Abstract
The purpose of the present study was to evaluate the impact of sleep disturbances on subsequent depressive symptomatology among a representative sample of patients following traumatic brain injury (TBI). Within a retrospective cohort design, our sample included 305 individuals from the Transforming Research and Clinical Knowledge in Traumatic Brain Injury Pilot (TRACK-TBI Pilot; NINDS-OD09-004) database. At 3-months post-TBI, symptoms of insomnia were reported by 34% of patients, and symptoms of hypersomnia were reported by 39% of patients. For the vast majority of individuals, sleep complaints were likely to persist through 6-month follow-up. Symptoms of hypersomnia but not insomnia at three months were associated with worsened depressive symptomatology at six months. These results highlight the importance of sleep disturbances in recovery from TBI and suggest targeted sleep treatments as a pathway to improve outcomes and quality of life following TBI.
Introduction
Disturbed sleep is among the most frequent complaints following traumatic brain injury (TBI), with 30 to 75% of TBI patients reporting poor sleep quality [1-4]. Although most sleep complaints resolve over time without intervention, approximately 25% of individuals report sleep complaints that persist for months or years following injury [1-4]. Among TBI patients, poor sleep is associated with worsened outcomes, including depression, pain, and cognitive impairment, which can prolong or exacerbate many common sequelae of TBI [1-4]. Further, the neurologic mechanisms underlying disturbed sleep are directly relevant to recovery from TBI [1-4]. As a result, there has been a dramatic increase in examining the role of sleep and sleep disorders following TBI, as sleep represents a modifiable target to enhance treatment outcomes.
In addition to disturbed sleep, psychological distress and depression increase after TBI and are associated with poorer functional and cognitive recovery [5-14]. Although disturbed sleep is a well-known antecedent and diagnostic criterion for many mood disorders [15], the temporal relationships between sleep disturbances and other TBI sequelae remain unclear. Several studies have suggested that sleep disturbances and comorbid psychological distress create a “vicious cycle,” with the effects of one of the comorbidities perpetuating and potentially exacerbating the others [15]. Understanding the temporal and mechanistic relationships between sleep and mood disturbances can help guide targeted treatment development efforts among patients with TBI.
We examined the temporal relationship between sleep disturbances and subsequent depressive symptomatology following TBI. We hypothesized that sleep disturbances would be positively related to subsequent depressive symptomatology. Specifically, we sought to evaluate whether difficulty sleeping at night (i.e., insomnia symptoms) as well as difficulty staying awake during the day (i.e., hypersomnia symptoms) would predict future depressive symptoms following TBI.
Methods
Participants
We conducted a retrospective cohort study using data from the Transforming Research and Clinical Knowledge in Traumatic Brain Injury Pilot (TRACK-TBI Pilot; NINDS-OD09-004) study database. TRACK-TBI Pilot was a prospective, multicenter study that enrolled participants with TBI from 2010-2012 [16]. Participants were recruited from emergency departments of three Level 1 trauma centers (University Medical Center Brackenridge, Austin, TX; University of Pittsburgh Medical Center; San Francisco General Hospital) within 24 hours of injury. Inclusion criteria were external force trauma to the head requiring a clinically indicated head computed tomography (CT) scan within 24 hours of injury and ability to understand and speak English. Exclusion criteria included pregnancy, incarceration, nonsurvivable physical trauma, debilitating mental health disorders or neurological disease, MRI contraindications (e.g., cardiac pacemakers, aneurism clips, insulin pumps), and pre-existing medical conditions which could interfere with neuropsychological assessments. Because we were interested in estimating the association between sleep disturbances at three months and depressive symptomology at six months, we excluded individuals <18 years old, those who died at discharge or within six months of injury, those with missing exposure information at three months, and those with missing outcome information at six months. This research was approved by multiple institutional review boards, including University of Maryland, Baltimore; University Medical Center Brackenridge, Austin, TX; University of Pittsburgh Medical Center; and San Francisco General Hospital. Data is available through the TRACK-TBI Consortium.
Measures
Patients were assessed at baseline, three months, and six months.
Demographics:
Demographic, clinical, and injury-related variables were obtained at baseline from the patient or proxy.
TBI Injury severity:
The Glasgow Coma Scale (GCS) score was used as a measure of TBI severity [17]. The GCS is a measure of neurologic deficit and is typically categorized as 13-15 (mild TBI), 9-12 (moderate TBI), and 3-8 (severe TBI).
Sleep complaints:
Sleep disturbances were measured at three and six months using four questions from the cognitive assessment [18]. We combined responses to two questions (“Do you have trouble falling asleep?” and “Are you sleeping less than usual?”) to form a face valid insomnia composite measure consisting of three independent scores (range 0-2). Similarly, we combined responses to two other questions (“Are you sleeping more than usual?” and “Are you feeling drowsier than usual?”) to form a face valid hypersomnia composite score (range 0-2).
Depressive symptomatology:
Sadness was measured at three months and six months using a single yes/no item. Depressive symptomology was measured at six months using the raw score from the depression subscale of the brief symptom inventory (BSI) assessed at six months (range 0-24) [18,19].
Analytic plan
We first examined the distribution and frequencies of all variables. Then, we compared all variables between participants who reported a positive response to at least one insomnia symptom and participants who reported no insomnia symptoms. Similar analyses were conducted for hypersomnia. Differences between categories were evaluated using chi square goodness of fit, Student’s t, or Wilcoxon rank sum tests.
The depression subscale at six months was highly skewed with numerous zero values. To accommodate over-dispersion of our outcome, we employed a zero-inflated negative binomial regression model, using study-site as the inflation factor to account for clustering. First, we ran unadjusted models. Next, potential confounders identified in bivariate analysis (p<0.05) were added simultaneously to the model. These were systematically removed if not significant (p<0.05) and the effect estimate was not changed by more than 10%. Two separate models were built for the hypersomnia and insomnia composite independent variables. Rate ratios (RR) and 95% confidence intervals (CI) are reported.
Sadness at three months was associated with depressive symptomology at six months, but technically could not be a confounder because it was measured at the same time as the exposures (insomnia and hypersomnia) [20,21]. Thus, we ran the regressions with and without the sadness variable.
Results
We initially identified 599 participants with TBI. After excluding those age <18 years (n=27), discharged dead (n=18), dead by six months (n=10), missing exposure at three months (n=129), or missing outcome at six months (n=110), we were left with 305 participants in our study sample. Average age was 42.7 ± 17.7 years (Table 1). The sample was predominantly male (70.5%) and white (82.5%). The majority had a GCS score of 13-15 (88.9%), suggesting mild TBI.
Table 1.
Characteristics of TRACK-TBI Pilot Study Participants Aged 18 and Older Who Survived to 6 Months Without Missing Exposure or Outcome Information, by Insomnia Status at 3 Months, N=305.
| Total Sample, n=305 |
No insomnia, n=201 |
Insomnia, n=104 |
p-value1 | |
|---|---|---|---|---|
| Age, mean (SD) | 42.7 (17.7) | 43.1 (18.4) | 42.1 (16.1) | 0.64 |
| Sex, n(%) | ||||
| Female | 90 (29.5) | 58 (28.9) | 32 (30.8) | 0.73 |
| Male | 215 (70.5) | 143 (71.1) | 72 (69.2) | |
| Race, n(%) | 0.31 | |||
| White | 241(82.5) | 164 (85.4) | 77 (77.0) | |
| Black | 25(8.6) | 13 (6.8) | 12 (12.0) | |
| Asian | 13(4.5) | 8 (4.2) | 5 (5.0) | |
| AI/AN/Hawaiian/Pacific | 13(4.4) | 7 (3.7) | 6 (6.0) | |
| Hispanic, n(%) | 39(12.8) | 20 (10.0) | 19 (18.3) | 0.04 |
| Education, n(%) | <0.001 | |||
| Less than HS | 16(5.5) | 4 (2.1) | 12 (11.9) | |
| High School | 148(50.5) | 90 (46.9) | 58 (57.4) | |
| Some College/Graduate | 93(31.7) | 71 (37.0) | 22 (21.8) | |
| Post Graduate | 36(12.3) | 27 (14.1) | 9 (8.9) | |
| Employment Status, n(%) | 0.43 | |||
| Full time | 130(44.1) | 90 (46.6) | 40 (39.2) | |
| Part time | 43(14.6) | 28 (14.5) | 15 (14.7) | |
| Unemployed | 122(41.4) | 75 (38.9) | 47 (46.1) | |
| Married, n(%) | 103(34.3) | 74 (37.8) | 29 (27.9) | 0.09 |
| Study Site, n(%) | 0.44 | |||
| UCSF | 196(64.3) | 126 (62.7) | 70 (67.3) | |
| UMC Brackenridge | 20(6.6) | 12 (6.0) | 8 (7.7) | |
| University of Pittsburgh | 89(29.2) | 63 (31.3) | 26 (25.0) | |
| Arrival GCS categories, n(%) | 0.23 | |||
| 13-15 | 271(88.9) | 183 (91.0) | 88 (84.6) | |
| 9-12 | 14(4.6) | 7 (3.5) | 7 (6.7) | |
| 3-8 | 20(6.6) | 11 (5.5) | 9 (8.7) | |
| Medical History, n(%) | ||||
| Anxiety | 47(15.4) | 21 (10.5) | 26 (25.0) | 0.001 |
| Depression | 76(24.9) | 40 (19.9) | 36 (34.6) | 0.01 |
| Sleep Disorder | 26(8.5) | 14 (7.0) | 12 (11.5) | 0.18 |
| Alcohol and/or Drug Use | 177(58.0) | 113 (56.2) | 64 (61.5) | 0.37 |
| Cardiovascular disease | 95(32.2) | 65 (32.3) | 30 (28.8) | 0.53 |
| Renal Disease | 26(8.5) | 18 (9.0) | 8 (7.7) | 0.71 |
| Diabetes | 21(6.9) | 15 (7.5) | 6 (5.8) | 0.58 |
| Seizures | 33(10.8) | 18 (9.0) | 15 (14.4) | 0.14 |
| Prior TBI | 65(21.3) | 32 (15.9) | 33 (31.7) | 0.001 |
| Pulmonary Disease | 68(22.3) | 48 (23.9) | 20 (19.2) | 0.36 |
| Sadness at 3 months, n(%) | <0.001 | |||
| Yes | 63(20.7) | 23 (11.5) | 40 (38.5) | |
| No | 241 (79.3) | 177 (88.5) | 64 (61.5) | |
| Hypersomnia at 6 months, n(%) | 149 (48.9) | 85 (42.3) | 64 (61.5) | 0.001 |
| Insomnia at 6 months, n(%) | 142 (42.6) | 52 (25.9) | 90 (86.5) | <0.001 |
| BSI2 Depression at 6 months, mean (SD) | 4.4 (5.1) | 3.2 (3.9) | 6.5 (5.6) | <0.001 |
| BSI2 Depression at 6 months, median (interquartile range) | 3 (0, 7) | 2 (0, 5) | 5 (1, 10) | <0.001 |
P-value reflects significance test between insomnia and no insomnia, using Chi Square Goodness of Fit, Student’s t-test, or Wilcoxon rank sum.
BSI – Brief Symptom Inventory
Prevalence of sleep complaints
At three months post-TBI, 104 (34.1%) participants reported at least one symptom of insomnia (Table 1). Of these, 55 (52.9%) reported one symptom and 49 (47.1%) reported two symptoms. These individuals were more likely to be Hispanic (18.3% vs 10.0%, X2 = 4.3, p=0.04) and to be less educated (X2 = 19.3, p<0.001 for all education levels). Individuals reporting at least one insomnia symptom were also more likely to report a history of anxiety (25.0% vs. 10.5%, X2 = 11.1, p=0.001), depression (34.6% vs. 19.9%, X2 = 7.9, p=0.01), and prior TBI (31.7% vs. 15.9%, X2 = 10.2, p=0.001). Among those reporting at least one symptom of insomnia at three months, 90 (86.5%) also reported at least one symptom of insomnia at six months, and 46 (61.5%) reported at least one symptom of hypersomnia at six months.
At three months post-TBI, 119 (39.0%) individuals reported at least one symptom of hypersomnia (Table 2). Of these, 83 (69.8%) reported one symptom and 36 (30.5%) reported two symptoms. Individuals with at least one symptom of hypersomnia were more likely to be women (36.1% vs 25.3%, X2 = 4.1, p=0.04) and to be less educated (X2 = 10.5, p=0.02 for all education levels). Individuals reporting at least one hypersomnia symptom were also more likely to report history of anxiety (23.5% vs. 10.2%, X2 = 9.9, p=0.002) and depression (35.3% vs. 18.3%, X2 = 11.2, p=0.001). Among those reporting at least one symptom of hypersomnia at three months, 89 (74.8%) also reported at least one symptom of hypersomnia at six months, and 71 (59.7%) reported at least one symptom of insomnia at six months.
Table 2.
Characteristics of TRACK-TBI Pilot Study Participants Aged 18 and Older Who Survived to 6 Months Without Missing Exposure or Outcome Information, by Hypersomnia Status at 3 Months, N=305.
| No Hypersomnia, n=186 | Hypersomnia, n=119 | p-value1 | |
|---|---|---|---|
| Age, mean (SD) | 43.2 (17.9) | 42.0 (17.3) | 0.57 |
| Sex, n(%) | |||
| Female | 47 (25.3) | 43 (36.1) | 0.04 |
| Male | 139 (74.7) | 76 (63.9) | |
| Race, n(%) | 0.40 | ||
| White | 146 (84.4) | 95 (79.8) | |
| Black | 12 (6.9) | 13 (10.9) | |
| Asian | 9 (5.2) | 4 (3.4) | |
| AI/AN/Hawaiian/Pacific | 6 (3.5) | 7 (5.9) | |
| Hispanic, n(%) | 23 (12.4) | 16 (13.5) | 0.78 |
| Education, n(%) | 0.02 | ||
| Less than HS | 8(4.4) | 8(7.1) | |
| High School | 80(44.2) | 68(60.7) | |
| Some College/Graduate | 68(37.0) | 26(23.2) | |
| Post Graduate | 26(14.4) | 10(8.9) | |
| Employment Status, n(%) | 0.42 | ||
| Full time | 85(46.7) | 45(39.8) | |
| Part time | 27(14.8) | 16(14.2) | |
| Unemployed | 70(38.5) | 52(46.0) | |
| Married, n(%) | 71(39.2) | 32(26.9) | 0.03 |
| Study Site, n(%) | 0.15 | ||
| UCSF | 123(66.1) | 73(61.3) | |
| UMC Brackenridge | 15(8.1) | 5(4.2) | |
| University of Pittsburgh | 48(25.8) | 41(34.5) | |
| Arrival GCS categories, n(%) | 0.19 | ||
| 13-15 | 170 (91.4) | 101 (84.9) | |
| 9-12 | 6 (3.2) | 8 (6.7) | |
| <9 | 10 (5.4) | 10 (8.4) | |
| Medical History, n(%) | |||
| Anxiety | 19(10.2) | 28(23.5) | 0.002 |
| Depression | 34(18.3) | 42(35.3) | 0.001 |
| Sleep Disorder | 14(7.5) | 12(10.1) | 0.44 |
| Alcohol and/or Drug Use | 107(57.5) | 70(58.8) | 0.82 |
| Arrhythmia | 121(65.1) | 96(80.7) | 0.003 |
| Cardiovascular disease | 58(31.2) | 37(31.1) | 0.99 |
| Renal Disease | 16(8.6) | 10(8.4) | 0.95 |
| Diabetes | 10(5.4) | 11(9.2) | 0.19 |
| Seizures | 15(8.1) | 18(15.1) | 0.05 |
| Prior TBI | 33(17.7) | 32(26.9) | 0.06 |
| Pulmonary Disease | 45(24.2) | 23(19.3) | 0.32 |
| Sadness at 3 months, n(%) | <0.001 | ||
| Yes | 25(13.5) | 38(31.9) | |
| No | 160(86.5) | 81(68.1) | |
| Hypersomnia at 6 months, n(%) | 60 (32.3) | 89 (74.8) | <0.001 |
| Insomnia at 6 months, n(%) | 71 (38.2) | 71 (59.7) | <0.001 |
| BSI2 Depression at 6 months, mean (SD) | 3.1 (4.3) | 6.4 (5.5) | <0.001 |
P-value reflects significance test between hypersomnia and no hypersomnia, using Chi Square Goodness of Fit, Student’s t-test, or Wilcoxon rank sum;
BSI – Brief Symptom Inventory
Depressive symptomatology
The mean raw score on the BSI depression subscale at 6 months was 4.4 ± 5.1.
Sleep complaints and future depression
Compared with individuals not reporting insomnia complaints, those who reported at least one insomnia symptom at three months scored significantly higher on the BSI at six months (6.5±5.6 vs. 3.2±3.9, p<0.001). Similarly, compared with individuals not reporting hypersomnia complaints, those who reported at least one hypersomnia symptom at three months scored significantly higher on the BSI at six months (6.4±5.5 vs. 3.1±4.3, p<0.001).
In an unadjusted model, insomnia symptoms at three months were significantly associated with increased depressive symptoms at six months (Table 3). This association was attenuated after adjusting for history of depression, prior TBI, and sadness at three months. When sadness was removed from the model, we once again observed a significant association between one symptom of insomnia at three months and depressive symptoms at six months (RR 1.33; 95% CI 1.00, 1.78). A similar association was also observed between two symptoms of insomnia at three months and depressive symptoms at six months (RR 1.45; 95% CI 1.10, 1.92).
Table 3.
Adjusted Rate Ratio (RR) Between Insomnia at Three Months and Raw Score on the Brief Symptom Inventory Depression Sub-Scale at Six Months.
| RR (95% CI) | |
|---|---|
| Unadjusted | |
| Insomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.54 (1.15, 2.07) |
| Two positive responses | 1.84 (1.37, 2.47) |
| Adjusted | |
| Insomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.21 (0.91, 1.60) |
| Two positive responses | 1.25 (0.91, 1.72) |
| History of depression | 1.78 (1.42, 2.23) |
| Prior TBI | 1.36 (1.06, 1.75) |
| Sadness at 3 months | 1.58 (1.24, 2.01) |
| Remove Sadness at 3 Months | |
| Insomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.33 (1.00, 1.78) |
| Two positive responses | 1.45 (1.10, 1.91) |
In an unadjusted model, hypersomnia symptoms at three months were significantly associated with increased depressive symptoms at six months (Table 4). After adjusting for female sex, marital status, history of depression and sadness at three months, we observed a significant association between one symptom of hypersomnia at three months and depressive symptoms at six months (RR 1.57; 95% CI 1.16, 2.12). We also observed a significant association between two symptoms of hypersomnia at three months and depressive symptoms at six months (RR 1.73; 95% CI 1.21, 2.48). Removing sadness at three months from the model did not significantly change the effect estimates.
Table 4.
Adjusted Rate Ratio (RR) Between Hypersomnia at Three Months and Raw Score on the Brief Symptom Inventory Depression Sub-Scale at Six Months.
| RR (95% CI) | |
|---|---|
| Unadjusted | |
| Hypersomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.73 (1.27, 2.36) |
| Two positive responses | 2.05 (1.44, 2.92) |
| Adjusted | |
| Hypersomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.57 (1.16, 2.12) |
| Two positive responses | 1.73 (1.21, 2.48) |
| Female | 0.72 (0.56, 0.92) |
| Married | 0.68 (0.52, 0.89) |
| History of depression | 1.87 (1.47, 2.39) |
| Sadness at 3 months | 1.78 (1.40, 2.28) |
| Remove Sadness at 3 Months | |
| Hypersomnia at 3 months | |
| No positive response | reference |
| One positive response | 1.58 (1.18, 2.12) |
| Two positive responses | 1.80 (1.28, 2.52) |
Discussion
In our sample of TRACK-TBI Pilot participants primarily with mild TBI (GCS 13-15), the prevalence of sleep disturbances was high among both men and women and increased over time. Further, the vast majority of individuals reporting symptoms of insomnia and hypersomnia at three months post-injury continued to report these difficulties at six-month follow-up, suggesting that sleep disturbances develop early post-injury and persist into the chronic phase. After adjusting for potential confounding variables, symptoms of hypersomnia but not insomnia at three months were predictive of depression at six months. These results add to a growing body of evidence regarding the role of sleep in recovery from TBI and highlight that disturbed sleep is an important precursor to worsened outcomes [1,2] .
In the broader TBI literature, insufficient and disturbed sleep are reported by between 30 to 75% of patients [1-4]. Insomnia, characterized by difficulty initiating or maintaining sleep with an associated daytime complaint, is the most common sleep disorder in this population and reported by up to 65% of individuals following TBI [1-4]. Further, in both healthy controls as well as non-TBI clinical samples, insomnia frequently persists for months or years and has been shown to be unlikely to remit on its own [22]. Results of the present study are consistent with these previous findings. Thirty-four percent of participants reported at least one symptom of insomnia at the three-month follow-up, and 86.5% of these individuals continued to report at least one symptom of insomnia at six months. However, our hypothesis regarding insomnia and depression was not supported. Although previous studies have demonstrated that insomnia is frequently prodromal for depression [13], insomnia at three months was not significantly associated with depression at six months, after adjusting for potential confounding variables. It is unclear whether this surprising finding is due to a unique characteristic of our sample or some other reason. The two items in our insomnia composite were scored yes/no, and neither item assessed difficulty maintaining sleep, a common symptom of insomnia. Certainly, our two-item insomnia scale cannot be considered a comprehensive assessment of sleep, and the screening accuracy and psychometric properties of our insomnia measure are thus unknown.
In addition to excessive wakefulness, excessive sleepiness is a common sequela of TBI [1-4]. In the present study, an even greater number (39%) of participants reported at least one symptom of hypersomnia at three months than reported insomnia, and nearly 75% of these individuals continued to report at least one symptom of hypersomnia at six months. Thus, like insomnia, hypersomnolence appears to develop early and persist for a minimum of six months following TBI. Further, hypersomnia at the three-month follow-up was associated with increased depressive symptomatology at six months, even after controlling for sex, marital status, history of depression, and self-reported sadness. Finally, it is also important to note that although TBI severity as measured by the GCS was not a significant predictor of depressive symptomatology in multivariate analyses, individuals reporting hypersomnolence were more likely to score in the moderate to severe TBI range on the GCS. Although studies of well-characterized TBI samples are generally lacking in the literature, extant data suggests increased rates of sleepiness following moderate and severe TBI [23]. This increase in somnolence post-TBI is likely multifactorial, involving not only more severe injury to sleep-related neural pathways but also extended hospital stays, which increased risk for sequelae related to delirium.
From a clinical perspective, it is important to understand the trajectory of post-TBI hypersomnolence; in this vein our results provide valuable insight into the role of sleepiness in the development of depressive symptoms post-TBI. At the same time, it is also necessary to understand the etiology of hypersomnia to enable targeted treatment development and delivery. Daytime sleepiness is a common consequence of many sleep disorders, including insomnia (prevalence in TBI: 29% [24]), obstructive sleep apnea (OSA; prevalence in TBI: 25% [24]), excessive sleep need (i.e., pleisomnia; prevalence in TBI: 22% [23]), post-traumatic narcolepsy (prevalence in TBI: 4% [24]), and post-traumatic hypersomnia (prevalence in TBI: 28% [24]). Compared with non-TBI controls, patients with TBI demonstrate elevated rates of each of these conditions. In addition to sleep disorders, Baumann and colleagues have observed low levels of CSF hypocretin-1 levels within four days of injury, which was associated with increased daytime sleepiness and which tended to normalize within six months [25]. Regardless, although the mechanisms for post-TBI sleepiness remain poorly understood, our results suggest that clinicians should be encouraged to screen proactively for sleep disturbances following TBI. In non-TBI samples, early sleep interventions have been shown to reduce subsequent depressive symptomatology, prevent relapse to clinical depression, and improve health-related quality of life.
Our study possesses several strengths. First, whereas prior studies examined TBI patients at a single time point or across relatively short intervals, our longitudinal design enabled assessment of baseline (i.e., less than four days) and post-acute (i.e., three month) factors influencing chronic (i.e., six month) TBI outcomes. Second, our TBI sample represents the full distribution of injury severity. Third, our sleep assessment measured both nighttime and daytime complaints and thus advances understanding of TBI sequelae across the 24-hour day.
Study Limitations
At the same time, several limitations must be noted. Most important, our assessment of sleep disturbances was limited, relying on four non-validated self-report items. Both subjective and objective measures of sleep are recommended when studying sleep in TBI populations [1,2]. Further, although the BSI was administered at six-month follow-up, no validated measure of depressive symptomatology was administered at three months. Thus, although we controlled for self-reported sadness at three months, we were unable to evaluate other markers of depressive onset during the transition from the post-acute to chronic phase. Finally, although the serial assessment of the TRACK-TBI Pilot is a strength, our six-month follow-up period is nonetheless relatively brief, and longer follow-up periods will enable greater understanding into the natural history and outcomes associated with sleep disturbances post-TBI. Finally, in part due to the multi-site nature of our study, a larger than expected number of patients were lost to follow-up and not included in the final sample. However, we compared individuals included and excluded from the study, and no differences were observed.
Conclusions
In this TRACK-TBI Pilot study analysis, sleep disturbances following traumatic brain injury were noted in more than threefourths of subjects, and sleep disturbances persisted through at least six months post-injury. Given the incidence and public health impact of TBI [1,2], targeted interventions for posttraumatic sleep disturbances represent a major opportunity to improve outcomes and improve quality of life following traumatic brain injury. Multimethod assessment [1,2] of post-TBI sleep and sleep complaints are likely to provide helpful guidance in initiating extant sleep interventions with potential to improve HrQOL and patient-centered outcomes post TBI.
Acknowledgements
TRACK-TBI investigators include (in alphabetical order): Opeolu M. Adeoye, MD, University of Cincinnati; Kimberly D. Boase, BA, University of Washington; Yelena Bodien-Guller, PhD, Harvard/Spaulding Rehabilitation Hospital; Malcolm R. Bullock, MD, PhD, University of Miami; Randall M. Chesnut, MD, FCCM, FACS, University of Washington; John D. Corrigan, PhD, Ohio State University; Karen L. Crawford, MILS, University of Southern California; Ramon Diaz-Arrastia, MD, University of Pennsylvania; Sureyya S. Dikmen, PhD, University of Washington; Ann-Christine Duhaime, MD, Harvard/Massachusetts General Hospital; Richard G. Ellenbogen, MD, FACS, University of Washington; Frank Ezekiel, University of California, San Francisco; Venkata R. Feeser, MD, Virginia Commonwealth University; Etienne Gaudette, PhD, University of Southern California; Joseph T. Giacino, PhD, Harvard/Spaulding Rehabilitation Hospital; Dana P. Goldman, PhD, University of Southern California; Wayne Gordon, PhD, Mount Sinai School of Medicine; Luis Gonzales, BA, TIRR Memorial Hermann; Shankar P. Gopinath, MD, Baylor College of Medicine; Rao P. Gullapalli, PhD, MBA, University of Maryland, Baltimore; Jesse C. Hemphill, MD, University of California, San Francisco; Gillian A. Hotz, PhD, University of Miami; Joel H. Kramer, PsyD, University of California, San Francisco; Harvey Levin, PhD, Baylor College of Medicine; Christopher J. Lindsell, PhD, Vanderbilt University; Joan Machamer, MA, University of Washington; Christopher Madden, MD, UT Southwestern; Alastair Martin, PhD, University of California, San Francisco; Bruce E. Mathern, MD, Virginia Commonwealth University; Thomas W. McAllister, MD, Indiana University; Michael A. McCrea, PhD, Medical College of Wisconsin; Randall E. Merchant, PhD, Virginia Commonwealth University; Pratik Mukherjee, MD PhD, University of California, San Francsico; Florence Noel, PhD, Baylor College of Medicine; David Okonkwo, MD PhD, University of Pittsburgh; Daniel P. Perl, MD, Uniformed Services University of the Health Sciences; Ava M. Puccio, RN, PhD, University of Pittsburgh; Miri Rabinowitz, PhD, University of Pittsburgh; Claudia S. Robertson, MD, Baylor College of Medicine; Jonathan Rosand, MD, MSC, Harvard/Massachusetts General Hospital; Angelle M. Sander, PhD, Baylor College of Medicine; Gabriela Satris, MSc, University of California, San Francisco; Seth A. Seabury, PhD, University of Southern California; Paulina Sergot, MD, FACEP, Baylor College of Medicine; Mark Sherer, PhD, TIRR Memorial Hermann; Deborah M. Stein, MD, MPH, University of Maryland, Baltimore; Murray B. Stein, MD, MPH, FRCPC, University of California, San Diego; Sabrina R. Taylor, PhD, University of California, San Francisco; Nancy R. Temkin, PhD, University of Washington; Arthur W. Toga, PhD, University of Southern California; L. Christine Turtzo, MD, PhD, Uniformed Services University of the Health Sciences; Mary J. Vassar RN MS, University of California, San Francisco; Paul M. Vespa, MD, University of California, Los Angeles; Kevin K. Wang, PhD, University of Florida; Ross Zafonte, DO, Harvard/Spaulding Rehabilitation Hospital; Zhiqun Zhang, MD, University of Florida.
Financial Support
This was an NIH-funded study (TRACK-TBI Pilot; NINDS-OD09-004).
Conflicts of Interest
EMW and JSA’s institution has received research funding from Merck, ResMed, and the AASM Foundation. EMW has served as a scientific consultant for DayZz, Merck, and Pudue and is an equity shareholder in WellTap. JSA is supported by Agency for Healthcare Quality and Research grant K01HS024560. No other authors have any conflicts to disclose.
References
- 1.Wickwire EM, Williams SG, Roth T, Capaldi VF, Jaffe M, et al. Sleep, Sleep Disorders, and Mild Traumatic Brain Injury. What We Know and What We Need to Know: Findings from a National Working Group. Neurotherapeutics. 2016; 13: 403–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Wickwire EM, Schnyer DM, Smith MT, Germain Al, Williams SG, et al. Sleep, sleep disorders, and circadian health following mild traumatic brain injury in adults: review and research agenda. J Neurotrauma. 2018; 35: 2615–2631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ouellet MC, Beaulieu-Bonneau S, Morin CM. Sleep-wake disturbances after traumatic brain injury. Lancet Neurol. 2015; 14: 746–757. [DOI] [PubMed] [Google Scholar]
- 4.Gilbert KS, Kark SM, Gehrman P, Bogdanova Y. Sleep disturbances, TBI and PTSD: implications for treatment and recovery. Clinical Psychology Review. 2015; 40: 195–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bombardier CH, Fann JR, Temkin NR, Esselman PC, Barber J, et al. Rates of major depressive disorder and clinical outcomes following traumatic brain injury. JAMA. 2010; 303: 1938–1945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bryant RA, O’Donnell ML, Creamer M, McFarlane AC, Clark CR, et al. The psychiatric sequelae of traumatic injury. Am J Psychiatry. 2010; 167: 312–320. [DOI] [PubMed] [Google Scholar]
- 7.Deb S, Lyons I, Koutzoukis C, Ali I, McCarthy G. Rate of psychiatric illness 1 year after traumatic brain injury. Am J Psychiatry. 1999; 156: 374–378. [DOI] [PubMed] [Google Scholar]
- 8.Dikmen SS, Bombardier CH, Machamer JE, Fann JR, Temkin NR. Natural history of depression in traumatic brain injury. Archives of physical medicine and rehabilitation. 2004; 85: 1457–1464. [DOI] [PubMed] [Google Scholar]
- 9.Jorge RE, Robinson RG, Moser D, Tateno A, Crespo-Facorro B, et al. Major depression following traumatic brain injury. Arch Gen Psychiatry. 2004; 61: 42–50. [DOI] [PubMed] [Google Scholar]
- 10.Rapoport MJ, Kiss A, Feinstein A. The impact of major depression on outcome following mild-to-moderate traumatic brain injury in older adults. J Affect Disord. 2006; 92: 273–276. [DOI] [PubMed] [Google Scholar]
- 11.Rogers JM, Read CA. Psychiatric comorbidity following traumatic brain injury. Brain Inj. 2007; 21: 1321–1333. [DOI] [PubMed] [Google Scholar]
- 12.Rosenthal M, Christensen BK, Ross TP. Depression following traumatic brain injury. Arch Phys Med Rehab. 1998; 79: 90–103. [DOI] [PubMed] [Google Scholar]
- 13.Whelan-Goodinson R, Ponsford J, Johnston L, Grant F. Psychiatric disorders following traumatic brain injury: their nature and frequency. J Head Trauma Rehabil. 2009; 24: 324–332. [DOI] [PubMed] [Google Scholar]
- 14.Albrecht JS, Kiptanui Z, Tsang Y, Khokhar B, Liu X, et al. Depression among older adults after traumatic brain injury: a national analysis. Am J Geriatr Psychiatry. 2015; 23: 607–614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Staner L. Comorbidity of insomnia and depression. Sleep Med Rev. 2010; 14: 35–46. [DOI] [PubMed] [Google Scholar]
- 16.Yue JK, Vassar MJ, Lingsma HF, Cooper SR, Okonkwo DO, et al. Transforming research and clinical knowledge in traumatic brain injury pilot: multicenter implementation of the common data elements for traumatic brain injury. J Neurotrauma. 2013; 30: 1831–1844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974; 2: 81–84. [DOI] [PubMed] [Google Scholar]
- 18.Derogatis LR, Melisaratos N. The Brief Symptom Inventory: an introductory report. Psychol Med. 1983; 13: 595–605. [PubMed] [Google Scholar]
- 19.Franke GH, Jaeger S, Glaesmer H, Barkmann C, Petrowski K, et al. Psychometric analysis of the brief symptom inventory 18 (BSI-18) in a representative German sample. BMC Med Res Methodol. 2017; 17: 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Rothman K, Greenland S, Lash T. Modern Epidemiology. Philadelphia, PA: Lippincott, Williams and Wilkins; 2008. [Google Scholar]
- 21.Schisterman EF, Cole SR, Platt RW. Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology. 2009; 20: 488–495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Morin CM, Bélanger L, LeBlanc M, Ivers H, Savard J, et al. The natural history of insomnia: a population-based 3-year longitudinal study. Arch Intern Med. 2009; 169: 447–453. [DOI] [PubMed] [Google Scholar]
- 23.Baumann CR. Sleep and traumatic brain injury. Sleep Med Clinics. 2016; 11: 19–23. [DOI] [PubMed] [Google Scholar]
- 24.Mathias JL, Alvaro PK. Prevalence of sleep disturbances, disorders, and problems following traumatic brain injury: a meta-analysis. Sleep Medicine. 2012; 13: 898–905. [DOI] [PubMed] [Google Scholar]
- 25.Baumann CR, Werth E, Stocker R, Ludwig S, Bassetti CL. Sleep-wake disturbances 6 months after traumatic brain injury: a prospective study. Brain. 2007; 130: 1873–1883. [DOI] [PubMed] [Google Scholar]