specific mortality in the UK Biobank: a prospective cohort study . Ann Intern Med. 2022. ; 175 ( 9 ): 1201 – 1211 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Zhou T , Yuan Y , Xue Q , et al . Adherence to a healthy sleep pattern is associated with lower risks of all-cause, cardiovascular and cancer-specific mortality . J Intern Med. 2022. ; 291 ( 1 ): 64 – 71 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Sambou ML , Zhao X , Hong T , et al . Investigation of the relationships between sleep behaviors and risk of healthspan termination: a prospective cohort study based on 323,373 UK-Biobank participants . Sleep Breathing. 2022. ; 26 ( 1 ): 205 – 213 . [DOI] [PubMed] [Google Scholar]
- 10. Li J , Yin J , Luo Y , et al . Association of healthy sleep pattern with the risk of cardiovascular disease and all-cause mortality among people with diabetes: a prospective cohort study . Diabetes Res Clin Pract. 2022. ; 186 : 109822 . [DOI] [PubMed] [Google Scholar]
- 11. Collop N . Sleep and sleep disorders in chronic obstructive pulmonary disease . Respiration. 2010. ; 80 ( 1 ): 78 – 86 . [DOI] [PubMed] [Google Scholar]
- 12. Penzel T , Wessel N , Riedl M , et al . Cardiovascular and respiratory dynamics during normal and pathological sleep . Chaos. 2007. ; 17 ( 1 ): 015116 . [DOI] [PubMed] [Google Scholar]
- 13. Myall KJ , West A , Kent BD . Sleep and interstitial lung disease . Curr Opin Pulm Med. 2019. ; 25 ( 6 ): 623 – 628 . [DOI] [PubMed] [Google Scholar]
- 14. Sudlow C , Gallacher J , Allen N , et al . UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age . PLoS Med. 2015. ; 12 ( 3 ): e1001779 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Palmer LJUK . UK Biobank: bank on it . Lancet. 2007. ; 369 ( 9578 ): 1980 – 1982 . [DOI] [PubMed] [Google Scholar]
- 16. Hirshkowitz M , Whiton K , Albert SM , et al . National Sleep Foundation’s sleep time duration recommendations: methodology and results summary . Sleep Health. 2015. ; 1 ( 1 ): 40 – 43 . [DOI] [PubMed] [Google Scholar]
- 17. Biobank UK . Mortality Data: Linkage to Death Registries. https://biobank.ndph.ox.ac.uk/ukb/ukb/docs/DeathLinkage.pdf . Accessed October 22, 2022.
- 18. Foster HME , Celis-Morales CA , Nicholl BI , et al . The effect of socioeconomic deprivation on the association between an extended measurement of unhealthy lifestyle factors and health outcomes: a prospective analysis of the UK Biobank cohort . Lancet Public Health. 2018. ; 3 ( 12 ): e576 – e585 . [DOI] [PubMed] [Google Scholar]
- 19. Fan M , Lyu J , He P . [Chinese guidelines for data processing and analysis concerning the International Physical Activity Questionnaire] . Zhonghua Liu Xing Bing Xue Za Zhi. 2014. ; 35 ( 8 ): 961 – 964 . [PubMed] [Google Scholar]
- 20. Vu TT , Van Horn L , Achenbach CJ , Rydland KJ , Cornelis MC . Diet and respiratory infections: specific or generalized associations? Nutrients. 2022. ; 14 ( 6 ): 1195 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Liu D , Li ZH , Shen D , et al . Association of sugar-sweetened, artificially sweetened, and unsweetened coffee consumption with all-cause and cause-specific mortality: a large prospective cohort study . Ann Intern Med. 2022. ; 175 ( 7 ): 909 – 917 . [DOI] [PubMed] [Google Scholar]
- 22. Lin Y , Liu S , Li S , Zuo H , Zhang B . Relationships between the changes in sleep patterns and sleep quality among Chinese people during the 2019 coronavirus disease outbreak . Sleep Med. 2022. ; 91 : 154 – 160 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Dumith SC , Meneghini KFD , Demenech LM . Who are the individuals with the worst perceived quality of sleep? A population-based survey in southern Brazil . Prev Med Rep. 2020. ; 21 : 101288 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Habib A , Alam MM , Hussain I , Nasir N , Almuthebi M . Erratic behavioral attitude leads to noncommunicable diseases: a cross-sectional study . BioMed Res Int. 2020. ; 2020 : 7257052 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Soler X , Siddall J , Small M , Stiegler M , Bogart M . The burden of nocturnal symptoms in patients with chronic obstructive pulmonary disease: results of a real-world survey in the USA . Pulm Ther. 2022. ; 8 ( 3 ): 269 – 282 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Singh DP , Jamil RT , Mahajan K . Nocturnal Cough . StatPearls. Treasure Island, FL: : StatPearls Publishing; ; 2022. . [PubMed] [Google Scholar]
- 27. Jayamaha AR , Jones AV , Katagira W , et al . Systematic review of physical activity, sedentary behaviour and sleep among adults living with chronic respiratory disease in low- and middle-income countries . Int J Chron Obstruct Pulmon Dis. 2022. ; 17 : 821 – 854 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Han H , Cao Y , Feng C , et al . Association of a healthy lifestyle with all-cause and cause-specific mortality among individuals with type 2 diabetes: a prospective study in UK Biobank . Diabetes Care. 2022. ; 45 ( 2 ): 319 – 329 . [DOI] [PubMed] [Google Scholar]
- 29. Locke BW , Lee JJ , Sundar KM . OSA and chronic respiratory disease: mechanisms and epidemiology . Int J Environ Res Public Health. 2022. ; 19 ( 9 ): 5473 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Irwin MR , Olmstead R , Carroll JE . Sleep disturbance, sleep duration, and inflammation: a systematic review and meta-analysis of cohort studies and experimental sleep deprivation . Biol Psychiatry. 2016. ; 80 ( 1 ): 40 – 52 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Clemente I , Sinatti G , Cirella A , Santini SJ , Balsano C . Alteration of inflammatory parameters and psychological post-traumatic syndrome in long-COVID patients . Int J Environ Res Public Health. 2022. ; 19 ( 12 ): 7103 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Malek F , Khalil Sayah S , Kia NS , Ghods E . The relationship between sleep quality and quality of life among patients with asthma . Cureus. 2022. ; 14 ( 3 ): e23402 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Driver HS , Taylor SR . Exercise and sleep . Sleep Med Rev. 2000. ; 4 ( 4 ): 387 – 402 . [DOI] [PubMed] [Google Scholar]
- 34. Hsu WY , Chiu NY , Chang CC , Chang TG , Lane HY . The association between cigarette smoking and obstructive sleep apnea . Tob Induc Dis. 2019. ; 17 ( April ): 27 . [DOI] [PMC free article] [PubMed] [Google Scholar]