Association of sleep with neurobehavioral impairments during inpatient rehabilitation after traumatic brain injury

Abstract

BACKGROUND:

Sleep disturbance is a common sequela after traumatic brain injury (TBI). Many of the impairments following TBI may be exacerbated by impaired sleep-wake cycle regulation.

OBJECTIVES:

To investigate the relationship between total sleep time (TST), measured by wrist actigraphy and observational sleep logs, and neurobehavioral impairments during inpatient rehabilitation after TBI.

METHODS:

Twenty-five subjects undergoing inpatient rehabilitation for traumatic brain injury were included. TST was measured using wrist actigraphy and observational sleep logs. Neurobehavioral impairments were assessed using the Neurobehavioral Rating Scale-Revised (NRS-R), a multidimensional clinician-based assessment.

RESULTS:

Of 25 subjects enrolled, 23 subjects completed the study. A significant negative correlation was found between total NRS-R and TST calculated by observational sleep logs (r = –0.28, p = 0.007). The association between total NRS-R and TST, as calculated by actigraphy, was not significantly correlated (R = –0.01, p = 0.921).

CONCLUSIONS:

Sleep disturbance during inpatient rehabilitation is associated with neurobehavioral impairments after TBI. TST measured by actigraphy may be limited by sleep detection algorithms that have not been validated in certain patient populations. Considerations should be made regarding the feasibility of using wearable sensors in patients with cognitive and behavioral impairments. Challenges regarding actigraphy for sleep monitoring in the brain injury population are discussed.

Keywords

Sleep actigraphy traumatic brain injury

1. Introduction

Sleep disturbance is common following traumatic brain injury (TBI) with 25–29% having a sleep disorder diagnosis (ie insomnia, hypersomnia, and apnea) (Mathias, & Alvaro 2012). Kempf et al. reported that 72% of TBI patients, six months after injury, had sleep-wake disturbances, most commonly manifesting as excessive daytime sleepiness. In healthy populations, in whom the prevalence of sleep disorder diagnoses is less, sleep disturbance affects mood and cognitive performance, measured by choice reaction time, vigilance reaction time, haptic sorting test and self-report mood state questionnaires (Orton, & Gruzelier, 1989).

Sleep disturbance in outpatient chronic TBI patients has shown similar associations with cognition and mood. Kempf et al. found that TBI patients with insomnia all reported depression, which was also commonly found in those with impaired vigilance. Ouellet et al. surveyed 452 TBI patients and found significantly depression and cognitive disturbance with insomnia using the Psychiatric Symptom Index. Sleep disturbance measured by self-report sleep diaries has also been associated with increased level of depression (Parcell, et al. 2006). TBI patients with poor sleep, as defined by index measurements and actigraphy, have also been shown to make more errors of commission on sustained attention tasks (Bloomfield, Espie, & Evans, 2009). There is clearly robust evidence showing the association of sleep disturbance with negative effects on cognition and mood in the chronic TBI patient.

There have been limited investigations into the association of sleep disturbance with neurobehavioral impairments in the acute period following TBI. Mild traumatic brain injuries have been studied in the acute period to identify pre- and post-injury sleep wake pattern changes after injury by self-report and separately to identify neuropsychological impairments following injury (Parsons, & Ver Beek 1982, Dikmen, McLean, & Temkin 1986). Actigraphy has been used to assess sleep disturbances one to three months after moderate to severe traumatic brain injury in the acute rehabilitation setting (Gardani, et al. 2015, Towns,et al. 2016). These studies have not measured the association of sleep disturbance using actigraphy with neurobehavioral and cognitive impairments during the acute phase after TBI.

Accurate monitoring and appropriate management of sleep disturbance during inpatient rehabilitation may improve participation and optimize recovery by reducing exacerbations of neurobehavioral impairments.

The evaluation of sleep in patients with TBI is, however, challenging during inpatient rehabilitation. Accuracy of self-report measures is often limited due to cognitive impairments, including post-traumatic amnesia, after moderate to severe TBI. Nazem et al. found poor agreement using sleep diary measures when comparing to actigraphy after moderate to severe brain injury. Given the unreliability of self-report measures, staff-reported sleep logs are utilized to quantify the amount of sleep during rehabilitation. This requires staff assessing at regular hourly intervals if a patient is awake or asleep based on visual observation. There are inherent limitations to this method, as eye closure does not confirm that the patient is asleep and sleep status can vary between intervals. Thus, there have been inconsistent results regarding the accuracy of this method (Krahn, et al. 1997, Fontaine, 1989, Edwards, & Schuring 1993, Freedman, Kotzer, &Schwab 1999).

Polysomnography (PSG) is the ‘gold standard’ for measuring sleep, however its use on an inpatient rehabilitation unit is limited due to cost, resources, behavioral issues, and medical complexity of patients. Wiseman-Hakes et al. found that ability to tolerate the materials involved with PSG and confusion or agitation limited the feasibility of performing PSG acutely after TBI. More recently the use of wrist actigraphy (ACG) has become a promising method to measure sleep in the TBI population. Wrist ACG uses a watch-like device containing an accelerometer to measure the amount of movement, which is typically described in terms of the sum of the acceleration magnitudes (activity counts) over a time window (epoch). A regression model is used to determine whether the subject is asleep or awake in each epoch based on the activity counts. ACG has been shown to correlate with PSG, the gold standard for sleep assessment. (Kushida, et al. 2001, Zinkhan, et al. 2014). A recent retrospective study demonstrated a moderate to strong correlation between ACG and PSG sleep measures for inpatients after TBI (Kamper, et al. 2016). Additionally, ACG has been reported to be a feasible method for measuring sleep in the inpatient rehabilitation setting after TBI, but additional studies are needed to add to previously published guidelines for its use after TBI (Towns, et al. 2016, Zollman, Cyborski, &Duraski 2010).

Research investigating the association between sleep disturbance and neurobehavioral impairments after TBI has been limited and primarily focused on the outpatient setting. This study was designed to investigate the relationship between total sleep time (TST) and neurobehavioral impairments during inpatient rehabilitation after TBI using wrist actigraphy and observational sleep logs. We hypothesize that TST will be inversely associated with neurobehavioral impairments, as measured by Neurobehavioral Rating Scale- Revised (NRS-R) total scores. Oursecondary hypothesis is that TST, as measured by the ACG, will have a stronger inverse association with neurobehavioral impairments than TST, as measured by sleep logs.

2. Methods

This prospective observational cohort study was conducted at an academic inpatient rehabilitation hospital with approval from the local Institutional Review Board. Subjects admitted to an inpatient brain injury medicine service were screened for eligibility. Subjects were included if they were≥18 years of age and had a traumatic, non-penetrating brain injury within the last 6 months. Subjects were excluded if they had a pre-existing sleep disorder, were unable to participate in study measures due to cognitive impairments or were not fluent in English. Enrolled subjects wore a wrist actigraph device (ActiGraph wGT3X+, ActiGraph LLC, Pensacola, FL, USA) for seven consecutive days. The actigraph was placed on the non-dominant wrist unless there was abnormal tone or weakness in that extremity. In cases of upper extremity weakness and/or abnormal tone, the least affected upper extremity was used. The actigraph was secured onto the patient’s wrist using the manufacturer’s hospital-style wrist straps. At the end of seven days, activity data was downloaded from the actigraph. TST was calculated by the Cole-Kripke algorithm (Cole, et al. 1992) available in the Actilife software and based on the activity data downloaded from the actigraph at the end of seven days. A structured interview, using the script provided from the NRS-R, was conducted on five out of seven days that the patient was wearing the actigraph. The Neurobehavioral Rating Scale-Revised (NRS-R) is a 29-item multidimensional clinician-based assessment instrument designed to measure neurobehavioral disturbances in the TBI population (McCauley, 2001). The NRS-R measures disturbances in five domains including intentional behavior (F1), lowered emotional state (F2), survival-oriented behavior/heightened emotional state (F3), arousal state (F4) and language (F5). Neurobehavioral impairments were rated by the interviewer using a scale of 0–3, absent, mild, moderate to severe, for all 29 items of the NRS-R, following scale instructions. TST from sleep logs recorded by nursing staff during nights preceding structured interviews was also collected. Nursing staff logged if the subject was asleep or awake by observation on an hourly basis overnight. TST from the nursing staff logs was calculated by the total hours the patient was logged asleep. A correlation analysis was performed using Pearson correlation coefficients to compare the total NRS-R scores, all NRS subscale scores except for F5, and TST as measured by the observational sleep logs and wrist actigraphy. Differences between TST measured with actigraphy and observational logs were compared with a paired t-test. Statistical significance was defined as a P value of less than 0.05 for any correlation.

3. Results

Of the 25 enrolled subjects, 23 subjects completed the study. One subject chose to withdraw from the study after the first interview. Another subject was withdrawn from the study after removing and damaging the wrist actigraph. Subject characteristics are shown in Table 1. The mean age of the 23 subjects who completed the study was 49.13 years (standard deviation (SD) 19.3). There were 15 males and 8 females. The mean time since injury was 1.53 (0.7) months. Most patients had a Rancho Los Amigos Scale (RLAS) 7 (n = 15, 65%) and the mean RLAS for the cohort was 6.5 (range 5–8). A significant negative correlation was found between total NRS-R and TST calculated by sleep log (r = –0.28, p = 0.007). No significant correlation was found between total NRS-R or NRS-R subscale scores and TST calculated by wrist actigraphy (r = –0.01, p = 0.921) (Fig. 1). TST as measured by sleep logs had a negative correlation with NRS-R subscales F1 (intentional behavior) (r = –0.25, p = 0.007) and F4 (arousal state) (r = –0.24, p = 0.011). In addition, there were no significant association between NRS-R and wake after sleep onset (WASO), number of awakenings (nAw) or fragmented sleep index as measured by ACG. There was a positive correlation with low agreement between TST as measured by ACG and TST as measured by sleep logs (r = 0.26, p = 0.005). Mean TST estimated by ACG was significantly higher than TST derived from observational logs (mean TST Actigraph = 8.9 h (1.3 h), mean TST Logs = 7.3 h (1.3 h); t-test, t = 3.99, p < 0.001), suggesting that actigraphy might have overestimated sleep time in this patient cohort.

Table 1
Patient Characteristics

Subject Age Sex RLAS Time since injury (m)

1 58 M 7 1.71

2 66 F 7 1.68

3 72 M 5 1.71

4 82 F 8 2.33

5 41 M 7 1.45

6 61 M 6 1.18

7 47 M 7 2.93

8 65 F 7 1.05

9 48 F 5 1.71

10 25 F 5 0.95

11 52 F 5 2.23

12 58 M 6 1.18

13 23 M 7 1.22

14 74 M 7 0.43

15 20 F 7 1.15

16 22 M 7 1.25

17 61 M 5 2.56

18 43 M 7 2.89

19 27 M 7 0.49

20 74 M 7 1.91

21 51 M 7 0.76

22 40 F 7 0.95

23 20 M 7 1.51

Subject	Age	Sex	RLAS	Time since injury (m)
1	58	M	7	1.71
2	66	F	7	1.68
3	72	M	5	1.71
4	82	F	8	2.33
5	41	M	7	1.45
6	61	M	6	1.18
7	47	M	7	2.93
8	65	F	7	1.05
9	48	F	5	1.71
10	25	F	5	0.95
11	52	F	5	2.23
12	58	M	6	1.18
13	23	M	7	1.22
14	74	M	7	0.43
15	20	F	7	1.15
16	22	M	7	1.25
17	61	M	5	2.56
18	43	M	7	2.89
19	27	M	7	0.49
20	74	M	7	1.91
21	51	M	7	0.76
22	40	F	7	0.95
23	20	M	7	1.51

M = male; F = female; RLAS = Rancho Los Amigo Scale; m = months.

Fig. 1.

Correlation between Total Sleep Time (TST) and Neurobehavioral Rating Scale –Revised (NRS-R). Scatter plot of NRS-R vs. total sleep time (TST) as measured by actigraphy (left) and sleep logs (right). Each dot represents a one-day measurement for a patient. Lines represent the least squares fit to the data and shaded areas are the 95% confidence interval of the regression.

4. Discussion

This study demonstrated that TST during inpatient rehabilitation after TBI, as measured by observational sleep logs, is negatively associated with neurobehavioral impairment. This is in line with previously published studies demonstrating a similar correlation in outpatients after TBI (Kempf, et al. 2010, Ouellet, Beaulieu-Bonneau, & Morin, 2006, Parcell, et al. 2006, Bloomfield, Espie, & Evans, 2009). Given the high prevalence of sleep disturbance after TBI and its correlation with neurobehavioral symptoms, appropriate evaluation and management during inpatient rehabilitation is important to improve participation and potential outcomes in rehabilitation programs.

Wrist actigraphy has been previously shown to be a feasible option for measuring sleep in the inpatient rehabilitation setting after TBI (Towns, et al. 2016). One potential limitation of using wrist actigraphy during inpatient rehabilitation, particularly seen in this study, is the risk of damage to devices, especially when recording from patients with cognitive and behavioral impairments. In this study, one actigraph was removed and lost even though the wrist straps provided by the manufacturer were used. This subject had a Rancho Los Amigo Scale of 7, was out of post traumatic amnesia, and displayed disinhibited behaviors. Those patients who show disinhibited, paranoid or perseverative behaviors may be at higher risk of damaging an actigraph even if secured properly. Thus, although a patient may be at a Ranchos Los Amigos level higher than III as suggested by Zollman et al., the risk for damage to equipment may increase. Use of wearable devices in this population in the inpatient setting may require additional cues or redirection from staff to allow monitoring to take place especially over several days. In this study one wearable device was lost during the time when the patient was not supervised. Management of the devices in the inpatient TBI setting does require a team approach for application and monitoring. Non-contact sensors, such as those using Wi-Fi signals (Liu, et al. 2014), may represent an interesting alternative for monitoring sleep in this particular population.

In our cohort, TST as measured by actigraphy did not have a significant correlation with neurobehavioral impairment measured by the NRS-R. In addition, TST measured by actigraphy was significantly higher than TST from observational logs. We hypothesize that TST measurement by wrist actigraphy may be limited by sleep detection algorithms that are not validated in TBI patients. Prior studies however have reported effective use of wrist actigraphy in patients with TBI (Sinclair, Ponsford, & Rajaratnam, 2014, Towns, et al. 2016, Kamper, et al. 2016, Duclos, et al. 2016). Kamper et al. found that actigraphy showed moderate to strong correlations between ACG and PSG in inpatient with severe TBI. While this is the only study in the literature reporting concordance of ACG with PSG in TBI patients, the Philips Respironics, Inc actigraphy system and its algorithm was used. Our study used the ActiGraph LLC system which has not been shown to be concordant with PSG in TBI patients. Further studies examining the concordance of multiple wearable systems with PSG are needed to be able to complete more valid sleep monitoring studies in the TBI population.

In addition, despite using the least impaired upper extremity, our cohort had multiple impairments that may have caused TST to be overestimated using wrist actigraphy. Figure 2 demonstrates the wide range of activity counts recorded by wrist actigraphy throughout the entire monitoring period that may account for this finding. The markedly low counts for certain subjects would suggest an underestimation of activity and overestimation of TST. Of the four subjects with the lowest activity counts, one subject had a hemiparetic arm and a fracture in the other arm through which he was non weight-bearing. Another patient was noted to have significant Parkinsonism, and a third was noted to be bradykinetic on admission to inpatient rehabilitation. Multiple injuries, trauma and impairments in TBI patient thus may limit the use of wrist ACG specifically.

Fig. 2.

Box plot showing the distribution of activity counts across subjects during the monitoring period.

In addition to those limitations previously mentioned, the validity of sleep data measured by wrist actigraphy and sleep logs is uncertain given the lack of measurement with PSG, which is the gold standard to measure TST. Additionally, TST was the only sleep measure accounted for in this study. Interruptions in sleep, time spent in different stages of sleep, and other sleep metrics may also have an influence on neurobehavioral impairments, but this was not accounted for in this study.

This study’s prospective design allowed for multiple points of assessment of both TST and neurobehavioral symptoms. The sample size in this study was similar to previous studies which examined sleep disturbance and neurobehavioral symptoms. This study also compared two different measures of TST that are used in TBI patients to assess their relationship with neurobehavioral symptoms. Administration and scoring of the NRS-R was standardized for all interviewers during the study, and application of the wrist actigraph followed the same protocol throughout the study. This study adds to the discussion of the limitations and challenges to using wrist actigraphy in individuals with significant cognitive and behavioral impairments. This study also is the first to compare the association of two commonly used methods of sleep measurement with neurobehavioral assessments in acute inpatient rehabilitation after traumatic brain injury.

Further investigations into the use of wrist actigraphy with validation by PSG in TBI patients are warranted. Additionally, larger studies to elucidate which neurobehavioral impairments are most affected by sleep disturbance would provide additional insight into the care of TBI patients. Further studies are needed to better clarify whether other sleep metrics (interruptions in sleep, time spent in different sleep stages, etc.) influence neurobehavioral impairments.

5. Conclusion

TST during inpatient rehabilitation as measured by observational sleep logs is negatively associated with neurobehavioral impairments after TBI. TST as measured by actigraphy is not correlated with neurobehavioral impairments and may be limited by sleep detection algorithms that have not been validated in more severely impaired TBI patients. Our findings suggest that sleep detection algorithms should be fitted using data from this specific population, in order for wrist actigraphy to be reliably used in an inpatient setting. In addition, observational sleep logs in the acute rehabilitation setting provide useful information regarding a patient’s TST and may be used to guide clinical decision making for management of sleep wake cycle regulation. Finally, considerations should be made to adequately secure wearable sensors on patients with cognitive and behavioral impairments who are most appropriate to be monitored.

Conflict of interest

None of the authors report any conflict of interest.

No direct or indirect support was provided for this study.

References

Bloomfield, I. L., Espie, C. A., & Evans, J. J. (2009). Do sleep difficulties exacerbate deficits in sustained attention following traumatic brain injury?Journal of the International Neuropsychological Society, 16(01), 17.

Dikmen, S., McLean, A., & Temkin, N. (1986). Neuropsycho-logical and psychosocial consequences of minor head injury. Journal of Neurology, Neurosurgery and Psychiatry, 49(11), 1227–1232.

Duclos, C., Dumont, M., Arbour, C., Paquet, J., Blais, H., Menon, D. K.,… Gosselin, N. (2016). Parallel recovery of consciousness and sleep in acute traumatic brain injury. Neurology, 88(3), 268–275.

Edwards, G. B., & Schuring, L. M. (1993). Pilot study: Validating staff nurses’ observations of sleep and wake states among critically ill patients, using polysomnography. American Journal of Critical Care2, 125–131.

Fontaine, D. K. (1989). Measurement of nocturnal sleep patterns in trauma patients. Heart and Lung, 18, 402–410.

Freedman, N. S., Kotzer

, & Schwab, R. J. (1999). Patient perception of sleep quality and etiology of sleep disruption in the intensive care unit. American Journal of Respiratory Critical Care Medicine, 159, 1155–1162.

Gardani, M., Morfiri, E., Thomson, A., O’Neill, B., & McMillan, T. (2015). Evaluation of sleep disorders in patients with severe traumatic brain injury during rehabilitation. Arch Phys Med Rehabil, 96, 1691–1697.

Kamper, J. E., Garofano, J., Schwartz, D. J., Silva, M. A., Zeitzer, J., Modarres, M.,… Nakase-Richardson, R. (2016). Concordance of Actigraphy With Polysomnography in Traumatic Brain Injury Neurorehabilitation Admissions. Journal of Head Trauma Rehabilitation, 31(2), 117–125.

Kempf, J., Werth, E., Kaiser, P. R., Bassetti, C. L., & Baumann, C. R. (2010). Sleep-wake disturbances 3 years after traumatic brain injury. J Neurol Neurosurg Psychiatry, 81(12), 1402–1405.

10.

Krahn, L. E., Lin, S. C., Wisbey, J., Rummans, T. A., & Ocon-nor, M. K. (1997). Assessing Sleep in Psychiatric Inpatients: Nurse and Patient Reports versus Wrist Actigraphy. Annals of Clinical Psychiatry, 9(4), 203–210.

11.

Kushida, C. A., Chang, A., Gadkary, C., Guilleminault, C., Carrillo, O., & Dement, W. C. (2001). Comparison of acti-graphic, polysomnographic, and subjective assessment of sleep parameters in sleep-disordered patients. Sleep Medicine, 2(5), 389–396.

12.

Liu, X., Cao, J., Tang, S., & Wen, J. (2014). Wi-Sleep: Contactless sleep monitoring via WiFi signals. In Real-Time Systems Symposium (RTSS), 2014 IEEE Real-Time Systems Symposium, 346–355.

13.

Mathias, J. L., & Alvaro, P. K. (2012). Prevalence of sleep disturbances, disorders, and problems following traumatic brain injury: A meta-analysis. Sleep Med, 13(7), 898–905.

14.

Mccauley, S. R. (2001). The neurobehavioral rating scale-revised: Sensitivity and validity in closed head injury assessment. Journal of Neurology, Neurosurgery & Psychiatry, 71(5), 643–651.

15.

Nazem, S., Forster

J. E.

, Brenner, L. A., & Matthew, E. E. (2016). Actigraphic and sleep diary measures in veterans with traumatic brain injury: Discepancy in selected sleep parameters. J Head Trauma Rehabil, 31(2), 136–146.

16.

Orton, D. I., & Gruzelier, J. H. (1989). Adverse changes in mood and cognitive performance of house officers after night duty. Bmj, 298(6665), 21–23.

17.

Ouellet, M., Beaulieu-Bonneau, S., &Morin, C. M. (2006). Insomnia in Patients with Traumatic Brain Injury. Journal of Head Trauma Rehabilitation, 21(3), 199–212.

18.

Parcell, D. L., Ponsford, J. L., Rajaratnam, S. M., & Redman, J. R. (2006). Self-Reported Changes to Nighttime Sleep After Traumatic Brain Injury. Archives of Physical Medicine and Rehabilitation, 87(2), 278–285.

19.

Parsons, L. C., & Ver Beek

(1982). Sleep-wake patterns following cerebral concussion. Nursing Research, 31, 260–264.

20.

Rao, V., Spiro, J., Vaishnavi, S., Rastogi, P., Mielke, M., Noll, K.,… Makley, M. (2008). Prevalence and types of sleep disturbances acutely after traumatic brain injury. Brain Injury, 22(5), 381–386.

21.

Sinclair, K. L., Ponsford, J., & Rajaratnam, S. M. (2013). Actigraphic Assessment of Sleep Disturbances following Traumatic Brain Injury. Behavioral Sleep Medicine, 12(1), 13–27.

22.

Towns, S. J., Zeitzer, J., Kamper, J., Holcomb, E., Silva, M. A., Schwartz, D. J., & Nakase-Richardson, R. (2016). Implementation of Actigraphy in Acute Traumatic Brain Injury (TBI) Neurorehabilitation Admissions: A Veterans Administration TBI Model Systems Feasibility Study. PM&R, 8(11), 1046–1054.

23.

Wiseman-Hakes, C., Duclos, C., Blais, H., Dumont, M., Bernard, F., Desautels, A., … Gosselin, N. (2016). Sleep in the acute phase of severe traumatic brain injury: A snapshot of polysomnography. Neurorehabilitation and Neural Repair, 30(8), 713–721.

24.

Zinkhan, M., Berger, K., Hense, S., Nagel, M., Obst, A., Koch, B.,… Stang, A. (2014). Agreement of different methods for assessing sleep characteristics: A comparison of two actigraphs, wrist and hip placement, and self-report with polysomnography. Sleep Medicine, 15(9), 1107–1114.

25.

Zollman, F., Cyborski, C., Duraski. S. (2010). Actigraphy for assessment of sleep in traumatic brain injury: Case series, review of the literature and proposed criteria for use. Brain Injury, 24(5), 748–754.