Associations between early sleep-disordered breathing following moderate-to-severe traumatic brain injury and long-term chronic pain status: a Traumatic Brain Injury Model Systems study

Abstract

Study Objectives:

To explore the relationship between polysomnography-derived respiratory indices and chronic pain status among individuals following traumatic brain injury (TBI).

Methods:

Participants (n = 66) with moderate to severe TBI underwent polysomnography during inpatient acute rehabilitation and their chronic pain status was assessed at 1- to 2-year follow-up as part of the TBI Model Systems Pain Collaborative Study. Pairwise comparisons across pain cohorts (ie, chronic pain, no history of pain) were made to explore differences on polysomnography indices.

Results:

Among our total sample, approximately three-quarters (74.2%) received sleep apnea diagnoses utilizing American Academy of Sleep Medicine criteria, with 61.9% of those endorsing a history of chronic pain. Of those endorsing chronic pain, the average pain score was 4.8 (standard deviation = 2.1), with a mean interference score of 5.3 (2.7). Pairwise comparisons revealed that those endorsing a chronic pain experience at follow-up experienced categorically worse indicators of sleep-related breathing disorders during acute rehabilitation relative to those who did not endorse chronic pain. Important differences were observed with elevations on central (chronic pain: 2.6; no pain: 0.8 per hour) and obstructive apnea (chronic pain: 15.7; no pain: 11.1 per hour) events, as well as oxygen desaturation indices (chronic pain: 19.6; no pain: 7.9 per hour).

Conclusions:

Sleep-disordered breathing appears worse among those who endorse chronic pain following moderate-to-severe TBI, but additional research is needed to understand its relation to postinjury pain. Prospective investigation is necessary to determine how clinical decisions (eg, opioid therapy) and intervention (eg, positive airway pressure) may mutually influence outcomes.

Clinical Trial Registration:

Registry: ClinicalTrials.gov; Name: Comparison of Sleep Apnea Assessment Strategies to Maximize TBI Rehabilitation Participation and Outcome (C-SAS); URL: https://clinicaltrials.gov/ct2/show/NCT03033901; Identifier: NCT03033901.

Citation:

Martin AM, Pinto SM, Tang X, et al. Associations between early sleep-disordered breathing following moderate-to-severe traumatic brain injury and long-term chronic pain status: a Traumatic Brain Injury Model Systems study. J Clin Sleep Med. 2023;19(1):135–143.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Sleep apnea, which is prevalent following traumatic brain injury (TBI), is associated with poor neuro-rehabilitation outcomes and may be associated with high rates of chronic pain that persists many years following TBI. Understanding the relationship between polysomnography-diagnosed sleep apnea and chronic pain may help improve treatment and rehabilitation in this at-risk population.

Study Impact: This study provides exploratory findings to support further investigation of the relationship between sleep apnea and chronic pain following moderate-to-severe TBI. Such findings underscore the potential value in exploring how early intervention for sleep-related breathing disorders may influence pain-related outcomes and highlights an area of necessary investigation in this clinical population.

INTRODUCTION

Traumatic brain injury (TBI), the resulting disruption of brain function and neuropathology following damage due to external force,¹ is associated with significant disturbances in sleep, as well as high rates of pain.^2,3 Sleep and pain comorbidities persist across the trajectory of TBI recovery.^4,5 A recently completed multicenter trial conducted during inpatient rehabilitation within the National Institute on Disability Independent Living and Rehabilitation Research and Veterans Affairs TBI Model Systems reported that two-thirds of the study cohort had polysomnography (PSG)-confirmed sleep apnea.⁶ Further, epidemiologic work in the TBI Model Systems has identified chronic pain (ie, pain greater than 3 months’ duration) and sleep disorders, such as sleep apnea, among the most prevalent medical comorbidities at 2 and 5 years post-TBI.^7,8 Additional understanding of the relationship between sleep and pain is especially needed among individuals following TBI, as increased sleep disturbance and pain are associated with higher odds of global disability in this at-risk population.^9,10

Cross-sectional investigations of adults with a TBI history have found associations between self-report measures of nonspecific sleep disturbance or insomnia and pain severity ratings.^11–14 Increased pain on visual analog rating scales and pain interference are seen among those defined as “poor sleepers” at 1 year following moderate to severe TBI.^15,16 Among individuals with mild to severe TBI, worst pain severity ratings independently contribute to self-report of overall sleep quality in multilinear models that account for depression and anxiety symptoms.¹⁷ Veterans with mild to moderate TBI experiencing post-traumatic headache report increased sleep disturbance and daytime fatigue compared with those without headache and non-TBI controls.¹⁸ These associations are also seen in longitudinal studies,¹⁹ where self-reported sleep complaints within 10 days of a mild TBI are associated with a 6.3-fold increase in headache at 6 weeks postinjury.

The relationship between sleep disturbance and pain is influenced by multiple factors in the context of acute post-TBI care,²⁰ where managing pain from musculoskeletal injury is balanced with concerns about the cognitive side effects of opioid medications and subsequent interference with active engagement in necessary rehabilitation.²¹ Yet, opioid medications are the most commonly administered class of psychotropic medications during recovery and are more likely to be prescribed with increasing TBI severity in the context of outpatient pain management despite significant risk.^22,23 Opioids may indirectly impact pain experience after moderate to severe TBI by contributing to the presence or exacerbation of sleep apnea symptoms, including increased apneas and worse oxygen desaturation.²⁴ Sleep apnea itself may contribute to systemic inflammation and have a hyperalgesia effect, while also potentiating the analgesic response to opioids.²⁵ Opioids are also likely to have an ongoing direct impact on pain experience following injury in the onset of physiological tolerance and the development of opioid-induced hyperalgesia.^26,27 Opioid-induced hyperalgesia is the nociceptive sensitization from exposure to opioids.²⁸ The iatrogenic result of treating pain with opioids is a paradoxical increase in the sensitivity to painful stimuli, with endogenous opioid dysregulation implicated in the high comorbidity between TBI and chronic pain.²⁹

Despite a significant degree of co-occurrence, the relationship between sleep apnea and chronic pain has received limited attention, likely due to the challenges in obtaining PSG during neurorehabilitation. This contributes to a dearth of literature examining the relationship between objectively diagnosed sleep-related breathing disorders and chronic pain following TBI. This is despite the fact that hypoxemia and sleep fragmentation, both components of sleep apnea, are known to enhance pain.³⁰ As such, sleep-related breathing disorders may represent an important intervention target for which the relationship with pain-related outcomes following TBI is unknown. To date, only 1 study has utilized PSG to examine the relationship between sleep apnea and pain following TBI. Khoury and colleagues³¹ reported increases in rapid electroencephalographic activity among 8 participants following mild TBI who endorsed acute pain, but apnea-hypopnea index (AHI) scores did not differ compared with healthy controls or those without postinjury acute pain. Additional information relevant to apnea diagnoses, such as oxygen desaturation, was not reported.

The current literature regarding the association between sleep disturbance and chronic pain following TBI is limited by self-report measures, which does not allow for focused investigation of sleep apnea and its association with the prominent long-term medical complaint of chronic pain. Currently, no studies have examined the relationship between sleep apnea and pain at a follow-up period greater than 3 months, the generally accepted temporal demarcation for chronic pain. Lengthier follow-up times are necessary to begin to examine the possible role of sleep disorders in pain chronification following TBI and to distinguish this relationship from the experience of acute pain (ie, <3 months postinjury). Additionally, the relationship between sleep apnea and pain following TBI may differ by severity of injury, yet the only prior objective examination of sleep apnea and pain occurred among those with mild TBI.³¹ It is possible that sleep apnea early after injury is differentially associated with pain status in a more severe TBI population, but this relationship has yet to be explored.

The current study leverages multiple datasets collected through the TBI Model Systems (TBIMS) and contributes to the extant literature by examining the relationship between objective PSG indices, the criterion standard for characterizing sleep apnea, and chronic pain in a multicenter cohort of individuals with a history of moderate to severe TBI. Additionally, the current study assesses pain at long-term follow-up (ie, up to 2 years postinjury), which captures a validly defined chronic pain status among a relatively large, well-characterized sample. The primary aim of this study is to address this gap by examining differences in PSG-derived indicators of sleep-related breathing disorders (ie, AHI, oxygenation) in those who do and do not report chronic pain at 1 to 2 years following injury. Additionally, the pain-related outcomes (ie, severity and interference) among participants following moderate to severe TBI will be described. The purpose of this analysis is to assess if a modifiable condition commonly seen early in the trajectory of TBI recovery—namely, sleep apnea—warrants further study to determine potential downstream effects in the development or maintenance of chronic pain.

METHODS

Participants

Data from the current study were taken from 2 separate studies that included participants enrolled in the TBIMS lifetime longitudinal study at 6 sites (Tampa, FL; Seattle, WA; Dallas, TX; Columbus, OH; Englewood, CO; and Elkins Park, PA). The first study was a Patient-Centered Outcomes Research Institute (PCORI)–funded diagnostic comparative effectiveness trial of sleep apnea screening tools compared against the criterion standard, level 1 PSG.^6,32–34 All participating sites received institutional review board approval for conduct of the study, with consent obtained by patients or designated medical proxy using TBIMS procedures. Inclusion in the TBIMS longitudinal study for moderate to severe injury relies on several clinical indices, including the Glasgow Coma Scale (GCS),³⁵ a brief early measure of altered consciousness typically obtained in the emergency department. The initial GCS score obtained upon presentation to the designated level 1 trauma center associated with each TBIMS site is the most consistently obtained marker of injury severity for the current study sample. GCS is not a stand-alone diagnostic tool and participants who met criteria for moderate to severe TBI based on other criteria such as post-traumatic amnesia over 24 hours and/or abnormal imaging findings were also included in the study.³⁶ Complete study inclusion and exclusion criteria for the PCORI trial are described in prior publications.^6,34 Of the 248 individuals who participated in the PCORI trial, 66 participated in the second study: the National Institute on Disability Independent Living and Rehabilitation Research–funded TBIMS Collaborative Study examining pain after TBI.

Procedure

In the PCORI study, consistent with American Academy of Sleep Medicine (AASM) guidelines, consecutive admissions were screened for eligibility which included at least 2 hours of sleep per night required for a valid sleep study, as well as medical stability (ie, no emergent medical issues precluding overnight PSG, minimal post-traumatic agitation).³⁷ Sleep duration to determine eligibility for PSG for this study was recorded using a wrist-worn accelerometer (Actiwatch Spectrum; Philips/Respironics, Bend, OR). Once determined eligible, an overnight PSG study was conducted by a registered polysomnographic technician (RPSGT) in the participant’s own hospital bed on the inpatient rehabilitation unit. Fully attended level 1 PSG was conducted in accordance with the AASM-recommended procedures.^37,38 Staff were instructed to allow participants their normal sleep period.

The lead center (James A. Haley Veterans’ Hospital, Tampa, FL) served as a centralized scoring center for all sleep studies. All de-identified studies were scored by 1 of 2 certified RPSGTs within 7 days of PSG completion and interpreted by a board-certified sleep medicine physician. Staff who scored and interpreted studies were masked to other sleep assessments. As previously described, the interrater agreement and scoring of sleep and respiratory events were strong to very strong, with most of the studies rated as interpretable/scoreable (97.2%).^6,34 Additionally, a subset of PCORI participants who had been diagnosed with sleep apnea were contacted to determine if treatment was sought, including verbal report of continuous positive airway pressure (CPAP) adherence.

A subset (n = 66) of those individuals who were also enrolled in the TBIMS and completed PSG were later invited to participate in the second study funded by the National Institute on Disability, Independent Living, and Rehabilitation Research on pain after TBI during their 1- or 2-year TBIMS follow-up. Additional criteria included being able to respond in English and able to complete the pain interview on their own (without a proxy). Participants who consented were asked if they were experiencing chronic pain to include headaches or pain anywhere in the body that occur more than half the days over a 3-month period (“current pain”), whether they had experienced chronic pain since their TBI but not currently (“past pain”), or had no chronic pain since injury (“no pain”). Individuals were offered to complete their relevant survey via interview, online, or via a mail-out survey.

Measures

Demographics, preinjury medical histories, and medical record abstraction were conducted by trained research assistants, following the TBIMS standard procedures.³³ The medical records of participants were reviewed and data regarding medications administered on the day of PSG were abstracted via previously published protocol.⁶ All follow-up data were completed via self-report. Pain survey items included location of pain, pain severity and interference, as well as sleep quality.

Polysomnography

Participants in the PCORI-funded study underwent administration of level 1 PSG (Philips Alice 6 LDx Diagnostic Sleep System with Philips Sleepware G3 version 3.8.1 software, Philips/Respironics, Bend, OR).³⁹ Sleep indices captured during PSG included the following: total sleep time, wake after sleep onset, arousal and awakening indices, and percentage of total sleep time spent in rapid eye movement (REM) sleep. Severity of sleep apnea was measured by the apnea-hypopnea index (AHI), using both AASM and Centers for Medicaid and Medicare services (CMS) scoring criteria with apneas defined as 90% or greater decrease in airflow for 10 seconds. Hypopneas were defined as a 30–90% reduction in airflow with at least a 3% decrease in oxygen saturation or an arousal (AASM) or 4% decrease in oxygen saturation and no consideration of arousals (CMS). Parameters collected for the purpose of this study included the following: central apnea events, obstructive apneas and hypopneas, and total AHI. A diagnosis of sleep apnea is determined by an AHI ≥ 5 events/h. Severity of sleep apnea is defined by AHI events per hour, with 5–14 denoting mild, 15–29 denoting moderate, and ≥ 30 indicating severe sleep apnea. The oxygen desaturation index for total sleep time reflects the number of desaturations by 3% per hour.

Pain severity and interference

The Brief Pain Inventory is an 11-item self-report measure of pain severity and the degree to which pain interferes with common domains of functioning impacted by pain. All items are rated on a 0 (“no pain”/“did not interfere”) to 10 (“pain as bad as you can imagine”/“completely interfered”) Likert-type scale.⁴⁰ It provides a mean-derived composite pain severity score of 4 somatic or sensory items (ie, worst, least, average, and current pain) and a composite interference score of 7 daily activities (ie, general activity, walking, work, mood, enjoyment of life, relations with others, and sleep).

Sleep quality

The Pittsburgh Sleep Quality Index (PSQI) is a 19-item measure evaluating sleep quality and disturbances over the past month.⁴¹ Scores are interpreted by a total score and component areas including sleep latency, duration, efficiency, use of sleeping medication, and daytime dysfunction. The PSQI has demonstrated good discriminative validity, differentiating those with sleep disorders from healthy controls with a PSQI total of greater than 5.

Analytical approach

Data were analyzed using statistical software R (R Foundation for Statistical Computing, Vienna, Austria). Summary statistics were expressed as mean (standard deviation [SD]) and quartiles (minimum, first quartile, median, third quartile, and maximum) for continuous variables and count (percentage) for categorical variables. Pairwise comparisons were made in each PSG index between participants with “no pain,” “past pain,” and “current pain,” as defined above. In the current study we did not perform any hypothesis testing (either parametric or nonparametric) as an a priori hypothesis was not formulated. Mean differences were estimated and a classification schema introduced in prior work leveraging the TBIMS national database⁴² was adopted to determine the clinical significance of each difference due to the exploratory nature of this study, relatively small sample size, and the number of comparisons made across PSG parameters (which may result in an inflated overall type 1 error rate from multiple hypothesis testing). The magnitude of differences was classified into 3 categories: mean differences < 25% of 1 SD were classified as insignificant (none); those ≥ 25% but < 50% of 1 SD were classified as minimal (minor); and those ≥ 50% of 1 SD were classified as important. The mean to SD ratio in this approach reflects the effect size difference between the groups. For comparisons to the “no pain” cohort, the SD of the “no pain” cohort was used as the reference SD. For comparisons between the “past pain” and “current pain” cohorts, the SD of the “past pain” cohort was used as the reference SD.

RESULTS

The final study sample consisted of 26 (39.4%) participants in the “no pain” cohort, 13 (19.7%) participants in the “past pain” cohort, and 27 (40.9%) participants in the “current pain” cohort. Demographics and characteristics related to the injury and PSG are summarized in Table 1 for the overall study sample and for each cohort. To fully characterize this sample, additional descriptive characteristics, including median and interquartile range, are presented in Table S1 ^{(38.6KB, pdf)} in the supplemental material. Study participants were predominantly male (81.8%), and the average age was 41.6 years. Overall, 60% of the study participants were White, while the “past pain” cohort included fewer White and more Hispanic participants as compared with the other 2 cohorts. Approximately 40% of the study participants were married. Injury severity was comparable among the 3 cohorts. The average time from injury to PSG was 66.8 days. A greater number of participants were administered opioids on the day of PSG in the “current pain” cohort than that in the “past pain” cohort and that in the “no pain” cohort (44.4% vs 30.8% vs 11.5%), which was consistent with any history of opioid prescription at follow-up (66.7% vs 61.5% vs 38.5%). The average time from injury to follow-up was 610 days.

Table 1.

Sample characteristics.

	Total (n = 66)	No Pain (n = 26)	Past Pain (n = 13)	Current Pain (n = 27)
Demographics
Age, mean (SD), y	41.6 (14.5)	39.9 (16.8)	47.2 (14.2)	40.4 (12)
Male, n (%)	54 (81.8)	23 (88.5)	12 (92.3)	19 (70.4)
Race/ethnicity, n (%)
White	39 (60)	16 (64)	5 (38.5)	18 (66.7)
Black	11 (16.9)	6 (24)	3 (23.1)	2 (7.4)
Hispanic	9 (13.8)	0 (0)	4 (30.8)	5 (18.5)
Other	6 (9.2)	3 (12)	1 (7.7)	2 (7.4)
Married, n (%)	26 (39.4)	12 (46.2)	5 (38.5)	9 (33.3)
At time of injury, mean (SD)
BMI, kg/m2	26 (4.7)	25.4 (4)	25.9 (6.1)	26.6 (4.9)
Severity of injury (GCS)	10.1 (4.7)	9.8 (4.6)	11.8 (4.4)	9.8 (5.1)
Time from injury to PSG	66.8 (62.7)	66.3 (74.3)	67.5 (39.4)	66.9 (61.6)
SRBD diagnoses, n (%)
Obstructive sleep apnea (AASM)	43 (67.2)	17 (68)	10 (76.9)	16 (61.5)
Central sleep apnea (AASM)	3 (4.7)	1 (4)	0 (0)	2 (7.7)
Obstructive sleep apnea (CMS)	39 (69.6)	14 (73.7)	9 (75)	16 (64)
Central sleep apnea (CMS)	2 (3.6)	0 (0)	0 (0)	2 (8)
At follow-up, mean (SD)
Injury to follow-up, days	610 (171)	566.7 (197.9)	682.6 (138.7)	615.1 (149.8)
PSQI total score	6.2 (4.3)	4.6 (3.2)	5.2 (4.2)	8.2 (4.7)
BMI, kg/m2	26.9 (5.1)	25.8 (3.8)	26.1 (5.4)	28.4 (5.8)

AASM = American Academy of Sleep Medicine, BMI = body mass index, CMS = Centers for Medicare and Medicaid Services, GCS = Glascow Coma Scale, PSQI = Pittsburgh Sleep Quality Index, PSG = polysomnography, SD = standard deviation, SRBD = sleep-related breathing disorder.

In the interim from PSG to pain follow-up, a subset of PCORI trial participants (n = 30) who had been diagnosed with sleep apnea were contacted to determine if treatment was sought following diagnosis and to what extent participants had been adherent to CPAP therapy. Adherence data were gathered, on average, 441 days following PSG, with only 6.1% (n = 4) of the total study sample endorsing adherence of any kind and no observable difference between “current pain,” “past pain,” and “no pain” cohorts. Of those with “current pain” 45.8% (n = 11) produced PSQI total scores at follow-up above the threshold for being considered poor sleepers, compared to 8.7% (n = 2) and 8.3% (n = 1) in the “no pain” and “past pain” cohorts, respectively. Those with “current pain” reported an average pain rating of 4.8 (2.1) with an average pain interference score of 5.3 (2.7). Those endorsing “past pain” reported lower overall pain and interference scores of 4.2 (1.4) and 4.8 (1.2), respectively.

Results from pairwise comparisons of PSG indices are displayed in Table 2. Important differences were found between the “current pain” and “no pain” cohorts on obstructive AHI (CMS; 12.8 vs 6.3), central events (2.6 vs 0.8), total AHI (CMS; 16.5 vs 7.8), and desaturation index (19.6 vs 7.9), with scores being higher among those in the “current pain” cohort. Similarly, the “past pain” cohort also had higher scores and demonstrated important differences relative to the “no pain” cohort on obstructive AHI (CMS; 17.3), total AHI (CMS; 18.6), and desaturation index (21.7). Additionally, important differences were found between “past pain” and “no pain” on obstructive (21.4 vs 11.1) and total AHI (22.8 vs 12.4) when considering the AASM scoring criteria. Differences between “current pain” and “no pain” cohorts using AASM scoring criteria were deemed minor. One important difference was seen in comparisons between the “current pain” and “past pain” cohorts with central event scores of 1.8 and 0.1, respectively. The results of pairwise comparisons of PSG-derived sleep indices are presented in Table S2 ^{(38.6KB, pdf)} , although all differences were determined to be minor across all cohorts. Although stricter scoring criteria of CMS (4% vs 3% O₂ saturation and no consideration of arousals) resulted in lower AHI scores, this may have highlighted a difference in both obstructive and total AHI between the “current” and “no pain” cohorts, which was considered minor when applying the AASM scoring criteria. Differences are likely a function of the categorical schema used for analyses, as a 50% or greater mean difference of 1 SD of the “no pain” group would be considered important. For example, while total AHI total (AASM) was categorized as a minor difference between the “current pain” and “no pain” cohorts, there was a trend, as this resulted in a 48% (ie, minor) mean difference.

Table 2.

Proportional differences between pain cohorts on polysomnography-derived respiratory indices.

	Mean (SD)			Category*
	No Pain (n = 26)	Past Pain (n = 13)	Current Pain (n = 27)	Difference (Past vs No)	Difference (Curr. vs No)	Difference (Curr. vs Past)
Obstructive AHI (AASM)	11.1 (10.8)	21.4 (24.9)	15.7 (19.8)	10.4	4.7	−5.7
	[1]	[0]	[1]	Important	Minor	None
Obstructive AHI (CMS)	6.3 (9.7)	17.3 (22.5)	12.8 (21)	11.1	6.5	−4.5
	[4]	[1]	[5]	Important	Important	None
Central events	0.8 (1.7)	0.9 (2)	2.6 (6.5)	0.1	1.8	1.7
	[1]	[0]	[1]	None	Important	Important
Total AHI (AASM)	12.4 (13.7)	22.8 (26.1)	19 (23.4)	10.4	6.6	−3.7
	[1]	[0]	[1]	Important	Minor	None
Total AHI (CMS)	7.8 (13.1)	18.6 (23.9)	16.5 (25.7)	10.8	8.7	−2.1
	[4]	[1]	[5]	Important	Important	None
SpO2 < 90% (minutes)	3 (13)	6.6 (16.8)	4.9 (11)	3.6	1.9	−1.7
				Minor	None	None
Oxygen desaturation index	7.9 (9.9)	21.7 (27.2)	19.6 (23.3)	13.7	11.7	−2
				Important	Important	None

Indices presented for total sleep time. Values in brackets “[ ]” indicate missing. *Category: for continuous comparisons, < 25% mean difference of 1 SD of the “no pain” cohort = immaterial difference (none); ≥ 25–49.99% mean difference of 1 SD = minor difference (minor); and > 50% mean difference of 1 SD = important difference (important). The “past pain” cohort serves as the reference group in final comparison. AHI = apnea-hypopnea index, AASM = American Academy of Sleep Medicine (criteria), CMS = Centers for Medicaid and Medicare Services (criteria), Curr. = current pain cohort, missing = missing participant data, SD = standard deviation, SpO₂ = oxygen saturation.

DISCUSSION

To the best of our knowledge, this is the first study to explore the relationship between objectively assessed indicators of sleep-disordered breathing and chronic pain status following moderate to severe TBI. Sixty-one percent of participants endorsed chronic pain at 1 or 2 years post-injury and important differences were found on respiratory measures between those who did and did not endorse chronic pain, including those who indicated a resolution of pain prior to long-term follow-up (“past pain”). Participants who endorsed any history of chronic pain (“past pain” and “current pain”) experienced greater oxygen desaturation, total AHI, and obstructive apneas compared with their “no pain” counterparts. Those endorsing “current pain” at follow-up also experienced increased central apneas relative to those who did not endorse pain or experienced resolution (“no pain” and “past pain”).

These findings are consistent with the general literature demonstrating a relationship between sleep apnea and chronic pain in which the underlying pathophysiological mechanisms have not been fully elucidated.³⁰ An overall worse profile of sleep-disordered breathing among those endorsing chronic pain is also consistent with findings of increased pain among those with self-reported sleep disturbances following TBI.⁹ However, earlier work that examined objectively measured differences in AHI and the presence of chronic pain within mild TBI did not find a relationship.³¹ Differences between the 2 studies may be related to TBI severity, timing of assessment of pain, and small sample size in the initial study. In earlier work by Khoury and colleagues³¹ the average AHI score was 6.25 events/h compared to 18.07 events/h in the current study sample, which may be related to injury severity or timing of assessment. Additionally, the current study examined chronic pain while Khoury et al focused on more acute pain, which may not necessarily capture the same experience of pain, which can change over time.^43,44

Given the heterogeneity of TBI severity and cause, the explanation for a high prevalence of sleep apnea following TBI remains unknown, although it may reflect an unrecognized premorbid condition, which ultimately leads to an increased risk of accidents resulting in TBI.⁴⁵ Additionally, given the prevalence of dysphagia following TBI, a loss of pharyngeal muscle tone secondary to neurologic injury may impact control of breathing post-injury. Further, earlier work established an association between receipt of opioid medications and sleep-disordered breathing following injury, particularly central apneas.²⁴ Opioids may be administered in the acute recovery period following traumatic injury for pain management. In the current study a greater proportion of those in the “current pain” and “past pain” cohorts received opioids on the day of PSG (see Table S3 ^{(38.6KB, pdf)} ) and endorsed historical receipt of opioids at long-term follow-up. Unfortunately, interim pharmacy data are not available for analysis in the current study to discern patterns of use, including dosage, aberrancy, and utilization of partial vs full opioid agonists for pain management. It has been reported that multiple independent risk factors associated with the adverse consequences of opioid use are present following TBI, including barriers to accessing pain care, post-injury misuse, neurobehavioral changes, and the potential for overprescribing,²⁶ highlighting the need for future prospective studies to explore opiate use as a potential moderator in the relationship between pain and sleep-disordered breathing. Given the high receipt of opioid medication among our sample, those in the “current pain” and “past pain” cohorts were potentially maintained on opioid medications following discharge, suggesting the possibility that worse sleep study metrics among those with a history of chronic pain in the current study could be attributed to opioid-induced hyperalgesia.

Although the impact of sleep apnea on pain remains obscured by factors unaccounted for in the current study, it is important to recognize that sleep-related disruptions to dopaminergic and opioidergic signaling and the influence of disrupted sleep on affect and maladaptive coping responses have been posed as potential mechanisms for the relationship between sleep disturbance and pain.⁴⁶ Hypoxemia, for instance, contributes to the sensitization of pain receptors and is associated with an increase in inflammatory markers.³⁰ Additionally, repeated arousals associated with apneas may contribute to sleep fragmentation and inefficiencies in sleep, encouraging aberrant glial activity which alters the nociceptive system, increasing susceptibility to maladaptive plasticity and central sensitization of pain.^47,48 Notably, the glymphatic system has been proposed as a potential shared neural pathway linking sleep-wake disorders and poorer outcomes after TBI, including greater post-TBI pain. The glymphatic system’s clearance of interstitial waste, which is active during sleep, becomes dysfunctional post-TBI.^49,50 This results in inefficiencies in the clearance of neuropeptides from the perivascular space, an important factor in the development of some chronic pain conditions.⁵¹ Thus, impaired sleep from untreated sleep apnea may contribute to disruption of the glymphatic system and, if left unaddressed, could have a compounding effect on hyperalgesia for those in a critical window of neural repair. This experience may further escalate reliance of respiratory suppressing medications (ie, opioids) among those with TBI, possibly contributing to a multitude of risk factors associated with poor opioid therapy outcomes.⁵²

Strengths and limitations

This study benefitted from a well-defined sample in an at-risk population and focused on an understudied relationship. The primary sleep-related outcomes were obtained using a criterion standard assessment for sleep apnea diagnosis; level 1 PSG. The accepted standard definition of chronic pain (ie, >3 months) was used to better determine course of pain vs reliance exclusively on a numerical rating scale. Moreover, the current study examined reports of pain beyond the post-acute window of recovery. These findings should be considered in light of limitations and qualifications. The current study leverages data from 2 separate studies for which understanding the relationship between sleep apnea and chronic pain was not the primary aim. As such, the study analyses are relational in nature and are not intended to determine causality. Additionally, PSG data at the time of follow-up were not available for analysis and participants were not followed prospectively, limiting determinations on a bidirectional relationship between sleep apnea and chronic pain. Sample selection was constrained to participants co-enrolled in 2 studies, which may have introduced selection bias. Differences in pain and sleep apnea management strategies may be a confounding factor in these findings and lack of treatment data is a limitation of the current study, and closer follow-up to determine the contribution of treatment to this relationship is needed. While opioid use associated with pain management is also likely to contribute to increased risk of sleep apnea, additional injury information at the time of PSG, as well as data for determining patterns of use at the time of follow-up, were unavailable to aid further interpretation. Effective treatment for sleep apnea, for example, has been shown to improve pain intensity as well as pain-related functional outcomes.^5,53 It is unclear if treatment approaches contribute to a resolution of pain in the “past pain” cohort despite having largely comparable PSG indices. Interim follow-up data in the current study are limited but indicate that many of those diagnosed with sleep apnea during acute recovery declined treatment or were CPAP-nonadherent across all study groups (ie, no, past, and current pain). All study participants had moderate to severe TBI; hence, findings may not generalize to individuals with mild TBI. Because of the time points used, several factors may have influenced the results over the 1- to 2-year follow-up period, including adherence to treatment for sleep apnea (eg, CPAP), pain treatments accessed, medication dosages, or additional injuries incurred.

Future directions

Sleep apnea is a prevalent, modifiable condition following TBI. While sleep disturbance has been linked to worse pain outcomes following TBI, the relative contribution of sleep apnea among many complex clinical factors requires further study in this high-risk population.⁵⁴ Prospective studies with multiple follow-up time points assessing nature of injury, receipt of treatment, and quantification of sleep apnea are necessary to better understand the relationship between sleep apnea and chronic pain. Closer follow-up that extends into the window of chronic recovery could yield more precise information about the relationship between sleep apnea and chronic pain, including how this relationship varies by management strategies for both conditions. Of particular importance would be the inclusion of opioid use–specific information, including adequate characterization of the pharmacotherapy profiles of individuals following TBI to address the potential presence of opioid-induced hyperalgesia adequately.

Given the prevalence of chronic pain following TBI, the detection of sleep disturbances during a critical period of neural recovery may lead to the development and implementation of earlier intervention, resulting in improved pain-related outcomes. Yet, the extent to which sleep apnea may contribute to long-term pain-related outcomes after TBI remains understudied. Future studies can better examine this relationship through the inclusion of PSG at follow-up time points and objectively assessed adherence to CPAP interventions, as treatment of sleep apnea via CPAP has been shown to reduce pain sensitivity in those without a history of TBI.⁵⁵ Experimental designs with greater controls would be better equipped to pose a more causative interpretation, which further underscores the importance of recent work to better understand factors associated with CPAP adherence following TBI, such as symptom severity.⁵⁶ Additional attention is needed to understand facilitators and barriers to accessing sleep and pain medicine care, particularly if the relationship between sleep-disordered breathing, respiration, and the development of chronic pain is found to be clinically relevant. As rates of pain and sleep apnea may vary, future research should also explore this relationship across all levels of injury severity.

CONCLUSIONS

The current study fills a research gap by examining the relationship between objectively measured sleep-related respiratory indices and chronic pain following moderate to severe TBI. While important differences were seen in sleep-disordered breathing across pain cohorts following moderate-to-severe TBI, further research is needed to determine the relative contribution of sleep apnea to pain experience in this clinical population. In addition, prospective investigation is warranted to determine the value of sleep apnea intervention on pain-related outcomes postinjury.

DISCLOSURE STATEMENT

The lead author’s affiliated location and all authors listed have contributed to, reviewed, and approved this manuscript. Work for this study was performed at James A. Haley Veterans’ Hospital in Tampa, Florida. Research reported in this article was funded through a Patient-Centered Outcomes Research Institute (PCORI) award (CER-1511-33005) and through the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR)–funded TBI Model Systems (TBIMS) Collaborative Study on Pain after TBI (90DPTB0017; Dr. Nakase-Richardson). This research was sponsored by VHA Central Office Veterans Affairs (VA) TBI Model Systems Program of Research; a subcontract from General Dynamics Information Technology (W91YTZ13-C-0015; HT0014-19-C-0004) from the Defense Health Agency TBI Center of Excellence; and NIDILRR (90DPTB0017; Department of Rehabilitation Medicine, University of Washington School of Medicine, 90DPTB0008, Dr. Hoffman; Carolinas TBI Model System Follow-Up site, 90DP0084-01-00, Dr. Pinto). The TBI Model Systems is a funded collaboration between the Department of Veterans Affairs and the Department of Health and Human Services: NIDILRR. Dr. Walker’s time was also supported by funding through grant 90DP0033, NIDILRR TBI Model Systems at Virginia Commonwealth University. This study was funded by the VA Headache Centers of Excellence at James A. Haley Veterans Hospital, Tampa, Florida. The authors report no conflicts of interest. Disclaimer: The views expressed in this publication are those of the authors and do not necessarily represent the official policy or position of the Defense Health Agency, Department of Defense, or any other US government agency. The statements presented in this publication are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors, or Methodology Committee. This work was prepared under contract HT0014-21-C-0012 with DHA Contracting Office (CO-NCR) HT0014 and, therefore, is defined as US government work under Title 17 U.S.C.§101. Per Title 17 U.S.C.§105, copyright protection is not available for any work of the US government. For more information, please contact DHA.TBICOEinfo@mail.mil. Unclassified.

ABBREVIATIONS

AASM

American Academy of Sleep Medicine

AHI

apnea-hypopnea index

CMS

Centers for Medicaid and Medicare Services

CPAP

continuous positive airway pressure

GCS

Glasgow Coma Scale

PCORI

Patient-Centered Outcomes Research Institute

PSG

polysomnography

PSQI

Pittsburgh Sleep Quality Index

SD

standard deviation

TBIMS

TBI Model Systems