What is the evidence for intra-articular PRP (Platelet rich plasma) injection for knee osteoarthritis?
10 small Level I randomized control trials looking at patients diagnosed with knee osteoarthritis based and the criteria described by the American College of Rheumatology.
The sample size of the studies ranged from 21 to 183 (total of 1069 patients; 1142 knees). Average age of the patient ranged from 46.6 to 66.5.
Injections were given to patients with mild to severe knee osteoarthritis.
Platelet-rich plasma (PRP) intraarticular injections. A variety of preparation (spinning technique, speed, length of centrifugation, activator), administration (frequency and volume) and preparation techniques were used.
Eight studies compared PRP with hyaluronic acid (HA; of different types and frequency). Three studies compared PRP with saline (placebo). Review inclusion criteria also included corticosteroid, exercise or no treatment but no articles found.
PRP vs HA
- At 6 months, 3 studies (n=339) showed no significant difference in WOMAC pain or function scores between the PRP and HA groups (P=0.27 and P=0.16, respectively).
- At 12 months, 3 studies (n=302) showed PRP was significantly more efficacious in pain relief compared with HA (MD -2.83, 95% CI -4.26 to -1.39, P=0.0001) and in functional improvement (MD: -12.53, 95% CI: -14.58 to -10.47, P < .00001). Overall effect size exceeded MCID of 20%.
- Adverse events – no significant difference between PRP and HA group (RR 0.63, 95% CI 0.20-1.98, p=0.43)
PRP vs Saline
- At 6 months and 12 months, 1 study showed PRP was significantly more efficacious in pain relief (MD -5.00, 95% CI -6.98 to -3.02, P<0.00001; MD -6.00, 95% CI -8.32 to -3.68, p <0.0001) and in functional improvement (MD: -24.00, 95% CI: -31.30 to -16.70, P < .00001; MD -24.00, 95% CI -30.01 to -17.99, p <0.00001) compared with saline. Overall effect size exceeded MCID of 20%.
- Adverse events – no significant difference between PRP and saline group (RR 2.63, 95% CI 0.04 to 158.93, p 0.64).
Subgroup analysis revealed that PRP preparation with platelet >5 x baseline, leukocyte rich PRP and using an activator, was significantly associated with less pain relief (as compared to HA) at 6 months.
versus Visco (HA) injection
Pain (6 months)
Pain (12 months)
Function (6 months)
Function (12 months)
Relevant Clinical Info
Five studies had a follow-up at 12 months, four studies had a follow-up at 6 months and one study had a follow-up at 3 months. The primary outcomes were the Western Ontario and McMaster Universities Arthritis Index (WOMAC) total score, sub scores or pain and function, the International Knee Documentation Committee (IKDC) Subjective Score, Lequesne score, and adverse events.
8/10 studies included were judged to have a high risk of bias.
Interestingly, the studies control group of HA is not recommended as routine intervention for knee OA.
Majority of study conclusions are based on 1-3 studies, therefore limited power.
their were 1069 people.
the sample size was 1142 knees
Mean age of 47-67 years.
their were 10 studies used.