Traditionally, glucocorticoid injection for the treatment of shoulder pain has been performed guided by anatomical landmarks alone. With the advent of readily available imaging tools such as ultrasound, image-guided injections have increasingly become accepted into routine care. While there is some evidence that the use of imaging improves accuracy, it is unclear from current evidence whether or not it improves patient-relevant outcomes. The aim of this review was to assess whether image-guided glucocorticoid injections improve patient-relevant outcomes compared to landmark-guided or systemic intramuscular injections in patients with shoulder pain. We searched the Cochrane Central Register of Controlled Trials (CENTRAL, via The Cochrane Library), MEDLINE (Ovid), and EMBASE (Ovid) to June 2011. We also searched the World Health Organisation International Clinical Trials Registry Platform (http://www.who.int/trialsearch/Default.aspx) to identify ongoing trials and screened reference lists of retrieved review articles and trials to identify potentially relevant studies. We included randomised controlled trials (RCTs) and quasi-randomised controlled clinical trials that compared image-guided glucocorticoid injection to landmark-guided or systemic intramuscular injection. Outcomes of interest included pain, function, range of motion, proportion of participants with overall improvement and adverse events. There were no restrictions on language or date of publication. Two review authors independently selected the studies for inclusion, extracted the data and performed a risk of bias assessment. Disagreement about inclusion or exclusion of individual studies and risk of bias was resolved by a third review author. Five studies (290 participants) were included in the review. The image-guided groups in all trials used ultrasound to guide injection. Four studies included participants with rotator cuff disease; in three the comparator was local landmarks to direct injection into the subacromial bursa and in the fourth the comparator was systemic intramuscular injection into the upper gluteal muscles in the buttock region. One study included participants with adhesive capsulitis and injection was directed into the glenohumeral joint by either ultrasound or anatomical landmark guidance.No significant differences between groups were observed with respect to reduction in pain at one to two weeks (two trials, 146 participants, standardized mean difference (SMD) -1.44, 95 CI -4.14 to 1.26), or function at one to two weeks (two trials, 146 participants, SMD 0.95, 95 confidence interval (CI) -1.29 to 3.20; back-translated to mean difference (MD) 4 points, 95 CI -5 to 13, on a 0 to 100 point scale, higher score means better function) or six weeks (three trials, 207 participants, SMD 0.63, 95 CI -0.06 to 1.33; back-translated to MD -3 points, 95 CI -11 to 5, on a 0 to 100 point scale) and the sensitivity analyses did not alter these results. While there was a significant difference between groups with respect to reduction in pain at six weeks favouring image guidance (three trials, 207 participants, SMD -0.80, 95 CI -1.46 to -0.14), there was considerable statistical heterogeneity and after removing trials with inadequate allocation concealment and inadequate blinding in a sensitivity analysis, the difference was no longer significant (one trial, 106 participants, MD -0.60 points, 95 CI -1.44 to 0.24 points on a 9-point scale).No statistical difference in adverse events between groups was identified (10/104 image-guided group versus 16/103 comparator; risk ratio (RR) 0.55, 95 CI 0.17 to 1.85). Minor adverse events reported included transient post-injection pain, facial redness and warmth.