Background/Purpose:
Pain sensitization is associated with pain severity in knee osteoarthritis
(OA), but its cause in humans is not well-understood. We recently demonstrated
that local inflammation at the joint could lead to pain sensitization. Whether
inflammation systemically could do the same is not known, and data from animal
models are sparse. Knee OA, considered to be a state of potentially low-grade
inflammation, is also associated with obesity, which itself can contribute to
systemic inflammation through elaboration of adipokines such as leptin and
adiponectin, which are expected to have opposite effects to one another. We
therefore evaluated whether systemically measured inflammatory cytokines and/or
adipokines may be related to pain sensitization in knee OA.

Methods: Participants
from the Multicenter Osteoarthritis (MOST) Study, a NIH-funded longitudinal
prospective cohort of persons with or at risk of knee OA, had plasma stored and
standardized pressure pain threshold (PPT) at the wrist and patellae assessed
at the 60-month visit. PPT was assessed with an algometer (1 cm² tip,
0.5 Kg/sec) as the point at which the subject felt the pressure change to
slight pain. Lower PPT indicates more sensitivity. We measured TNF-α
(Linco multiplex), leptin and total adiponectin (R&D ELISA for both) from a
random sample of participants who were free of knee replacement.  We evaluated
the sex-specific cross-sectional relation of each inflammatory marker to PPT
using linear regression (PPT was normally distributed) with GEE to adjust for
correlations between two knees within a person, adjusting for relevant
potential confounders (age, race, BMI, pain medications, depressive symptoms,
radiographic knee OA (ROA) in either knee). We repeated analyses stratified by
ROA status.

Results: There
were 750 subjects randomly selected from the 60-month MOST study visit among
those who had bilateral native knees (mean age (SD) 66.8 (7.6), mean BMI 30.3
(5.8), 58% female, 44.8% knees with radiographic knee OA). Plasma levels of TNF-α
were not associated with PPT at either site (Table). Leptin and total adiponectin
were both associated with lower PPT at the patella (more sensitized), while
there was no association at the wrist. The results were similar among those
with and without ROA.

Conclusion:
Although a general marker of systemic inflammation (TNF-α) does not
appear to be associated with pain sensitization, adipokines may be associated
with increased pain sensitization and/or sensitivity at the knee. While we
expected adiponectin to have an association opposite to that of leptin, total
adiponectin may not be an accurate reflection of adiponectin’s biologically
active form. Nonetheless, the finding of similar effects of these two
adipokines on pain sensitization requires further study.