Sleep before and after work-related concussion: Sex differences in effects and functional outcomes

Abstract

BACKGROUND:

Recent studies have started disentangling components of disturbed sleep as part of the post-concussive syndrome, but little is known about the workers with an injury’ perspectives on post-injury sleep changes or what causes these changes.

OBJECTIVES:

To determine the effects of work-related concussion/mild traumatic brain injury (wr-mTBI) on perceptions of refreshing sleep in workers with an injury and to identify the relevant factors responsible for sleep changes.

METHODS:

We studied post-concussive changes in sleep in 66 adults (50% male workers, 42% aged 30–50 years, median post-injury days: 155) who had sustained wr-mTBI and experienced functional limitations long after the injury. We collected sociodemographic, occupational and health status data and identified variables related to post-concussive changes in refreshing sleep.

RESULTS:

Forty-seven workers with wr-mTBI (79% of male workers, 64% of female workers) perceived their sleep as being refreshing before injury and unrefreshing afterwards (χ2 = 67.70 for change, χ2 = 27.6 for female and χ2 = 41.1 for male workers, p < 0.0001). Post-concussive losses in refreshing sleep were associated with socio demographic, occupational, and health status data variables. Sex stratification revealed differences between male and female workers.

CONCLUSIONS:

Workers with wr-mTBI experience clinically meaningful changes in refreshing sleep that are associated with modifiable variables. The observed differences in functional outcomes between male and female workers warrant further study.

Keywords

Fatigue Global Assessment of Functioning insomnia mild traumatic brain injury occupational injury

1 Introduction

Concussion is the most common form of mild traumatic brain injury (mTBI) [1], affecting all occupational sectors from construction to health care and social services [2]. The incidence of work-related mTBI (wr-mTBI) is increasing, with the proportion of workplace concussions having undergone a drastic increase over the past decade [3]. Since the introduction of diagnostic criteria for concussion/mTBI by the American Congress of Rehabilitation Medicine in 1993 [4], the complexity of recovery following such injury has been recognised, given the persistence of symptoms in 14–26% of persons at six months post-injury and in approximately 23% of persons at one year post-injury [5]. Known risk factors for delayed recovery following mTBI include: psychological (i.e., depression, anxiety) [8, 9], pain [5] and sleep problems [5 –7], or a combination of these factors. With regard to the general population, a large body of work exists on the adverse effects of disturbed sleep on mood, including the development of major depressive disorder and decreased depression treatment response [10 –16], greater pain, and degraded family and societal functioning, and work capacity [19, 20]. Sleep disturbances are common in both the acute and chronic phases post-injury, and may worsen in some persons over time [6, 7]. In fact, insomnia is one of the most commonly cited problems in persons with mTBI [21, 22], and it is also the second most widely recognised sleep disturbance in the TBI population, after snoring, with a conservatively estimated 53% of patients chronically suffering from it [23]. Recent clinical studies highlighted association between problems with sleep in mTBI and advanced age, female sex, lower education, and several medical and psychiatric disorders presenting post-injury [7 –12]. To date, the degree to which pre-injury medical and psychiatric disorders, and pre-injury sleep disturbances associate with sleep disturbances post- injury have not been studied.

To address the gap in knowledge, this study aimed to investigate processes underlying post-injury sleep changes in persons with work-related mTBI (wr-mTBI). This was accomplished by simultaneously considering a priori hypothesised variables relevant to change in refreshing sleep, as reported by workers with an injury, falling under psychosocial [13], pre- and post-injury clinical, and brain-injury-related categories, through sex lens. The variables within these categories, investigated in the context of wr-mTBI, were mapped (Fig. 1). The structure of the map served as a guide, aiding research design, and constituting a basis for interpretation of the results. It was hypothesised that: (1) wr-mTBI will negatively affect perceptions of refreshing sleep in male and female workers with an injury; (2) this change in perception of refreshing sleep will be influenced by workers’ pre- and post-injury health status and psychosocial variables; and (3) variables which influence change in refreshing sleep will differ between male and female workers. To identify the consequences of post-mTBI changes in refreshing sleep, their associations with functional outcomes, the physician-determined Global Assessment of Functioning (GAF) score and patient-reported fatigue severity, were studied.

Fig. 1

Hypothesized relationships encompassing the construct of refreshing sleep and its potential modification by injury.

2 Methods and materials

The study protocol was approved by the ethics committees at the clinical and academic institutions that the authors are affiliated with. All participants provided written informed consent for participation and use of their clinical and insurer’s health and injury-related information.

2.1 Participants

Since 1998, the neurology service program of a large rehabilitation teaching hospital in Canada has had an agreement with the provincial workers’ compensation board to provide expert diagnostic opinions, and continued management for persons who have, or are suspected to have sustained, neurological injuries at work. Professionals within psychiatry, neurology, and neuropsychology, as well as occupational therapy and physiotherapy, performed full clinical assessments, including referrals for neuroimaging (i.e. magnetic resonance imaging [MRI]) and requests for other medical opinions, as necessary. Much of this information (e.g. MRI findings, neurocognitive data, and neurological signs) ultimately contributes to a diagnosis for each person assessed. The diagnosis of TBI, where established, is provided to the insurer, and therefore indicates liability.

Between 2014 and 2016, a purposive sample of 102 adults (18–65 years of age) who sustained a head injury at work at least three months prior were recruited at the neurology services clinic [20]. The participants underwent comprehensive specialty investigations and neuroimaging scans to establish TBI diagnoses. The researchers were blinded to the participants’ diagnoses until medical charts became available for review following completion of assessments. Consistent with the authors’ previous research in the same setting [7], 82 participants (80%) were diagnosed with mTBI/concussion, and 20 (20%) were diagnosed with moderate or severe TBI or did not receive a brain injury diagnosis. Data from participants with established mTBI/concussion diagnoses were used in the present study, but only from those participants who had complete data on the hypothesised outcome variables (n = 66, 50% male and 50% female workers).

2.2 Characteristics of respondents and non-respondents

The likelihood of obtaining complete outcome data about sleep pre- and post-injury was significantly lower for participants aged 20–24 years, those with body mass indices (BMIs) of ≥25 kg/m², and those with a history of previous head injury. The 66 enrolled participants and 16 excluded candidates did not significantly differ with regard to injury mechanism, time since injury (TSI), education level, sex, diagnoses of malingering, or occupational variables such as industry, occupational category, or work shift (i.e. day versus night) during the injury. Fig 2 depicts the study sample derivation.

Fig. 2

Flowchart depicting process of selection of participating individuals’ data for analysis.

2.3 Instruments and measures

2.3.1 Measure of exposure: Refreshing sleep trajectory from pre- to post-injury

The primary outcome variable was the absolute perceived change in refreshing sleep from pre- to post-injury, as determined by the answers to the questions “Pre-injury: I found my sleep refreshing” and “Post-injury: (i.e. now) I find my sleep refreshing”. The response options for both questions were “always”, “sometimes”, and “never”. The “sometimes” option represented 5% and 8% of responses to the first and second questions, respectively. There was a dilemma as to whether “sometimes” fit better into a positive or negative response category while preserving statistical power and maintaining conceptual accuracy. It was determined that “sometimes” would be placed within the “always” category, mainly due to the clinical implications of a ‘sometimes’ answer. If the person is capable of experiencing refreshing sleep at least sometimes across the time continuum, this suggests that the construct of non-refreshing sleep [17, 18] sleep is dichotomous, and not continuous, which limits the clinical significance of such results (i.e. “sometimes” answer), pointing to the likelihood that behavioural, as opposed to physiological, factors are at play.

2.3.2 Categories of variables associated with change in sleep from pre- to post-injury

The socio demographic and health status variables covering the period from pre- to post-injury were collected from the medical and insurer files for each participant. Occupational variables were gathered from the employer’s report of injury/illness and other relevant forms. To quantify the independent association between changes in refreshing sleep from pre- to post-injury and other factors, the selection of variables was based on the available literature and a priori-defined hypotheses (Fig. 1). The results of bivariate analyses (i.e. associations between variables and changes in refreshing sleep) for the whole sample, and male and female workers, separately, are presented in Table 1.

Table 1
Variables associated with change in refreshing sleep pre- to post-injury

Variables N(% N)^* Change in refreshing sleep pre- to post-injury, overall Change in refreshing sleep pre- to post-injury, male workers Change in refreshing pre- to sleep post-injury, female workers

χ²^‡/Phi p ^‡ χ²^†/Phi p ^† χ²^†/Phi p ^†

Sociodemographic and occupational factors

Sex

Male 33 (50) 1.847/0.167 0.174 NA NA NA NA

Female 33 (50)

Age category

<30 9 (14) 0.844/0.113 0.655 1.344/0.202 0.510 0.651/0.141 0.722

30–50 28 (42)

>50 29 (44)

Tension with injury employer or insurer (n = 65)^∧

Yes 12 (19) 3.60/0.237 0.057 1.748/0.234 0.186 5.775/0.418 0.016

No 53 (82)

Work uncertainty (n = 65)^∧

Yes 27 (42) 2.01/0.176 0.156 0.055/-0.041 0.815 3.42/0.322 0.064

No 38 (59)

Brain injury-related factors

Mechanism of injury: Struck by inanimate object

Yes 14 (21) 3.900/0.243 0.048 2.49/0.275 0.115 1.66/0.224 0.198

No 52 (79)

Mechanism of injury: Struck by another person

Yes 10 (15) 0.440/-0.082 0.505 1.245/0.194 0.265 3.367/-0.319 0.067

No 56 (85)

Localization of injury: Impact occipital

Yes 22 (33) 2.365/0.189 0.124 5.80/0.419 0.016 0.000 1.00

No 44 (67)

Localization of injury: Impact temporal

Yes 14 (21) 0.416/0.079 0.519 1.673/0.225 0.195 0.029/-0.030 0.865

No 52 (79)

Previous head injury

Yes 27 (41) 0.461/0.084 0.497 0.436/-0.115 0.509 1.954/0.243 0.162

No 39 (59)

Health-related data

Depression

Yes 24 (36) 5.345/0.285 0.021 1.079/-0.181 0.290 12.92/0.626 < 0.001

No 42 (64)

Substance use disorder

Yes 21 (32) 2.00/0.174 0.158 5.31/0.401 0.021 0.012/-0.020 0.909

No 45 (68)

Pain disorder

Yes 14 (21) 0.416/0.080 0.519 2.392/-0.269 0.122 4.721/0.378 0.030

No 52 (79)

Chronic fatigue syndrome

Yes 10 (15) 0.008/0.011 0.927 1.974/-0.245 0.160 2.936/0.298 0.087

No 56 (85)

Mood disorder (n = 65)^∧

Yes 24(47) 5.07/0.279 0.024 0.091/-0.052 0.763 9.406/0.542 0.002

No 41(63)

Medication intake at the time of investigation (from clinical files)

Sleeping pills

Yes Pre: 5(8)

Post: 25(38) 9.471/0.379 0.009 5.077/0.392 0.079 4.310/0.361 0.038

No Pre: 61(92)

Post: 41(62)

Opioid analgesics

Yes 7 (11) 0.756/0.107 0.385 0.005/-0.013 0.943 3.726/0.366 0.054

No 59 (89)

Behavioural variables

Napping (n = 57)^∧

Yes 33(58) 0.696/0.106 0.404 4.637/0.387 0.031 1.176/-0.195 0.278

No 24(42)

Body mass index^£ All: 27.59±6.87 5.53^£ 0.022 2.050^£ 0.212 11.21^£ 0.002

M: 28.3±6.71

F: 26.8±7.04

Total hours in bed^£ All: 7.80±1.38 0.010^£ 0.927 4.580^£ 0.009 2.26^£ 0.870

M: 7.64±1.43

F: 7.94±1.35

Functional states at the time of investigation

Fatigue, FSS score^£ All: 50.57±10.2 3.06^£ 0.085 6.440^£ < 0.001 0.05^£ 0.817

M: 50.83±11.13

F: 50.3±9.22

Global Assessment of All: 65.8±4.4 0.011^£ 0.906 2.670^£ 0.007 1.20^£ 0.777

Functioning, GAF score^£ M: 65.4±4.0

F: 66.1±4.8

Variables	N(% N)^*	Change in refreshing sleep pre- to post-injury, overall	Change in refreshing sleep pre- to post-injury, male workers	Change in refreshing pre- to sleep post-injury, female workers
Sociodemographic and occupational factors
Sex
Male	33 (50)	1.847/0.167	0.174	NA	NA	NA	NA
Female	33 (50)
Age category
<30	9 (14)	0.844/0.113	0.655	1.344/0.202	0.510	0.651/0.141	0.722
30–50	28 (42)
>50	29 (44)
Tension with injury employer or insurer (n = 65)^∧
Yes	12 (19)	3.60/0.237	0.057	1.748/0.234	0.186	5.775/0.418	0.016
No	53 (82)
Work uncertainty (n = 65)^∧
Yes	27 (42)	2.01/0.176	0.156	0.055/-0.041	0.815	3.42/0.322	0.064
No	38 (59)
Brain injury-related factors
Mechanism of injury: Struck by inanimate object
Yes	14 (21)	3.900/0.243	0.048	2.49/0.275	0.115	1.66/0.224	0.198
No	52 (79)
Mechanism of injury: Struck by another person
Yes	10 (15)	0.440/-0.082	0.505	1.245/0.194	0.265	3.367/-0.319	0.067
No	56 (85)
Localization of injury: Impact occipital
Yes	22 (33)	2.365/0.189	0.124	5.80/0.419	0.016	0.000	1.00
No	44 (67)
Localization of injury: Impact temporal
Yes	14 (21)	0.416/0.079	0.519	1.673/0.225	0.195	0.029/-0.030	0.865
No	52 (79)
Previous head injury
Yes	27 (41)	0.461/0.084	0.497	0.436/-0.115	0.509	1.954/0.243	0.162
No	39 (59)
Health-related data
Depression
Yes	24 (36)	5.345/0.285	0.021	1.079/-0.181	0.290	12.92/0.626	< 0.001
No	42 (64)
Substance use disorder
Yes	21 (32)	2.00/0.174	0.158	5.31/0.401	0.021	0.012/-0.020	0.909
No	45 (68)
Pain disorder
Yes	14 (21)	0.416/0.080	0.519	2.392/-0.269	0.122	4.721/0.378	0.030
No	52 (79)
Chronic fatigue syndrome
Yes	10 (15)	0.008/0.011	0.927	1.974/-0.245	0.160	2.936/0.298	0.087
No	56 (85)
Mood disorder (n = 65)^∧
Yes	24(47)	5.07/0.279	0.024	0.091/-0.052	0.763	9.406/0.542	0.002
No	41(63)
Medication intake at the time of investigation (from clinical files)
Sleeping pills
Yes	Pre: 5(8)
	Post: 25(38)	9.471/0.379	0.009	5.077/0.392	0.079	4.310/0.361	0.038
No	Pre: 61(92)
	Post: 41(62)
Opioid analgesics
Yes	7 (11)	0.756/0.107	0.385	0.005/-0.013	0.943	3.726/0.366	0.054
No	59 (89)
Behavioural variables
Napping (n = 57)^∧
Yes	33(58)	0.696/0.106	0.404	4.637/0.387	0.031	1.176/-0.195	0.278
No	24(42)
Body mass index^£	All: 27.59±6.87	5.53^£	0.022	2.050^£	0.212	11.21^£	0.002
	M: 28.3±6.71
	F: 26.8±7.04
Total hours in bed^£	All: 7.80±1.38	0.010^£	0.927	4.580^£	0.009	2.26^£	0.870
	M: 7.64±1.43
	F: 7.94±1.35
Functional states at the time of investigation
Fatigue, FSS score^£	All: 50.57±10.2	3.06^£	0.085	6.440^£	< 0.001	0.05^£	0.817
	M: 50.83±11.13
	F: 50.3±9.22
Global Assessment of	All: 65.8±4.4	0.011^£	0.906	2.670^£	0.007	1.20^£	0.777
Functioning, GAF score^£	M: 65.4±4.0
	F: 66.1±4.8

Abbreviations: FSS = Fatigue severity scale; GAF = Global Assessment of Functioning. ^*Mean (SD) for last category (behavioural, other related and functional states); ^∧Missing values were not included in the test since they were not noted and assumed to be non-differential; ^†Fisher’s exact test p-values where the cell value of a contingency table was less than 5; ^‡Chi-square test p-value; bold values indicate significance at p≤0.05; ^£t-test results. N = 66 unless otherwise specified.

2.3.3 Sociodemographic and occupational category

The binary variables in sociodemographic category included sex and education, and the categorical variable was age. The binary occupational variables were occupation at injury, and tension with the employer or insurer as reported in the medical files. Since the construct of refreshing sleep in working persons in general, as well as in workers with an injury with delayed recovery from wr-mTBI, is currently not defined, occupational variables including work uncertainty and shift work at time of injury were considered.

2.3.4 Health status characteristics

Binary variables included the presence of pre- and post-injury disorders, including mood, substance-related, sleep pain, and chronic fatigue disorders, and self-reported anxiety or depression and napping, as well as use of medications, including benzodiazepines, opioid analgesics, non-steroidal anti-inflammatory, and any self-reported sleeping pill intake pre- and post-injury. Continuous variables comprised of self-reported total time in bed before and after injury and BMI.

2.3.5 Brain injury-related characteristics

Binary variables included previous history of head trauma, severity of brain injury as determined by the presence of loss of consciousness (LOC), retrograde or anterograde post-traumatic amnesia (PTA), mechanism and localisation of injury.

2.4 Functional evaluation by a physician and the worker with an injury

The GAF, which is included in the DSM-IV-TR, involves physicians’ evaluations of patients’ functional capacity (i.e. social, occupational, and psychological) [21]. Scores range from 1 (severely impaired functioning) to 100 (extremely high functioning) [21].

On the fatigue severity scale (FSS), participants rate their level of agreement with nine statements regarding the severity of their fatigue and its impact on their everyday activities [22]. The total score is the mean rating of all statements, and higher scores indicate greater fatigue. For further information on the psychometric properties of the GAF and the FSS, the reader is referred to previous studies [21 –23].

2.5 Statistical analysis

Descriptive statistics were presented as count and percentage for binary and categorical variables and mean and standard deviation (SD) for continuous variables. To test the differences between the groups pre- and post-injury on binary and categorical variables, and to examine where the differences lie, the Mantel-Haenszel test or Fisher’s exact test (where the cell value of a contingency table was < 5) were utilised, and phi statistics were reported to indicate the direction of the relationship. To examine the differences between the means of two groups on continuous variables, the t-test was utilised. All analyses were conducted on the whole sample, as well as on male and female workers separately (Supplementary Table 1).

The strongest variables within each a priori-hypothesised category of predictors of outcome for which near-significant differences (two-sided P≤.2) were observed between pre- and/or post-injury values, and which were demonstrated to have clinical or functional relevance, were further studied to assess the association with the perceived post-concussive change in refreshing sleep (Table 1). All analyses were conducted using SAS software version 9.3 (SAS Institute Inc., Cary, NC, USA).

2.6 Sample size calculation

A working guideline for modelling studies holds that if k is the number of independent variables and p is the proportion of either negative or positive cases in the population (whichever is smaller), then the minimum sample size is 10×k/p. In previous studies of this population by the authors, it was found that the percentage of persons with sleep disorders increased from 10% pre-injury to 69% post-injury [7]. By taking the minimum proportion value and setting k equal to four (i.e. number of categories of hypothesised variables), the minimum sample size for this study was calculated to be 10×4/0.6≈67. For group comparisons (pre- to post-injury), setting α= 0.05 and power at 80% (β= 0.20), and placing percent increase in sleep adversity from pre- to post injury at 60% for both sexes, calculations showed it would take 15 male and 22 female workers to detect differences between groups. The actual sample size was thus adequate to examine the hypotheses of interest.

3 Results

3.1 Sociodemographic and occupational sample characteristics

At the time of assessment, 86% of the participants were ≥30 years of age, and 80% had completed at least some post-secondary training or qualification, such as a trade-related certification or higher. Among the 65 participants with available data, 19% had experienced at least some degree of tension with their pre-injury employer or insurer in dealing with their compensation claim, and 42% felt uncertain about their work in terms of availability during their recovery.

3.2 Health status characteristics

Common documented pre-injury disorders inclu-ded mood disorders (primarily depression) (36%), anxiety (41%), substance abuse (32%), sleep disorders (29%), pain (21%), and chronic fatigue (15%). Tension with an employer or insurer within DSM-IV-TR Axis IV was documented in files of 19% of participants.

The use of medications with possible sleep-re-lated side effects [7] was reported by 87% of the participants, the most common being non-steroidal anti-inflammatory medications (82%). The use of sleeping aid medications pre- and post-injury was reported by 8% and 38% of the participants, respectively. DSM-IV-TR-defined substance use-related disorders were diagnosed in 19% of the participants.

3.3 Brain injury-related characteristics

The TSI distribution was skewed (mean 299 days, SD: 501 days, median: 155 days) in both male (mean±SD: 378±643 days, median: 179 days) and female workers (mean±SD: 208±229 days, median: 144 days). The major causes of wr-mTBI included ground-level falls (23%), object strikes (21%), assaults by other persons (15%), and falls from elevated positions (9%). The initial impact sites included the frontal (47%), occipital (33%), temporal (21%), and parietal (20%) head regions. Previous head injuries, LOC, and PTA were documented in 41% , 14% , and 18% of the participants, respectively.

3.4 Perceived post-concussive change in refreshing sleep and associated variables

The frequencies of refreshing sleep pre- and post-injury and associations are presented in Supplementary Table 1. There were no consistently significant variables associated with non- refreshing sleep pre- and post- injury in either male or female workers with the exception of difficulty with employers and insurer, and substance use disorder. In female workers, job uncertainty, sleeping pill intake, and higher BMI were associated with non-refreshing sleep both pre- and post-injury, and in male participants, the mechanism of injury of being struck by an inanimate object and occipital impact were the associated factors with non-refreshing sleep post injury (Supplementary Table 1).

3.5 Post-concussive change in refreshing sleep

Forty-seven participants (79% of male, 64% of female workers) perceived their sleep to be refreshing pre-injury and non-refreshing post-injury ( $χ_{MH}^{2}$ = 67.70 change overall; $χ_{MH}^{2}$ = 27.6 change for female and $χ_{MH}^{2}$ = 41.1 change for male workers, both P < 0.0001), while 19 (21% of male workers, 36% of female workers) perceived no change.

3.5.1 Variables associated with the perceived post-concussive change in refreshing sleep, overall

Compared to participants who did not perceive a change in refreshing sleep, those who did were significantly more likely to: (1) have been injured by inanimate objects ( $χ_{MH}^{2}$ = 3.90, r_φ = 0.243, P =0.048); (2) have a post-injury DSM-IV-TR-based mood disorder diagnosis ( $χ_{MH}^{2}$ = 5.07, r_φ = 0.279, P = 0.024); (3) take sleeping aid medications post-injury ( $χ_{MH}^{2}$ = 5.78, r_φ = 0.379, P = 0.009); (3) and have a higher BMI (t = 5.53, P = 0.022). The two groups did not significantly differ in any other variables (Table 1). The unadjusted odd ratios for variables hypothesised to be associated with the perceived post-concussive change in refreshing sleep are presented in Table 1.

3.5.2 Variables associated with the perceived change in refreshing sleep in male workers

When results were stratified by sex, bivariate analyses revealed that, in male workers, the perceived change in refreshing sleep was associated with: (1) initial blows to the occipital area ( $χ_{MH}^{2}$ = 5.80, r_φ = 0.419, P = .016); (2) history of substance use disorder ( $χ_{MH}^{2}$ = 5.31, r_φ = 0.401, P = 0.021); (3) napping during the day ( $χ_{MH}^{2}$ = 4.64, r_φ = 0.387, P = 0.031); and (4) duration of nocturnal time in bed (t = 4.58, P = 0.009) (Table 1).

3.5.3 Variables associated with the perceived change in refreshing sleep in female workers

Female workers who perceived a change in refreshing sleep were more likely to: (1) be experiencing difficulties with their employers or insurer ( $χ_{MH}^{2}$ =5.78, r_φ = 0.418, P = 0.016); (2) have depressive ( $χ_{MH}^{2}$ = 12.92, r_φ = 0.626, P = 0.0003) or pain disorders ( $χ_{MH}^{2}$ = 4.72, r_φ = 0.378, P = 0.030); (3) have diagnosis of mood disorder ( $χ_{MH}^{2}$ = 9.41, r_φ = 0.542, P = 0.002); (4) use sleeping aid medications ( $χ_{MH}^{2}$ = 4.31, r_φ = 0.361, P = 0.038); and (5) have higher BMIs (t = 11.21, P = 0.002).

3.6 Post-concussive changes in refreshing sleep and associations with functional outcomes

The mean total GAF score for all participants was 65.8±4.4; 65.4±4.0 for male and 66.1±4.8 for female workers (scores in the range of 61–70 indicate some difficulty in social, occupational, or school functioning). The mean total FSS score was 50.57±10.2 for the entire sample, 50.83±11.13 for male workers, and 50.3±9.22 for female workers (scores > 37 indicate excessive fatigue). The post-concussive change in refreshing sleep was significantly associated with the GAF score and the FSS score in male workers (t = 2.70, P = 0.007 and t = 6.44, P < 0.001, respectively). In female workers, these associations were not significant (t = 1.20, P = 0.777 and t = 0.05, P = 0.817, respectively) (Table 1).

4 Discussion

In this study, among persons with delayed recovery from mTBI/concussion, 12% of male and 27% of female workers experienced non-refreshing sleep before their injury. After the injury, non-refreshing sleep was pervasive, affecting 91% of male and female workers. These numbers are significantly higher than those reported by previous research in longitudinal analysis of 346 adults with mTBI in all stages of recovery, wherein < 10% experienced sleep difficulties pre-injury, 65% did so within the first 2 weeks post-injury, and 41% continued to experience sleep difficulties one year later [25]. Sex stratification of the present results provided a more precise indication of the associative value of the hypothesised variables within the socio-demographic/psychosocial, clinical/medical, brain injury-related, and medication/sub-stance use categories, and highlighted that refreshing sleep is constructed differently in male and female workers, before and after their injuries.

4.1 Post-concussive sleep change in male workers

The present analysis of post-concussive change in refreshing sleep in male workers identified occipital impact head injury, history of substance use disorder, diagnoses of malingering, time spent in bed, and napping during the day as significant covariates. These relationships were a priori hypothesised. The external force in antero-posterior blows has been demonstrated to induce basal ganglia hypoperfusion, which is associated with a range of symptoms including sleeplessness in patients with mild to moderate TBI [26]. Similarly, following injury in an axial direction, development of sleep pathologies in persons with a shallow tentorial angle has been reported [27], wherein direct impact between the tentorium and the pineal gland during an event can injure the gland, thereby disrupting melatonin homeostasis [27] and resulting in feeling unrefreshed upon awakening. However, the sex-specificity of this association currently remains unclear, and should be further investigated.

Substance use disorder has long been considered to be associated with sleep pathology [28], which reinforces the significance of the present findings and the observed sex differences. Substance abuse and dependence is common, both pre- and post-injury. Alcohol use is estimated to be a contributing factor in mTBIs, with 33% of patients admitted to trauma centres being intoxicated at the time of injury, which is more common in male persons than in female (78% versus 60% ; P<0.001) [29]. In sleep, alcohol exacerbates obstructive sleep apnoea, precipitates sleep-related breathing disorders, and increases the risk of periodic limb movements [30, 31], thus disrupting the sleep architecture and manifesting as the perception of poor sleep upon awakening. Abnormal sleep patterns (e.g. fragmented sleep) can persist for up to three years from abstinence onset [32]. The lack of a significant relationship between these two entities (i.e. sleep and substance use) in female persons with mTBI in the present study may be more related to the low frequencies of the disorder, rather than a difference in the susceptibility to sleep problems. Further investigation of this aspect is required.

The presence of malingering disorder (present in 17% of the cohort) was found to be associated with changes in refreshing sleep in male workers. According to the DSM-IV-TR, a diagnosis of malingering is appropriate when two or more of the following criterions are met: (1) presentation of symptoms in a medico-legal context; (2) discordance between the individual’s stated disability and objective data; (3) uncooperative behaviour during evaluation; and (4) presentation of antisocial personality disorder. All of the diagnosed participants in the present study met the first two criteria, given that they were referred by the insurer for evaluation and because wr-mTBI remains a clinical diagnosis with no currently available imaging technique to support diagnosis [33]. Earlier research in persons under litigation suggested that over-reporting of symptoms should not be viewed as malingering, but rather deficits in articulating internal experiences, where older age and sleep problems can exacerbate inattentive responses [34]. Another study highlighted that emotional awareness, social attachment, and interpersonal relating plays a role in an individual’s ability to accurately report, which is subject to sex differences [35]. Recently, the use of antidepressants was demonstrated to be associated with difficulty in identifying feelings [36]. It is apparent that consideration of the full range of factors in performance validity research in persons under litigation is needed to bring more clarity to the topic.

4.2 Post-concussive sleep change in female workers

The present results highlight independent associations among the change in refreshing sleep and tension with an employer or insurer, pain and mood disorders, sleeping pill intake, and BMI in female persons. These associations have been previously described in mTBI and more severe brain injury [37]. Previous research has proposed that a higher BMI is a risk factor for obstructive sleep apnoea, a form of sleep-related breathing disorder, which remains largely undiagnosed in female persons due to variations in clinical presentation (i.e. in female persons, it manifests as symptoms of depressive mood, insomnia, and morning headaches versus snoring and visible apnoea in male persons) as well as higher tolerance to symptoms [38]. Similarly, consistent findings were observed in the associations among sleep, sleeping pill intake, depression, and pain [9]. It is difficult to explain the sex-specific nature of these associations. This hypothesis may have been red in male persons due to insufficient statistical power within these variables, whereas identification of these associations in female persons was possible due to a higher probability of undiagnosed, and therefore untreated, sleep disorders.

4.3 The meaning of refreshing sleep and the significance of post-concussive changes in the functional context

The purpose of this research was to delineate factors associated with changes in refreshing sleep from pre-to post-injury in male and female persons with wr-mTBI and to provide information concerning the significance of such changes in the occupational context. In this regard, we were concerned with whether changes implied functional deficits, which would have implications for a broader health and safety management system, in particular regarding release to return to work after injury. Both male and female persons with wr-mTBI experienced losses of refreshing sleep from pre- to post-injury, and these losses were associated with several aspects of workers’ health and their environment before, at the time of, and after the injury. Our results indicate that changes in sleep from pre- to post-injury in workers with wr-mTBI are a product of diverse environmental influences brought to bear on compromised cerebral function (Fig. 1). The evidence on the association between refreshing sleep loss and modifiable factors such as difficulties with an employer or insurer is interesting, in that it provides a potential avenue for intervention. This is particularly relevant for workers with limited employability after the injury due to their pre-injury health status or job availability, but who should be given coverage for injury proven to be attributable to a specific cause (i.e. wr- mTBI in our case). The latter is embedded in the principle of ‘historical compromise’, where the insurance system is funded exclusively by employers, and, in exchange, employers receive protection from lawsuits that could by brought by workers with an injury on the basis of tort law [39, 40].

The degraded functioning of male workers who experienced non-refreshing sleep post-injury raises questions regarding the safety of allowing these workers to return to occupations where their performance is critical to maintaining their own and their co-workers’ safety. Existing labour regulations do not define the minimum acceptable functioning levels for returning to work post-concussion, and our participants of both sexes obtained GAF scores indicative of some difficulty in social, occupational, or school functioning. The factors determining susceptibility to functional impairments from non-refreshing sleep require further investigation. Longitudinal studies should investigate how and to what extent post-concussive refreshing sleep change exerts sex-specific influences on functioning. It also worth investigating whether gendered expectations impose greater pressure on male workers to return to work despite functional impairment related to wr-mTBI and non-refreshing sleep. The lack of significant findings in female workers warrants further exploration of elements of functioning affected by non-refreshing sleep.

Current approaches to fatigue following TBI incorporate the scientific understanding of sleep dynamics and recovery to address a person’s inability to function at the pre-injury level due to excessive fatigue. Unexpectedly, experiencing post-concussive loss of refreshing sleep was related to greater fatigue in male but not in female workers. In contrast, earlier studies have reported that both sexes with insomnia report greater fatigue, impaired alertness [40], and overall more severe disability compared to their well-rested counterparts [41]. It is difficult to explain these results and they should be interpreted with caution given the small sample size, and the fact that 27% of female persons compared to 12% of male persons in our cohort experienced non-refreshed sleep pre-injury and only 9% of both sexes reported refreshing sleep post-injury. Given that perceptions are shaped by personal experiences, a greater number of female workers with pre-injury non-refreshing sleep in the present cohort may have adjusted their daytime activity demands to the level of their capacity compared to male workers. Alternatively, the function-related hypothesis, particularly concerning post-concussive changes in female workers, may have been red due to insufficient statistical power within it. Larger longitudinal studies are needed to determine how to manage changes in the restorative qualities of sleep and their relationship with daytime functioning. Nonetheless, these observations lend further credibility to future observations of sex-specificity in the factors associated with post-concussive change in refreshing sleep.

4.4 Strengths and limitations

This study combines sociodemographic, occupational, and health status data for workers with mTBI/concussion with their corresponding pre- and post-injury sleep data. Although the results generally confirmed earlier reports that brain injury, even mild in severity, disturbs sleep function in both male and female persons [43], a qualitatively enriched profile on the construct of refreshing sleep was generated, highlighting the following: (1) the relevance of refreshing sleep to the workers with an injury, both male and female; (2) the magnitude of post-concussive change in refreshing sleep; (3) correlations among this change and persons’ health status and their environment; and (4) sex-specific variables associated with sleep changes. Finally, it is demonstrated that the perception of post-concussive change in refreshing sleep has both clinical and functional significance in male workers, and clinical significance in female. By testing null hypotheses and analysing correlations, the value of the research design and the logic behind the research hypotheses in the sample of workers with an injury are presented. Simultaneously analysing person- and clinician-reported data provided results that advance previous mTBI/concussion research and practice.

This work has several limitations. The sample was recruited from a rehabilitation hospital clinic that receives referrals for persons whose symptoms were not resolved within three months of their head injury; thus, the results may not be transferable to the broader TBI-affected population with a speedy recovery. The absence of longitudinal data on post-concussive changes raises questions regarding the reliability of the participants’ recall, since some had sustained their injuries long before. Another possible source of recall bias is the inclusion of cross-sectional concurrently- and cross-sectional retrospectively-recalled data. While no associations between perceived change in refreshing sleep and TSI were observed, we cannot r the possibility of a causal link between refreshing sleep and time.

5 Conclusions

The present work indicates that both male and female workers reported post-injury changes in refreshing sleep, and these changes were associated with modifiable factors requiring rehabilitation intervention. Associations of post-wr-mTBI loss of refreshing sleep with daytime functional capacity both self- and clinician-assessed, advances our understanding of post-injury recovery and functioning, and uncovers, once again, the unity of brain circuits controlling sleep and those that subserve daytime functioning. The sex specificity of the results with respect to covariates of refreshing sleep loss supports previous research, and suggests that female persons may face and react to adverse life events differently than do male persons. Nonetheless, both male and female workers with wr-mTBI have specific needs for their sleep and multiple comorbid disorders and psychosocial factors to be dealt with simultaneously, pointing to the value of multidisciplinary collaborations and evaluations of all aspects of workers’ health and their environment (including psychosocial) when faced with persistent post-concussive symptoms [44].

Conflicts of interest

The authors report no conflicts of interest.

Funding

The study was supported by the Ontario Ministry of Labour (#13-R-056) grant and, in part, by the Canada Research Chairs Program.. During the work on the project the first author was supported by the 2013/2015 Frederick Banting and Charles Best Doctoral Research Award from the Canadian Institutes of Health Research, and the postdoctoral research fellowships from the University of Toronto and the Alzheimer’s Association (AARF-16-442937). Angela Colantonio was supported by the Canadian Institutes for Health Research Grant–Institute for Gender and Health (#CGW-126580). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Footnotes

The supplementary material is available in the electronic version of this article: .

References

World Health Organization. Neurological disorders: Public health challenges. 2010. Retrieved December 27, 2019 from https://www.who.int/neurology/en.

Chang

, Guerriero

, Colantonio

. Epidemiology of work-related traumatic brain injury: a systematic review. Am J Ind Med. 2015;58(4):353–77.

Workplace Safety and Insurance Board. By the Numbers: WSIB Statistical Report. Retrieved December 27, 2019 at: https://www.wsib.on.ca/WSIBPortal/faces/WSIBDetailPage?cGUID=WSIB066165&rDef=WSIB_RD_ARTICLE

Harrington

, Malec

, Cicerone

, Katz

. Current perceptions of rehabilitation professionals towards mild traumatic brain injury. Arch Phys Med Rehabil. 1993;74:579–86.

Cassidy

, Cancelliere

, Carroll

, Côté

, Hincapié

, Holm

, Hartvigsen

, Donovan

, Nygren-de Boussard

, Kristman

, Borg

. Systematic review of self-reported prognosis in adults after mild traumatic brain injury: results of the International Collaboration on Mild Traumatic Brain Injury Prognosis. Arch Phys Med Rehabil. 2014;95(3 Suppl):S132–151.

Chan

, Feinstein

. Persistent Sleep Disturbances Independently Predict Poorer Functional and Social Outcomes 1 Year After Mild Traumatic Brain Injury. J Head Trauma Rehabil. 2015;30(6):E67–75.

Mollayeva

, Mollayeva

, Shapiro

, Cassidy

, Colantonio

. Insomnia in workers with delayed recovery from mild traumatic brain injury. Sleep Med. 2016;153–61.

Laker

. Epidemiology of concussion and mild traumatic brain injury. PM R. 2011;3(10 Suppl 2):S354–358.

Bernstein

, Calamia

. Characteristics of a Mild Traumatic Brain Injury Sample Recruited Using Amazon’s Mechanical Turk. PM R. 2018;10(1):45–55.

10.

Zhang

, Wing

. Sex differences in insomnia: a meta-analysis. Sleep. 2006;29:85–93.

11.

Pandi-Perumal

, Abumuamar

, Spence

, Chattu

, Moscovitch

, BaHammam

. Racial/Ethnic and Social Inequities in Sleep Medicine: The Tip of the Iceberg? J Natl Med Assoc. 2017;109(4):279–86.

12.

Friedman

, Love

, Rosenkranz

, et al. Socioeconomic status predicts objective and subjective sleep quality in aging women. Psychosom Med. 2007;69:682–91.

13.

Mokarami

, Kalteh

, Varmazyar

, Poursdaeghiyan

, Marioryad

. The effect of work-related and socio-demographic factors on Work Ability Index (WAI) among Iranian workers. Work. 2019. doi: 10.3233/WOR-193066. [Epub ahead of print].

14.

Mollayeva

, D’Souza

, Mollayeva

. Sleep and Psychiatric Disorders in Persons With Mild Traumatic Brain Injury. Curr Psychiatry Rep. 2017;19(8):47.

15.

Stein

, Phillips

, Bolton

, Fulford

, Sadler

, Kendler

. What is a mental/psychiatric disorder? From DSM-IV to DSM-V. Psychol Med. 2010;40(11):1759–65.

16.

Scheenen

, de Koning

, van der Horn

, Roks

, Yilmaz

, van der Naalt

, Spikman

. Acute alcohol intoxication in patients with mild traumatic brain injury: Characteristics, recovery, and outcome. Neurotrauma. 2016;33(4):339–45.

17.

Wilkinson

, Shapiro

. Nonrestorative sleep: symptom or unique diagnostic entity? Sleep Med. 2012;13(6):561–9.

18.

Libman

, Fichten

, Creti

, Conrod

, Tran

, Grad

, Jorgensen

, Amsel

, Rizzo

, Baltzan

, Pavilanis

, Bailes

. Refreshing Sleep and Sleep Continuity Determine Perceived Sleep Quality. Sleep Disord. 2016;2016:7170610.

19.

Health Canada. Health Portfolio Sex and Gender-Based Analysis Policy. Online Retrieved December20, 2018, from hc-sc.gsc.ca/hl-vs/pubs/women-femmes/sgba-policy-politique-ags-eng.php.

20.

Sharma

, Nowrouzi

, Mollayeva

, Lewko

, Grigorovich

, Kontos

, Liss

, Gibson

, Mantis

, Colantonio

. Work-related traumatic brain injury: The injured workers’ perspective on job and health and safety training, supervision, and injury preventability. Work. 2019;62(2):319–25.

21.

Söderberg

, Tungström

, Armelius

. Reliability of global assessment of functioning ratings made by clinical psychiatric staff. Psychiatr Serv. 2005;56(4):434–8.

22.

Krupp

, laRossa

, Muir-Nash

, Steinberg

. The fatigue severity scale. Application to patients with multiple sclerosis and systemic lupus erythematosus. Arch Neurol. 1989;46(10):1121–3.

23.

Tyson

, Brown

. How to measure fatigue in neurological conditions? A systematic review of psychometric properties and clinical utility of measures used so far. Clin Rehabil. 2014;28(8):804–16.

24.

Peduzzi

, Concato

, Kemper

, Holford

, Feinstein

. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373–9.

25.

Theadom

, Cropley

, Parmar

, Barker-Collo

, Starkey

, Jones

, Feigin

; BIONIC Research Group. Sleep difficulties one year following mild traumatic brain injury in a population-based study. Sleep Med. 2015;16:926–32.

26.

Abdel-Dayem

, Abu-Judeh

, Kumar

, Atay

, Naddaf

, El-Zeftawy

, Luo

. SPECT brain perfusion abnormalities in mild or moderate traumatic brain injury. Clin Nucl Med. 1998;23(5):309–17.

27.

Yaeger

, Alhilali

, Fakhran

. Evaluation of tentorial length and angle in sleep-wake disturbances after mild traumatic brain injury. AJR Am J Roentgenol. 2014;202:614–8.

28.

Kendler

, Prescott

. Genes, environment, and psy-chopathology: understanding the causes of psychiatric and substance use disorders. New York, NY: Guilford Press, 2006.

29.

Scheenen

, de Koning

, van der Horn

, Roks

, Yilmaz

, van der Naalt

, Spikman

. Acute alcohol intoxication in patients with mild traumatic brain injury: Characteristics, recovery, and outcome. Neurotrauma. 2016;33(4):339–45.

30.

Martindale

, Hurley

, Taber

. Chronic alcohol use and sleep homeostasis: Risk factors and neuroimaging of recovery. J Neuropsychiatry Clin Neurosci. 2017;29(1):A6–5.

31.

Sharma

, Sahota

, Thakkar

. A single episode of binge alcohol drinking causes sleep disturbance, disrupts sleep homeostasis and downregulates equilibrative nucleoside transporter 1. J Neurochem. 2018;146(3):304–21.

32.

Schrader

, Mickeviciene

, Gleizniene

, Jakstiene

, Surkiene

, Stovner

, Obelieniene

. Magnetic resonance imaging after most common form of concussion. BMC Med Imaging. 2009;9:11.

33.

Merckelbach

, Prins

, Boskovic

, Niesten

, À Campo

. Alexithymia as a potential source of symptom over-reporting: An exploratory study in forensic patients and non-forensic participants. Scand J Psychol. 2018;59(2):192–7.

34.

Brooks

, Iverson

, Atkins

, Zae

, Berkner

. Sex Differences and Self-Reported Attention Problems During Baseline Concussion Testing. Appl Neuropsychol Child. 2016;5(2):119–26.

35.

Kajanoja

, Scheinin

, Karukivi

, Karlsson

. Is antidepressant use associated with difficulty identifying feelings? A brief report. Exp Clin Psychopharmacol. 2018;26(1):2–5.

36.

Duan-Porter

, Goldstein

, McDuffie

, Clowse

, Hughes

, Klap

, Masilamani

, Allen LaPointe

, Williams

Jr . Mapping the evidence: Sex effects in high-impact conditions for women veterans – depression, diabetes, and chronic pain. Washington (DC): Department of Veterans Affairs (US); 2015 VA Evidence-based Synthesis Program Reports.

37.

Association of Workers’ Compensation Boards of Canada (AWCBC). (2011). A national resource on workers’ compensation. Retrieved December 27, 2019 from https://www.awcbc.org/en/

38.

Lippel

. Preserving workers’ dignity in workers’ compensation systems: An international perspective. Am J Industr Med. 2012;55(6):519–36.

39.

Mollayeva

, Shapiro

, Cassidy

, Mollayeva

, Colantonio

. Assessment of mild traumatic brain injury/concussion-related fatigue, alertness, and daytime sleepiness: A diagnostic modeling study. Neuropsychiatry (Lon-don). 2017;7(2):86–104.

40.

Mollayeva

, Pratt

, Mollayeva

, Shapiro

, Cassidy

, Colantonio

. The relationship between insomnia and disability in workers with mild traumatic brain injury/concussion: Insomnia and disability in chronic mild traumatic brain injury. Sleep Med. 2016;20:157–66.

41.

Minen

, Boubour

, Walia

, Barr

. Post-Concussive Syndrome: a Focus on Post-Traumatic Headache and Related Cognitive, Psychiatric, and Sleep Issues. Curr Neurol Neurosci Rep. 2016;16(11):100.

42.

Mollayeva

, Colantonio

, Cassidy

, Vernich

, Moineddin

, Shapiro

. Sleep stage distribution in persons with mild traumatic brain injury: a polysomnographic study according to American Academy of Sleep Medicine standards. Sleep Med. 2017;34:179–192.

43.

Vargo

, Vargo

, Gunzler

, Fox

. Interdisciplinary Rehabilitation Referrals in a Concussion Clinic Cohort: An Exploratory Analysis. PM R. 2016;8(3):241–8.