A preliminary fMRI study of analgesic treatment in chronic back pain and knee osteoarthritis

Abstract

The effects of an analgesic treatment (lidocaine patches) on brain activity in chronic low back pain (CBP) and in knee osteoarthritis (OA) were investigated using serial fMRI (contrasting fMRI between before and after two weeks of treatment). Prior to treatment brain activity was distinct between the two groups: CBP spontaneous pain was associated mainly with activity in medial prefrontal cortex, while OA painful mechanical knee stimulation was associated with bilateral activity in the thalamus, secondary somatosensory, insular, and cingulate cortices, and unilateral activity in the putamen and amygdala. After 5% lidocaine patches were applied to the painful body part for two weeks, CBP patients exhibited a significant decrease in clinical pain measures, while in OA clinical questionnaire based outcomes showed no treatment effect but stimulus evoked pain showed a borderline decrease. The lidocaine treatment resulted in significantly decreased brain activity in both patient groups with distinct brain regions responding in each group, and sub-regions within these areas were correlated with pain ratings specifically for each group (medial prefrontal cortex in CBP and thalamus in OA). We conclude that the two chronic pain conditions involve distinct brain regions, with OA pain engaging many brain regions commonly observed in acute pain. Moreover, lidocaine patch treatment modulates distinct brain circuitry in each condition, yet in OA we observe divergent results with fMRI and with questionnaire based instruments.

Figure 1

Pain ratings and treatment effects on pain. A. An example of pain ratings in CBP and OA. CBP patients rated spontaneous fluctuations in their pain intensity in the absence of an overt experimental stimulus (top). OA patients rated the fluctuations in their pain intensity in response to a pressure stimulus (gray) applied on the affected knee. B. CBP patients exhibited a significant decrease in pain after 2 weeks of topical Lidocaine treatment as measured by VAS (p < 0.01) and average pain during the scan (p < 0.05) (right bar graphs). OA patients did not show any decrease in pain after treatment as measured by VAS (p = 0.5), and a borderline decrease in pain in response to pressure stimulation (computed as average pain during scan divided by average pressure in standardized units) (p = 0.06) (left bar graphs). Error bars are standard deviations.

Figure 2

Brain activity maps for CBP and OA. CBP and OA group average brain activity maps (fixed effects analysis) for session1 (prior to treatment), session 2 (2 weeks after application of Lidoderm) and contrast between before and after treatment (scan 1 > scan 2). Each scan map is a subtraction between the pain rating task and the control visual rating task.

Figure 3

Covariate analysis of brain activity with pain intensity in CBP and OA. A. Covariate map shows brain regions that encoded pain intensity in CBP. It identifies the rostral ACC and mPFC. Scatter plots show the relationship between pain intensity and z-values (averaged within 3 × 3 cluster around the peak) in mPFC and right thalamus (identified from OA covariate analysis) in CBP. B. Covariate map for OA. It shows many brain regions, none overlap with the region seen in CBP. Scatter plots are in OA for the same regions shown in A.