3 AIMS OF THE STUDY
5.1 Neuroradiological Factors
Study III was designed to address significant gaps in the literature relating diffusion tensor imaging (DTI) and functional outcome following MTBI. Specifically, the aim was to examine the association between white matter integrity and functional outcome in comprehensive way (subjective post-concussion symptom reporting, mental health, return to work, cognitive outcome) following uncomplicated MTBI. Only patients with uncomplicated MTBIs who underwent 3T MRI were included in this study.
5.1.1 Exploratory Diffusion Tensor Imaging Analyses
Exploratory analyses revealed no significant differences between the uncomplicated MTBI and neuroimaging control group on 24 of 26 DTI measures. Correcting for multiple comparisons, there would be no statistically significant differences between groups in any ROI. There were significant differences for ADC in the genu of the corpus callosum (p=.022, d=.58, medium effect size) and FA in splenium of the corpus callosum (p=.027, d=.56, medium effect size). For both ROIs, there was increased ADC (genu) and increased FA (splenium) in the MTBI group compared to controls.
Increased FA in the splenium is the opposite of what is expected based on the literature.
Previous reports have suggested that high FA in subacute mTBI appears to be related to the post injury interval (Bazarian et al., 2007; Wilde et al., 2008; Henry et al., 2011).
Therefore, a scatter plot was created to visualize the association between Total number of low FA scores and time post injury (Figure 3). The scatterplot reveals no correlation between time post injury and number of low scores.
Also, we ran exploratory regression analyses to examine whether time post injury predicts FA scores. Regression analysis was run with total number of low FAs as a dependent factor and time post injury as an independent factor. Time post injury did not predict total number of low FAs (R= .141, R2=.020, p=.339).
5.1.2 Multivariate Region of Interest Analysis
To explore whether uncomplicated MTBI patients will have reduced white matter integrity in a greater number of regions of interest compared to healthy controls, multivariate ROI analysis was used. For these analyses, the 16 ROIs for FA and 10 ROIs for ADC were considered simultaneously. To examine the prevalence of low or high scores, when all ROIs were considered simultaneously, a cut-off score for each ROI was set at 1.28 SDs below or above the mean of control values. The 1.28 SDs below the mean for each FA score for each ROI was selected as a cutoff score for abnormally low FA scores (i.e., 10th percentile) and 1.28 SDs above the mean for each ADC score for each ROI was selected as a cutoff score for abnormally high ADC scores (i.e., 90th percentile). The 10th and 90th percentiles were selected because the control sample Figure 3. Total number of low FA scores and time post injury.
was relatively small and this would create more variability, and mediate the effects of possible outliers, in the control sample.
Overall, there were a greater number of low FA scores in the MTBI group compared to the control group, i.e. patients with uncomplicated MTBIs had reduced white matter integrity in a greater number of regions of interest compared to healthy controls.
Chi-square analyses revealed that there was a significantly greater number of low FA scores when using 2 or more low scores as the criterion (p=.003, 66.7% MTBI, 29.2%
controls). Similarly, there were also a greater number of high ADC scores in the MTBI group compared to the control group. Chi-square analyses revealed that there was a significantly greater number of high ADC scores when using 2 or more high scores (p=.011, 47.9% MTBI, 16.7% controls) and 3 or more high scores (p=.007, 33.3%
MTBI, 4.2% controls) as the criterion.
5.1.3 Diffusion Tensor Imaging and Clinical Outcome
To examine the relation between DTI abnormalities and clinical outcome, the MTBI sample was divided into two groups based on the presence or absence of multiple areas of abnormally low FA values or abnormally high ADC values: (a) broadly normal white matter (WM) group (n=23, 47.9%), and (b) multifocal abnormal WM group (n=25, 52.1%). The multifocal abnormal WM group was defined as follows: 4 or more areas of abnormally low FA values OR 3 or more areas of abnormally high ADC values. The broadly normal WM group was defined as follows: <4 areas of abnormally low FA values AND <3 areas of abnormally high ADC values. Based on this criterion, 52.1%
of the MTBI group and only 12.5% of the control group showed evidence of multifocal white matter findings (χ2=10.55, p=.002; OR=13.1, 95% CI=3.5–50.0).
For the demographic and injury-related variables, there were no significant differences between groups for age, education, gender, RTA, or LOC (all p > .05).
However, the broadly normal WM group had a significantly longer duration of PTA compared to the multifocal abnormal WM group (p=.034, d=.75, large effect size).
The two groups were compared on the neurocognitive measures and number of days to return to work. There were no significant differences between groups for the majority of measures, with the exception of the 15” and 30” retention interval trials on the FWSTMT. For these two measures, the multifocal abnormal WM group had higher scores (performed better) compared to the broadly normal WM group (15” retention trial, p=.035; d=.64; 30” retention trial, p=.026, d=.68). The multifocal abnormal white matter group did not take longer to return to work than the broadly normal white matter group (p=.939).
5.1.4 Diffusion Tensor Imaging and Post-concussion Symptoms
To examine the relation between self-reported post-concussion symptoms and neuropsychological and DTI measures, the MTBI sample (Study III) was divided into two groups based on International Classification of Diseases (ICD-10; World Health Organization, 1992) Category C symptom criteria for Postconcussional Syndrome (PCS): (a) PCS-Present (n=11), and (b) PCS-Absent (n=37). PCS was classified using symptoms endorsed as moderate or higher on the RPSQ.
There were no significant differences between groups on all demographic (gender, age, education) and injury related variables (LOC, PTA, RTA), and for the majority of neurocognitive measures. There were significant differences and very large effect sizes between groups on both measures of fatigue (FIS total score, p<.01; d=1.60; BNI-FS total score, p <.01; d=1.49), depression (BDI-II total score, p <.01; d=1.55), and general health (EQ-5D™ VAS score, p = .028, d = .79). For the DTI measures, there were no significant differences between the two groups for all ROIs for FA and ADC (all p>.05, see Table 12). In addition, when all ROIs were considered simultaneously, the prevalence of low FA scores or high ADC scores did not differ between groups (Table 13).
Table 12. Exploratory comparisons of apparent diffusion coefficient and fractional anisotropy of post-concussion absent (PCS -) and post-post-concussion present (PCS +) groups.
PCS – (n=37) PCS + (n=11) Cohen’s Effect Size (d)
M SD M SD p
ADC (10–3 mm2/sec)
Basilar Pons right .685 .068 .716 .069 .197 0.45
Basilar Pons left .717 .105 .730 .078 .690 0.13
Internal Capsule right .688 .036 .695 .030 .524 0.20
Internal Capsule left .677 .042 .689 .023 .225 0.14
Corona Radiata posterior right .667 .044 .652 .024 .285 0.38
Corona Radiata posterior left .646 .081 .658 .029 .655 0.17
Uncinate fasciculus right .776 .043 .776 .049 .982 0.00
Uncinate fasciculus left .780 .043 .753 .060 .092 0.58
Corpus Callosum: Genu .792 .083 .810 .091 .532 0.21
Corpus Callosum: Splenium .710 .068 .716 .043 .774 0.10
FA
Basilar Pons right .636 .075 .594 .087 .125 0.54
Basilar Pons left .626 .083 .611 .078 .589 0.18
Cerebral peduncle right .857 .052 .857 .057 .998 0.00
Cerebral peduncle left .857 .060 .846 .046 .608 0.19
Internal Capsule right .725 .045 .727 .045 .892 0.04
Internal Capsule left .724 .046 .712 .027 .425 0.29
Corona Radiata posterior right .440 .065 .457 .098 .517 0.23
Corona Radiata posterior left .517 .073 .513 .092 .902 0.05
Anterior Corona Radiata right .556 .070 .549 .082 .771 0.10
Anterior Corona Radiata left .546 .082 .548 .065 .923 0.03
Uncinate fasciculus right .545 .066 .509 .080 .140 0.52
Uncinate fasciculus left .541 .066 .506 .089 .156 0.49
forceps Minor right .543 .090 .566 .073 .430 0.27
forceps Minor left .569 .091 .561 .102 .802 0.09
Corpus Callosum: Genu .834 .058 .800 .058 .090 0.57
Corpus Callosum: Splenium .884 .039 .856 .055 .148 0.66
Note: ADC = apparent diffusion coefficient. FA = fractional anisotropy; * p < 0.05
Table 13. Cumulative frequency distribution of 16 low FA and 10 high ADC scores considered simultaneously by PCS group.
ScoresLow
fA
p High
Scores
ADC
p PCS -
(n = 37) PCS +
(n = 11) PCS -
(n = 37) PCS +
(n = 11)
f cp f cp f cp f cp
6 0 - 1 9.1 .229 6 0 - 0 -
-5 5 13.5 0 - 1.00 5 1 2.7 0 - 1.00
4 3 21.6 1 18.2 1.00 4 4 13.5 1 - 1.00
3 5 35.1 4 45.6 .248 3 9 37.8 1 4.2 .293
2 10 62.1 3 81.9 .293 2 4 48.6 3 16.7 .852
1 9 86.4 2 100.0 .576 1 9 72.9 4 66.7 .705
0 5 100.0 0 - 0 10 100 2 100
Note: f = Frequency; cp = Cumulative Percentage; p based on χ2-test
In study IV, the relation between self-reported post-concussion symptoms and DTI measures was further examined in a larger sample of patients with MTBIs (including those with complicated MTBIs) and prospectively (at one month and one year post injury). Multifocal abnormal WM was found in 12.5% of the control group (3/24) and 50.7% of the MTBI group (36/71). Patients in the MTBI group were significantly more likely to show evidence of multifocal diminished white matter than participants in the control group [χ2(1,95)=10.82, p=.001; RR=4.06, 95% CI (1.44–16.01)]. However, the presence of multifocal diminished white matter was not significantly associated with the presence or absence of ICD-10 PCS. In sum, there were no significant differences in DTI measures between those who met ICD-10 criteria for PCS and those who did not meet criteria for PCS.