Abstract
Study Objectives:
Insomnia is a diagnosis with broad health and economic implications that has been increasingly recognized in military service members. This trend was concurrent with an increase in traumatic wartime injuries. Accordingly, we sought to determine longitudinal predictors of persistent insomnia in combat veterans who sustained traumatic injuries.
Methods:
Retrospective cohort study of service members deployed to conflict zones from 2002 to 2016, with longitudinal follow-up in the Veterans Affairs and Military Health Systems. Two cohorts were derived: (1) service members who sustained traumatic injuries and (2) an age-, sex-, and service component–matched cohort of uninjured service members who deployed to a combat zone. Insomnia was defined using International Classification of Diseases, Ninth Revision or International Classification of Diseases, 10th Revision-Clinical Modification codes.
Results:
The final population of 17,374 service members was followed from date of injury (or date of matched participant’s injury) for a median of 8.4 (interquartile range, 5.3–10.7) years. Service members with traumatic injury were at significantly greater risk of developing insomnia than uninjured service members (hazard ratio = 1.43; 95% confidence interval, 1.30–1.58) after adjustment. Traumatic brain injury was associated with insomnia compared with patients without traumatic brain injury in the multivariable model: mild/unclassified traumatic brain injury (hazard ratio = 2.07; 95% confidence interval, 1.82–2.35) and moderate/severe/ penetrating traumatic brain injury (hazard ratio = 2.43; 95% confidence interval, 2.06–2.86). Additionally, burn injury (hazard ratio = 1.95; 95% confidence interval, 1.47–2.59) and amputation (hazard ratio = 1.61; 95% confidence interval, 1.26–2.06) significantly increased the risk of a diagnosis.
Conclusions:
Traumatic injuries significantly predicted a diagnosis of insomnia after controlling for mental health disorders. Our findings strongly suggest the need for long-term surveillance of sleep disorders in trauma survivors.
Citation:
Haynes ZA, Collen JF, Poltavskiy EA, et al. Risk factors of persistent insomnia among survivors of traumatic injury: a retrospective cohort study. J Clin Sleep Med. 2021;17(9):1831–1840.
Keywords: insomnia, trauma, injury, veterans, TBI, burns, amputation, PTSD, depression, anxiety
BRIEF SUMMARY
Current Knowledge/Study Rationale: Insomnia and disordered sleep are strongly related to mental health disorders, posttraumatic stress disorder, and traumatic brain injury, although the longitudinal development of insomnia among veterans with combat-related traumatic injuries is not well defined. An understanding of time frames and which types of traumatic injuries are most strongly associated with insomnia could improve the long-term management of trauma survivors.
Study Impact: Our study examines the development of insomnia over time in survivors of combat-related traumatic injuries while controlling for comorbid mental health disorders and participant demographics. This analysis allows for the identification of discrete risk factors and suggestions for when insomnia screening should optimally begin.
INTRODUCTION
Insomnia symptoms affect up to 50% of individuals annually, eventually becoming a chronic disorder in 15%–20% of the US population.1–3 The impact of insomnia on quality of life can be more disabling than other medical conditions, including depression and chronic arthritic pain.4 Insomnia is also a risk factor for the development of depression, cardiovascular disease, stroke, and diabetes.5,6 Equally concerning are the economic effects of the disease, as recent data noted more than $63,000 in additional health care costs in an 11-month period for patients with insomnia, which was attributable primarily to increased numbers of emergency department and inpatient encounters.7
Risk for insomnia increases with older age, female sex, lower socioeconomic status, coexisting medical disorders, trait sleep reactivity (increased likelihood of sleep disruption in response to a stressful event), and comorbid psychiatric disease.1,8 Traumatic brain injures (TBI) have also been associated with insomnia; however, studies on this exposure have been limited by small samples, a lack of longitudinal multivariable analysis, and sampling biases.9–13 There is a paucity of data on the risk of insomnia among survivors of various types of traumatic injury, to include burns and amputations.
Insomnia is increasingly recognized within military populations. One recent large Veterans Affairs Healthcare System (VA) cohort found self-reported insomnia symptoms in 57.2% of veterans serving in the post-9/11 era.14 Additionally, insomnia diagnosis rates in veterans since 2001 have risen by more than 300% and exceeded rates observed in civilian populations.15–19 During this same time, more than 2.77 million US service members served in combat operations in Iraq and Afghanistan, with 53,119 service members experiencing life-threatening wounds.20 Traumatic injuries were often the result of high-energy explosive munitions.21 Resuscitative care in theater and definitive care at home stations encompassed damage control surgeries, intensive care unit admission, and complications, including life-threatening hemorrhage and hospital-acquired infections.22–24 Given the widespread use of improvised explosive devices, many consider burns, amputations, TBI, and posttraumatic stress disorder (PTSD) to be signature wounds of the war.25–27
The primary purpose of the current study was to evaluate the relationship between traumatic injury and the development of persistent insomnia. We hypothesized that combat-related traumatic injuries would be independent risk factors for persistent insomnia in a cohort of wounded service members and aimed to clarify the risk of insomnia associated with various types of traumatic injury. Accordingly, we present the first study examining the longitudinal risk factors for persistent insomnia among combat-deployed service members treated within the military health system and VA.
METHODS
This was a retrospective cohort study of US service members involved in combat operations in Iraq or Afghanistan. Protocols were reviewed and approved by the David Grant US Air Force Medical Center Institutional Review Board, the University of Utah Institutional Review Board, and the Research Review Committee of the VA Salt Lake City Health Care System.
Cohort development
Two cohorts were derived: (1) service members who deployed with a documented traumatic combat injury and (2) service members who deployed without injury. A third control cohort of nondeployed service members was not included, because these individuals have higher rates of baseline comorbidities that preclude them from deploying.28 Our injured cohort was derived from a random sample of 10,000 service members who were wounded during combat operations in Iraq or Afghanistan between February 1, 2002 and June 14, 2016. Their data were extracted from the US Department of Defense Trauma Registry (DoDTR), a database used to document all traumatic combat injuries severe enough to require admission to a theater hospital. A group of uninjured service members was matched 1:1 with the injured cohort for birth year (±1 year), branch of service (Army, Navy, Air Force, Marines, and Coast Guard), and sex using the Veterans Affairs/DoD Identity Repository. Participants in the uninjured cohort had neither a documented combat injury in DoDTR nor were they separated from military service for combat injuries. Although the uninjured cohort participants did not incur an injury severe enough to require admission to a theater hospital, they could still have a less severe injury while deployed or a severe injury once returning from overseas, and these injuries were included in our analysis.
Participants were excluded from analysis if they could not be matched, died within 90 days of index, had no documented encounters within the study period or after the index date, had a preexisting diagnosis of insomnia, or if they had missing data on a variable of interest. Index dates for participants in the injured cohort were the date of injury and the matched uninjured participants shared the same index date as their injured counterpart. Participants were followed until a diagnosis of insomnia occurred, death, they were lost to follow-up, or the study period ended.
Data collection and organization
Longitudinal data were obtained from several databases. Birth year, service branch, rank, and sex were obtained from DoDTR (for the injured cohort) and Veterans Affairs/DoD Identity Repository (for the uninjured cohort). Given the small sample of Coast Guard participants, they were pooled with the Navy for analysis. Rank was categorized into junior enlisted, senior enlisted, and officers as a surrogate for socioeconomic status. The Defense Manpower Data Center, a database used to consolidate military administrative personnel records, was used as the primary source of race/ethnicity data. The Military Health System Data MART, a database managed by the Defense Health Agency for collecting metrics on health care utilization, provided marital status and was a secondary source of race/ethnicity data. If race/ethnicity was missing from both Defense Manpower Data Center and Military Health System Data MART, it was derived from Veterans Affairs/DoD Identity Repository. The Joint VA-DoD Suicide Data Repository National Death Index provided mortality data. Injury characteristics, including injury type, mechanism, and Injury Severity Score were obtained from DoDTR. Injury Severity Score is an anatomically based, validated quantitative measure of injury severity that was used to demonstrate the median injury severity in the injured cohort, with scores ranging from 1 to 75 and higher scores representing more severe injury.29
Outcomes and covariates
Insomnia was defined using International Classification of Diseases, Ninth Revision or International Classification of Diseases, 10th Revision-Clinical Modification (ICD-9/ICD-10-CM) diagnosis codes. Diagnosis codes for insomnia and covariates of interest were obtained from Military Health System Data MART and the Veterans Informatics and Computing Infrastructure. Diagnosis codes documented before a participant’s index date were included in our analysis, with the exception of prior insomnia diagnoses, which led to exclusion of the participant from the analysis. Participants were considered to have the outcome of persistent insomnia if they had 2 or more outpatient encounters within 2 years or 1 inpatient discharge documenting an ICD-9/ICD-10-CM diagnosis corresponding to insomnia (307.42, 327.09, 780.52, F51.01, G47.XX), which is based on previously published case definitions of chronic conditions.30,31 Covariates included age at index date, obesity, PTSD, anxiety disorders (generalized anxiety, panic, phobic, and obsessive-compulsive disorders), depression (major depression, dysthymia, and other persistent mood disorders), alcohol dependence, and opioid dependence, in addition to the injury subtypes of TBI, burn, and amputation. Armed Forces Health Surveillance Branch ICD-9/ICD-10-CM case definitions were used to define obesity, PTSD, anxiety disorders, depression, alcohol dependence, and opioid dependence.32 TBI, burns, and amputations were defined by ICD-9/ICD-10-CM coding reflecting the injury subtypes extracted from DoDTR, Military Health System Data MART, and Veterans Informatics and Computing Infrastructure.
Analysis plan
Descriptive statistics were used to compare the baseline characteristics of the study population and the 2 cohorts. Categorical variables (demographics, insomnia, and injury incidence) are presented by percentages and compared through the use of a χ2 test. Continuous variables (age at index, health care utilization, Injury Severity Score) are presented as medians with interquartile ranges and compared through use of Wilcoxon rank sum tests.
Our primary analyses used Fine and Gray competing risk models to account for the competing risk of death. Stratified models were used to examine our cohorts. These included univariate models for each covariate and 3 nested multivariable models: (1) demographics (race/ethnicity, marital status, rank, active/reserve, age), (2) health behaviors and mental health (obesity, PTSD, depression, anxiety disorders, opioid dependence, alcohol dependence) considered as time-dependent variables, and (3) injury factors (injury cohort, TBI, burn, amputation). Data are presented graphically by using cumulative incidence functions.
The 3 nested models were used in an effort to further describe relationships between covariates. Changes in the hazard ratios observed for each covariate with the introduction of another nested model can be used to infer information about the relationship between those covariates and the development of insomnia. Demographic variables were included in the models as they are factors that may influence a participant’s long-term health outcomes, access to medical care, or interaction with military or combat stressors. Health behaviors and mental health covariates were included because they are factors that have been previously associated with sleep disorders. Finally, injury factors were included because they are the variables of interest in this study.
RESULTS
Participants excluded from the analysis can be seen in Figure 1. The final population contained 8,687 matched pairs (17,374 study participants).
Figure 1. Exclusion process to derive final matched cohorts.
The characteristics of the study population are presented in Table 1. The injured cohort participants were more likely to be senior enlisted (34.8% vs 26.6%, P < .001) and married (49.0% vs 46.1%, P < .001) than their uninjured counterparts. Most participants were non-Hispanic White. Median age at index date was 24 (interquartile range [IQR], 22–29) years. The median follow-up time after index date was 8.4 (IQR, 5.3–10.7) years, which differed significantly between the injured (median, 8.8; IQR, 5.7–10.9) and uninjured (7.8; IQR, 4.9–10.4; P < .001) groups. For the injured cohort, median Injury Severity Score was 6 (IQR, 2–13), which is a score considered to be consistent with mild injury. Injured patients had significantly higher rates of insomnia (36.6% vs 14.7%, P < .001). Although the participants in the uninjured cohort could not have an injury severe enough to be admitted to a theater hospital, injuries were still observed through longitudinal follow-up. These injuries mainly comprised mild/unclassified TBI (17.9%) and moderate/severe/penetrating TBI (4.0%), although there were several burns (0.1%) and amputations (0.1%).
Table 1.
Demographics and initial cohort characteristics.
| Total (n = 17,374) | Injured (n = 8,687) | Uninjured (n = 8,687) | P | |
|---|---|---|---|---|
| Sex, n (%) | 1.000 | |||
| Male | 17,032 (98.0) | 8,516 (98.0) | 8,516 (98.0) | |
| Female | 342 (2.0) | 171 (2.0) | 171 (2.0) | |
| Age at index, median (IQR) | 24 (22–29) | 24 (22–29) | 24 (22–29) | 1.000 |
| Race/ethnicity, n (%) | < .001 | |||
| NH White | 12,815 (73.8) | 6,586 (75.8) | 6,229 (71.7) | |
| NH Black | 1,863 (10.7) | 736 (8.5) | 1,126 (13.0) | |
| Hispanic | 1,830 (10.5) | 931 (10.7) | 899 (10.4) | |
| Asian/PI | 577 (3.3) | 289 (3.3) | 288 (3.3) | |
| Other | 290 (1.7) | 145 (1.7) | 145 (1.7) | |
| Marital status, n (%) | < .001 | |||
| Single | 9,120 (52.5) | 4,435 (51.1) | 4,685 (53.9) | |
| Married | 8,254 (47.5) | 4,252 (49.0) | 4,002 (46.1) | |
| Service branch, n (%) | 1.000 | |||
| Army | 12,752 (73.4) | 6,376 (73.4) | 6,376 (73.4) | |
| Air Force | 308 (1.8) | 154 (1.8) | 154 (1.8) | |
| Marine Corps | 3,826 (22.0) | 1,913 (22.0) | 1,913 (22.0) | |
| Navy | 488 (2.8) | 244 (2.8) | 244 (2.8) | |
| Military component, n (%) | < .001 | |||
| Active duty | 12,924 (74.4) | 7,522 (86.6) | 5,402 (62.2) | |
| Reserve/guard | 4,450 (25.6) | 1,165 (13.4) | 3,285 (37.8) | |
| Rank, n (%) | < .001 | |||
| Junior enlisted (E1–E5) | 10,795 (62.1) | 5,093 (58.6) | 5,702 (65.6) | |
| Senior enlisted (E5–E9) | 5,329 (30.7) | 3,020 (34.8) | 2,309 (26.6) | |
| Officer | 1,250 (7.2) | 574 (6.61) | 676 (7.8) | |
| Injury subtypes, n (%) | ||||
| Burn | 930 (5.4) | 921 (10.6) | 9 (0.1) | < .001 |
| Amputation | 1,126 (6.5) | 1,114 (12.8) | 12 (0.1) | < .001 |
| Traumatic brain injury, n (%) | < .001 | |||
| Mild/unclassified | 5,810 (33.4) | 4,251 (48.9) | 1,559 (17.9) | |
| Mod/sev/penetrating | 2,587 (14.9) | 2,242 (25.8) | 345 (4.0) | |
| Insomnia diagnosis in study period, n (%) | < .001 | |||
| Yes | 4,454 (25.6) | 3,178 (36.6) | 1,276 (14.7) | |
| No | 12,920 (74.4) | 5,509 (63.4) | 7,411 (85.3) | |
| Injury Severity Score, median (IQR) | — | 6 (2–13) | — | – |
| Health care utilization | ||||
| Years of follow-up, median (IQR) | 8.4 (5.3–10.7) | 8.8 (5.7–10.9) | 7.84 (4.9–10.4) | < .001 |
| Annual encounters, median (IQR) | 10.0 (4.4–22.4) | 17.7 (8.8–35.6) | 5.7 (2.7–11.7) | < .001 |
P value of 1.000 denotes variable used for cohort matching. Dashes represent data which was unavailable for analysis, as subjects in the uninjured cohort were not registered in DoDTR. IQR = interquartile range, Mod = moderate, NH = non-Hispanic, PI = Pacific Islander, sev = severe.
The results are presented graphically in Figure 2 for traumatic injury (A), TBI (B), amputation (C), and burn (D). Compared with the uninjured service members, patients with traumatic injury were at significantly greater risk of insomnia (hazard ratio [HR] = 3.33; 95% confidence interval [CI], 3.12–3.55, P < .001) in the univariate analysis (Table 2). Compared with patients without TBI, mild/unclassified TBI (HR = 4.44; 95% CI, 4.00–4.92; P < .001) and moderate/severe/penetrating TBI (HR = 6.51; 95% CI, 5.70–7.44; P < .001) were also associated with insomnia. Burns (HR = 6.09; 95% CI, 4.76–7.78; P < .001) and amputation (HR = 5.98; 95% CI, 4.85–7.38; P < .001) also increased the risk for insomnia in the unadjusted model.
Figure 2. Graphical representation of the risk of insomnia among injury covariates.
CI = confidence interval, concus = concussion, HR = hazard ratio, mod = moderate, penetr = penetrating, sev = severe, TBI = traumatic brain injury, unclas = unclassified.
Table 2.
Comparison of predicted risk of developing insomnia among covariates.
| Univariate Models | |||
|---|---|---|---|
| HR | 95% CI | P | |
| Index age | 0.94 | 0.89–0.99 | .019 |
| Race/ethnicity | |||
| NH White | 1.00 | — | — |
| Hispanic | 1.14 | 1.02–1.27 | .022 |
| NH Black | 0.78 | 0.70–0.87 | < .001 |
| Asian/PI | 0.89 | 0.73–1.08 | .247 |
| Other | 1.03 | 0.80–1.33 | .818 |
| Rank | |||
| Junior enlisted | 1.00 | — | — |
| Senior enlisted | 1.35 | 1.23–1.48 | < .001 |
| Officer | 0.56 | 0.49–0.70 | < .001 |
| Marital status | |||
| Single | 1.00 | — | — |
| Married | 1.37 | 1.28–1.50 | < .001 |
| Mental health/health behaviors | |||
| PTSD | 4.25 | 3.83–4.70 | < .001 |
| Depression | 3.41 | 3.08–3.77 | < .001 |
| Anxiety | 4.51 | 4.03–5.05 | < .001 |
| Alcohol dependence | 2.13 | 1.92–2.37 | < .001 |
| Opioid dependence | 4.34 | 3.20–5.90 | < .001 |
| Obesity | 1.49 | 1.36–1.63 | < .001 |
| Cohort | |||
| Uninjured | 1.00 | — | — |
| Injured | 3.33 | 3.12–3.55 | < .001 |
| Injury subtypes | |||
| Traumatic brain injury | |||
| No TBI | 1.00 | — | — |
| Mild/unclassified | 4.44 | 4.00–4.92 | < .001 |
| Mod/sev/penetrating | 6.51 | 5.70–7.44 | < .001 |
| Burn | 6.09 | 4.76–7.78 | < .001 |
| Amputation | 5.98 | 4.85–7.38 | < .001 |
Hazard ratio of 1.00 denotes a baseline reference variable. Dashes represent values which could not be calculated for reference variables. CI = confidence interval, HR = hazard ratio, Mod = moderate, NH = non-Hispanic, PI = Pacific Islander, PTSD = posttraumatic stress disorder, sev = severe, TBI = traumatic brain injury.
When adjusting for all covariates in our multivariable model (Table 3), we found attenuation of risk across all mental health disorders (model 2) and injury (model 3) covariates compared with our univariate model results. After incorporating all covariates, traumatic injury (HR = 1.43; 95% CI, 1.30–1.58; P < .001) remained statistically significant, with TBI ranking as the strongest predictor: mild/unclassified TBI (HR = 2.07; 95% CI, 1.82–2.35; P < .001) and moderate/severe TBI (HR = 2.43; 95% CI, 2.06–2.86; P < .001) compared with participants without TBI. Amputation (HR = 1.61; 95% CI, 1.26–2.06; P < .001) and burns (HR = 1.95; 95% CI, 1.47–2.59; P < .001) were additionally found to be predictors of insomnia. Other significant covariates predicting an insomnia diagnosis can be seen in Table 3.
Table 3.
Comparison of predicted multivariable risk of developing insomnia among covariates.
| Model 1 | Model 2 | Model 3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | P | HR | 95% CI | P | HR | 95% CI | P | |
| Index age | 0.99 | 0.93–1.06 | .782 | 0.96 | 0.88–1.03 | .236 | 1.03 | 0.94–1.12 | .531 |
| Race/ethnicity | |||||||||
| NH White | 1.00 | — | — | 1.00 | — | — | 1.00 | — | — |
| Hispanic | 1.07 | 0.95–1.21 | .280 | 1.08 | 0.93–1.25 | .298 | 1.16 | 0.98–1.37 | .085 |
| NH Black | 0.70 | 0.62–0.79 | < .001 | 0.75 | 0.64–0.86 | < .001 | 0.98 | 0.83–1.17 | .834 |
| Asian/PI | 0.78 | 0.62–0.98 | .031 | 0.90 | 0.68–1.18 | .436 | 0.89 | 0.97–1.20 | .453 |
| Other | 0.95 | 0.72–1.26 | .726 | 1.12 | 0.82–1.54 | .476 | 1.10 | 0.77–1.59 | .593 |
| Rank | |||||||||
| Junior enlisted | 1.00 | — | — | 1.00 | — | — | 1.00 | — | — |
| Senior enlisted | 1.14 | 1.03–1.27 | .015 | 1.23 | 1.07–1.40 | .003 | 1.01 | 0.87–1.17 | .924 |
| Officer | 0.48 | 0.40–0.59 | < .001 | 0.68 | 0.53–0.86 | .001 | 0.61 | 0.46–0.81 | < .001 |
| Marital status | |||||||||
| Single | 1.00 | — | — | 1.00 | — | — | 1.00 | — | — |
| Married | 1.27 | 1.13–1.35 | < .001 | 1.24 | 1.12–1.38 | <.001 | 1.25 | 1.11–1.40 | < .001 |
| Mental health/health behaviors | |||||||||
| PTSD | 2.42 | 2.14–2.72 | < .001 | 1.41 | 1.22–1.62 | < .001 | |||
| Depression | 1.56 | 1.37–1.79 | < .001 | 1.65 | 1.41–1.92 | < .001 | |||
| Anxiety | 2.57 | 2.26–2.92 | < .001 | 2.19 | 1.90–2.53 | < .001 | |||
| Alcohol dependence | 1.15 | 1.00–1.34 | .053 | 1.26 | 1.07–1.49 | .005 | |||
| Opioid dependence | 1.83 | 1.25–2.66 | .002 | 1.54 | 1.02–2.31 | .039 | |||
| Obesity | 1.21 | 1.06–1.38 | .006 | 1.18 | 1.01–1.37 | .043 | |||
| Cohort | |||||||||
| Uninjured | 1.00 | — | — | ||||||
| Injured | 1.43 | 1.30–1.58 | < .001 | ||||||
| Injury subtypes | |||||||||
| Traumatic brain injury | |||||||||
| Mild/unclassified | 2.07 | 1.82–2.35 | < .001 | ||||||
| Mod/sev/penetrating | 2.43 | 2.06–2.86 | < .001 | ||||||
| Burn | 1.95 | 1.47–2.59 | < .001 | ||||||
| Amputation | 1.61 | 1.26–2.06 | < .001 | ||||||
Hazard Ratio of 1.00 denotes value used as a baseline reference variable for the multivariable analysis. Dashes represent values which could not be calculated for reference variables. CI = confidence interval, HR = hazard ratio, Mod = moderate, NH = non-Hispanic, PI = Pacific Islander, PTSD = posttraumatic stress disorder, sev = severe, TBI = traumatic brain injury.
After accounting for the interaction with time, traumatic injury was associated with a HR of 3.77 (95% CI, 2.48–5.74) for the first 2 years after injury. The risk was reduced at time points after year 2 (HR = 0.71; 95% CI, 0.62–0.80). Similar patterns were seen with mild/unclassified TBI (HR = 2.99; 95% CI, 1.81–4.90) and moderate/severe/penetrating TBI (HR = 12.50; 95% CI, 5.21–29.98) within the first 2 years. After year 2, the risk in the mild/unclassified TBI participants (HR = 1.98; 95% CI, 1.69–2.31) was noted to continue to be elevated to a greater degree than the participants with moderate/severe/penetrating TBI (HR = 1.50; 95% CI, 1.22–1.84). This pattern was also observed in participants with amputations before year 2 (HR = 4.96; 95% CI, 1.74–14.19) and after year 2 (HR = 0.84; 95% CI, 0.62–1.14). These clearly demonstrate the nonproportionality of the impact of these variables on the development of insomnia through time, with risk primarily concentrated in the first 2 years after the index date.
DISCUSSION
We present the first large, longitudinal study focused on the exposure of traumatic injury and associated risk factors on the subsequent development of persistent insomnia in US military veterans. Our results suggest that traumatic injury increases a patient’s lifetime risk for persistent insomnia, as we saw a higher disease prevalence within the injured cohort than is reported in the general population.1,2 We also found that burn injuries and amputation are risk factors for insomnia, independently increasing estimated risk by 95% and 61%, respectively. Furthermore, all severity classes of TBI more than doubled the risk of developing insomnia. The greatest risk imposed appears to be present in the first 2 years from the time of injury across our 3 injury groupings (TBI, burn, amputation), displaying the nonproportionality of trauma’s effect on insomnia development. The data also show a slight variation in this pattern between the participants with mild/unclassified and moderate/severe/penetrating TBI, where most risk continues to be imparted within the first 2 years, but the overall risk for disease increases to a greater degree over the study period in mild/unclassified TBI.
TBI may lead to sleep disturbances through numerous mechanisms, including disrupting sleep-related neuronal circuitry, dysregulation of orexin and melatonin signaling, and circadian misalignment.33,34 Sleep-wake disturbances have been reported in up to 85% of patients with TBI in prior studies, and this population is known to have higher rates of insomnia diagnoses.10,12,28,35,36 Unfortunately, prior studies are limited by selection bias and lack of longitudinal analysis controlling for confounders, because most patients were referred to a sleep disorders center for evaluation of preexisting sleep complaints. Thus, our data fill a gap in the literature with a large sample of patients who have robust follow-up from outside of a sleep disorders center. Among the patients with TBI, our data suggest a continually increasing risk of insomnia over time in the mild/unclassified TBI population relative to the moderate/severe/penetrating TBI subgroup. This may be caused, in part, by an increased propensity for patients with mild TBI (mTBI) to seek a diagnosis from a sleep clinic after their injury, as they are often the most widely represented TBI population seen in sleep disorders centers.11,12,28,37 Additionally, those with higher severity TBI may be limited in their ability to conform to traditional methods of diagnosis and treatment by increased morbidity, which may alter diagnostic outcomes and the trajectory of their disease course. Regardless, this finding has implications for civilian populations, because rates of mTBI exceed 1.7 million cases annually in the United States and are suspected to be highly under-reported.38 Aside from sleep disorders, TBI can lead to cognitive impairment and mood disorders, which both can be worsened by impaired sleep.39 This wide effect signals a need for careful screening, assessment, and management of sleep complaints in patients with mTBI, even years after the initial injury.
Prior studies involving insomnia in patients with burns and amputees are limited in number and quality and were generally focused on the effects of pain on sleep quality after injury. Studies within patients with burns failed to characterize the prevalence or risk of persistent insomnia and associated risk factors, although they noted complaints of generally poor quality sleep on patient-completed questionnaires.40,41 Similarly, amputees have been observed to have improvement in insomnia-related symptoms, measured by Insomnia Severity Index scores, with reductions in the pain experienced from phantom limb syndrome, although they did not establish the injury type as a risk factor.42 Among survivors of burns and amputations, chronic pain can be a challenging complication that widely affects quality of life.43,44 Pain has previously been shown to predispose patients to insomnia in numerous studies, and this factor may be leading to the increased risk seen in our population as well.45,46 Although we do not have information on pain levels in our participants, our dataset defines burns and amputations as discrete risk factors for insomnia, therefore suggesting that pain management may be critical to preventing the development of sleep disturbances in these patients. Further future research into the effects of pain levels in these populations may better characterize the relationship between burns and amputations and the development of persistent insomnia.
Sleep and mental health disorders are common comorbid conditions, particularly in veterans. PTSD, anxiety, and depression were associated with insomnia in our analysis, underscoring the importance of comorbid psychiatric diseases and their effect on insomnia. However, injury variables remained significant in the fully adjusted models, suggesting that traumatic injury increases the risk of insomnia development independent of these previously recognized mental health risk factors. There is an established interplay between sleep and mood disturbances, with each issue having a propensity to worsen the other.47,48 This relationship can have dramatic downstream effects, as both mental health and sleep disorders substantially increase the risk of suicidality, and insufficient and poor quality sleep have been linked to completed suicide attempts.49–51 This highlights the importance of heightened awareness toward sleep complaints in patients with mental health disorders who have experienced traumatic injuries.
Overall, our findings provide a unique opportunity to inform both military and civilian care given the incidence of these exposures in the population at large and the relative paucity of adequately powered research related to the long-term complications of trauma. Recent measures on the prevalence of amputees in this country were estimated at more than 1.6 million patients, which is expected to more than double by 2050, and emergency department visits for TBI have exceeded 2.5 million visits annually.52,53 Burns have been estimated to complicate as many as 5.8% of civilian trauma cases, and PTSD has been observed to develop in 8.3% of trauma patients.54,55 Such large incidences in our community and their broad implications underscore the importance of our findings and the need for further sleep disorder surveillance in affected populations. Presently, and as these populations age, insomnia can significantly impact morbidity, such as incident hypertension, and overall health care costs.7,56 By directing screening toward the populations outlined here, it may be possible to intervene on impaired sleep before it creates a significant patient and societal burden. Unfortunately, the worsening shortage of sleep medicine specialists relative to patient volume in the United States will only complicate these mitigation efforts.57
Limitations
Our data have some limitations. First, the insomnia and covariate diagnoses were based on ICD-9/ICD-10-CM codes from electronic medical records, which can be prone to some degree of diagnostic error. The use of ICD-9/ICD-10-CM coding also raises concern for underreporting of insomnia prevalence, as 4 of 5 of patients treated with insomnia pharmacotherapy do not have a documented ICD-9/ICD-10-CM insomnia diagnosis.58 Second, our population is nearly all young males, which may limit generalizability. Third, service members may have greater access to medical care than the general population, which may be particularly true for wounded service members who often undergo prolonged and detailed multispecialty evaluations. This may have facilitated an increased rate of diagnoses for insomnia and its associated risk factors, as we saw significantly higher median annual encounters in the injured cohort. Additionally, our dataset does not have information on prescription drug utilization, which could have been used as a surrogate for pain levels in patients with chronic pain and further described the relationship between pain and insomnia. Finally, our analysis did not control for the presence of obstructive sleep apnea, which may have led to a falsely elevated insomnia prevalence. Prior research has demonstrated that obstructive sleep apnea and insomnia are often comorbid and can be challenging to differentiate given their overlapping symptoms.59
CONCLUSIONS
Traumatic injuries, including TBI, burns, and amputations, and mental health disorders are risk factors for insomnia. Survivors of traumatic injury are diagnosed with persistent insomnia at a higher rate than the general population, and this could potentially be caused by the injury directly altering neuronal sleep circuitry, residual pain, or increased interaction with medical providers. Within traumatic injuries, most of the heightened risk is imparted during the first 2 years after injury, although patients with mTBI see a progressively increasing risk of insomnia for up to 12 years after injury. These findings have broad implications for civilian and military providers, and sleep disorder surveillance within the first 2 years after traumatic injury appears to be warranted. Particularly for those with mTBI, this surveillance should extend beyond the first 2 years. Given the known effects of insomnia on behavioral health, comorbid disease, hospital admission, and health expenditures, proper recognition and treatment of injured veterans and civilians could have a significant impact on overall health outcomes.
ACKNOWLEDGMENTS
The authors thank the Department of Defense Trauma Registry, the Defense Health Agency, the Defense Suicide Prevention Office, the VA Informatics and Computing Infrastructure, the VA/DoD Identity Repository, and the US Department of Veterans Affairs for providing data for this study. Kevin Chung introduced the authors’ respective research groups, assisted in study design, and provided administrative research support.
ABBREVIATIONS
- CI,
confidence interval
- DoDTR,
Department of Defense Trauma Registry
- HR,
hazard ratio
- ICD,
International Classification of Diseases
- IQR,
interquartile range
- mTBI,
mild traumatic brain injury
- PTSD,
posttraumatic stress disorder
- TBI,
traumatic brain injury
- VA,
Veterans Affairs Healthcare System
DISCLOSURE STATEMENT
All authors have seen and approved the final version of this manuscript. This study was funded by the US Air Force Headquarters, Office of the Surgeon General and supported by VA Center of Innovation Award I50HX001240 from the Office of Research and Development of the US Department of Veterans Affairs. EMW’s institution received prior research support from the American Academy of Sleep Medicine Foundation, the US Department of Defense, Merck, and ResMed. EMW has served as a scientific consultant for DayZz, Eisai, Merck, and Purdue and is an equity shareholder in WellTap. The authors report no conflicts of interest. The views expressed in this manuscript are those of the author and do not reflect the official policy of the Uniformed Services University, the US Department of Defense, the US Department of the Army/Navy/Air Force, the US Department of Veterans Affairs, or the US government. This work was prepared by a military or civilian employee of the US government as part of the individual’s official duties and therefore is in the public domain and does not possess copyright protection.
REFERENCES
- 1. Buysse DJ . Insomnia . JAMA. 2013. ; 309 ( 7 ): 706 – 716 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Morin CM , Benca R . Chronic insomnia . Lancet. 2012. ; 379 ( 9821 ): 1129 – 1141 . [DOI] [PubMed] [Google Scholar]
- 3. Levenson JC , Kay DB , Buysse DJ . The pathophysiology of insomnia . Chest. 2015. ; 147 ( 4 ): 1179 – 1192 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Olfson M , Wall M , Liu SM , Morin CM , Blanco C . Insomnia and impaired quality of life in the United States . J Clin Psychiatry. 2018. ; 79 ( 5 ): 17m12020 . [DOI] [PubMed] [Google Scholar]
- 5. Winkelman JW . Insomnia disorder . N Engl J Med. 2015. ; 373 ( 15 ): 1437 – 1444 . [DOI] [PubMed] [Google Scholar]
- 6. Wu MP , Lin HJ , Weng SF , Ho CH , Wang JJ , Hsu YW . Insomnia subtypes and the subsequent risks of stroke: report from a nationally representative cohort . Stroke. 2014. ; 45 ( 5 ): 1349 – 1354 . [DOI] [PubMed] [Google Scholar]
- 7. Wickwire EM , Tom SE , Scharf SM , Vadlamani A , Bulatao IG , Albrecht JS . Untreated insomnia increases all-cause health care utilization and costs among Medicare beneficiaries . Sleep. 2019. ; 42 ( 4 ): zsz007 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Drake CL , Pillai V , Roth T . Stress and sleep reactivity: a prospective investigation of the stress-diathesis model of insomnia . Sleep. 2014. ; 37 ( 8 ): 1295 – 1304 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Castriotta RJ , Lai JM . Sleep disorders associated with traumatic brain injury . Arch Phys Med Rehabil. 2001. ; 82 ( 10 ): 1403 – 1406 . [DOI] [PubMed] [Google Scholar]
- 10. Castriotta RJ , Atanasov S , Wilde MC , Masel BE , Lai JM , Kuna ST . Treatment of sleep disorders after traumatic brain injury . J Clin Sleep Med. 2009. ; 5 ( 2 ): 137 – 144 . [PMC free article] [PubMed] [Google Scholar]
- 11. Castriotta RJ , Murthy JN . Sleep disorders in patients with traumatic brain injury: a review . CNS Drugs. 2011. ;25(3):175–185. [DOI] [PubMed] [Google Scholar]
- 12. Castriotta RJ , Wilde MC , Lai JM , Atanasov S , Masel BE , Kuna ST . Prevalence and consequences of sleep disorders in traumatic brain injury . J Clin Sleep Med. 2007. ; 3 ( 4 ): 349 – 356 . [PMC free article] [PubMed] [Google Scholar]
- 13. Collen J , Orr N , Lettieri CJ , Carter K , Holley AB . Sleep disturbances among soldiers with combat-related traumatic brain injury . Chest. 2012. ; 142 ( 3 ): 622 – 630 . [DOI] [PubMed] [Google Scholar]
- 14. Colvonen PJ , Almklov E , Tripp JC , Ulmer CS , Pittman JOE , Afari N . Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare . Sleep. 2020. ; 43 ( 12 ): zsaa119 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Insomnia, Active Component, U.S. Armed Forces, January 2000-December 2009. MSMR. 2010;17(5):12–15 .
- 16. Caldwell JA , Knapik JJ , Lieberman HR . Trends and factors associated with insomnia and sleep apnea in all United States military service members from 2005 to 2014 . J Sleep Res. 2017. ; 26 ( 5 ): 665 – 670 . [DOI] [PubMed] [Google Scholar]
- 17. Caldwell JA , Knapik JJ , Shing TL , Kardouni JR , Lieberman HR . The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation . Sleep. 2019. ; 42 ( 8 ): zsz112 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mysliwiec V , Gill J , Lee H , et al. Sleep disorders in US military personnel: a high rate of comorbid insomnia and obstructive sleep apnea . Chest. 2013. ; 144 ( 2 ): 549 – 557 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Mysliwiec V , McGraw L , Pierce R , Smith P , Trapp B , Roth BJ . Sleep disorders and associated medical comorbidities in active duty military personnel . Sleep. 2013. ; 36 ( 2 ): 167 – 174 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. US Department of Defense. Immediate Release Casualty Status. https://www.defense.gov/casualty.pdf . Accessed January 20, 2020 .
- 21. Undurraga Perl VJ , Johnson C , Kelley S , Woods P , Martin K , Zonies D . Missed injuries in combat casualties: lessons from Iraq and Afghanistan . Injury. 2019. ; 50 ( 5 ): 1138 – 1142 . [DOI] [PubMed] [Google Scholar]
- 22. Pidcoke HF , Aden JK , Mora AG , et al . Ten-year analysis of transfusion in Operation Iraqi Freedom and Operation Enduring Freedom: increased plasma and platelet use correlates with improved survival . J Trauma Acute Care Surg. 2012. ; 73 ( 6 , Suppl 5) : S445 – S452 . [DOI] [PubMed] [Google Scholar]
- 23. Blackbourne LH . Combat damage control surgery . Crit Care Med. 2008. ; 36( 7 , Suppl) : S304 – S310 . [DOI] [PubMed] [Google Scholar]
- 24. Weintrob AC , Murray CK , Xu J , et al . Early infections complicating the care of combat casualties from Iraq and Afghanistan . Surg Infect (Larchmt). 2018. ; 19 ( 3 ): 286 – 297 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Melcer T , Walker J , Sechriest VF II , et al . A retrospective comparison of five-year health outcomes following upper limb amputation and serious upper limb injury in the Iraq and Afghanistan conflicts . PM R. 2019. ; 11 ( 6 ): 577 – 589 . [DOI] [PubMed] [Google Scholar]
- 26. Williams SG , Collen J , Orr N , Holley AB , Lettieri CJ . Sleep disorders in combat-related PTSD . Sleep Breath. 2015. ; 19 ( 1 ): 175 – 182 . [DOI] [PubMed] [Google Scholar]
- 27. Collen J , Orr N , Lettieri CJ , Carter K , Holley AB . Sleep disturbances among soldiers with combat-related traumatic brain injury . Chest. 2012. ; 142 ( 3 ): 622 – 630 . [DOI] [PubMed] [Google Scholar]
- 28. Granado NS , Smith TC , Swanson GM , et al .; Millennium Cohort Study Team . Newly reported hypertension after military combat deployment in a large population-based study . Hypertension. 2009. ; 54 ( 5 ): 966 – 973 . [DOI] [PubMed] [Google Scholar]
- 29. Baker SP , O’Neill B , Haddon W Jr , Long WB . The injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care . J Trauma. 1974. ; 14 ( 3 ): 187 – 196 . [PubMed] [Google Scholar]
- 30. Yoon J , Chow A . Comparing chronic condition rates using ICD-9 and ICD-10 in VA patients FY2014-2016 . BMC Health Serv Res. 2017. ; 17 ( 1 ): 572 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Robitaille C , Dai S , Waters C , et al . Diagnosed hypertension in Canada: incidence, prevalence and associated mortality . CMAJ. 2012. ; 184 ( 1 ): E49 – E56 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Health.mil. Surveillance case definitions . https://health.mil/Military-Health-Topics/Combat-Support/Armed-Forces-Health-Surveillance-Branch/Epidemiology-and-Analysis/Surveillance-Case-Definitions . Accessed April 28, 2020.
- 33. Shekleton JA , Parcell DL , Redman JR , Phipps-Nelson J , Ponsford JL , Rajaratnam SM . Sleep disturbance and melatonin levels following traumatic brain injury . Neurology. 2010. ; 74 ( 21 ): 1732 – 1738 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Baumann CR , Bassetti CL , Valko PO , et al . Loss of hypocretin (orexin) neurons with traumatic brain injury . Ann Neurol. 2009. ; 66 ( 4 ): 555 – 559 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Kempf J , Werth E , Kaiser PR , Bassetti CL , Baumann CR . Sleep-wake disturbances 3 years after traumatic brain injury . J Neurol Neurosurg Psychiatry. 2010. ; 81 ( 12 ): 1402 – 1405 . [DOI] [PubMed] [Google Scholar]
- 36. Verma A , Anand V , Verma NP . Sleep disorders in chronic traumatic brain injury . J Clin Sleep Med. 2007. ; 3 ( 4 ): 357 – 362 . [PMC free article] [PubMed] [Google Scholar]
- 37. Cassidy JD , Carroll LJ , Peloso PM , et al .; WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury . Incidence, risk factors and prevention of mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury . J Rehabil Med. 2004. ; 36 ( 43 , Suppl) : 28 – 60 . [DOI] [PubMed] [Google Scholar]
- 38. Pfister T , Pfister K , Hagel B , Ghali WA , Ronksley PE . The incidence of concussion in youth sports: a systematic review and meta-analysis . Br J Sports Med. 2016. ; 50 ( 5 ): 292 – 297 . [DOI] [PubMed] [Google Scholar]
- 39. Timmer ML , Jacobs B , Schonherr MC , Spikman JM , van der Naalt J . The spectrum of long-term behavioral disturbances and provided care after traumatic brain injury . Front Neurol. 2020. ; 11 : 246 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Longley L , Simons T , Glanzer L , et al . Evaluating sleep in a surgical trauma burn intensive care unit: an elusive dilemma . Dimens Crit Care Nurs. 2018. ; 37 ( 2 ): 97 – 101 . [DOI] [PubMed] [Google Scholar]
- 41. Smith MT , Klick B , Kozachik S , et al . Sleep onset insomnia symptoms during hospitalization for major burn injury predict chronic pain . Pain. 2008. ; 138 ( 3 ): 497 – 506 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Trevelyan EG , Turner WA , Summerfield-Mann L , Robinson N . Acupuncture for the treatment of phantom limb syndrome in lower limb amputees: a randomised controlled feasibility study . Trials. 2016. ; 17 ( 1 ): 519 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Wang Y , Beekman J , Hew J , et al . Burn injury: challenges and advances in burn wound healing, infection, pain and scarring . Adv Drug Deliv Rev. 2018. ; 123 : 3 – 17 . [DOI] [PubMed] [Google Scholar]
- 44. Modest JM , Raducha JE , Testa EJ , Eberson CP . Management of post-amputation pain . R I Med J. 2020. ; 103 ( 4 ): 19 – 22 . [PubMed] [Google Scholar]
- 45. Morphy H , Dunn KM , Lewis M , Boardman HF , Croft PR . Epidemiology of insomnia: a longitudinal study in a UK population . Sleep. 2007. ; 30 ( 3 ): 274 – 280 . [PubMed] [Google Scholar]
- 46. Tang NK , McBeth J , Jordan KP , Blagojevic-Bucknall M , Croft P , Wilkie R . Impact of musculoskeletal pain on insomnia onset: a prospective cohort study . Rheumatology (Oxford). 2015. ; 54 ( 2 ): 248 – 256 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Borders A , Rothman DJ , McAndrew LM . Sleep problems may mediate associations between rumination and PTSD and depressive symptoms among OIF/OEF veterans . Psychol Trauma. 2015. ; 7 ( 1 ): 76 – 84 . [DOI] [PubMed] [Google Scholar]
- 48. Dinis J , Bragança M . Quality of sleep and depression in college students: a systematic review . Sleep Sci. 2018. ; 11 ( 4 ): 290 – 301 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Tubbs AS , Perlis ML , Basner M , et al . Relationship of nocturnal wakefulness to suicide risk across months and methods of suicide . J Clin Psychiatry. 2020. ; 81 ( 2 ): 19m12964 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Nock MK , Stein MB , Heeringa SG , et al .; Army STARRS Collaborators . Prevalence and correlates of suicidal behavior among soldiers: results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) . JAMA Psychiatry. 2014. ; 71 ( 5 ): 514 – 522 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Lin HT , Lai CH , Perng HJ , et al . Insomnia as an independent predictor of suicide attempts: a nationwide population-based retrospective cohort study . BMC Psychiatry. 2018. ; 18 ( 1 ): 117 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Ziegler-Graham K , MacKenzie EJ , Ephraim PL , Travison TG , Brookmeyer R . Estimating the prevalence of limb loss in the United States: 2005 to 2050 . Arch Phys Med Rehabil. 2008. ; 89 ( 3 ): 422 – 429 . [DOI] [PubMed] [Google Scholar]
- 53. Taylor CA , Bell JM , Breiding MJ , Xu L . Traumatic brain injury-related emergency department visits, hospitalizations, and deaths: United States, 2007 and 2013 . MMWR Surveill Summ. 2017. ; 66 ( 9 ): 1 – 16 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Santaniello JM , Luchette FA , Esposito TJ , et al . Ten year experience of burn, trauma, and combined burn/trauma injuries comparing outcomes . J Trauma. 2004. ; 57 ( 4 ): 696 – 700 . [DOI] [PubMed] [Google Scholar]
- 55. Kilpatrick DG , Resnick HS , Milanak ME , Miller MW , Keyes KM , Friedman MJ . National estimates of exposure to traumatic events and PTSD prevalence using DSM-IV and DSM-5 criteria . J Trauma Stress. 2013. ; 26 ( 5 ): 537 – 547 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Howard JT , Stewart IJ , Kolaja CA , et al . Hypertension in military veterans is associated with combat exposure and combat injury . J Hypertens. 2020. ; 38 ( 7 ): 1293 – 1301 . [DOI] [PubMed] [Google Scholar]
- 57. Watson NF , Rosen IM , Chervin RD ; Board of Directors of the American Academy of Sleep Medicine . The past is prologue: the future of sleep medicine . J Clin Sleep Med. 2017. ; 13 ( 1 ): 127 – 135 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Albrecht JS , Wickwire EM , Vadlamani A , Scharf SM , Tom SE . Trends in insomnia diagnosis and treatment among Medicare beneficiaries, 2006-2013 . Am J Geriatr Psychiatry. 2019. ; 27 ( 3 ): 301 – 309 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Bahr K , Cámara RJA , Gouveris H , Tuin I . Current treatment of comorbid insomnia and obstructive sleep apnea with CBTI and PAP-therapy: a systematic review . Front Neurol. 2018. ; 9 : 804 . [DOI] [PMC free article] [PubMed] [Google Scholar]