Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 6 , Pages 956-965, June 2009

Baseline Predictors of Fatigue 1 Year After Mild Head Injury

Presented to the James W. Rae Scientific Day, Department of Physical Medicine and Rehabilitation, University of Michigan Health System, May 9, 2008, Ann Arbor, MI.

  • Marita B. de Leon, PhD

      Affiliations

    • Department of Physical Medicine and Rehabilitation, University of Michigan, Ann Arbor, MI
    • ,
  • Ned L. Kirsch, PhD

      Affiliations

    • Department of Physical Medicine and Rehabilitation, University of Michigan, Ann Arbor, MI
    • Corresponding Author InformationCorrespondence to Ned L. Kirsch, PhD, Dept of Physical Medicine and Rehabilitation, University of Michigan Health Systems, 355 Briarwood Circle, Ann Arbor, MI 48109-5742
    • ,
  • Ronald F. Maio, DO

      Affiliations

    • Department of Emergency Medicine and the Injury Research Center, University of Michigan, Ann Arbor, MI
    • ,
  • Cheribeth U. Tan-Schriner, PhD

      Affiliations

    • Michigan Public Health Institute, Okemos, MI
    • ,
  • Scott R. Millis, PhD

      Affiliations

    • Department of Physical Medicine and Rehabilitation, Wayne State University, Detroit, MI
    • ,
  • Shirley Frederiksen, RN, MS

      Affiliations

    • Department of Emergency Medicine, University of Michigan, Ann Arbor, MI
    • ,
  • Clare L. Tanner, PhD

      Affiliations

    • Michigan Public Health Institute, Okemos, MI
    • ,
  • M. Lynn Breer, PhD

      Affiliations

    • Michigan Public Health Institute, Okemos, MI

Article Outline

Abstract 

de Leon MB, Kirsch NL, Maio RF, Tan-Schriner CU, Millis SR, Frederiksen S, Tanner CL, Breer ML. Baseline predictors of fatigue 1 year after mild head injury.

Objective

To compare reports of fatigue 12 months after minor trauma by participants with mild head injury (MHI) with those with other injury, and identify injury and baseline predictors of fatigue.

Design

An inception cohort study of participants with MHI and other nonhead injuries recruited from and interviewed at the emergency department (ED), with a follow-up telephone interview at 12 months.

Setting

Level II community hospital ED.

Participants

Participants (n=58) with MHI and loss of consciousness (LOC) of 30 minutes or less and/or posttraumatic amnesia (PTA) less than 24 hours, 173 with MHI but no PTA/LOC, and 128 with other mild nonhead injuries. Inclusion criteria: age 18 years or older, within 24 hours of injury, Glasgow Coma Scale score of 13 or higher, and discharge from the ED.

Interventions

Not applicable.

Main Outcome Measure

Medical Outcomes Study 36-Item Short-Form Health Survey Vitality subscale.

Results

Significant predictors of fatigue severity at 12 months were baseline fatigue, having seen a counselor for a mental health issue, medical disability, marital status, and in some stage of litigation. Injury type was not a significant predictor.

Conclusions

Fatigue severity 12 months after injury is associated with baseline characteristics and not MHI. Clinicians should be cautious about attributing persisting fatigue to MHI without comprehensive consideration of other possible etiologic factors.

Key Words: Fatigue, Rehabilitation

List of Abbreviations: AIS, Abbreviated Injury Scale, ANOVA, analysis of variance, ED, emergency department, GOAT, Galveston Orientation and Amnesia Test, ISS, Injury Severity Score, LOC, loss of consciousness, MHI, mild head injury, MTBI, mild traumatic brain injury, PCS, postconcussion syndrome, PTA, posttraumatic amnesia, SF-36, Medical Outcomes Study 36-Item Short-Form Health Survey

 

FATIGUE IS AMONG the most frequently cited problems by individuals with either MHI1, 2, 3, 4, 5, 6 or other minor injury.7, 8 Occurrence rates for fatigue among MHI groups have ranged from 21% to 91%, with symptom persistence reported to be from 2 weeks to as long as 5 years.1, 2, 3, 5, 9, 10, 11 The literature regarding fatigue among individuals with other minor injury is limited, but reports have suggested occurrence rates of 31% to 35%.7, 8 There is considerable overlap on reported rates of fatigue across these 2 populations. However, review of available studies that specifically include data concerning fatigue as a symptomatic expression of MHI revealed methodologic and conceptual issues that prevent formulation of clear and coherent conclusions. These issues are reviewed.

Back to Article Outline

 

Issues Concerning Type of Control Group 

Studies that compared MHI with healthy controls have generally found a higher incidence of fatigue for the MHI groups.4, 5, 10 Similar findings were reported in studies that compared participants with mild to severe traumatic brain injury to a noninjured group.12, 13, 14 However, it is not possible to determine from such comparisons whether fatigue is attributable directly to head injury, to trauma in general, or independently to both types of injury mechanisms. A more appropriate comparison group would be participants who have had other (ie, nonhead) injuries.

When an other-injury control group has been reported, results were more varied than for studies that only used healthy controls. For example, Stulemeijer et al9 found more participants with MHI reporting fatigue compared with those with ankle or wrist distortions (32% vs 12%). In contrast, Kraus et al3 reported very similar occurrence rates between MHI and an other-injury group (42.6% and 43.0%, respectively). In this study, a variety of nonhead injuries was sustained by the comparison group. These included fractures, contusions, or lacerations to the extremities, chest, and abdomen; sprains or strains to the neck or back; and facial contusions or lacerations. Ponsford et al15 showed higher fatigue frequency ratings at 1 week postinjury for MHI participants than a comparison group with unspecified minor injuries, but not at 3 months. These findings raise concerns about whether fatigue is reliably and specifically associated with MHI.

Issues Concerning the Use of Cross-Sectional and Descriptive Survey Designs 

Studies that have used cross-sectional5, 9 or descriptive survey designs10, 11 have reported fatigue at specific intervals after minor head injury, but were not able to describe the process of either symptom progression or resolution over time. Ponsford et al15 is the only study, to our knowledge, that prospectively followed fatigue for both patients with MHI who were evaluated in the ED and discharged without admission, and other injury participants, with follow-up phases at 1 week and 3 months postinjury. Other prospective studies of fatigue among head injury of varying severity, including MHI, have had limitations such as not reporting other injury controls,1, 2 only following the MHI group over time, but not the control group,4 and not having baseline or early measures for either group to demonstrate change over time.3

Studies that have investigated outcome measures other than fatigue have followed participants for up to 1 year post-MHI.16, 17, 18 The methodologic approach used by these studies serves as a model, suggesting that a prospective design with a nonhead injury control and follow-up at 1 year or longer would be ideal. Such a design would be more likely to reliably assess (1) whether fatigue persists beyond the time of expected recovery, regardless of injury type, and (2) whether the incidence or severity of persisting fatigue for MHI groups is greater than for other injury groups.

Issues Concerning Mediators and Moderators of Fatigue 

Studies that investigated post-MHI fatigue have often included demographic variables as correlates of interest. Several of these studies have shown that age and sex were not associated with fatigue after MHI9 or head injury of varying severity.11, 12, 14, 19 However, Cantor et al13 have reported an effect for sex across both head injury and no injury control groups. The relationship between education and fatigue is less clear. In one study, educational level was not significantly related to fatigue,9 but in another, fatigue was predicted by higher education.14

The impact of postinjury functional changes on persistence of symptoms has also been widely investigated. A recent well designed study20 of MHI among soldiers after deployment to Iraq found that MHI no longer significantly predicted persistent symptoms, including fatigue, after accounting for postinjury depression and posttraumatic stress disorder. The authors suggested that the higher rates of physical and postconcussion symptoms reported by the participants with MHI were mediated by decline in psychologic health. Similarly, other studies have found postinjury anxiety and depression,11, 14, 21 decline in social activity,9, 22, 23 and health-related quality of life13 to be associated with fatigue.

These studies are distinctive because they report the impact on outcome of changes in functioning subsequent to the injury itself, rather than participant characteristics that were already present at time of injury. However, identifying preinjury characteristics associated with fatigue is also important in order to screen patients who, at the time of injury, might already be at risk for persisting fatigue. There are studies that have implicated the role of these preinjury factors for head injury–related outcomes other than fatigue. For example, the study by Dikmen et al24 of hospitalized patients with MHI and other trauma controls showed that demographic variables and pre-existing conditions were predictors of performance on neuropsychologic tests, more so than having had a head injury. Other studies have reported associations between, on one hand, preinjury health status including physical and mental health, pain, substance abuse, and prior head injury and, on the other hand, a variety of head-injury related outcomes such as postconcussion symptoms, cognitive functioning, disability, and employment.25, 26, 27, 28, 29, 30, 31 However, in our review, no indicators of preinjury functioning other than demographic variables have been investigated as predictors, mediators, or moderators of fatigue.

Issues Concerning Inconsistency in Terminology and Criteria for Determining Mild Head Injury 

The terms mild traumatic brain injury, mild head injury, and concussion have been used almost interchangeably in the literature. Kay et al32 recommended making a distinction between MTBI and MHI, so that MHI refers to any trauma to the head and face, while MTBI refers to a MHI with damage to the brain. McLean et al33 recommended using neutral terms because of the uncertainty of pathogenesis of symptoms after head injuries (eg, minor head injury rather than MTBI) and descriptive rather than definitive classification terms. Most commonly, however, physical evidence of brain injury is not obtained in MHIs. Therefore, brain injury must be deduced using criteria such as presence and length of LOC or PTA and symptomatic signs of neurologic or neuropsychologic dysfunction.34, 35 Cases involving 1 or a combination of these criteria most often tend to be methodologically grouped together under the general category of MTBI, often without consideration of the degree to which the extent and number of such symptoms might reflect injury severity.

Consistent with these concerns, Dikmen et al36 and Rohling et al,37 in their replication of Dikmen's36 study, reported a dose-response relationship between head injury severity and residual cognitive impairment. For example, the Dikmen36 study demonstrated that the mildest injury subgroup, as defined by “time to follow commands” of less than 1 hour, a proxy for coma length, had considerable overlap with trauma controls on various neuropsychologic measures. However, differences between the study groups became more pronounced with increasing head injury severity. Hoge et al,20 although they used the term MTBI, also deliberately classified head injury participants into 2 distinct groups based on presence or absence (ie, only altered mental status) of reported LOC, comparing each of these groups separately with a control condition of other injury. Although the study does not specifically report statistical inferential tests of differences between the 2 head injury groups on physical and other postconcussion symptoms (table 2, p 458),20 visual inspection of the reported data do suggest more symptomatic complaints among participants with LOC than those with only altered mental status. Therefore, we further examined differences between the 2 Hoge et al20 MHI groups by performing multiple chi-square analyses of the percentage occurrence by group of each reported postinjury symptom. These analyses revealed statistically significant differences between the MTBI groups for 5 of the 19 reported symptoms, including headache, chest pain, heart pounding, constipation, and fatigue. These apparent relationships, consistent with the earlier reports by Dikmen et al36 and Rohling et al,37 suggest a constellation of dose-response effects for even MHI that would not otherwise be evident based on analysis of a combined group, regardless of severity.

Overview of the Present Study 

The current literature review presents evidence that fatigue is a persisting symptomatic concern after MHI, but suggests the hypothesis that fatigue is not a response specific to MHI itself.3, 9 For the present study, consistent with our review, several methodologic decisions were made to test this hypothesis. First, we compared reports of fatigue by participants with MHI to those with other injuries. Second, data are reported from collection waves at baseline (ie, in the ED) and prospectively at 12 months after trauma. Data were also collected at 1 month and 3 months postinjury, but are not reported here because the focus of the present report is specifically assessment of long-term outcome relative to baseline status. Third, the term MHI is used rather than MTBI both for its more neutral connotation and because no medical information was available in our data set about whether demonstrable damage to the brain had actually occurred.32 For the study from which the present data were derived, most cases were considered medically too mild even to entertain imaging studies based on ED presentation, or when such studies were available (n=88), no acute, accident-related intracranial lesions were found. However, as a technical consideration, it should be noted that there are strong reasons, based on neuropsychologic findings,38 not to consider participants with demonstrable acute lesions as even having sustained injuries classifiable as mild. For this reason, had such cases been identified among the present sample, they would have been excluded from the study. Fourth, the current method differentiates MHI status based on presence or absence of brief LOC and/or PTA, so that a possible dose-response relationship between MHI and fatigue can be examined. Finally, this study explores both injury and preinjury factors that may additionally or alternatively account for persisting fatigue.

Back to Article Outline

Methods 

Data for this study were derived from a larger dataset that primarily investigated PCS and other outcomes for an ED-based prospective/inception cohort of patients with minor injury. Initial findings were reported by McLean et al.33 The study protocol was approved by the Institutional Review Boards of the University of Michigan Medical School, Saint Joseph Mercy Health System, and the Michigan Public Health Institute.

Participants 

The sample size was determined based on estimates of frequency of PCS for MHI and non-MHI groups. A study of ED patients with MHI and a control group of patients with minor orthopedic injuries reported frequencies of PCS for the MHI group of 58% at 1 month, 43% at 3 months, and 25% at 6 months.39 We assumed a frequency of 58%, a precision of ±5%, and an alpha of .10, and estimated sample sizes at 264 for MHI and 243 for non-MHI.

Recruitment was accomplished in several stages, each of which successively determined eligibility for the study. Patients were recruited who presented directly to the ED within 24 hours of minor injury, regardless of method of transport. The ED site was a Level II trauma center at a community teaching hospital. Patients were recruited 7 days a week and 8 hours a day over a 15-month period. Shifts were sampled as follows: 70% of the shifts from 3 pm to 11 pm, 20% from 7 am to 3 pm, and 10% from 11 pm to 7 am. Based on prior review of ED admission patterns conducted by a subgroup of the authors (C.U.T.-S., C.L.T., M.L.B.), these shift sample percentages were roughly representative of the frequency distribution of ED trauma patients when stratified by time of arrival.

Patients were initially identified as potential participants based on presenting information in the ED, including the following: age 18 years or older, within 24 hours of injury, Glasgow Coma Scale score of 13 or higher, did not meet the institution's criteria for activation of the adult trauma team, and were to be discharged directly from the ED. Exclusion criteria were transfer from another hospital, inability to speak English, being incarcerated, medical evaluation resulting in admission, evidence of still being in a state of PTA at the time of recruitment, assessed LOC of 30 minutes or more, or LOC not attributable to trauma. Based on these criteria, 945 patients were invited to participate.

Eligible patients were then assessed to determine if they were competent to provide informed consent. Criteria were Mini-Mental State Examination40 of 18 or higher and the ability to describe to the research assistant the essential elements of the study. Of the 945 patients invited to participate, 535 (56.6%) gave consent. Of the 410 who did not give consent, reasons for refusal were no reason given or recorded (44%), physical symptoms preclude participation (22%), and concern that participation would delay discharge (12%). Consenting participants were next interviewed to determine assignment to study groups.

Head injury groups 

Patients were assigned to one of the head injury groups based on self-report and evidence in the medical record of having sustained a head injury. All discrepancies were resolved by the attending physician or nurse. Patients with a confirmed head injury and LOC of 30 minutes or less and/or resolved PTA of any duration were assigned to a head injury with PTA and/or LOC group. Those with confirmed head injury with neither LOC nor PTA were assigned to a head injury only group. To assure sample integrity, the physician or nurse advised about whether or not any reported LOC was potentially attributable to medical conditions independent of head injury. The GOAT41 was also administered to ensure that participants were not currently in a state of PTA, and participants were excluded if they scored below 76. Based on these combined procedures, a total of 28 participants were excluded from the study, 11 because of eventual hospital admission, 4 because of primary language other than English, 2 because of failure to obtain or record confirmation of head injury status from the treating physician, 1 for a GOAT score below the study criterion, and 10 based on subsequent discovery of exclusion criteria after meticulous re-examination of participant records.

Other injury group 

Participants confirmed to have no head injury were assigned to the other injury group. Three participants in this group were excluded from the analyses because of reported LOC or PTA despite not having had documented head trauma. The most common injuries sustained by participants in this group included sprains (25%) to the extremities, neck, and back; contusions (24%) or lacerations (16%) to the extremities or upper body; and closed fractures (21%) to the extremities, vertebra, or ribs. The remaining types of injuries included shoulder dislocation, crushing injuries to fingers, minor eye injury, and burn to hand.

The final sample size at the ED was 504 participants, with 94 participants in the head injury with PTA and/or LOC group, 245 in the head injury only group, and 165 in the other injury group. The sample size at 12 months was 359, with an overall retention rate of 71.2%. Figure 1 illustrates the procedures for determining group membership, and figure 2 summarizes the sampling and retention data.

  • View full-size image.
  • Fig 1. 

    Procedure for determining group membership. *Participants assigned to the other injury group were also screened for LOC and/or PTA that were not injury-related (eg, a syncopal episode that caused their accident). All participants with noninjury-related PTA/LOC were excluded from the study (n=3). Abbreviation: DK, don't know. Legend: , Possible head injury; , Head injury with PTA &/or LOC; , Head inury only; , Other injury.

Measures 

Items and measures were selected from the larger data set, to be used as the following: (1) the dependent fatigue variable, (2) covariates of fatigue, and (3) baseline predictors of fatigue at 12 months. For the predictor variables, 4 data types were used: (1) preinjury health status variables, (2) medical variables including information regarding diagnosis and injury characteristics, (3) demographic variables, and (4) a derived dichotomous variable addressing litigation status. A description follows of all variables. They are organized based on their respective roles in the various analyses.

Sample characteristics 

Demographic and injury variables were used to characterize the sample subgroups. Age, sex, education, marital status, employment status, and ethnicity were obtained from the structured interview and cross-validated against the medical record where possible. Cause of injury and injury severity as indicated by the ISS42 were obtained from medical records. The ISS is an index of anatomical injury that provides an overall score for patients with multiple injuries. Each injury is assigned an AIS score, which ranges from 1 to 6, with 1 being minor, and 6 being lethal or unsurvivable. The AIS score for each injury is allocated to 1 of 6 body regions. The ISS is defined as the sum of squares of the highest AIS score in the 3 most severely injured body regions. The ISS ranges from 0 to 75.

Primary outcome 

Fatigue at 12 months was assessed with the 4-item SF-36 Vitality subscale, a component of the SF-36,43 which was administered in full. Consistent with the standard instructions for this measure, participants were asked to “please give the 1 answer that comes closest to the way you were feeling during the past month.” Two of the 4 subscale items concern energy (eg, “How much of the time during the past month did you have a lot of energy?”) and 2 concern fatigue (eg, “How much of the time during the past month did you feel worn out?”) Each item is rated on a 6-point scale ranging from 1 (all of the time) to 6 (none of the time). The 2 items reflecting energy are reverse-scored. Low scores on the Vitality subscale indicate more fatigue—that is, less vitality. Scores for the SF-36 Vitality subscale administered at 12 months (as well as the same subscale administered at baseline) were converted to T scores for ease of interpretation.44 The Vitality subscale has good internal consistency (Cronbach α=.86).45 Validity of the Vitality subscale has been established with multiple clinical populations.46 The Vitality subscale has also specifically distinguished between MHI and healthy or no disability comparison groups.47, 48

Covariates 

The 4-item SF-36 Vitality subscale was administered at baseline using the same scoring procedures as described for the 12-month administration of the subscale. The slightly modified instruction was, “Please give the answer that comes closest to the way you were feeling during the month prior to your current injury.” This covariate was used to control for possible response bias resulting from repeated use of the same instrument. By introducing baseline fatigue as a covariate, the relationship between the predictors and fatigue at 12 months could be more accurately determined. Litigation status at 12 months was also added as a covariate because it has been consistently found to be associated with persistent symptoms after MHI.49, 50 Litigation status was broadly determined as positive if any of the following were true: the participant was actively involved in litigation, had received services from a lawyer or a paralegal, had consulted with an attorney, or was considering consulting an attorney, regarding the injury.

Independent or predictor variables 

Participants were classified into 1 of 3 injury groups: head injury with PTA and/or LOC, head injury only, and other injury. This variable was used as a dummy grouping variable, predicting the primary outcome measure.

Other predictors of fatigue at 12 months were cause of injury (ie, motor vehicle crash vs nonmotor vehicle crash), demographic characteristics (age, sex, education, employment status, marital status), and baseline health status. Having a medical disability (regardless of cause) was used as the indicator of baseline health status and was assessed with a question regarding reasons for not working. Additional components of baseline health status were assessed through responses to single questions regarding history of psychologic or mental health problems. These questions were, “Have you ever seen a counselor, such as a psychologist or psychiatrist for a mental health issue?” and “Have you ever been treated for a drug or alcohol problem?”

Procedure 

The baseline interview, approximately 30 minutes, was administered, if possible, between medical examinations and other treatments. At the conclusion of each interview, contact information was verified for follow-up. The participants were subsequently interviewed by telephone at 1, 3 (both not reported for this study), and 12 months. Halfway between 3 and 12 months, participants were contacted by mail to remind them about an impending telephone follow-up interview at 12 months. At each of the follow-up time points, a maximum of 10 attempts was made to contact each participant. They were given $25 after completion of the interview at the ED, 1 month, and 3 months, and $75 at 12 months to maximize retention rates.

Data Analysis 

Statistical analyses were performed with SPSS version 14.a For all analyses, 2-tailed tests of significance with alpha level of .05 were used.

Differences between study groups in categorical baseline and injury characteristics (eg, sex, employment, cause of injury) and fatigue status at 12 months were assessed with Pearson chi-square tests of independence. The phi coefficient was calculated as an estimate of effect size. Group differences in the main outcome variable, SF-36 Vitality subscale administered at 12 months, and interval-level baseline variables (eg, age, ISS, SF-36 Vitality subscale administered at baseline) were assessed with a series of univariate ANOVA, with Bonferroni correction to adjust for multiple pairwise comparisons. Partial eta squared was the measure of effect size.

Hierarchical linear regression analysis was done to identify variables associated with fatigue severity at 12 months. These variables included selected preinjury characteristics as well as injury group membership. Condition indices and variance proportions were examined to assess for possible multicollinearity between predictor variables.

Back to Article Outline

Results 

Participant Baseline Characteristics 

Pearson chi-square tests of independence and ANOVA (where appropriate) indicated no significant differences between the groups for sex, race, education, marital status (table 1), and ISSs. For the entire study sample, more than half of the participants were female (57%), 80% were white, 93% had at least high school education, and 45% were married. ANOVA yielded no significant group differences in mean ISSs (F=.071, P=.931), with the scores further verifying the minor nature of the injuries for each of the groups (head injury with PTA and/or LOC: mean ± SD, 2.00±1.58; head injury only: mean ± SD, 1.91±2.06; other injury: mean ± SD, 1.89±1.50).

Table 1. Characteristics of Study Sample (% of Cases a Category a Group)
Baseline CharacteristicsHI-P/L (n=58)HIO (n=173)OI (n=128)Total (N=359)P⁎
Female55.259.056.357.4.836
Race .428
White84.581.577.380.5
Black15.513.917.215.3
Hispanic/Asian/other0.04.65.54.2
Less than high school education10.58.73.97.3.167
Currently employed67.269.481.373.3.037
With medical disability6.95.83.15.0.448
Marital status .554
Married/in relationship35.145.648.444.9
Divorced, separated, widowed21.119.318.019.1
Never married43.935.133.636.0
Motor vehicle collision37.933.518.829.0.005

NOTE. Partial total percentages may not equal 100% because of rounding.

Abbreviations: HI-P/L, MHI with PTA and/or LOC; HIO, MHI without PTA or LOC; OI, mild nonhead injury.

⁎P based on Pearson χ2 tests of independence.

The 3 study groups differed significantly in cause of injury, employment status, age, and baseline Vitality score. Pearson chi-square analysis yielded a significant difference between groups in cause of injury. More of the head injury with PTA and/or LOC (37.9%) and head injury only (33.5%) groups and fewer of the other injury group (18.8%) sustained their injuries from a motor vehicle collision (χ2=10.51, P=.005, φ coefficient=.17). Employment status was also significantly different for the study groups. More of the other injury group were employed at the time of injury (81.3%) compared with a lower proportion for the head injury with PTA and/or LOC (67.2%) and head injury only (69.4%) groups (χ2=6.58, P=.037, φ coefficient=.14).

ANOVA revealed a significant difference in age (F2,356=5.525, P=.004; partial η2=.03), with the head injury only group significantly older than the head injury with PTA and/or LOC (Bonferroni-corrected P=.016) and other injury groups (Bonferroni-corrected P=.030). Mean age and SDs for each group are as follows: head injury only, 43.0±18.95; head injury with PTA and/or LOC, 35.7±16.17; and other injury, 37.8±14.47. The 3 groups were also compared on the SF-36 Vitality subscale administered at baseline (F2,356=3.83, P=.023; partial η2=.02). Mean SF-36 Vitality subscale administered at baseline, expressed as T scores, and SDs for the head injury with PTA and/or LOC, head injury only, and other injury groups, respectively, were 52.8±9.53, 50.4±10.48, and 53.4±8.64. The head injury only group scored significantly lower (ie, had greater fatigue severity) than the other injury group (Bonferroni-corrected P=.026). No other pairwise comparisons were significant.

Participants lost to follow-up (n=145, 28.8%) were compared with those who completed the study. A larger percentage (38.3%) of the head injury with PTA and/or LOC group had no follow-up data compared with 29.4% of the head injury only and 22.4% of the other injury groups (χ2=7.45, P=.024, φ coefficient=.12). In addition, more of the participants with the following characteristics were lost to follow-up: males, 18 to 34 years old (youngest age group classification), with less than high school education, and whose cause of injury was assault.

Group Differences in Fatigue Severity at 12 Months 

The primary objective of this study was to assess differences in reported fatigue between participants with and without head injury 12 months after minor trauma. As an initial attempt to determine these differences, scores on the SF-36 Vitality subscale administered at 12 months for the 3 groups were compared. Mean scores of the SF-36 Vitality subscale administered at 12 months, expressed as T scores, and SDs for the head injury with PTA and/or LOC, head injury only, and other injury groups, respectively, were 52.3±12.22, 49.6±11.83, and 53.0±10.37. There were significant differences between the groups (F2,356=3.77, P=.024; partial η2=.02). Pairwise comparisons revealed the head injury only group to have a lower mean score at 12 months than the other injury group (Bonferroni-corrected P=.027). No other pairwise comparisons were significant.

Contributions of Injury Status and Preinjury Characteristics to Fatigue Severity 

Because of the group differences noted for baseline measures, as summarized in the section on participant baseline characteristics, including the SF-36 Vitality subscale administered at baseline, and the negligible effect size for injury group, a hierarchical regression analysis was performed in order to determine significant predictors of the SF-36 Vitality subscale administered at 12 months. This analysis successively introduced 3 blocks of variables. The first block included the covariates SF-36 Vitality subscale administered at baseline and litigation status at 12 months, which were introduced to remove the possible confounding effects of these 2 variables. The second block included cause of injury, demographic, and baseline health status variables. Finally, the third block included head injury status to verify its contribution to prediction of fatigue after accounting for these variables. The regression model is summarized in table 2. Examination of the collinearity diagnostics indicated that multicollinearity was not an issue (maximum condition index=18.2, maximum variance proportion=.35).

Table 2. Hierarchical Regression Analysis Predicting SF-Vitality Score at 12 Months
Summary of Coefficients for Model 3Part rPBSE of Bβ
Constant 34.4445.841
Block 1: covariates
12-mo litigation–.088.044–6.6703.307–.092
Baseline SF-Vitality.388.000.4980.056.427
Block 2: baseline characteristics
Sex.013.762.6602.176.014
Age–.029.507–.0560.084–.042
Education.078.077.9920.559.085
Married–.052.237–3.2382.737–.070
Separated/divorced/widowed–.122.006–9.6003.444–.164
Working–.012.791–0.7312.760–.014
With medical disability–.110.012–13.6935.417–.128
Prior counseling–.105.016–5.3992.236–.112
Prior drug/alcohol treatment.038.3823.7124.240.041
Motor vehicle crash–.034.444–1.7712.309–.035
Block 3: injury groups
HI-P/L (MHI with PTA or LOC).012.7920.8183.090.013
HIO (MHI without PTA or LOC)–.035.427–1.8552.331–.040
Model SummaryFdfPAdjusted R2R2 ChangeF for R2 ChangeP for R2 Change
Model 1 (block 1 only)68.912352.000.277.28168.91.000
Model 2 (blocks 1 and 2)15.2712,342.000.326.0683.55.000
Model 3 (all 3 blocks)13.1314,340.000.324.0020.55.580

As expected, the SF-36 Vitality subscale administered at baseline was a significant predictor and accounted for the largest proportion of the variance of SF-36 Vitality subscale administered at 12 months. It was not possible to determine from our data whether the variance in SF-36 Vitality subscale administered at 12 months accounted for by the SF-36 Vitality subscale administered at baseline is exclusively attributable to a substantive relationship between the SF-36 Vitality subscale administered at baseline and the SF-36 Vitality subscale administered at 12 months (ie, that fatigue at 12 months is clinically associated with reports of fatigue at baseline) or to response bias attributable to use of the same instrument. Although our impression is that the relationship is, in fact, substantive, not artifactual, we wanted to extract from the regression analysis both of these possible sources of variance in order to establish the association between the SF-36 Vitality subscale administered at 12 months and other baseline variables more clearly. In addition, being involved in some stage of the litigation process at 12 months was also associated with fatigue severity. After accounting for the contribution of SF-36 Vitality subscale administered at baseline and litigation status at 12 months, the significant predictors of SF-36 Vitality subscale administered at 12 months were marital status (ie, separated, divorced, or widowed), having a medical disability, having seen a counselor for a mental health issue, and being in some stage in the process of litigation. Having had a head injury was not significantly associated with fatigue severity at 12 months after accounting for baseline characteristics.

Back to Article Outline

Discussion 

This study examined rates and severity of fatigue at 12 months postinjury for participant groups (evaluated in and discharged directly from the ED) characterized by 2 degrees of MHI severity, compared with a group having other mild nonhead injuries. The data yielded several critical findings.

First, in contrast with earlier work examining the relationship between MHI and fatigue, our findings indicate that fatigue is not a significant problem for the current sample of very mildly injured participants when assessed 12 months after injury. Normative scoring of the SF-36 Vitality subscale indicates that mean Vitality scores for all the study groups approximate the population mean for this measure. In addition, SF-36 Vitality subscale scores for this study's MHI subgroups are higher than what has been reported in studies using the same measure at 3 to 4 months51 and at least 1 year48 after MHI. This difference may be a result of the very mild nature of injury severity among our cohort compared with samples in other studies. For example, our MHI subgroups had lower mean ISS scores and the head injury with PTA and/or LOC group had shorter mean PTA length than the MHI sample of Paniak et al.51 Other studies reporting fatigue among participants with MHI also included hospitalized cases in their sample,4, 9 in contrast with our nonhospitalized cohort.

Second, compared with the results of the above referenced studies,4, 9, 48, 51 our findings of less fatigue for a nonhospitalized MHI cohort might suggest that injury severity affects later fatigue, even for head injuries of the mildest kind. However, despite this possibility, comparisons of our 3 study groups yielded no significant difference between the MHI subgroups in fatigue severity at 12 months. Although the less severe head injury group (ie, with no accompanying LOC or PTA) did report more fatigue than the other injury group, our regression analysis indicates that this seemingly paradoxical finding is a result of differences in initial participant characteristics. Specifically, our findings indicate that poor medical health, mental health, and marital status at baseline (being widowed, divorced, or separated), and not type of minor injury, were the significant predictors of fatigue severity at 12 months. Regarding marital status, post hoc examination of the data further indicates that significantly more of those who were widowed, divorced, or separated were living alone at the time of injury, even compared with those who reported being single at the time of injury (χ2=66.9, P=.001). Although we do not have data to support a specific theoretical interpretation of this finding, the data do suggest that disrupted social support is a critical variable that affects reports of fatigue and, hypothetically, feelings of general well being.

This finding, that preinjury status, rather than injury type, predicts later fatigue severity, emphasizes the importance of including preinjury characteristics when examining the etiology of fatigue after MHI or, more generally, mild trauma of any kind. Our data suggest that in the absence of information about baseline functioning, fatigue may be misattributed to the occurrence of an injury when, in fact, the persistence of fatigue may simply reflect an individual's health and psychosocial status as of the time they report to the ED.

Third, involvement in the litigation process is an important covariate of fatigue severity. This finding represents partial support for the consistently reported role of compensation issues in poor recovery after MHI.49, 50

Study Limitations 

The generalizability of findings from the present sample of nonhospitalized cases to the general MHI population may be limited. Comparison of our findings to other studies raises the possibility that injury severity may, in fact, influence fatigue (although we did not find such a difference specifically for our sample of very mildly injured participants). This issue could be clarified by studies that incorporate samples of participants for comparison across a broader range of injury severity subclasses.

Another limitation of the present study is the possibility of sample bias, as reflected by several baseline differences between the study groups. First, the higher percentage of participants with MHI with injuries from a motor vehicle crash may be indicative of more severe trauma. However, this appears unlikely because the ISSs indicate very minor injury severity for all groups. Second, the higher percentage of employed participants in the other injury group may reflect better functioning for this group than the MHI subgroups. However, if higher rates of employment among other injury participants resulted in less fatigue, there would be a significant difference in reported fatigue between the other injury and PTA/LOC groups. This was not the case. Third, the head injury only group reports poorer baseline health status than the other 2 groups, as well as greater fatigue. The mechanisms responsible for these findings are unclear and require further investigation. However, there are several hypotheses worth considering. Individuals with characteristics such as poor baseline health may be more likely to seek emergency care after minor head injury if they are sensitized to the possibility that additional insults might compromise their functional status. Additionally, as suggested by Mittenberg et al,52 some individuals may misattribute baseline levels of fatigue to their injury and continue to do so over longer periods.

Finally, a third study limitation is the relatively high rate of observed attrition. The initial refusal rate at the ED was large (43%). It is possible that refusal to participate is a nonrandom phenomenon that resulted in a biased sample. Unfortunately, we were unable to verify the possibility of such a bias. The absence of information from those who refused to participate prevented analysis of initial differences between consenting and nonconsenting patients. Attrition at 12 months may be an additional source of bias. We have no data about the reasons participants had for dropping out. It is equally plausible that participants dropped out because they had recovered or, conversely, were experiencing more problems.

Back to Article Outline

Conclusions 

Our findings indicate that fatigue was not a significant problem 12 months after minor trauma for nonhospitalized participants with mild head or other injury and that there were no significant differences between the study groups in fatigue severity after accounting for baseline characteristics. The finding that baseline characteristics are significant predictors of fatigue severity at 12 months suggests that it may be inadvisable to attribute persisting fatigue to MHI without careful and comprehensive consideration of other possible etiologic factors. The compensation context is additionally an important issue affecting reported recovery after MHI that must be considered. The treatment of patients who present with such problems may be best shaped by these considerations, including the use of interventions that directly address the baseline issues that affect functioning, rather than the minor trauma itself.

Supplier

Back to Article Outline

References 

  1. Emanuelson I, Andersson Holmkvist E, Bjorklund R, Stalhammar D. Quality of life and post-concussion symptoms in adults after mild traumatic brain injury: a population-based study in western Sweden. Acta Neurol Scand. 2003;108:332–338
  2. Haboubi NH, Long J, Koshy M, Ward AB. Short-term sequelae of minor head injury (6 years experience of minor head injury clinic). Disabil Rehabil. 2001;23:635–638
  3. Kraus J, Schaffer K, Ayers K, Stenehjem J, Shen H, Afifi AA. Physical complaints, medical service use, and social and employment changes following mild traumatic brain injury: a 6-month longitudinal study. J Head Trauma Rehabil. 2005;20:239–256
  4. Lundin A, de Boussard C, Edman G, Borg J. Symptoms and disability until 3 months after mild TBI. Brain Inj. 2006;20:799–806
  5. Paniak C, Reynolds S, Phillips K, Toller-Lobe G, Melnyk A, Nagy J. Patient complaints within 1 month of mild traumatic brain injury: a controlled study. Arch Clin Neuropsychol. 2002;17:319–334
  6. Smith-Seemiller L, Fow NR, Kant R, Franzen MD. Presence of post-concussion syndrome symptoms in patients with chronic pain vs mild traumatic brain injury. Brain Inj. 2003;17:199–206
  7. Andersson AL, Bunketorp O, Allebeck P. High rates of psychosocial complications after road traffic injuries. Injury. 1997;28:539–543
  8. van der Sluis CK, Eisma WH, Groothoff JW, ten Duis HJ. Long-term physical, psychological and social consequences of a fracture of the ankle. Injury. 1998;29:277–280
  9. Stulemeijer M, van der Werf S, Bleijenberg G, Biert J, Brauer JP. Recovery from mild traumatic brain injury: a focus on fatigue. J Neurol. 2006;253:1041–1047
  10. Sundström A, Nilsson LG, Cruts M, Adolfsson R, Van Broeckhoven C, Nyberg L. Fatigue before and after mild traumatic brain injury: pre-post-injury comparisons in relation to apolipoprotein E. Brain Inj. 2007;21:1049–1054
  11. Oullet MC, Morin CM. Fatigue following traumatic brain injury: frequency, characteristics, and associated factors. Rehabil Psychol. 2006;51:140–149
  12. Borgaro SR, Baker J, Wethe JV, Prigatano GP, Kwasnica C. Subjective reports of fatigue during early recovery from traumatic brain injury. J Head Trauma Rehabil. 2005;20:416–425
  13. Cantor JB, Ashman T, Gordon W, et al. Fatigue after traumatic brain injury and its impact on participation and quality of life. J Head Trauma Rehabil. 2008;23:41–51
  14. Ziino C, Ponsford J. Measurement and prediction of subjective fatigue following traumatic brain injury. J Int Neuropsychol Soc. 2005;11:416–425
  15. Ponsford J, Willmott C, Rothwell A, et al. Factors influencing outcome following mild traumatic brain injury in adults. J Int Neuropsychol Soc. 2000;6:568–579
  16. Bernstein DM. Recovery from mild head injury. Brain Inj. 1999;13:151–172
  17. Carroll LJ, Cassidy JD, Peloso PM, et al. Prognosis for mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med. 2004;36(Suppl 43):84–105
  18. Petchprapai N, Winkelman C. Mild traumatic brain injury: determinants and subsequent quality of life: a review of the literature. J Neurosci Nurs. 2007;39:260–272
  19. Bushnik T, Englander J, Wright J. The experience of fatigue in the first 2 years after moderate-to-severe traumatic brain injury: a preliminary report. J Head Trauma Rehabil. 2008;23:17–24
  20. Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Castro CA. Mild traumatic brain injury in U.S. soldiers returning from Iraq. N Engl J Med. 2008;358:453–463
  21. Frénisy MC, Bénony H, Chahraoui K, et al. Brain injured patients versus multiple trauma patients: some neurobehavioral and psychopathological aspects. J Trauma. 2006;60:1018–1026
  22. Brown M, Gordon W, Spielman L. Participation in social and recreational activity in the community by individuals with traumatic brain injury. Rehabil Psychol. 2003;48:266–274
  23. Dumont C, Gervais M, Fougeyrollas P, Bertrand R. Toward an explanatory model of social participation for adults with traumatic brain injury. J Head Trauma Rehabil. 2004;19:431–444
  24. Dikmen SS, Machamer J, Temkin NR. Mild head injury: facts and artifacts. J Clin Exp Neuropsychol. 2001;23:729–738
  25. Binder LM. A review of mild head trauma, part II: clinical implications. J Clin Exp Neuropsychol. 1997;19:432–457
  26. Bohnen N, Twijnstra A, Jolles J. Post-traumatic and emotional symptoms in different subgroups of patients with mild head injury. Brain Inj. 1992;6:481–487
  27. Harvey AG, Bryant RA. Predictors of acute stress following mild traumatic brain injury. Brain Inj. 1998;12:147–154
  28. Meares S, Shores EA, Taylor AJ, et al. Mild traumatic brain injury does not predict acute postconcussion syndrome. J Neurol Neurosurg Psychiatry. 2008;79:300–306
  29. Mooney G, Speed J, Sheppard S. Factors related to recovery after mild traumatic brain injury. Brain Inj. 2005;19:975–987
  30. Thornhill S, Teasdale GM, Murray GD, McEwen J, Roy CW, Penny KI. Disability in young people and adults one year after head injury: prospective cohort study. BMJ. 2000;320:1631–1635
  31. Wagner AK, Hammond FM, Sasser HC, Wiercisiewski D. Return to productive activity after traumatic brain injury: relationship with measures of disability, handicap, and community integration. Arch Phys Med Rehabil. 2002;83:107–114
  32. Kay T, Newman B, Cavallo M, Ezrachi O, Resnick M. Toward a neuropsychological model of functional disability after mild traumatic brain injury. Neuropsychology. 1992;6:371–384
  33. McLean SA, Kirsch NL, Tan-Schriner CU, et al. Health status not head injury predicts concussion symptoms after minor injury. Am J Emerg Med. 2009;27:182–190
  34. American Congress of Rehabilitation Medicine. Definition of mild traumatic brain injury. J Head Trauma Rehabil. 1993;8:86–87
  35. National Center for Injury Prevention and Control. Report to Congress on mild traumatic brain injury in the United States: steps to prevent a serious public health problem. Atlanta: Centers for Disease Control and Prevention; 2003;
  36. Dikmen SS, Machamer J, Winn HR, Temkin NR. Neuropsychological outcome at 1-year post head injury. Neuropsychology. 1995;9:80–90
  37. Rohling ML, Meyers JE, Millis SR. Neuropsychological impairment following traumatic brain injury: a dose-response analysis. Clin Neuropsychol. 2003;17:289–302
  38. Kashluba S, Hanks RA, Casey JE, Millis SR. Neuropsychologic and functional outcome after complicated mild traumatic brain injury. Arch Phys Med Rehabil. 2008;89:904–911
  39. Bazarian JJ, Wong T, Harris M, Leahey N, Mookerjee S, Dombovy M. Epidemiology and predictors of post-concussive syndrome after minor head injury in an emergency population. Brain Inj. 1999;13:173–189
  40. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–198
  41. Levin HS, O'Donnell VM, Grossman RG. The Galveston Orientation and Amnesia Test: a practical scale to assess cognition after head injury. J Nerv Ment Dis. 1979;167:675–684
  42. Baker SP, O'Neill B, Haddon WJ, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14:187–196
  43. Ware JD, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): conceptual framework and item selection. Med Care. 1992;30:473–483
  44. Ware JE, Kosinski M, Dewey JE, Lincoln RI. How to score Version Two of the SF-36 Health Survey. Lincoln: Quality Metric Inc; 2000;
  45. Ware J. SF-36 Health Survey update. Spine. 2000;25:3130–3139
  46. O'Connor PJ. Evaluation of four highly cited energy and fatigue mood measures. J Psychosom Res. 2004;57:435–441
  47. Paniak C, Phillips K, Toller-Lobe G, Durand A, Nagy J. Sensitivity of three recent questionnaires to mild traumatic brain injury-related effects. J Head Trauma Rehabil. 1999;14:211–219
  48. Findler M, Cantor J, Haddad L, Gordon W, Ashman T. The reliability and validity of the SF-36 health survey questionnaire for use with individuals with traumatic brain injury. Brain Inj. 2001;15:715–723
  49. Carroll LJ, Cassidy JD, Holm L, Kraus J, Coronado VG. Methodological issues and research recommendations for mild traumatic brain injury: the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med. 2004;43(Suppl):113–125
  50. Binder LM, Rohling ML. Money matters: a meta-analytic review of the effects of financial incentives on recovery after closed head injury. Am J Psychiatry. 1996;153:7–10
  51. Paniak C, Toller-Lobe G, Durand A, Nagy J. A randomized trial of two treatments for mild traumatic brain injury. Brain Inj. 1998;12:1011–1023
  52. Mittenberg W, DiGiulio DV, Perrin S, Bass AE. Symptoms following mild head injury: expectation as aetiology. J Neurol Neurosurg Psychiatry. 1992;55:200–204
  • a SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.

 Supported by the Centers for Disease Control (grant no. R49/CCR523223-01); the National Institutes of Health; the National Institute of Child Health and Human Development; and the National Center for Medical Rehabilitation Research (grant no. 5-T32-HD007422-17).

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.

 Reprints are not available from the author.

PII: S0003-9993(09)00159-2

doi:10.1016/j.apmr.2008.12.016

Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 6 , Pages 956-965, June 2009

Article Tools

* Image Usage & Resolution