Abstract
This secondary data analysis aimed to determine the nature of the relationship between obstructive sleep apnea (OSA) risk, biological age acceleration, and nonspecific chronic low back pain (CLBP). In total, 199 adults aged 18–82 years who filled both STOPBANG and pain questionnaires subsetted for secondary analysis. Based on the STOPBANG questionnaire, 104 had a low OSA risk and 95 had an intermediate or high OSA risk. Dunedin Pace of Aging Computed from the Epigenome (DunedinPACE), Horvath’s, Hannum’s, PhenoAge, and GrimAge clocks were used to determine biological age and pace of biological aging. Individuals with low OSA risk reported increased DunedinPACE compared to those with intermediate/high OSA risk (p < 0.001). There was a significant correlation between the risk for OSA and biological age acceleration measured by PhenoAge as well as pace of biological aging (p < 0.05). Mediation analysis detected indirect effects of OSA risk on chronic pain outcomes through the pace of biological aging. Targeted interventions addressing OSA risk offers a promising therapeutic strategy. This could be particularly valuable for aging populations where both accelerated biological aging and chronic pain conditions are prevalent, offering a more holistic approach to improving nonspecific chronic pain outcomes through quality of sleep and restfulness.
Keywords: DunedinPACE, obstructive sleep apnea, nonspecific chronic low back pain, pain disparities, pace of biological aging
Introduction
Obstructive sleep apnea (OSA) is one of the most common sleep disorders, affecting approximately 39 million adults in the United States. 1 It is characterized by recurrent episodes of partial (hypopnea) or complete (apnea) upper airway collapse during sleep, which leads to intermittent hypoxia, sleep fragmentation, nocturnal dysrhythmias, depression, fatigue, and cardiopulmonary dysregulation. 2 OSA is associated with increased risk for major cardiovascular events, hypertension, ischemic heart disease, cognitive decline, neurodegenerative diseases, and decreased quality of life. 3 Individuals with OSA often present with chronic pain, including headaches, temporomandibular disorders, fibromyalgia, rheumatoid arthritis, and musculoskeletal pain. 4 It is estimated that about 55.4% to 76.2% of individuals with OSA experience from chronic pain, 5 with a significant number of patients with OSA reporting high-impact chronic pain. 4 Findings from a systematic review suggest that OSA is associated with increased pain sensitivity and decreased pain tolerance. 6 Like most sleep disorders, the relationship between OSA and chronic pain is believed to be bidirectional, with increasing evidence suggesting that sleep problems precede chronic pain. However, the specific mechanisms linking OSA with chronic pain remain unknown.
Emerging evidence suggests that sleep deprivation and intermittent hypoxia can increase inflammatory cytokines and nociceptive sensitization, leading to persistent pain in individuals with OSA. 7 Other investigators have suggested that OSA-induced hypoxemia can directly influence the formation of oxygen-free radicals, which can sensitize nociceptors. Alterations or damage in the signal transmission in nociceptors have been associated with chronic pain. 8
Some physiological changes induced by OSA resemble the hallmarks of aging, including systemic inflammation, genomic instability, and telomere attrition. 9 Emerging evidence from clinical and experimental studies suggests that OSA can accelerate the aging process.10,11 This accelerated rate of biological aging may be attributed to intermittent hypoxia, which promotes oxidative stress, low-grade inflammation, telomere shortening, and epigenetic modifications. 9 Epigenetic clocks have emerged as a promising measure of biological aging. These clocks estimate the biological age by assessing the DNA methylation levels at specific CpG sites, which can be compared to the chronological age to determine biological age acceleration or deceleration. In fact, biological age acceleration derived from recent epigenetic clocks such as GrimAge 12 and DunedinPACE (Dunedin Pace of Aging Computed from the Epigenome) 13 shows a stronger association with age-related chronic conditions, including chronic pain. Of note, no study has examined the role of biological age acceleration on OSA in patients with chronic pain.
Among chronic pain conditions, chronic low back pain (CLBP) is a highly prevalent, costly, and disabling condition that is also associated with biological age acceleration, nociceptive sensitization, and sleep problems, including OSA and insomnia symptoms. 14 It has been suggested that the resulting poor sleep quality and daytime fatigue associated with OSA amplifies pain perception and reduces physical activity, potentially worsening CLBP symptoms. 15 Similarly, OSA-induced accelerated biological aging might impair tissue repair mechanisms and pain modulation systems, leading to poorer outcomes in CLBP patients.16,17 In fact, despite being an age-related chronic condition, the pace of biological aging (DunedinPACE) is a stronger predictor of nonspecific CLBP than chronological age. 16 Also, we recently found that the pace of biological aging mediates the relationship between insomnia and CLBP severity, 17 suggesting a possible relationship between sleep disruption, biological aging, and CLBP. Furthermore, a bidirectional relationship exists between CLBP and OSA, suggesting that CLBP can disrupt sleep patterns and potentially increase the risk of developing or worsening OSA, thereby accelerating biological aging processes. 15 However, no study has investigated relationships between OSA, the pace of biological aging, and CLBP outcomes.
This study examines the relationship between the risk of OSA, the pace of biological aging, and nonspecific CLBP outcomes (pain severity and pain interference). To accomplish this goal, we will test the relationship between the risk of OSA and biological age acceleration using five DNA methylation clocks to explore the relationship extensively. Then, we will investigate whether the pace of biological aging mediates the relationship between the risk of OSA and nonspecific CLBP outcomes (Figure 1). We hypothesize that a higher risk for OSA correlates with a faster pace of biological aging, which is associated with greater CLBP severity and interference.
Figure 1.
The theoretical model of the relationship between risk for OSA, DunedinPACE, and CLBP outcomes. The relationship between risk for OSA and CLBP outcomes is mediated by DunedinPACE.
Note: DunedinPACE: Dunedin Pace of Aging Computed from Epigenome; CLBP: chronic low back pain.
Materials and methods
Participants and settings
This secondary data analysis (R01AR079178 and R01MD010441) evaluates gene expression signatures and epigenetics of racial and socioeconomic status differences in nonspecific CLBP within a wider cross-sectional study. Adults aged 18 years or older who self-identified as African American/non-Hispanic Black (NHBs) or Caucasian/non-Hispanic White (NHWs), had nonspecific CLBP or were pain-free controls, could read and write in English, and provided written informed consent were eligible participants. Electronic medical records verified the CLBP diagnosis. 18 Participants were ineligible if their CLBP was caused by a compression fracture, ankylosing spondylitis, infection, cancer, or other trauma; if they had uncontrolled hypertension, neurological disorders, or serious psychiatric disorders requiring hospitalization within the past year; or if they were pregnant.
Procedures
Briefly, flyers distributed at pain clinics around Birmingham, AL were used to enlist a convenient sample of participants. To establish their eligibility to participate in the study, potential participants completed an initial telephone screening. We also recruited healthy individuals for comparison, who fulfilled all eligibility requirements but did not have a diagnosis of CLBP. In the UAB laboratory, participants completed two sets of pain assessments separated by one week. During the initial appointment, informed consent was obtained, measurements of blood pressure, heart rate, temperature, weight, height, and circumference of the neck and waist were taken, questionnaires were filled out, and clinical pain ratings were documented. During their second visit, participants underwent functional performance activities to assess movement evoked pain, and blood samples were taken for DNA extraction and the calculation of biological age. The study was approved by the Institutional Review Board at the University of Alabama at Birmingham (IRB-170119003).
Pain intensity and interference
Pain intensity and interference were assessed using the graded chronic pain scale (GCPS). 19 In line with earlier research, the average of three pain rating items multiplied by 100 was used to calculate pain severity. Higher ratings indicated more severe pain. The pain intensity scores varied from 0 to 100. The average of the seven items pertaining to interference with daily activities was multiplied by 100 to estimate the degree of pain interference. The pain interference score runs from 0 to 100, where higher numbers indicate greater levels of interference with day-to-day activities and life satisfaction. To fully capture the impact of CLBP, participants were asked, “How many days in the last 6 months have you been kept from your usual activities because of pain.” Together with the pain intensity and interference scores, the disability days were used to compute the GCPS grades as Grade 0 = pain-free controls (PFCs); Grades 1–2 = low-impact pain; and Grades ≥3 = high-impact pain, as previously described. 16 In our sample, GCPS has excellent internal consistency reliability (Cronbach’s α = 0.97).
Dietary caffeine consumption
Caffeine consumption data was collected using a comprehensive 7-day daily caffeine dairy. The diary was designed to capture detailed caffeine intake across different times of day: morning (7:00 AM to 11:59 PM), afternoon (12:00 PM to 5:59 PM), evening (6:00 PM to 1:59 AM), and late night (2:00 AM to 6:59 AM). Participants recorded their caffeine intake, documenting consumption across multiple categories of drinks containing caffeine including coffee, tea, soft drinks, energy drinks, chocolate, and caffeine-containing medications. For each item, participants noted the specific product name, serving sizes in ounces, and the number of servings consumed. This caffeine dairy was adapted from the Caffeine Consumption Questionnaire, a scientifically validated tool established for use in adult populations. 20 Previous research has established the reliability of self-reported caffeine intake as an accurate representation of actual consumption. 21 To quantify caffeine consumption, standardized milligram (mg) values per ounce were applied to each consumed substance, drawing from established literature and manufacturer nutritional information. These values were obtained from previously published literature on the topic and/or manufacturer’s nutritional information, with consideration for preparation style (e.g., brewed coffee, instant coffee, espresso), type (e.g., dark chocolate, milk chocolate), and brand (e.g., Coca-Cola, Mountain Dew). 22 Total dietary caffeine consumption in milligrams was calculated per day, summed across the 7 days, and then averaged.
Home sleep monitoring
Objective sleep data was collected using the Actiwatch2 (Respironics, Bend, OR), a wrist-worn actigraphic device equipped with a solid-state accelerometer that samples movement at 32 Hz to measure ambulatory activity and sleep-wake patterns. The Actiwatch2 has demonstrated good reliability and criterion validity in sleep research applications. 23 Participants received instructions to activate an event marker button on the Actiwatch2 at bedtime and upon morning awakening. These objective markers were cross-referenced with corresponding daily sleep diary entries to ensure data accuracy and completeness. Sleep-wake patterns were analyzed using Actiware Sleep software (v 6.0.8), which processes data in 30-second epochs using algorithms based on movement amplitude and frequency. The following actigraphic parameters were extracted:
Total Sleep Time: Duration of sleep (in minutes) from sleep onset to sleep offset.
Sleep Onset Latency: Time required (in minutes) to transition from wakefulness to sleep.
Wake After Sleep Onset (WASO): Total minutes of wakefulness occurring between sleep onset and final awakening.
Sleep Efficiency: Ratio of total sleep time to total time in bed, calculated as (total sleep time/time in bed) × 100, where values approaching 100 indicate optimal sleep efficiency.
This comprehensive approach provided objective, quantitative assessment of participants’ sleep patterns throughout the monitoring period, complementing self-reported sleep diary reports.
Sleep diary
Participants completed sleep diaries each day of actigraphic recording. 24 Questions regarding their sleep habits (e.g., “what time did you get into bed,” “how many times did you wake up not counting your final awakening,” etc.) were gathered from the sleep diaries. Individuals were also asked to rate their self-reported quality of sleep (1-very poor to 5-very good) and how rested or refreshed they felt upon awakening (1- not at all to 5-very well-rested).
Insomnia severity index
Insomnia Severity Index (ISI) assessed the severity and impact of insomnia symptoms. The ISI is a brief self-report questionnaire consisting of seven items that measure the perceived severity of insomnia symptoms such as difficulty falling asleep, maintaining sleep, and early morning awakening. 25 Using a scale from 0 to 4, each participant rated the severity of each symptom over the last month. We summed each item to calculate a total score that ranged from 0 to 28, with higher scores indicating greater insomnia severity. In the sample under study, the ISI demonstrated excellent internal consistency (Cronbach’s α = 0.94).
Obstructive sleep apnea risk
Participants completed the STOP-BANG questionnaire. 26 The four-item STOP questionnaire consists of the following four questions: S—“Do you snore loudly (louder than talking or loud enough to be heard through closed doors)?” T—“Do you often feel tired, fatigued, or sleepy during daytime?” O—“Has anyone observed you stop breathing during your sleep?” P—“Do you have or are you being treated for high blood pressure?” and were scored (0 – no, 1 – yes). The information concerning BMI, age, neck circumference, and gender (BANG portion of the questionnaire) were binarily scored using a cutoff value (0— equal to or below the cutoff, 1—above the cutoff; 0—female, 1—male) and combined to determine the risk of OSA. Scores range from 0 to 8 with high scores reflecting more risk for OSA: 0–2 is low risk and 3–8 is intermediate/high risk for OSA.
Epigenetic aging and pace of biological aging
The peripheral veins were used to draw blood samples, and genomic DNA was isolated in accordance with earlier instructions. 27 After centrifugation, the Puregen DNA isolation kit (Qiagen, Valencia, CA, USA) was used to extract DNA from the buffy coat. A NanoDrop UV spectrophotometer was used to quantify isolated DNA, and it was standardized to a concentration of 50 ng per microliter. The University of Minnesota Genomic Center received the extracted DNA on dry ice, and the Illumina Infinium Human MethylationEPIC BeadChip v2.0 arrays (Illumina, Inc., San Diego, CA) were used to measure DNAm. Quality control procedures were carried out, and all samples were scanned using the Illumina iScan (Illumina, Inc., San Diego, CA). Five samples were chosen at random for quality control and were run twice. Sample replicates were checked to ensure a correlation value of at least 0.99.
According to the manufacturer’s instructions, the Infinium MethylationEPIC v2.0 kit (Illumina, Inc., San Diego, CA) at the University of Minnesota Genomics Center targeted over 935,000 CpG sites in the human methylome. The manifest file, “IlluminaHumanMethylationEPICv2manifest”, was utilized for minfi (v1.48.0) 28 to determine differential methylation loci (https://github.com/jokergoo/IlluminaHumanMethylationEPICv2manifest). The raw data was quantile normalized using “preprocessQuantile”. 29 For each CpG site, the raw beta values were extracted and submitted with the sex and age of participants to calculate DNA methylation age, biological age acceleration, and biological age. Biological age acceleration, including the Horvath, 30 Hannum, 31 PhenoAge, 32 and GrimAge2 12 were calculated using the regression method in the online DNA Methylation Age Calculator developed by the Horvath laboratory. 33 DunedinPACE scores were calculated from the beta values using the R-software package DunedinPACE (v0.99.0) and the default proportion of probes required for EPICv2 data as described by Belsky et al. 13
Statistical analysis
Initial screening and cleaning of the data were completed prior to analysis using IBM SPSS (Windows v29.0.2.0) and visualization with R® (v4.3.3). Descriptive analyses compared risk for OSA, chronological age, BMI, pain outcomes, biological age, DundinPACE, and biological age acceleration between pain impact groups (no pain, low impact pain, and high impact pain), using one-way ANOVA to determine average group differences. One-way ANCOVA was used to determine differences in DunedinPACE between pain impact status with chronological age as a covariate. A post-hoc pairwise analysis was performed using Bonferroni corrections. The chi-squared test was utilized to detect differences in categorical variables (sex and race) among the pain impact groups. Spearman’s correlations were used to evaluate the association between biological age acceleration, the pace of biological aging, the risk for OSA, and pain outcomes. SPSS PROCESS Macro (v4.2) 34 Model 4 was utilized to test whether the pace of biological aging mediates the relationship between the risk for OSA and pain outcomes with race, caffeine intake, insomnia severity, and medication usage (opioids, muscle relaxers, benzodiazepines) as covariates (Figure 1). Since covariates such as sex, age, and BMI are already included in the STOP-BANG total, we did not control for these variables. A bootstrapping approach with a random sample of 5000 was used for the mediation analysis, and a p-value < 0.05 was statistically significant.
Results
Characteristics of study participants
In total, 199 NHW and NHB participants who filled out the STOP-BANG questionnaire were included: 104 individuals with low OSA risk and 95 individuals with intermediate/high OSA risk. The distribution of demographic and health characteristics according to the pain impact group is shown in Table 1. The demographics include NHW (47.20%) and NHB (52.80%) adults that self-identified as male (48.70%) or female (51.30%). Men were more likely to report intermediate/high OSA risk at the time of DNAm sampling than women (p < 0.001), but no significant differences were detected between NHB and NHW participants between OSA risk groups (p = 0.051). Individuals with low OSA risk reported a lower pain severity and pain interference compared to those with intermediate/high OSA risk (p < 0.001). In terms of biological aging, individuals with low OSA risk had a significantly younger Horvath age, Hannum age, PhenoAge, and GrimAge relative to those with intermediate/high OSA risk (p < 0.05). Overall, significant differences in biological age acceleration between the low risk (−1.023 ± 0.613) and intermediate/high risk (-3.712 ± 0.893) for OSA were observed for PhenoAgeAccel (p = 0.013), and the pace of biological aging was significantly faster in those with intermediate/high risk (1.081 ± 0.014) than those with low risk (1.025 ± 0.013) for OSA (p < 0.001). Participant caffeine consumption was not different between OSA risk groups (p > 0.05).
Table 1.
Baseline variables among cases based on OSA risk.
| Variables | Low risk (n = 104) | Intermediate/high risk (n = 95) | Total (n = 199) | Missing cases | p-value |
|---|---|---|---|---|---|
| Sex (n, %) | 0 (0.000) | <0.001 * | |||
| Men | 33 (31.70) | 64 (67.40) | 97 (48.70) | ||
| Women | 71 (68.30) | 31 (32.60) | 102 (51.30) | ||
| Race (n, %) | 0 (0.000) | 0.051 * | |||
| Black | 48 (46.20) | 57 (60.00) | 105 (52.80) | ||
| White | 56 (53.80) | 38 (40.00) | 94 (47.20) | ||
| Pain severity | 25.48 ± 2.797 | 49.11 ± 3.547 | 36.70 ± 2.382 | 1 (0.503) | <0.001 # |
| Pain interference | 16.84 ± 2.347 | 40.04 ± 3.544 | 27.85 ± 2.238 | 1 (0.503) | <0.001 # |
| Chronological age | 34.52 ± 1.243 | 47.76 ± 1.491 | 40.84 ± 1.070 | 0 (0.000) | <0.001 # |
| BMI | 30.15 ± 1.219 | 31.77 ± 0.702 | 30.93 ± 0.719 | 4 (2.010) | 0.260 # |
| Daily caffeine intake | 0.726 ± 0.032 | 0.786 ± 0.033 | 0.755 ± 0.023 | 10 (5.025) | 0.205 # |
| Horvath age | 38.10 ± 1.509 | 50.73 ± 1.513 | 44.13 ± 1.157 | 0 (0.000) | <0.001 # |
| Hannum age | 28.31 ± 2.062 | 41.55 ± 2.130 | 34.63 ± 1.551 | 0 (0.000) | <0.001 # |
| Phenoage | 33.50 ± 1.290 | 44.05 ± 1.375 | 38.53 ± 1.011 | 0 (0.000) | <0.001 # |
| GrimAge2 | 46.66 ± 1.199 | 58.10 ± 1.345 | 52.02 ± 0.986 | 9 (4.523) | <0.001 # |
| DunedinPACE | 1.025 ± 0.013 | 1.081 ± 0.014 | 1.052 ± 0.010 | 0 (0.000) | <0.001 $ |
| Accelerated Epigenetic Aging | |||||
| ΔHorvath | 3.576 ± 0.856 | 2.968 ± 0.914 | 3.286 ± 0.624 | 0 (0.000) | 0.628 # |
| ΔHannum | −6.214 ± 1.622 | −6.205 ± 1.732 | −6.210 ± 1.181 | 0 (0.000) | 0.997 # |
| PhenoAgeAccel | −1.023 ± 0.613 | −3.712 ± 0.893 | −2.307 ± 0.540 | 0 (0.000) | 0.013 # |
| GrimAgeAccel | −0.487 ± 0.441 | 0.301 ± 0.554 | −0.118 ± 0.350 | 9 (4.523) | 0.262 # |
Note: All values are represented as mean ± SEM. p-value was obtained from chi-squared test*, independent t-test#, or one-way ANCOVA$ via SPSS where appropriate, with significant p-values italicized. ΔHorvath: difference between biological age by Horvath’s and chronological age; ΔHannum: difference between biological age by Hannum’s and chronological age; PhenoAgeAccel: difference between phenotypic biological age and chronological age; GrimAgeAccel: difference between biological age measure by GrimAge and chronological age. Daily caffeine intake is an average value of whether participants ingested caffeine each day over the course of 7 days. Values closer to 1 indicate caffeine was ingested for most days during the session.
Sleep metrics
Lower sleep quality and restfulness occurred in individuals with intermediate/high risk for OSA compared to the participants with low risk (p < 0.001). Interestingly, no differences were detected between OSA risk groups when measuring sleep efficiency, sleep latency, WASO, time in bed, time awake, and total sleep time. Mean descriptive characteristics of sleep metrics are provided in Table 2.
Table 2.
Objective and self-reported sleep metrics among cases based on OSA Risk.
| Low risk (n = 104) | Intermediate/high risk (n = 95) | Total (n = 199) | Missing cases | p-value | |
|---|---|---|---|---|---|
| Objective | |||||
| Sleep efficiency (%) | 80.89 ± 1.019 | 80.46 ± 1.284 | 80.68 ± 0.810 | 28 (14.07) | 0.788 |
| WASO (minutes) | 39.77 ± 1.864 | 40.77 ± 2.401 | 40.25 ± 1.502 | 28 (14.07) | 0.741 |
| Time awake (minutes) | 34.84 ± 1.429 | 33.86 ± 1.596 | 34.37 ± 1.065 | 28 (14.07) | 0.648 |
| Sleep latency (minutes) | 31.75 ± 3.118 | 30.67 ± 3.863 | 31.23 ± 2.456 | 28 (14.07) | 0.827 |
| Time in bed (hh:mm) | 08:17 ± 00:07 | 08:17 ± 00:13 | 08:17 ± 00:07 | 28 (14.07) | 0.970 |
| Total sleep time (hh:mm) | 06:43 ± 00:07 | 06:41 ± 00:14 | 06:43 ± 00:08 | 28 (14.07) | 0.920 |
| Self-reported | |||||
| Quality of sleep | 3.528 ± 0.059 | 3.223 ± 0.067 | 3.385 ± 0.046 | 9 (4.523) | <0.001 |
| Rest/refreshed sleep | 3.210 ± 0.073 | 2.828 ± 0.078 | 2.68 ± 0.055 | 9 (4.523) | <0.001 |
Note: Different letters indicate means that differ between pain impact groups, p < 0.05. All values are represented as mean ± SEM. P-value was obtained from independent t-test via SPSS.
WASO: Wake After Sleep Onset = Total Time in Bed − Total Spent Asleep.
Associations between obstructive sleep apnea risk, sleep metrics, pain outcomes, and markers of biological age
Table 3 shows unadjusted bivariate correlation analyses between key study variables. The risk for OSA was positively correlated with pain interference (rs = 0.442, p < 0.001) and pain severity (rs = 0.431, p < 0.001). Pain severity was positively correlated with WASO (rs = 0.168, p < 0.05) but negatively correlated with sleep quality (rs = −0.393, p < 0.001) and restfulness (rs = −0.439, p < 0.001). Additionally, pain interference was negatively correlated with sleep quality (rs = −0.427, p < 0.001) and restfulness (rs = −0.467, p < 0.001). Also, a faster pace of biological aging (higher DunedinPACE scores) significantly correlated with pain severity (rs = 0.412, p < 0.001), pain interference (rs = 0.405, p < 0.001), risk for OSA (rs = 0.279, p < 0.001), WASO (rs = 0.212, p < 0.01), and sleep latency (rs = 0.169, p < 0.05); however, DunedinPACE was inversely correlated with sleep efficiency (rs = −0.203, p < 0.01). Notably, the risk for OSA negatively correlated with PhenoAgeAccel (rs = −0.179, p < 0.05) but not with other measures of biological age acceleration.
Table 3.
Spearman’s correlations among obstructive sleep apnea risk, pain outcomes, markers of biological age, and sleep metrics.
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Pain Severity | 1.000 | ||||||||||
| 2 | Pain Interference | 0.937 *** | 1.000 | |||||||||
| 3 | STOP-BANG | 0.442 *** | 0.431*** | 1.000 | ||||||||
| 4 | Sleep Efficiency | −0.086 | −0.100 | 0.009 | 1.000 | |||||||
| 5 | WASO | 0.168* | 0.143 | 0.003 | −0.645*** | 1.000 | ||||||
| 6 | Time Awake | 0.044 | 0.062 | −0.065 | −0.410*** | 0.651*** | 1.000 | |||||
| 7 | Sleep Latency | 0.034 | 0.063 | −0.088 | −0.769*** | 0.313*** | 0.227** | 1.000 | ||||
| 8 | Time in Bed | 0.051 | 0.061 | −0.125 | −0.033 | 0.236** | 0.368*** | 0.216** | 1.000 | |||
| 9 | Total Sleep Time | −0.024 | −0.024 | −0.119 | 0.549*** | −0.196* | 0.073 | −0.298*** | 0.741*** | 1.000 | ||
| 10 | Quality of Sleep | −0.393*** | −0.427*** | −0.280*** | 0.147 | −0.125 | −0.045 | −0.039 | 0.151 | 0.185* | 1.000 | |
| 11 | Refreshed Sleep | −0.439*** | −0.467*** | −0.315*** | 0.107 | −0.099 | 0.002 | −0.006 | 0.224** | 0.225** | 0.849*** | 1.000 |
| 12 | Horvath EAA | −0.128 | −0.110 | −0.009 | 0.070 | 0.000 | −0.159* | −0.120 | −0.106 | −0.029 | −0.070 | −0.088 |
| 13 | Hannum EAA | −0.061 | −0.040 | −0.003 | 0.005 | 0.044 | −0.144 | −0.068 | −0.065 | −0.029 | −0.127 | −0.119 |
| 14 | PhenoAgeAccel | −0.012 | −0.007 | −0.179* | −0.085 | −0.033 | 0.168* | 0.142 | 0.043 | 0.0151 | 0.075 | 0.097 |
| 15 | GrimAgeAccel | 0.226** | 0.208** | 0.102 | −0.176* | 0.124 | −0.049 | 0.166* | 0.015 | −0.123 | −0.123 | −0.086 |
| 16 | DunedinPACE | 0.412*** | 0.405*** | 0.279*** | −0.203** | 0.212** | 0.006 | 0.169* | 0.039 | −0.130 | −0.124 | −0.149* |
p < 0.05; ** p < 0.01; *** p < 0.001.
Mediating role of the pace of biological aging between the risk for obstructive sleep apnea and chronic pain outcomes
Figures 2 and 3 illustrate the mediation of the pace of biological aging between the risk for OSA and both pain severity and pain interference, which are summarized in Table 4. The total effect of OSA risk on pain severity (β = 3.766 (2.726), p = 0.007) and pain interference (β = 3.163 (1.187), p = 0.008) were significant after controlling for race, caffeine intake, insomnia severity, and medication usage (opioids, muscle relaxers, benzodiazepines). The indirect effects of risk of OSA through the pace of biological aging on both pain severity (β = 0.565 (0.322), Boot 95% CI [0.015, 1.270]) and pain interference (β = 0.566 (0.307), Boot 95% CI [0.046, 1.235]) were significant, supporting the mediating role of the pace of biological aging in the relationship between the risk for OSA and pain outcomes. The model including STOP-BANG, DunedinPACE, CLBP outcomes, and covariates accounted for 50.1% and 58.8% of the variance in pain severity and pain interference, respectively.
Figure 2.
The mediating model for risk for OSA, DunedinPACE, and pain severity.
Note: DunedinPACE: Dunedin Pace of Aging Computed from Epigenome.
Figure 3.
The mediating model for risk for OSA, DunedinPACE, and pain interference.
Note: DunedinPACE: Dunedin Pace of Aging Computed from Epigenome.
Table 4.
Bootstrapping indirect effects for the mediating model.
| Model paths | Bootstrap total effect | Bootstrap indirect effect | ||||
|---|---|---|---|---|---|---|
| Beta (SE) | LLCI | ULCI | Beta | LLCI | ULCI | |
| STOP-BANG → DunedinPACE → Pain severity | 3.766 (2.726) | 1.040 | 6.492 | 0.565 (0.322) | 0.015 | 1.270 |
| STOP-BANG → DunedinPACE → Pain interference | 3.163 (1.187) | 0.821 | 5.504 | 0.566 (0.307) | 0.046 | 1.235 |
LLCI: Lower Limit 95% Confidence Interval; ULCI: Upper Limit 95% Confidence Interval; SE: standard error. All models included race, caffeine intake, insomnia severity, and medication usage (opioids, muscle relaxers, benzodiazepines) as covariates.
Discussion
Given the restorative functions of sleep and adequate oxygenation, the intricate relationship between OSA risk, biological aging, and CLBP presents a complex interplay of physiological processes that can significantly impact health outcomes. Several authors have documented that poor sleep quality and insomnia can accelerate biological aging and exacerbate chronic pain.15,17 However, to our knowledge, this is the first study examining the relationship between OSA risk, pace of biological aging, and CLBP. Our findings highlight the enduring issue of OSA in biological aging and CLBP but also suggest that indirect factors could explain the relationship between the risk of OSA and self-reported outcomes of CLBP.
Previous studies have documented the reciprocal relationship between chronic pain and sleep quality, where poor sleep exacerbates chronic pain and chronic pain disrupts sleep.17,35 Frequently, self-reported measures of an individual’s sleep experience are used to determine sleep quality and restfulness. While other sleep metrics showed variations, sleep quality and restfulness were significantly lower in the high-impact pain group compared to both no and low-impact pain groups. These differences in sleep quality and restfulness align with broader research demonstrating the bidirectional relationship between chronic pain and sleep disturbances. 35 The compromised sleep experience in individuals with chronic pain can contribute to increased pain catastrophizing 36 and accelerated biological aging. 17 These findings highlight the complex and nuanced nature of sleep disruption in chronic pain populations to improve sleep quality and restfulness and the potential implications for pain management and biological aging.
The main finding concerning the link between OSA risk and CLBP outcome includes that the pace of biological aging, measured by DunedinPACE, partially mediates the relationship between OSA risk and CLBP severity and interference. This mediating role of the pace of biological aging remained significant after controlling covariates. This is notable, given that higher BMI is a strong predictor of OSA and the prevalence of OSA among young adults. 37 The results are consistent with previous studies that greater risk for OSA is associated with greater pain intensity 38 and pain interference. 4 While no study has directly examined the OSA-chronic pain link, previous works demonstrated that biological age acceleration associated with OSA is reversible after proper interventions are established. 39 This mediation pathway underscores not only the potential long-term health consequences of untreated OSA on biological aging and CLBP but also a potential target for therapeutic intervention. These findings suggest that early detection and management of OSA can improve sleep quality, mitigate accelerated biological aging, and reduce the effect of chronic pain.
The multifaceted relationship between OSA and chronic pain involves multiple causal mechanisms. 40 The intermittent hypoxia associated with sleep apnea may exacerbate pain sensitivity and contribute to central sensitization, a key feature in many chronic pain conditions. 41 Additionally, the fragmented sleep along with daytime fatigue resulting from sleep apnea can impair pain-coping mechanisms and amplify pain perception. 42 Both sleep apnea and chronic pain lead to accelerated aging via telomere shortening, 43 epigenetic alterations including aberrant DNA methylation and histone modifications, 44 and mitochondrial dysfunction resulting in disrupted ATP production and increased oxidative stress. 45 OSA risk was also associated with differential methylated genes involved with inflammation (ABCA1, ABCG1, CD36, FABP4,HMOX, NOS2, PEPCK, ADIPOQ, L1R2, AR, NPR2, SP140, ACTA1, HDAC2, SUMO1, FOXP3, IRF1)46,47 and cardiovascular dysfunction (eNOS). 44 These biological mechanisms are deeply intertwined and are better measured by DunedinPACE, which measures the pace of aging, than other epigenetic clocks, as it captures the rate of system decline across multiple physiological systems affected by both sleep conditions and chronic pain conditions. 17 Future research should focus on interventions addressing sleep apnea to improve sleep quality and pain management, which may potentially slow the pace of biological aging.
Our findings provide valuable insights for healthcare providers, researchers, and policymakers, as the cost of treating OSA and CLBP is heavy on public health. Collectively, the cost of sleep disorders and CLBP are estimated to exceed $200 billion annually.48,49 We can, therefore, expect that by understanding the nature of the relationship between OSA risk and CLBP, its above-mentioned effect can be mitigated. Knowing the biological link between OSA risk and CLBP can inform future interventions for pain management through optimal sleep quality. While our results do not estimate causality, they suggest that physiological processes involved in biological aging are influenced by OSA risk, such that greater risk for OSA positively correlates with a faster biological aging process and worse CLBP outcomes. Given that the pace of biological aging, measured with epigenetic clocks, represents a dynamic and reversible process, environmental, lifestyle, and behavioral modifications can have a positive impact. 40 Further longitudinal research is needed to establish temporal precedence, uncover underlying mechanisms, and identify relevant targets for interventions managing OSA, biological aging, and CLBP.
Strengths and limitations
This study has several strengths and limitations. Mediation analysis offers a robust and statistically powerful method to elucidate the complex causal relationships between OSA risk, biological aging, and CLBP outcomes. The use of well-phenotyped samples, coupled with valid and reliable instruments, enhances the generalizability of the findings. We also used a clinically relevant tool (STOP-BANG) to assess the risk for OSA, which enhances the translation of our findings into clinical practice. Another key strength of this study is the use of multiple epigenetic clocks, providing a comprehensive understanding of the underlying biological aging phenomena. Furthermore, the case-control study design enables direct comparisons between participants experiencing varying levels of pain impact, while the focus on NHB and NHW groups allows for distinct racial comparisons.
Our findings must be interpreted within the context of some limitations. First, even though the mediation model assumes a causal relationship, we used cross-sectional data, which does not allow us to draw direct causal effects, and a potential inverse causality should be considered: CLBP can increase the risk for OSA. Therefore, further longitudinal studies are needed. Second, using self-reported indicators in a cross-sectional study might artificially inflate variables’ covariance, which could impact variance estimates. However, all the instruments used in this study are valid and widely used in the literature. Additionally, while STOP-BANG is a validated tool, a diagnosis through formal sleep study would confirm the presence of OSA rather than assessing the likelihood based on risk factors. Lastly, the community context of the study, centered in Birmingham, Alabama, and surrounding areas, may limit the generalizability of the findings to the broader US population.
Conclusion
In conclusion, this study provides insight into the relationship between OSA risk, biological aging pace, and chronic low back pain outcomes. By employing epigenetic analysis and comprehensive pain assessments, we demonstrate that greater OSA risk is linked to faster biological aging, which correlates with higher pain severity and interference. Our mediation analysis further clarified this relationship by establishing that biological aging significantly mediates the association between OSA risk and pain outcomes. These findings continue to build our understanding of pain mechanisms, moving the field beyond traditional focus on sleep disturbance as solely detrimental to pain experience. This work establishes a foundation for novel clinical approaches that could target biological aging itself to mitigate chronic pain, creating an entirely new avenue for intervention beyond conventional sleep-focused treatments. Future research should build upon this mechanistic understanding to longitudinally assess the relationship between OSA, biological aging, and chronic pain outcomes over time.
Footnotes
Author contributions: The study was conceptualized by R.E.S., B.R.G., and E.N.A. Sample collection was performed by F.B.A.T.A., A.M.W., D.S.O., and S.S.G. Sample processing and data analysis was primarily performed by T.L.Q. and K.W.F. Initial manuscript preparation by K.W.F. with review and revision by F.B.A.T.A., K.R.K., S.J.T., W.W.F., S.L.C., M.A.O., R.E.S., B.R.G., and E.N.A. All authors approved the final manuscript.
Availability of data and material: Data will be provided upon reasonable requests and approval by ethical committee.
Consent for publication: All participants provided written informed consent prior to enrollment in the study.
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institutes of Health (Grant No. R01AR079178 and R01MD010441).
Ethics approval and consent to participate: The study was approved by the Institutional Review Board at the University of Alabama at Birmingham (IRB-170119003). All participants provided written informed consent prior to enrollment in the study. This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.
ORCID iDs: Khalid W Freij 
https://orcid.org/0009-0001-1795-2927
Fiona BAT Agbor
https://orcid.org/0009-0006-9169-5064
Philemon Domoyeri
https://orcid.org/0009-0000-1822-0712
Edwin N Aroke
https://orcid.org/0000-0002-8355-0870
References
- 1. Ling V. Sleep apnea statistics and facts you should know. National Council on Aging, 2024. [Google Scholar]
- 2. Yamashiro Y, Kryger MH. Why should sleep apnea be diagnosed and treated? Clin Pulm Med 1994; 1(4): 250–259. [Google Scholar]
- 3. Mitra AK, Bhuiyan AR, Jones EA. Association and risk factors for obstructive sleep apnea and cardiovascular diseases: a systematic Review. Diseases 2021; 9(4): 88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Larsen DB, Bendix L, Abeler K, et al. Obstructive sleep apnea is common in patients with high-impact chronic pain—an exploratory study from an interdisciplinary pain center. Scand J Pain 2022; 22: 106–117. [DOI] [PubMed] [Google Scholar]
- 5. Mundt JM, Eisenschenk S, Robinson ME. An examination of pain’s relationship to sleep fragmentation and disordered breathing across common sleep disorders. Pain Med 2018; 19: 1516–1524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Charokopos A, Card ME, Gunderson C, et al. The association of obstructive sleep apnea and pain outcomes in adults: a systematic review. Pain Med 2018; 19: S69–S75. [DOI] [PubMed] [Google Scholar]
- 7. Chivers SB, Andrade MA, Hammack RJ, et al. Peripheral macrophages contribute to nociceptor priming in mice with chronic intermittent hypoxia. Sci Signal 2024; 17: eadn8936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Oehler B, Brack A, Blum R, et al. Pain control by targeting oxidized phospholipids: functions, mechanisms, perspectives. Front Endocrinol 2021; 11: 613868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lin W, Zhang Z, Wang C, et al. Genetic overlap between obstructive sleep apnea and ischemic stroke: a large-scale genome-wide cross-trait analysis. Nat Sci Sleep 2025; 17: 413–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Li X, Joehanes R, Hoeschele I, et al. Association between sleep disordered breathing and epigenetic age acceleration: evidence from the Multi-Ethnic Study of Atherosclerosis. EBioMedicine 2019; 50: 387–394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Pinilla L, Santamaria-Martos F, Benítez ID, et al. Association of obstructive sleep apnea with the aging process. Ann Am Thorac Soc 2021; 18: 1540–1547. [DOI] [PubMed] [Google Scholar]
- 12. Lu AT, Quach A, Wilson JG, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging 2019; 11: 303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Belsky DW, Caspi A, Corcoran DL, et al. DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife 2022; 11: e73420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Tsatsaraki E, Bouloukaki I, Kontakis G, et al. Associations between combined psychological and lifestyle factors with pain intensity and/or disability in patients with chronic low back pain: a cross-sectional study. Healthcare (Basel) 2023; 11(22): 2928. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Liu L, Li X, Xue P, et al. Subjective sleep disruption and mood disorders are associated with the risk of chronic pain in patients with obstructive sleep apnea. Nat Sci Sleep 2022; 14: 2023–2032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Aroke EN, Srinivasasainagendra V, Kottae P, et al. The pace of biological aging predicts nonspecific chronic low back pain severity. J Pain 2024; 25(4): 974–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Aroke EN, Wiggins AM, Hobson JM, et al. The pace of biological aging helps explain the association between insomnia and chronic low back pain. Mol Pain 2023; 19: 17448069231210648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chou R, Qaseem A, Snow V, et al. Diagnosis and treatment of low back pain: a joint clinical practice guideline from the American College of Physicians and the American Pain Society. Ann Intern Med 2007; 147: 478–491. [DOI] [PubMed] [Google Scholar]
- 19. Von Korff M, DeBar LL, Krebs EE, et al. Graded chronic pain scale revised: mild, bothersome, and high-impact chronic pain. Pain 2020; 161: 651–661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Shohet KL, Landrum RE. Caffeine consumption questionnaire: a standardized measure for caffeine consumption in undergraduate students. Psychol Rep 2001; 89: 521–526. [DOI] [PubMed] [Google Scholar]
- 21. Addicott MA, Yang LL, Peiffer AM, et al. The effect of daily caffeine use on cerebral blood flow: how much caffeine can we tolerate? Hum Brain Mapp 2009; 30: 3102–3114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Frary CD, Johnson RK, Wang MQ. Food sources and intakes of caffeine in the diets of persons in the United States. J Am Diet Assoc 2005; 105: 110–113. [DOI] [PubMed] [Google Scholar]
- 23. Wood AC, Kuntsi J, Asherson P, et al. Actigraph data are reliable, with functional reliability increasing with aggregation. Behav Res Methods 2008; 40: 873–878. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Carney CE, Buysse DJ, Ancoli-Israel S, et al. The consensus sleep diary: standardizing prospective sleep self-monitoring. Sleep 2012; 35: 287–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bastien CH, Vallieres A, Morin CM. Validation of the insomnia severity index as an outcome measure for insomnia research. Sleep Med 2001; 2: 297–307. [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] [PubMed] [Google Scholar]
- 27. Aroke EN, Jackson P, Meng L, et al. Differential DNA methylation in Black and White individuals with chronic low back pain enrich different genomic pathways. Neurobiol Pain 2022; 11: 100086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Aryee MJ, Jaffe AE, Corrada-Bravo H, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics 2014; 30: 1363–1369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Touleimat N, Tost J. Complete pipeline for Infinium(R) Human Methylation 450K BeadChip data processing using subset quantile normalization for accurate DNA methylation estimation. Epigenomics 2012; 4: 325–341. [DOI] [PubMed] [Google Scholar]
- 30. Horvath S, Raj K. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet 2018; 19: 371–384. . [DOI] [PubMed] [Google Scholar]
- 31. Hannum G, Guinney J, Zhao L, et al. Genome-wide methylation profiles reveal quantitative views of human aging rates. Mol Cell 2013; 49: 359–367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Levine ME, Lu AT, Quach A, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging 2018; 10: 573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Horvath S. DNA methylation age of human tissues and cell types. Genome Biol 2013; 14: R115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Hayes AF. Introduction to mediation, moderation, and conditional process analysis. Guilford Press, 2018. [Google Scholar]
- 35. Seiger AN, Penzel T, Fietze I. Chronic pain management and sleep disorders. Cell Rep Med 2024; 5: 101761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Gerhart JI, Burns JW, Post KM, et al. Relationships between sleep quality and pain-related factors for people with chronic low back pain: tests of reciprocal and time of day effects. Ann Behav Med 2017; 51: 365–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Zasadzińska-Stempniak K, Zajączkiewicz H, Kukwa A. Prevalence of obstructive sleep apnea in the young adult population: a systematic review. J Clin Med 2024; 13: 1386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Athar W, Card ME, Charokopos A, et al. Obstructive sleep apnea and pain intensity in young adults. Ann Am Thorac Soc 2020; 17: 1273–1278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Cortese R, Sanz-Rubio D, Kheirandish-Gozal L, et al. Epigenetic age acceleration in obstructive sleep apnoea is reversible with adherent treatment. Eur Respir J 2022; 59(4): 2103042. [DOI] [PubMed] [Google Scholar]
- 40. Sanders AE, Greenspan JD, Fillingim RB, et al. Associations of sleep disturbance, atopy, and other health measures with chronic overlapping pain conditions. J Oral Facial Pain Headache 2020; 34: S73–S84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Kaczmarski P, Karuga FF, Szmyd B, et al. The role of inflammation, hypoxia, and opioid receptor expression in pain modulation in patients suffering from obstructive sleep apnea. Int J Mol Sci 2022; 23(16): 9080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Nikolaou K, Voulgaris A, Steiropoulos P. Obstructive sleep apnea and pain. Curr Sleep Med Rep 2019; 5: 95–103. [Google Scholar]
- 43. Kim KS, Kwak JW, Lim SJ, et al. Oxidative stress-induced telomere length shortening of circulating leukocyte in patients with obstructive sleep apnea. Aging Dis 2016; 7: 604–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Kheirandish-Gozal L, Khalyfa A, Gozal D, et al. Endothelial dysfunction in children with obstructive sleep apnea is associated with epigenetic changes in the eNOS gene. Chest 2013; 143: 971–977. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Mesarwi OA, Loomba R, Malhotra A. Obstructive sleep apnea, hypoxia, and nonalcoholic fatty liver disease. Am J Respir Crit Care Med 2019; 199: 830–841. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Chen YC, Chen TW, Su MC, et al. Whole genome DNA methylation analysis of obstructive sleep apnea: IL1R2, NPR2, AR, SP140 methylation and clinical phenotype. Sleep 2016; 39: 743–755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Cortese R, Zhang C, Bao R, et al. DNA methylation profiling of blood monocytes in patients with obesity hypoventilation syndrome: effect of positive airway pressure treatment. Chest 2016; 150: 91–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Faria A, Allen AH, Fox N, et al. The public health burden of obstructive sleep apnea. Sleep Sci 2021; 14: 257–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Dutmer AL, Schiphorst Preuper HR, Soer R, et al. Personal and societal impact of low back pain: the Groningen spine cohort. Spine 2019; 44: E1443–E1451. [DOI] [PubMed] [Google Scholar]