SUMMARY
Sleep apnea is increasing in prevalence and an important cause of cardiometabolic diseases and mortality worldwide. Its only established modifiable risk factor is obesity; however, up to half of all sleep apnea cases may occur in non-obese persons, and hence there is a pressing need to identify other modifiable risk factors to facilitate more effective prevention. We sought to examine, for the first time, cardiorespiratory fitness (CRF) in relation to risk of sleep apnea, independent of obesity. A national cohort study was conducted to examine CRF in all 1,547,478 Swedish military conscripts during 1969–1997 (97–98% of all 18-year-old men) in relation to risk of sleep apnea through 2012 (maximum age 62 years). CRF was measured as maximal aerobic workload in Watts, and sleep apnea was identified from nationwide outpatient and inpatient diagnoses. A total of 44,612 (2.9%) men were diagnosed with sleep apnea in 43.7 million person-years of follow-up. Adjusting for age, height, weight, socioeconomic factors, and family history of sleep apnea, low CRF at age 18 years was associated with a significantly increased risk of sleep apnea in adulthood (lowest vs. highest CRF tertile: incidence rate ratio [IRR], 1.44; 95% CI, 1.40–1.49; P<0.001; continuous CRF per 100 Watts: IRR, 0.71; 95% CI, 0.70–0.73; P<0.001). An increased risk was observed even among men with normal body mass index (lowest vs. highest CRF tertile: IRR, 1.30; 95% CI, 1.26–1.35; P<0.001). These findings identify low CRF early in life as a new modifiable risk factor for development of sleep apnea in adulthood.
Keywords: exercise, obesity, obstructive sleep apnea, physical fitness, sleep apnea
INTRODUCTION
Sleep apnea is common worldwide and an important cause of cardiometabolic diseases and mortality (Senaratna et al., 2017). Its prevalence in the general population has been estimated to range from 9% to 38% (Senaratna et al., 2017), and has increased over the past two decades (Namen et al., 2016). Sleep apnea has been linked with multiple other health risks including hypertension (Nieto et al., 2000), type 2 diabetes (Botros et al., 2009), coronary heart disease (Gottlieb et al., 2010), heart failure (Gottlieb et al., 2010), arrhythmias (Mehra et al., 2006), stroke (Yaggi et al., 2005), and premature mortality (Punjabi et al., 2009, Yaggi et al., 2005). Given its enormous disease burden and economic costs (Tarasiuk and Reuveni, 2013, American Academy of Sleep Medicine, 2016), more effective prevention of sleep apnea is an important public health priority.
The most common (~84%) and widely studied form of sleep apnea is obstructive sleep apnea (OSA), whereas <1% of cases are central sleep apnea (CSA) and ~15% are mixed (obstructive and central) (Morgenthaler et al., 2006). Male sex, advancing age, and obesity are well-established risk factors for the development of OSA (Young et al., 2004). Because obesity is the only established risk factor that is modifiable, and up to half of all OSA cases may occur in non-obese persons (Young et al., 2005), there is a pressing need to identify other modifiable risk factors that may help guide better prevention. Self-reported high physical activity or exercise has been associated with improved symptoms and outcomes in patients with existing OSA, irrespective of weight loss (Mendelson et al., 2018, Aiello et al., 2016, Iftikhar et al., 2014). However, few studies have examined physical activity in relation to the risk of developing OSA (Murillo et al., 2016, da Silva et al., 2017), and none has examined objectively measured cardiorespiratory fitness (CRF). Self-reported physical activity is a poor proxy for CRF, the underlying physiologic factor affected by physical activity that may more directly influence sleep apnea risk (Lee et al., 2011). To our knowledge, no studies have examined CRF in relation to the risk of sleep apnea.
We sought to provide the first population-based risk estimates for sleep apnea associated with CRF, using prospectively ascertained data in a large national cohort. Specifically, our aim was to examine CRF at age 18 years in relation to the long-term risk of sleep apnea in adulthood, independent of body weight and other potential confounders. The results may identify a new modifiable early-life risk factor that would inform efforts to prevent sleep apnea and its large disease burden across the life course.
METHODS
Study Population
We identified 1,547,478 men (age 18 years) who underwent a military conscription examination in Sweden during 1969–1997. This examination was compulsory for all ~18-year-old men nationally each year except for 2–3% who either were incarcerated or had severe chronic medical conditions or disabilities documented by a physician. This study was approved by the ethics committee of Lund University in Sweden (No. 2010/476), and conformed to the principles embodied in the Declaration of Helsinki.
Cardiorespiratory Fitness, Muscular Strength, Height, and Weight Measurement
CRF, muscular strength, height, and weight measurements were obtained using the Swedish Military Conscription Registry, which contains information from a 2-day standardized physical and psychological examination required for all conscripts starting in 1969 (Crump et al., 2017). CRF was measured as the maximal aerobic workload in Watts, using a well-validated electrically-braked stationary bicycle ergometer test (Nordesjo and Schele, 1974). Following a warm-up period, each conscript performed 5–10 minutes of exercise on a stationary bicycle at a starting work rate of 75 to 175 Watts (determined by a sliding scale based on body weight), and increasing by 25 Watts/minute until volitional exhaustion. Maximal aerobic workload was calculated as the power output in Watts before the last intensity increase, plus the prorated output for the last stage. Maximal aerobic workload is highly correlated with maximal oxygen uptake (VO2 max; correlation ~0.9) (Patton et al., 1982), and its measurement using this bicycle ergometer test is highly reproducible, with a test-retest correlation of 0.95 (Andersen, 1995). CRF measured in this manner was examined alternatively as a continuous variable or categorical variable in tertiles (Watts: low [<240], medium [240–288], high [≥289]).
In addition to CRF, we examined muscular strength as a secondary predictor of interest because it represents a different aspect of physical fitness. Muscular strength was measured in Newtons using well-validated isometric dynamometer tests, and calculated as the weighted sum of maximal knee extension (weighted × 1.3), elbow flexion (weighted × 0.8), and hand grip (weighted × 1.7) (Hook and Tornvall, 1969). Each dynamometer test was performed three times and the maximum value was recorded for analysis, except when the last value was highest, in which case testing was repeated until strength values stopped increasing. All testing equipment was calibrated daily (Nordesjo and Schele, 1974, Hook and Tornvall, 1969). Muscular strength was examined alternatively as a continuous variable or categorical variable in tertiles (Newtons: low [<1900], medium [1900–2170], high [≥2171]).
Height and weight were measured using standard protocols and modeled alternatively as continuous variables or categorical variables in approximate quintiles. Body mass index (BMI) also was examined as an alternative to height and weight. BMI was calculated as body weight in kilograms divided by the square of height in meters, and examined alternatively as a continuous or categorical variable using Centers for Disease Control and Prevention (CDC) definitions for children and adolescents aged 2 to 19 years to facilitate comparability with US studies: “overweight” is defined as ≥85th and <95th percentile and “obesity” as ≥95th percentile on the CDC’s 2000 sex-specific BMI-for-age growth charts, which correspond to BMI ≥25.6 and <29.0 and BMI ≥29.0, respectively, for 18-year-old males (Ogden and Flegal, 2010). In the present study, “normal BMI” refers to <85th percentile on the CDC’s 2000 sex-specific BMI-for-age growth charts, which corresponds to BMI <25.6 for 18-year-old males (Ogden and Flegal, 2010).
Sleep Apnea Ascertainment
The study cohort was followed up for the earliest diagnosis of sleep apnea from the date of the military conscription examination through December 31, 2012. Sleep apnea was identified using International Classification of Diseases (ICD) diagnosis codes in the Swedish Hospital Registry and Swedish Outpatient Registry (ICD-9: 327.2, 780.51, 780.53, 780.57; ICD-10: G47.3). The Swedish Hospital Registry contains all primary and secondary hospital discharge diagnoses from six populous counties in southern Sweden starting in 1964, and with nationwide coverage starting in 1987; and the Swedish Outpatient Registry contains outpatient diagnoses from all specialty clinics nationwide starting in 2001. More specific coding to distinguish obstructive vs. central sleep apnea was unavailable; however, prior data have indicated that the large majority of cases are obstructive (e.g., 84% OSA, <1% CSA, 15% mixed) (Morgenthaler et al., 2006).
Adjustment Variables
Other variables that may be associated with CRF and sleep apnea were obtained from the Swedish Military Conscription Registry and national census data, which were linked using an anonymous personal identification number. The following were used as adjustment variables: year of the military conscription examination (modeled simultaneously as a continuous and categorical [1969–1979, 1980–1989, 1990–1997] variable to account for follow-up time and attained age); highest attained education level during the study period (<12, 12–14, ≥15 years); neighborhood socioeconomic status (SES) at baseline (included because neighborhood characteristics have been associated with sleep apnea (Billings et al., 2016) and with physical activity and BMI (Stoddard et al., 2013); composed of an index that includes low education level, low income, unemployment, and social welfare receipt, as previously described (Crump et al., 2011), and categorized as low [>1 SD below the mean], medium [within 1 SD from the mean], or high [>1 SD above the mean]); and family history of sleep apnea in a parent or sibling (yes or no, identified from diagnoses in the Swedish Hospital Registry during 1964–2012 and the Swedish Outpatient Registry during 2001–2012 using the same diagnostic codes noted above).
Missing data for each variable were imputed using a standard multiple imputation procedure based on the variable’s relationship with all other covariates and sleep apnea (Rubin, 1987). Missing data were relatively infrequent for CRF (5.7%), muscular strength (5.0%), height (7.2%), weight (7.3%), education level (0.4%), and neighborhood SES (9.1%). Data were complete for all other variables.
Statistical Analysis
Poisson regression with robust standard errors was used to compute incidence rate ratios (IRRs) and 95% confidence intervals (CIs) for associations between CRF or other variables and subsequent risk of sleep apnea (Zou, 2004). Two different adjusted models were performed: the first was adjusted for year of the military conscription exam (to account for follow-up time and attained age), and the second additionally included CRF, muscular strength, height and weight (or alternatively BMI), education level, neighborhood SES, and family history of sleep apnea, with continuous variables modeled as such. As a secondary aim, the covariates also were examined in relation to the risk of sleep apnea to identify or confirm other potential risk factors. Poisson model goodness-of-fit was assessed using deviance and Pearson chi-squared tests, which showed a good fit in all models. In sensitivity analyses, we repeated the main analyses after: (1) restricting to individuals with complete data for all variables (N=1,361,083; 88.0%), as an alternative to multiple imputation; and (2) restricting the follow-up period to 2001 and later when both outpatient and inpatient data were available.
Potential interaction between CRF and BMI on either the additive or multiplicative scale was examined in relation to the risk of sleep apnea. Additive interactions were assessed using the “relative excess risk due to interaction” (RERI), which is computed for binary variables as: RERIIRR = IRR11 − IRR10 – IRR01 + 1 (Vanderweele and Knol, 2014). Multiplicative interactions were assessed using the ratio of IRRs: IRR11 / (IRR10 IRR01) (Vanderweele and Knol, 2014). All statistical tests were 2-sided and used an α-level of 0.05. All analyses were conducted using Stata version 14.
RESULTS
Among the 1,547,478 men in this cohort, 44,612 (2.9%) were subsequently diagnosed with sleep apnea in 43.7 million person-years of follow-up (mean follow-up, 28.2 years). The median age at the end of follow-up was 47.2 years (mean 47.4, SD 7.9, range 19.0 to 62.0). The median age at diagnosis of sleep apnea was 46.4 years (mean 45.7, SD 7.5, range 18.4 to 62.0). A total of 10,241 (23.0%) cases were diagnosed at ages <40 years, 20,076 (45.0%) at ages 40–49 years, and 14,295 (32.0%) at ages 50–62 years. Most (83.3%) cases were diagnosed in outpatient settings, consistent with usual clinical practice in Sweden, whereas 16.7% were diagnosed during a hospitalization. Table 1 shows sleep apnea incidence rates by CRF level and other factors.
Table 1.
Characteristics of 18-year-old men who were or were not subsequently diagnosed with sleep apnea, Sweden, 1969–2012.
| No sleep apnea n (%) | Sleep apnea n (%) | Ratea | |
|---|---|---|---|
| Total | 1,502,866 (100.0) | 44,612 (100.0) | 98.9 |
| CRF, Watts (mean ± SD) | 267. 3 ± 54.5 | 253.3 ± 49.5 | |
| Lowest tertile | 492,447 (32.8) | 18,903 (42.4) | 110.3 |
| Middle tertile | 504,671 (33.6) | 15,876 (35.6) | 102.8 |
| Highest tertile | 505,748 (33.6) | 9,833 (22.0) | 78.6 |
| Muscular strength, Newtons (mean ± SD) | 1983.1 ± 454.8 | 2058.7 ± 378.2 | |
| Lowest tertile | 498,373 (33.2) | 12,552 (28.1) | 87.9 |
| Middle tertile | 507,579 (33.7) | 15,973 (35.8) | 99.7 |
| Highest tertile | 496,914 (33.1) | 16,087 (36.1) | 108.8 |
| Height, cm (mean ± SD) | 177.9 ± 7.6 | 177.5 ± 7.9 | |
| <172.0 | 322,337 (21.4) | 10,525 (23.6) | 104.6 |
| 172.0–176.9 | 276,148 (18.4) | 8,029 (18.0) | 97.6 |
| 177.0–179.9 | 336,418 (22.4) | 9,638 (21.6) | 96.1 |
| 180.0–183.9 | 288,558 (19.2) | 8,342 (18.7) | 97.3 |
| ≥184.0 | 279,405 (18.6) | 8,078 (18.1) | 98.5 |
| Weight, kg (mean ± SD) | 68.7 ± 10.3 | 71.4 ± 12.4 | |
| <60.5 | 303,990 (20.2) | 7,656 (17.2) | 79.1 |
| 60.5–64.9 | 317,198 (21.1) | 8,104 (18.2) | 83.3 |
| 65.0–69.9 | 318,908 (21.2) | 8,115 (18.2) | 86.2 |
| 70.0–75.9 | 278,498 (18.5) | 8,135 (18.2) | 99.8 |
| ≥76.0 | 284,272 (18.9) | 12,602 (28.2) | 155.0 |
| Body mass index (mean ± SD) | 21.6 ± 2.8 | 22.6 ± 3.5 | |
| Normal | 1,389,691 (92.5) | 37,824 (84.8) | 90.3 |
| Overweight | 80,332 (5.3) | 4,336 (9.7) | 189.0 |
| Obese | 32,843 (2.2) | 2,452 (5.5) | 268.3 |
| Education (years) | |||
| <12 | 228,719 (15.2) | 8,125 (18.2) | 110.2 |
| 12–14 | 662,789 (44.1) | 20,838 (46.7) | 107.6 |
| ≥15 | 611,358 (40.7) | 15,649 (35.1) | 85.3 |
| Neighborhood SES | |||
| Low | 231,734 (15.4) | 7,660 (17.2) | 107.3 |
| Medium | 991,868 (66.0) | 30,367 (68.0) | 101.1 |
| High | 279,264 (18.6) | 6,585 (14.8) | 83.3 |
| Family history of sleep apnea | |||
| No | 1,414,451 (94.1) | 40,067 (89.8) | 94.1 |
| Yes | 88,415 (5.9) | 4,545 (10.2) | 181.0 |
Sleep apnea incidence rate per 100,000 person-years.
CRF = cardiorespiratory fitness, IRR = incidence rate ratio, SES = socioeconomic status.
Main Effects of CRF and Other Factors
Low CRF at age 18 years was associated with subsequent increased risk of sleep apnea in adulthood, after adjusting for height, weight, and other potential confounders (Table 2, adjusted model 2, lowest vs. highest tertile: IRR, 1.44; 95% CI, 1.40–1.49; P<0.001). When CRF was examined as a continuous variable, there was a strong inverse linear trend in relation to risk of sleep apnea (adjusted model 2, trend test per additional 100 Watts: IRR, 0.71; 95% CI, 0.70–0.73; P<0.001). Muscular strength and height were positively associated with sleep apnea risk before but not after adjusting for weight. In the fully adjusted model, low muscular strength was associated with a slightly increased risk of sleep apnea (lowest vs. highest tertile: IRR, 1.06; 95% CI, 1.03–1.09; P<0.001), but without a significant linear trend (trend test per additional 1000 Newtons: IRR, 0.99; 95% CI, 0.96–1.02; P=0.44).
Table 2.
Associations between CRF or other factors among 18-year-old men and subsequent risk of sleep apnea, Sweden, 1969–2012.
| Adjusted Model 1a | Adjusted Model 2b | |||||
|---|---|---|---|---|---|---|
| IRR | 95% CI | P | IRR | 95% CI | P | |
| CRF (tertiles) | ||||||
| Low | 1.14 | 1.11, 1.17 | <0.001 | 1.44 | 1.40, 1.49 | <0.001 |
| Medium | 1.14 | 1.11, 1.18 | <0.001 | 1.25 | 1.21, 1.28 | <0.001 |
| High | 1.00 | 1.00 | ||||
| Per 100 Watts (trend test) | 0.94 | 0.92, 0.95 | <0.001 | 0.71 | 0.70, 0.73 | <0.001 |
| Muscular strength (tertiles) | ||||||
| Low | 0.78 | 0.77, 0.80 | <0.001 | 1.06 | 1.03, 1.09 | <0.001 |
| Medium | 0.87 | 0.85, 0.89 | <0.001 | 1.03 | 1.01, 1.06 | 0.006 |
| High | 1.00 | 1.00 | ||||
| Per 1000 Newtons (trend test) | 1.35 | 1.32, 1.39 | <0.001 | 0.99 | 0.96, 1.02 | 0.44 |
| Height (cm) | ||||||
| <172.0 | 1.00 | 1.00 | ||||
| 172.0–176.9 | 0.97 | 0.95, 1.00 | 0.07 | 0.83 | 0.81, 0.86 | <0.001 |
| 177.0–179.9 | 0.96 | 0.94, 0.99 | 0.007 | 0.75 | 0.73, 0.77 | <0.001 |
| 180.0–183.9 | 0.98 | 0.95, 1.01 | 0.15 | 0.70 | 0.68, 0.72 | <0.001 |
| ≥184.0 | 1.00 | 0.97, 1.03 | 0.95 | 0.62 | 0.60, 0.64 | <0.001 |
| Per 5 cm (trend test) | 1.01 | 1.00, 1.02 | 0.002 | 0.88 | 0.87, 0.89 | <0.001 |
| Weight (kg) | ||||||
| <60.5 | 1.00 | 1.00 | ||||
| 60.5–64.9 | 1.11 | 1.07, 1.14 | <0.001 | 1.34 | 1.29, 1.38 | <0.001 |
| 65.0–69.9 | 1.18 | 1.15, 1.22 | <0.001 | 1.62 | 1.56, 1.68 | <0.001 |
| 70.0–75.9 | 1.39 | 1.35, 1.43 | <0.001 | 2.05 | 1.97, 2.13 | <0.001 |
| ≥76.0 | 2.21 | 2.15, 2.27 | <0.001 | 3.37 | 3.25, 3.50 | <0.001 |
| Per 5 kg (trend test) | 1.16 | 1.15, 1.17 | <0.001 | 1.21 | 1.20, 1.22 | <0.001 |
| Body mass index | ||||||
| Normal | 1.00 | 1.00 | ||||
| Overweight | 2.20 | 2.14, 2.27 | <0.001 | 2.12 | 2.06, 2.19 | <0.001 |
| Obesity | 3.18 | 3.06, 3.31 | <0.001 | 2.98 | 2.86, 3.10 | <0.001 |
| Per 1 BMI unit (trend test) | 1.08 | 1.07, 1.09 | <0.001 | 1.08 | 1.07, 1.09 | <0.001 |
| Education (years) | ||||||
| <12 | 1.16 | 1.13, 1.19 | <0.001 | 1.03 | 1.00, 1.06 | 0.05 |
| 12–14 | 1.28 | 1.25, 1.30 | <0.001 | 1.16 | 1.14, 1.19 | <0.001 |
| ≥15 | 1.00 | 1.00 | ||||
| Per higher category (trend) | 0.91 | 0.89, 0.92 | <0.001 | 0.97 | 0.95, 0.98 | <0.001 |
| Neighborhood SES | ||||||
| Low | 1.20 | 1.16, 1.24 | <0.001 | 1.09 | 1.06, 1.13 | <0.001 |
| Medium | 1.15 | 1.12, 1.18 | <0.001 | 1.07 | 1.04, 1.10 | <0.001 |
| High | 1.00 | 1.00 | ||||
| Per higher category (trend) | 0.91 | 0.90, 0.93 | <0.001 | 0.96 | 0.94, 0.97 | <0.001 |
| Family history of sleep apnea | ||||||
| No | 1.00 | 1.00 | ||||
| Yes | 2.05 | 1.99, 2.11 | <0.001 | 1.96 | 1.90, 2.02 | <0.001 |
Adjusted for year of military conscription examination (to account for follow-up time and attained age).
Adjusted for year of military conscription examination, CRF, muscular strength, height, weight, education level, neighborhood SES, and family history of sleep apnea. BMI was modeled as an alternative to height and weight in a separate model. The reference category for all variables is indicated by an IRR of 1.00.
CRF = cardiorespiratory fitness, IRR = incidence rate ratio, SES = socioeconomic status.
Weight or BMI was positively associated with risk of sleep apnea in the fully adjusted model. High weight and low height were each independently associated with increased risk (Table 2, adjusted model 2), but high weight was a much stronger risk factor (Pheterogeneity<0.001). When BMI was modeled as an alternative to height and weight, overweight or obese men had more than a 2-fold or nearly 3-fold risk of sleep apnea, respectively, relative to those with normal BMI (Table 2, adjusted model 2).
Education level and neighborhood SES were each inversely related to sleep apnea risk (i.e., high education level and high neighborhood SES were modestly protective) (Table 2; Ptrend<0.001). A first-degree family history of sleep apnea was associated with nearly a 2-fold risk of sleep apnea (Table 2, adjusted model 2: IRR, 1.96; 95% CI, 1.90–2.02; P<0.001).
The sensitivity analyses yielded similar risk estimates as the main analyses and the conclusions were unchanged. For example, after restricting to men with complete data (N=1,361,083; 88.0% of the entire cohort), the fully adjusted IRR for sleep apnea comparing lowest vs. highest CRF tertile was 1.43 (95% CI, 1.38–1.48; P<0.001), and for continuous CRF per 100 Watts was 0.72 (95% CI, 0.70–0.74; P<0.001). The corresponding IRRs after restricting the follow-up period to 2001 and later (when outpatient as well as inpatient data were available) were 1.42 (95% CI, 1.37–1.46; P<0.001) and 0.72 (95% CI, 0.70–0.74; P<0.001).
Interaction Between CRF and BMI
Table 3 shows the interaction between CRF and BMI in relation to risk of sleep apnea. After adjusting for other factors, low CRF was associated with increased risk of sleep apnea among men with normal BMI (lowest vs. highest CRF tertile: IRR, 1.30; 95% CI, 1.26–1.35; P<0.001; Table 3, first row, right-most column). However, there was no significant association between CRF and risk of sleep apnea among overweight or obese men (Table 3, second and third rows). Low CRF and high BMI had a significant negative interaction on both the additive and multiplicative scale (P<0.001) (i.e., the combined effect of these factors was less than the sum or product of their separate effects). Figure 1 shows the estimated probability of being diagnosed with sleep apnea for men at the 25th, 50th, and 75th percentiles of CRF across the full distribution of BMI, from the fully adjusted model.
Table 3.
Interaction between CRF and BMI in 18-year-old men in relation to subsequent risk of sleep apnea, Sweden, 1969–2012.a
| CRF (tertiles) | IRRs (95% CI) for medium CRF within strata of BMI | IRRs (95% CI) for low CRF within strata of BMI | ||||||
|---|---|---|---|---|---|---|---|---|
| High | Medium | Low | ||||||
| No. cases/ total | IRR (95% CI) | No. cases/ total | IRR (95% CI) | No. cases/ total | IRR (95% CI) | |||
| BMI | ||||||||
| Normal | 7,347/ 457,985 |
1.00 | 13,242/ 479,789 |
1.23 (1.19, 1.27); P<0.001 |
17,235/ 489,741 |
1.30 (1.26, 1.35); P<0.001 |
1.23 (1.19, 1.27); P<0.001 |
1.30 (1.26, 1.35); P<0.001 |
| Overweight | 1,574/ 41,323 |
2.38 (2.25, 2.51); P<0.001 |
1,728/ 28,688 |
2.48 (2.35, 2.61); P<0.001 |
1,034/ 14,657 |
2.56 (2.40, 2.74); P<0.001 |
1.04 (0.97, 1.11); P=0.24 |
1.08 (0.99, 1.16); P=0.07 |
| Obesity | 912/ 16,273 |
3.63 (3.40, 3.89); P<0.001 |
906/ 12,070 |
3.36 (3.14, 3.60); P<0.001 |
634/ 6,952 |
3.39 (3.13, 3.67); P<0.001 |
0.92 (0.84, 1.01); P=0.07 |
0.93 (0.84, 1.02); P=0.15 |
| IRRs (95% CI) for overweight within strata of CRF | 2.38 (2.25, 2.51); P<0.001 |
2.01 (1.92, 2.11); P<0.001 |
1.97 (1.85, 2.09); P<0.001 |
|||||
| IRRs (95% CI) for obesity within strata of CRF | 3.63 (3.40, 3.89); P<0.001 |
2.73 (2.55, 2.91); P<0.001 |
2.60 (2.40, 2.80); P<0.001 |
|||||
| Interaction on additive scale: RERI (95% CI) | −0.55 (−0.89, −0.21); P<0.001 | |||||||
| Interaction on multiplicative scale: IRR ratio (95% CI) | 0.72 (0.64, 0.79); P<0.001 | |||||||
IRRs are adjusted for year of military conscription exam, muscular strength, education level, neighborhood SES, and family history of sleep apnea.
BMI = body mass index, CRF = cardiorespiratory fitness, IRR = incidence rate ratio, RERI = relative excess risk due to interaction
Figure 1.
Probability of sleep apnea by CRF and BMI at age 18 years (with follow-up to median attained age 47 years, maximum 62 years), adjusted for height, weight, and other covariates, Sweden, 1969–2012.
DISCUSSION
In this large national cohort study, low CRF at age 18 years was associated with higher risk of sleep apnea in adulthood, after adjusting for height, weight, sociodemographic factors, and family history of sleep apnea. High BMI was the strongest risk factor for sleep apnea in this cohort. However, we found that low CRF also was independently associated with increased risk of sleep apnea even among men with normal BMI. In contrast, muscular strength had minimal association with sleep apnea risk. These findings identify low CRF early in life as a new modifiable risk factor for development of sleep apnea in adulthood.
To our knowledge, this is the first study to examine CRF in relation to the risk of sleep apnea. Prior studies have suggested that high physical activity or exercise levels are associated with reduced symptoms and improved outcomes in patients with existing OSA, irrespective of body weight or BMI (Mendelson et al., 2018, Aiello et al., 2016, Iftikhar et al., 2014, da Silva et al., 2017, Murillo et al., 2016). For example, three meta-analyses of small clinical trials reported that exercise training was independently associated with reduced severity of apnea symptoms in patients with OSA (Mendelson et al., 2018, Aiello et al., 2016, Iftikhar et al., 2014). The underlying mechanisms are not established but may potentially involve improved strength of pharyngeal dilator muscles, decreased overnight rostral fluid redistribution, and stabilization of sympathetic outflow (Iftikhar et al., 2014, Redolfi et al., 2015). In addition, a cohort study of 14,087 Hispanic/Latino adults found that meeting the recommended levels of moderate or vigorous physical activity was linked with reduced risk of OSA, independent of BMI (e.g., adjusted odds ratio for moderate to severe OSA, 0.76; 95% CI, 0.59–0.98) (Murillo et al., 2016). A cross-sectional study of 5,453 Brazilian adults reported that structured physical exercise was associated with 23% lower odds of moderate OSA and 34% lower odds of severe OSA, after adjusting for BMI (da Silva et al., 2017). All of these studies examined physical activity or exercise, but not CRF, and most studies have examined effects on symptom severity in patients with existing OSA. Because CRF is more objectively and accurately measured, it has been shown to be a better indicator of habitual physical activity than self-reported physical activity (Swift et al., 2013). Importantly, we found that low CRF was associated with increased risk of sleep apnea among men with normal BMI, suggesting that interventions to enhance CRF may help prevent sleep apnea even in persons who are not overweight or obese.
Consistent with prior studies, we also found that high BMI was a strong risk factor for sleep apnea, with a 2-fold and nearly 3-fold risk among overweight and obese men, respectively (Young et al., 2004). To our knowledge, muscular strength has not been previously examined in relation to sleep apnea. We found that muscular strength had minimal association with sleep apnea risk after adjusting for BMI and other factors. Specifically, low muscular strength was associated with only a slightly increased risk of sleep apnea, but without a significant linear trend. However, muscle strength was measured as a composite of knee extension, elbow flexion, and hand grip, and did not include oropharyngeal strength, which may be more directly involved in sleep apnea mechanisms (Javaheri et al., 2017). Also consistent with other studies, a first-degree family history was associated with nearly a 2-fold risk of developing sleep apnea (Young et al., 2004). This may be related in part to shared lifestyle factors, although some evidence has also supported a genetic predisposition for sleep apnea (Varvarigou et al., 2011, Redline and Tishler, 2000).
Given the high prevalence and disease burden of sleep apnea in the general population (Senaratna et al., 2017), the identification of CRF as a new modifiable risk factor has important implications for prevention. Behavioral interventions to enhance CRF are low cost and relatively easy to adopt, and may help reduce the risk of sleep apnea regardless of patients’ BMI or ability to achieve recommended weight loss. Patient counseling on physical activity has been shown to be clinically and cost effective (Orrow et al., 2012, Grandes et al., 2009), yet remains under-utilized in the health care system (Vuori et al., 2013). Such counseling should play a key role in the prevention and treatment of sleep apnea, and may help reduce its high disease burden and economic costs which exceed $100 billion annually in the US (American Academy of Sleep Medicine, 2016).
Strengths of this study include objective measurement of CRF and prospective ascertainment of sleep apnea in a large national cohort. This study design helped minimize potential selection bias and enabled the first population-based risk estimates for sleep apnea associated with CRF. The use of registry data with prospectively measured CRF, covariates, and sleep apnea prevented self-reporting bias. We were able to adjust for height, weight, and other potential confounders, including family history and both individual- and neighborhood-level socioeconomic factors, which also were prospectively ascertained and not self-reported.
Limitations include the measurement of CRF and weight at only one age (~18 years), and hence we were unable to examine changes in these factors over time. We were unable to distinguish different types of sleep apnea, although the large majority of cases are obstructive (Morgenthaler et al., 2006), nor to assess sleep apnea severity or temporal changes in diagnosis. Polysomnography or other clinical data to validate diagnoses were unavailable. Because this study was based on Swedish military conscripts, the cohort consisted entirely of men. However, other studies have found that associations between BMI or physical activity and sleep apnea severity are similar in women compared to men (Young et al., 2004, Aiello et al., 2016, Iftikhar et al., 2014, da Silva et al., 2017, Murillo et al., 2016). Sleep apnea may have been substantially under-diagnosed in the present cohort. Prior studies have suggested that a large majority (>75%) of adults with sleep apnea are undiagnosed (Kapur et al., 2002). Outpatient diagnoses in the present study were unavailable before 2001, which contributed to under-reporting. However, under-diagnosis or other misclassification of sleep apnea is likely to be non-differential with respect to CRF level, and therefore to influence our results conservatively (i.e., toward the null hypothesis). Finally, this study cohort was relatively young, with a median age of 47 years (maximum 62) at the end of follow-up. As a result, the mean age at diagnosis was approximately 10 years younger than in the typical OSA patient population captured in the Swedish Sleep Apnea Registry (SESAR Annual Report, 2017). Additional follow-up will be needed to examine early-life CRF in relation to development of sleep apnea at older ages when sleep apnea is more common.
In summary, this large national cohort study provides the first population-based risk estimates for sleep apnea associated with CRF. We found that low CRF and high BMI at age 18 years were independently associated with higher risk of sleep apnea in adulthood. Low CRF was associated with an increased risk even among men with normal BMI. These findings identify low CRF in late adolescence as a risk factor for sleep apnea later in life. Further investigations will be needed to elucidate whether interventions to enhance CRF may help prevent the development of sleep apnea across the life course.
ACKNOWLEDGMENTS
This work was supported by the National Heart, Lung, and Blood Institute at the National Institutes of Health (R01 HL116381); the Swedish Research Council; and ALF project grant, Region Skåne/Lund University, Sweden. The funding agencies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Footnotes
There were no conflicts of interest.
REFERENCES
- Aiello KD, Caughey WG, Nelluri B, Sharma A, Mookadam F and Mookadam M Effect of exercise training on sleep apnea: A systematic review and meta-analysis. Respir Med, 2016, 116: 85–92. [DOI] [PubMed] [Google Scholar]
- American Academy of Sleep Medicine Hidden Health Crisis Costing America Billions In, Mountain View, CA, 2016. [Google Scholar]
- Andersen LB A maximal cycle exercise protocol to predict maximal oxygen uptake. Scand J Med Sci Sports, 1995, 5: 143–6. [DOI] [PubMed] [Google Scholar]
- Billings ME, Johnson DA, Simonelli G et al. Neighborhood Walking Environment and Activity Level Are Associated With OSA: The Multi-Ethnic Study of Atherosclerosis. Chest, 2016, 150: 1042–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Botros N, Concato J, Mohsenin V, Selim B, Doctor K and Yaggi HK Obstructive sleep apnea as a risk factor for type 2 diabetes. Am J Med, 2009, 122: 1122–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Crump C, Sundquist J, Winkleby MA and Sundquist K Interactive Effects of Aerobic Fitness, Strength, and Obesity on Mortality in Men. Am J Prev Med, 2017, 52: 353–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Crump C, Sundquist K, Sundquist J and Winkleby MA Neighborhood deprivation and psychiatric medication prescription: a Swedish national multilevel study. Ann Epidemiol, 2011, 21: 231–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Da Silva RP, Martinez D, Pedroso MM et al. Exercise, Occupational Activity, and Risk of Sleep Apnea: A Cross-Sectional Study. J Clin Sleep Med, 2017, 13: 197–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gottlieb DJ, Yenokyan G, Newman AB et al. Prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure: the sleep heart health study. Circulation, 2010, 122: 352–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grandes G, Sanchez A, Sanchez-Pinilla RO et al. Effectiveness of physical activity advice and prescription by physicians in routine primary care: a cluster randomized trial. Arch Intern Med, 2009, 169: 694–701. [DOI] [PubMed] [Google Scholar]
- Hook O and Tornvall G Apparatus and method for determination of isometric muscle strength in man. Scand J Rehabil Med, 1969, 1: 139–42. [PubMed] [Google Scholar]
- Iftikhar IH, Kline CE and Youngstedt SD Effects of exercise training on sleep apnea: a meta-analysis. Lung, 2014, 192: 175–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Javaheri S, Barbe F, Campos-Rodriguez F et al. Sleep Apnea: Types, Mechanisms, and Clinical Cardiovascular Consequences. J Am Coll Cardiol, 2017, 69: 841–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kapur V, Strohl KP, Redline S, Iber C, O’connor G and Nieto J Underdiagnosis of sleep apnea syndrome in U.S. communities. Sleep Breath, 2002, 6: 49–54. [DOI] [PubMed] [Google Scholar]
- Lee DC, Sui X, Ortega FB et al. Comparisons of leisure-time physical activity and cardiorespiratory fitness as predictors of all-cause mortality in men and women. Br J Sports Med, 2011, 45: 504–10. [DOI] [PubMed] [Google Scholar]
- Mehra R, Benjamin EJ, Shahar E et al. Association of nocturnal arrhythmias with sleep-disordered breathing: The Sleep Heart Health Study. Am J Respir Crit Care Med, 2006, 173: 910–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mendelson M, Bailly S, Marillier M et al. Obstructive Sleep Apnea Syndrome, Objectively Measured Physical Activity and Exercise Training Interventions: A Systematic Review and Meta-Analysis. Front Neurol, 2018, 9: 73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morgenthaler TI, Kagramanov V, Hanak V and Decker PA Complex sleep apnea syndrome: is it a unique clinical syndrome? Sleep, 2006, 29: 1203–9. [DOI] [PubMed] [Google Scholar]
- Murillo R, Reid KJ, Arredondo EM et al. Association of self-reported physical activity with obstructive sleep apnea: Results from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Prev Med, 2016, 93: 183–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Namen AM, Chatterjee A, Huang KE, Feldman SR and Haponik EF Recognition of Sleep Apnea Is Increasing. Analysis of Trends in Two Large, Representative Databases of Outpatient Practice. Ann Am Thorac Soc, 2016, 13: 2027–34. [DOI] [PubMed] [Google Scholar]
- Nieto FJ, Young TB, Lind BK et al. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA, 2000, 283: 1829–36. [DOI] [PubMed] [Google Scholar]
- Nordesjo L and Schele R Validity of an ergometer cycle test and measures of isometric muscle strength when predicting some aspects of military performance. Swedish J Defence Med, 1974, 10: 11–23. [Google Scholar]
- Ogden CL and Flegal KM Changes in terminology for childhood overweight and obesity. Natl Health Stat Report, 2010: 1–5. [PubMed]
- Orrow G, Kinmonth AL, Sanderson S and Sutton S Effectiveness of physical activity promotion based in primary care: systematic review and meta-analysis of randomised controlled trials. BMJ, 2012, 344: e1389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Patton JF, Vogel JA and Mello RP Evaluation of a maximal predictive cycle ergometer test of aerobic power. Eur J Appl Physiol Occup Physiol, 1982, 49: 131–40. [DOI] [PubMed] [Google Scholar]
- Punjabi NM, Caffo BS, Goodwin JL et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med, 2009, 6: e1000132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Redline S and Tishler PV The genetics of sleep apnea. Sleep Med Rev, 2000, 4: 583–602. [DOI] [PubMed] [Google Scholar]
- Redolfi S, Bettinzoli M, Venturoli N et al. Attenuation of obstructive sleep apnea and overnight rostral fluid shift by physical activity. Am J Respir Crit Care Med, 2015, 191: 856–8. [DOI] [PubMed] [Google Scholar]
- Rubin DB Multiple Imputation for Nonresponse in Surveys Wiley, New York, 1987. ( [Google Scholar]
- Senaratna CV, Perret JL, Lodge CJ et al. Prevalence of obstructive sleep apnea in the general population: A systematic review. Sleep Med Rev, 2017, 34: 70–81. [DOI] [PubMed] [Google Scholar]
- SESAR Annual Report, 2017. Swedish Sleep Apnea Registry; [Available from: https://stratum.registercentrum.se/#!page?id=1243 accessed January 10, 2019]. [Google Scholar]
- Stoddard PJ, Laraia BA, Warton EM et al. Neighborhood deprivation and change in BMI among adults with type 2 diabetes: the Diabetes Study of Northern California (DISTANCE). Diabetes Care, 2013, 36: 1200–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Swift DL, Lavie CJ, Johannsen NM et al. Physical activity, cardiorespiratory fitness, and exercise training in primary and secondary coronary prevention. Circ J, 2013, 77: 281–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tarasiuk A and Reuveni H The economic impact of obstructive sleep apnea. Curr Opin Pulm Med, 2013, 19: 639–44. [DOI] [PubMed] [Google Scholar]
- Vanderweele TJ and Knol MJ A tutorial on interaction. Epidemiol Methods, 2014, 3: 33–72. [Google Scholar]
- Varvarigou V, Dahabreh IJ, Malhotra A and Kales SN A review of genetic association studies of obstructive sleep apnea: field synopsis and meta-analysis. Sleep, 2011, 34: 1461–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vuori IM, Lavie CJ and Blair SN Physical activity promotion in the health care system. Mayo Clin Proc, 2013, 88: 1446–61. [DOI] [PubMed] [Google Scholar]
- Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM and Mohsenin V Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med, 2005, 353: 2034–41. [DOI] [PubMed] [Google Scholar]
- Young T, Peppard PE and Taheri S Excess weight and sleep-disordered breathing. J Appl Physiol (1985), 2005, 99: 1592–9. [DOI] [PubMed] [Google Scholar]
- Young T, Skatrud J and Peppard PE Risk factors for obstructive sleep apnea in adults. JAMA, 2004, 291: 2013–6. [DOI] [PubMed] [Google Scholar]
- Zou G A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol, 2004, 159: 702–6. [DOI] [PubMed] [Google Scholar]