Intra-Articular Corticosteroid Injections for Symptomatic Knee Osteoarthritis

What the Orthopaedic Provider Needs to Know

Cody L. Martin, MD; James A. Browne, MD


J Am Acad Orthop Surg. 2019;27(17):e758-e766. 

In This Article

Location for Injection and Likelihood of Injecting Within the Joint

Multiple approaches exist for an intra-articular knee injection.[22] Common injection locations include the standard anterolateral and anteromedial arthroscopic portal sites (Figure 1). In addition, there are peripatellar sites, including the medial or lateral mid patella and medial or lateral superior patella (Figure 2).

Figure 1.

Photograph showing AL and AM injection sites used with the knee in flexion. AL = anterolateral, AM = anteromedial

Figure 2.

Photograph showing SL, SM, LMP, and MMP injection sites used with the knee in extension. LMP = lateral mid-patellar, MMP = medial mid-patellar, SL = superolateral, SM = superomedial

No benchmark location exists for injection, and various success rates of actual intra-articular injection versus extra-articular injection have been published on the different locations varying from 55% to 93% success.[22,23] The peripatellar injection sites are performed with the knee extended, and their drawback is potential injury to the patellar cartilage, especially in the mid-patellar sites because the needle must pass under the patella to access the knee joint. The superior patellar sites have the advantage of the suprapatellar pouch, and the needle does not have to pass under the patella. The anterior injection sites are performed with the knee bent, so the patient can be sitting. Choosing the injection site is up to the provider, but patients can be offered the option of sitting or lying, and anterior or peripatellar injection sites can be used respectively depending on the patient's choice.

Methods for increasing the accuracy of intra-articular knee injections include aspirating fluid before injecting or using imaging such as ultrasonography or fluoroscopy. Aspiration before injection is the cheapest and quickest technique to confirm that the needle is intra-articular; however, if the knee does not have a substantial effusion, this technique may be ineffective and unreliable. Recently, Hussein[24] described a technique improving successful intra-articular knee injections even in the absence of effusion and without the need of imaging. The technique involves using an anterolateral portal with the knee in full flexion and the needle being directed toward the lateral compartment instead of the femoral notch. With femoral roll back occurring in high flexion and the tibia internally rotating, the lateral compartment volume is larger and allows for a 97.1% accuracy (P = 0.000).[24]

Of the imaging modalities used for intra-articular knee injections, ultrasonography guidance has the benefit of using nonionizing radiation, costing relatively little, and being easy to use in an outpatient setting. A review of ultrasonography-guided knee injection studies has shown an improved accuracy with ultrasonography use compared with tactile, nonimaged injections (95.8% versus 77.8%, respectively; P < 0.001).[25] The improved intra-articular accuracy may result in improved patient outcomes. Sibbitt et al[26] performed a randomized controlled trial of 94 osteoarthritic knees without effusion comparing intra-articular corticosteroid treatment with either ultrasonography guidance or anatomic guidance and found that ultrasonography guidance resulted in 48% less needle introduction pain (P = 0.0003), 36% increase in therapeutic duration (P = 0.01), and 42% less knee pain at 2 weeks (P = 0.025) but no difference at 6 months (P = 1.0).

Although accuracy in attaining an intra-articular corticosteroid injection for the management of symptomatic osteoarthritis inherently seems important, some data challenge the idea. Hirsch et al[27] performed an observational study comparing the accuracy of intra-articular steroid injections with short-term pain relief. They found that approximately 60% of patients at 3 weeks and 45% of patients at 9 weeks had a 40% or greater reduction in Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain subscale scores from baseline. However, an accurately placed intra-articular corticosteroid injection versus extra-articular had no association with the reduction in pain scores (P = 0.389 at 3 weeks and P = 0.365 at 9 weeks).[27]