Introduction Patellofemoral pain (PFP) is common and impacts health-related quality of life. Despite numerous published guidelines, understanding of intervention efficacy is limited.

Materials and Methods We registered with PROSPERO (CRD42019152252) and searched multiple databases to April 2021 and included only high-quality randomised controlled trials (RCTs) scoring >7 on the PEDro scale. We pooled methodologically homogenous pain (e.g., numerical rating scale) and function (e.g., kujala scale) data using random effects models at short-, medium- and long-term (<3, >3-<12,>12 months respectively). Interventions demonstrated primary efficacy if there was a significant comparison with sham, placebo, or wait-and-see control in an adequately powered RCT (n>23) or data pooling. Secondary efficacy or superiority was determined by a significant comparison to an intervention with primary efficacy. The GRADE criteria determined evidence certainty (very low to high).

Results We identified 61 high-quality RCTs involving 3,543 participants. Four interventions demonstrated primary efficacy for short-term pain and function: knee-targeted exercise (high/moderate), multi-modal physiotherapy (low/low), foot orthoses (low/absent), and lower-quadrant manual therapy (absent/moderate). Hip- and knee-targeted exercise demonstrated secondary efficacy for pain and function in the short (low/moderate), medium (moderate/moderate), and long-term (moderate/moderate). Knee-targeted exercise combined with perineural dextrose injection demonstrated secondary efficacy in the short-term (moderate/moderate). Multi-modal physiotherapy demonstrated superiority for pain and function in the short-term (very low/very low) compared to knee-targeted exercise alone.

Conclusions Wait-and-see, sham, or placebo should not be used for PFP, as several interventions have proven efficacy. Future effectiveness studies should include long-term follow-up. Our synthesis will inform a future updated best practice guide.