Mediterr J Rheumatol 2016;27(4):179-90
Efficacy of ultrasound-guided versus landmark-guided injections in rheumatology: A systematic review
Authors Information
1: Department of Rheumatology, The Dudley Group of Hospitals NHS Foundation Trust, Russells Hall Hospital, Dudley, West Midlands, United Kingdom
2: NIHR Manchester Musculoskeletal Biomedical Research Unit, Central Manchester NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK

Background: There is good evidence that ultrasound (US) increases the accuracy of needle placement for joint aspiration and injection. However, considerable uncertainty remains as to whether US-guided injections (USGI) achieve better clinical outcome in rheumatic conditions compared with landmark-guided injections (LGI), and how any effect on outcome relates to accuracy.

Methods: We conducted a literature search using PubMed, yielding 810 references. After applying randomised controlled trials (RCTs), human studies and relevance to rheumatic conditions as eligibility criteria, abstract review produced 26 studies. Study design quality was assessed using the Jadad scale.

Results: The median sample size of the 26 studies was 58.5 (range 20-244) patients. Median Jadad score was low at 2.5 (1-5). Only 6 RCTs had an effective double blinding method, with sham-US/blindfolding of patients and a blinded outcome assessor. Sixteen of 26 (61.5%) trials showed superior outcome of USGI compared to LGI for one or more outcomes measuring pain, function or range of movement. The 3 RCTs with the lowest risk for bias failed to demonstrate clear USGI superiority for the main clinical outcomes. In 2 of the 4 studies that assessed both accuracy and efficacy, USGI did not show greater clinical efficacy overall, although they had superior accuracy and accurate injections were associated with better outcome of at least one clinical criterion.

Conclusion: The majority of trials suggest US guidance associates with superior outcome of injections. However, higher quality RCTs and trials with similar accuracy between the two techniques fail to support this. There is need for further trials with better design and adequate sample power.