Forty-one studies (2,281 participants) were included in the review: 23 case-control studies, 15 case only studies, two RCTs and one uncontrolled trial. Full quality assessment results were not reported.
Thirteen of the 41 included studies reported that people with osteoarthritis were more sensitive than normal controls to painful stimuli (eight studies reported pressure pain threshold, two studies reported punctuate pain threshold, one study reported mechanical and thermal pain thresholds, one study reported thermal pain thresholds and one study reported chemical pain rating). Two further studies applied electrocutaneous stimuli and reported that the threshold to elicit flexor withdrawal reflex was significantly lower in the osteoarthritis groups than in healthy controls. Three studies reported no significant difference between the osteoarthritis and control groups in piston pressure pain ratings, finger pressure pain ratings and heat and cold pain. Sensory detection thresholds in people with osteoarthritis were either higher (two studies) or similar (three studies) to those of healthy controls.
Pooled SMD for pressure pain threshold calculated by selecting the anatomical test site with the smallest SMD from each study in people with osteoarthritic compared with healthy controls was -0.87 (95% CI -1.08 to -0.66; seven studies; Ι²=5%). SMDs ranged from -0.47 (95% CI -1.00 to 0.06) to -3.04 (95% CI -3.77 to -2.31) depending on the anatomical site tested.
Where the smallest SMD for the affected joint, distal and the remote anatomical test sites were selected from each study and pooled, the SMD was larger for the affected joint sites (-1.24, 95%CI -1.54 to -0.93) than for remote sites (-0.88, 95%CI -1.11 to -0.65).
There was no evidence of publication bias.
Data on test-retest reliability from three studies were reported.