Targeting inflammation as a treatment modality for neuropathic pain in spinal cord injury: a randomized clinical trial

All participants from both the treatment and control group completed the entire 3-month
duration of the study and were included in the analysis. No adverse events were reported.
The participants’ overall compliance to the diet was assessed based on the average
of the three diet records during the study (1 month, 2 months, and 3 months). One
participant completed all three testing sessions but failed to produce the 2-month
and the 3-month diet record. This participant had a dietary compliance over the first
month of 92 %. All other participants completed each of the required diet records,
and overall compliance ranged from 70 to 100 %, with a mean compliance of 89 %. A
detailed analysis regarding specific diet adherence data will be presented elsewhere.

Change in pain scores

Changes in sensory, affective, and sensitivity pain scores are shown in Table 2. There was a significant group?×?time interaction for the sensory component of the
self-reported neuropathic pain scores (p??0.01; Cohen’s d?=?1.29). Post hoc analysis showed a significant reduction in sensory scores in the
treatment group from both baseline to 1 month, as well as baseline to 3 months (p?=?0.00 and p?=?0.01, respectively). This included reductions of greater than 30 % for six participants
and a reduction between 20 and 30 % for another three participants. Post hoc analysis
showed a significant increase in sensory scores in the control group from baseline
to 1 month (p?=?0.04) but no significant change from baseline to 3 months (p?=?0.210). No significant group?×?interaction was found for the affective component
of the self-reported neuropathic pain scores (p?=?0.17; Cohen’s d?=?0.63). Of note, the change in affective neuropathic pain scores were significantly
correlated to the previously reported scores related to changes in mood in this population
14]. In regard to the non-parametric analysis of sensitivity scores, the Mann-Whitney
test indicated that change scores of sensitivity pain were not significantly different
between the treatment group and control group (U?=?36.0, p?=?0.35). The Friedman test showed that there was no significant change in sensitivity
pain scores for either the treatment group (x²
?=?3.38, p?=?0.19) or control group (x²
?=?0.09, p?=?0.96) across testing sessions.

Table 2. Changes in neuropathic pain questionnaire scores

Change in inflammatory mediators

Changes in serum levels of inflammatory mediators are shown in Table 3. When considering a proinflammatory composite score (average of IL-2, IL-6, IL-1?,
TNF-?, and IFN-y), the Mann-Whitney test indicated that the change scores (3 months???baseline)
were significantly different between the treatment group and the control group (U?=?13.0, p?=?0.01). The Friedman test showed that there was a statistically significant reduction
in the proinflammatory composite scores in the treatment group (x²
?=?6.50, p?=?0.04), but no significant change in the control group (x²
?=?5.25, p?=?0.07). Post hoc analysis performed with the Wilcoxon signed-rank test showed significant
reductions in the treatment group from both baseline to 1 month and baseline to 3 months
(z?=??2.197, p?=?0.03; z?=??2.275, p?=?0.02 respectively). When analyzing each cytokine separately, the Mann-Whitney test
indicated that the change scores (3 months???baseline) were significantly different
between the treatment group and the control group for IFN-y (U?=?13.0, p?=?0.01.), IL-1? (U?=?14.0, p?=?0.01), and IL-2 (U?=?12.0, p?=?0.01) and showed a trend for CRP (U?=?27.0, p?=?0.10). The Friedman test showed that in the treatment group, there was a statistically
significant reduction in IFN-y (x²
?=?8.67, p?=?0.01), IL-1? (x²
?=?17.78, p??0.01), IL-6 (x²
?=?6.17, p??0.05), and a trend for CRP (x²
?=?4.5, p?=?0.10). The Friedman test showed no statistically significant reductions for any
inflammatory mediator in the control group. Post hoc analysis performed with the Wilcoxon
signed-rank test showed significant reductions in the treatment group for IFN-y from
baseline to 1 month and baseline to 3 months (z?=??2.275, p?=?0.02; z?=??2.510, p?=?0.01, respectively), as well as significant reductions in the treatment group for
IL-1? from baseline to 1 month and baseline to 3 months (z?=??3.059, p??0.01; z?=??2.934, p??0.01, respectively), and a significant reduction in the treatment group for IL-6
from baseline to 1 month, and a trend from baseline to 3 months (z?=??2.275, p?=?0.02; z?=??1.726, p?=?0.08, respectively). Two-way repeated measures ANOVA were performed for the normally
distributed mediator’s TNF-? and PGE2 and showed trends towards group?×?time interactions
(p?=?0.10; p?=?0.07 respectively).

Table 3. Changes in inflammatory mediators

Relationship between inflammatory mediators and indices of pain

Results from the multiple regression are shown in Table 4. To help elucidate a potential mechanism between the reduction in neuropathic pain
scores and inflammatory mediators, a stepwise backward multiple regression analysis
was performed. When assessing the change in sensory pain score as the outcome variable,
results from the regression analysis provided partial confirmation of the research
hypothesis that change in sensory neuropathic pain is a function of the change in
proinflammatory cytokines and eicosanoids. The three-variable model included the proinflammatory
cytokines IL-2 and IFN-y, as well as the eicosanoid PGE2 (R?=?0.689, R²
?=?0.474). The overall F statistic for the model was 4.811, df?=?3.16, p?=?0.01. Standardized beta weights were ?0.730 for IL-2, 0.544 for IFN-y, and 0.526
for PGE2. When assessing the change in affective pain score as the outcome variable,
only PGE2 remained in the model (R?=?0.558, R²
?=?0.312). The overall F statistic for the one-variable model was 8.145, df?=?1.18, p?=?0.01, and the standardized beta weight was 0.558. Lastly, when assessing the change
in sensitivity score as the outcome variable, a three-variable model including the
proinflammatory cytokines IL-1B and IL-2 as well as the eicosanoid PGE2 remained (R?=?0.715, R²
?=?0.511). The overall F statistic for the three-variable model was 5.580, df?=?3.16, p?=?0.008. Standardized beta weights were 0.491 for IL-1B, ?0.666 for IL-2, and 0.378
for PGE2.

Table 4. Backward stepwise multiple regression analysis of the relationship between the change
in neuropathic pain scores and the change in inflammatory mediators

In order to account for the potential influence of depression on sensory neuropathic
pain scores, an additional multiple regression analysis was performed while controlling
for CES-D scores of depression. In this analysis, model 1 included CES-D scores of
depression and model 2 included the inflammatory mediators IL-2, IFN-y, and PGE2.
Model 1 resulted in a non-significant R square change score of 0.131 (p?=?0.12) while model 2 resulted in a significant R square change score of 0.394 (p?=?0.03). This suggests that the change in inflammatory mediators IL-2, IFN-y, and
PGE2 explains an additional 39 % of the variability in the change in sensory neuropathic
pain scores when the change in CES-D scores of depression has been statistically controlled
for. When assessing standardized beta weights, CES-D scores were not significant (beta?=?0.244,
p?=?0.23), while IL-2 and PGE2 remained significant (beta?=??0.731, p?=?0.01; beta?=?0.495, p?=?0.03, respectively), and IFN-y showed a trend towards significance (beta?=?0.462,
p?=?0.07). The relationship between neuropathic pain scores and inflammatory mediators
was also assessed at baseline by means of a stepwise backward elimination multiple
regression analysis. When assessing baseline sensory neuropathic pain scores, IL-4
and PGE2 remained in the model (R?=?0.672, R²
?=?0.451). The overall F statistic for the model was 4.385, df?=?3.16, p?=?0.02. Standardized beta weights were ?0.783 and 0.472 for IL-4 and PGE2, respectively.
When assessing baseline affective neuropathic pain scores, IL-2, IFN-y, PGE2, and
IL-4 remained in the model (R?=?0.914, R²
?=?0.836). The overall F statistic for the model was 14.295, df?=?5.14, p??0.01. Standardized beta weights were ?0.580 for IL-2, 0.756 for IFN-y, 0.965 for
PGE2, and ?0.832 for IL-4. When assessing baseline sensitivity pain scores, IL-4 and
LTB4 remained in the model (R?=?0.909, R²
?=?0.826). The overall F statistic for the model was 13.313, df?=?5.14, p??0.01. Standardized beta weights were ?0.567 for IL-4 and 0.850 for LTB4.