Background/Purpose: Quantitative sensory testing (QST)
can be used to assess pain processing mechanisms, and
theoretically identify a subset of chronic pain patients that has central
sensitization or centralized pain.  However,
measures of “evoked pain” from experimental stimuli obtained via QST typically are
only marginally related to clinical pain reports.  This study looked the relationship
between manual algometry pressure-pain thresholds (PPT) at multiple sites and
clinical pain measures in knee osteoarthritis (KOA) patients.

Methods:  
Fifty-four subjects meeting ACR criteria for osteoarthritis of the knee  (mean age: 62.04 years; range: 37 –
83 years; 24 male, 34 female) presented for QST and a clinical pain battery
including WOMAC, PainDETECT, Michigan Body Map (MBM), and Brief Pain Inventory
(BPI).  PPT was measured three times
(Wagner FDX 25: 0.5 kg/s ramp rate, 20 inter-stimulus interval) bilaterally at
four body sites: lateral patella (knee), tibialis
anterior, trapezius and wrist.  Associations
between PPTs and clinical pain were assessed by Pearson’s product-moment correlations (r) and partial
correlations controlling for gender were assessed for significance using SPSS
22.

Results:  WOMAC was significantly correlated to left and right knee
PPT, r = -0.37 and -0.52 (p < 0.05), respectively.  WOMAC pain, stiffness and functional
limitation subscales were all significantly correlated to right knee PPT, r =
-0.32, -0.39 and -0.52 (all p < 0.05), respectively; however, only
functional limitation was significantly correlated to left knee PPT, r = -0.38,
p < 0.05.  PainDETECT was
significantly correlated to all sites (knee, trapezius, tibialis
anterior and wrist) and correlation coefficients for left and right side were r
= -0.52, -0.51, -0.45, -0.34 and -0.59, -0.59. -0.50, -0.37
(all p <0.05), respectively. 
MBM, a measure of clinical pain distribution, correlated with both knee PPT
(r = -0.33, p < 0.05) and trapezius PPT (r = -0.43, p < 0.01) on the
right side; however, the left side was not significantly correlated.  BPI severity and interference were not
correlated to PPT at any body site (all p > 0.31). All correlations remained
significant when controlling for gender.

Conclusion: These data suggest that the MBM and painDETECT assess pain quality associated with augmented
central pain processing in patients with symptomatic knee OA.  These results provide a foundation for
future assessment of correlations between evoked and clinical pain measures.  Additional studies are needed to
determine the specific sites best suited for assessment of OA pain.