Article Text

Ultrasound-guided hip joint injections are more accurate than landmark-guided injections: a systematic review and meta-analysis
  1. Shane Hoeber1,
  2. Abdel-Rahman Aly1,
  3. Nigel Ashworth1,
  4. Sathish Rajasekaran1,2
  1. 1Division of Physical Medicine and Rehabilitation, University of Alberta, Edmonton, Alberta, Canada
  2. 2Department of Orthopaedics and Rehabilitation, University of Iowa Sports Medicine, Iowa City, Iowa, USA
  1. Correspondence to Dr Sathish Rajasekaran, Department of Orthopaedics and Rehabilitation, University of Iowa Sports Medicine, 2701 Prairie Meadow Drive, Iowa City, IA 52242, USA; sathish.k.rajasekaran{at}


Aim To compare the accuracy of ultrasound (US)-guided versus landmark-guided hip joint injections.

Methods PubMed, Medline and Cochrane libraries were searched up to 31 July 2014. Two independent authors selected studies assessing accuracy of intra-articular hip injections based on predetermined inclusion and exclusion criteria. Selected papers were then evaluated for quality and a meta-analysis of accuracy was performed using random effects models.

Results 4 US-guided (136 hip injections) and 5 landmark-guided (295 hip injections) studies were reviewed. The weighted means for US-guided and landmark-guided hip injection accuracies were 100% (95% CI 98% to 100%) and 72% (95% CI 56% to 85%), respectively. US-guided hip injection accuracy was significantly higher than landmark-guided accuracy (p<0.0001).

Summary This is the first systematic review and meta-analysis of the accuracy of US-guided versus landmark-guided hip joint injections that has revealed that US-guided injections are significantly more accurate than those that are landmark guided. Future studies should compare US with fluoroscopic-guided hip joint injections for accuracy, efficacy, safety profile, cost-effectiveness and patient satisfaction.

  • Ultrasound
  • Hip

Statistics from


Intra-articular hip injections offer a means to provide targeted, local therapy in patients suffering from various hip pathologies. Accurate intra-articular placement of the needle is thought to increase efficacy and improve patient safety, and has become even more important with the widespread use of hyaluronic acid and orthobiologicals, as they lack the theoretical systemic effect of corticosteroids.1 Landmark and image guidance are used by clinicians to perform intra-articular hip injections, with the latter likely being more prevalent. Fluoroscopic guidance was the mainstay for image guidance, but ultrasound (US) guidance has become increasingly prevalent in clinical practice due to its similar accuracy to fluoroscopic-guided injections, with less associated cost, and no radiation exposure.2–5 Although the use of image guidance has been advocated for most peripheral joint injections, debate still exists as to its superiority in regard to accuracy and efficacy over landmark guidance.2

Although a recent position statement reviewed the literature for the accuracy of landmark-guided and US-guided hip injections, no paper has systematically reviewed the literature and performed a meta-analysis of the available literature.6 As no current studies have directly compared landmark to US guidance, the aim of this study was to identify studies that assessed accuracy of either technique using an accepted gold-standard (fluoroscopic contrast, CT contrast, MRI arthrogram, or direct visualisation of injectate during surgery). We systematically compared studies of the accuracy of landmark-guided and US-guided intra-articular hip injections, and performed a meta-analysis of the available data.

Material and methods

Search strategy

The search strategy and the systematic review protocol were performed in compliance with the PRISMA statement.7 The electronic databases of PubMed, Medline and Cochrane were searched up to 31 July 2014. The MeSH terms used were: hip AND injection. Identified systematic review bibliographies were also searched for relevant articles.2 ,8

Study selection and critical appraisal

Two authors independently assessed the titles and abstracts from the combined electronic search database for eligibility (box 1). Disagreements between the two authors were resolved through a discussion with the third author, and all three authors agreed on the included studies. The studies were evaluated for bias risk using the Cochrane Collaboration tool for assessing risk of bias. Authors expanded on the ‘other bias’ section of this tool and evaluated studies for additional limitations related to participants and selection, confounders and data analysis. Each trial was scored on 15 items, with a three-point scale ranging from 0 to 2 (0=high risk; 1=unclear risk and 2=low risk), and a total score of 0–30 (lowest–highest quality study).1 The trials were then graded according to score: low quality (<18), moderate quality (18–24) and high quality (>24).1

Box 1

Eligibility criteria

  • A. Inclusion criteria

    • 1. English language papers

    • 2. Randomised, quasi-randomised and/or prospective clinical trials

    • 3. Papers that assess accuracy of ultrasound-guided or landmark-guided intra-articular injections of the hip

    • 4. Live human subjects

  • B. Exclusion criteria

    • 1. Incomplete or pilot studies

    • 2. Animal or cadaveric subjects

    • 3. Retrospective studies

    • 4. Papers without any data analysis

    • 5. Non-English language papers

Measurement of accuracy

Random effects models were used to analyse the accuracies between the techniques due to the heterogeneity in the methodology of the included studies. The statistic Q, its two-sided p value and I2 were reported to support this decision by creating a random effects model using all studies without categorical variable grouping. Proportions were transformed using the Freeman-Tukey double arcsine transformation.9 Using a single mixed-effects model applying a dummy coded factor, the categorical moderator variables ‘landmark-guidance’ and ‘US-guidance’ were used to group the studies and compare results using the Q value and its associated two-sided p value. The inverse of the Freeman-Tukey double arcsine transformation was used to calculate CIs for each categorical variable, and the harmonic mean of the proportions was used with this transformation to calculate the weighted mean for each categorical variable.10 Statistical significance was set at p<0.05. All statistical analysis and forest plots were created using R 2.15 (R Foundation for Statistical Computing) Metafor package.


Search results

Fifteen articles were assessed for eligibility, and subsequently nine studies (431 injections performed in 349 patients) were included in the systematic review (figure 1).11–19 There was total agreement (κ=1) between the two authors who selected and evaluated the clinical trials. Four studies (136 injections on 111 patients) assessing the accuracy of US-guided and five studies (295 injections on 238 patients) assessing the accuracy of landmark-guided injections were reviewed (table 1).

Table 1

Summary table of selected studies

Critical appraisal

Bias risk assessment

(1–3) All studies except Kantarci et al's paper were single-arm investigations with no randomisation, concealed allocation, or blinding (table 2(i)).11 ,12 ,14–19 Kantarci et al13 randomised patients into two arms (comparing two different US-guided approaches) and adequately blinded participants, study personnel and outcome assessors, and also concealed allocation. (4) One of the eight single-arm studies had an independent investigator who did not perform the injection accuracy assessment.18 (5) One study had four participant drop-outs (1, needle phobia; 3, MRI claustrophobia); these four patients were eliminated from the analysis.13 (6) All studies reported the accuracy of the injections, and none had selective reporting.

Table 2

Critical appraisal of selected studies

Limitations related to patient selection

(1) A small sample size was noted in one study (14 patients) (table 2(ii)).11 The remainder of the studies had unclear risk, as no formal power analysis was performed.12–19 (2) Three of the papers failed to mention eligibility criteria.11 ,14 ,16 (3) Three studies did not list appropriate patient demographic information.11 ,16 ,19 (4) One study noted standardised preimaging.19

Limitations related to confounders

(1) Four of the papers specifically stated the number of years of experience the individual performing the intervention had (table 2 (iii)).12 ,13 ,16 ,18 One study did not state how many individuals performed injections, or their experience levels.17 This study grouped injectors as ‘consultants’ and ‘trainees’.17 (2) All papers described the approach/technique used to inject patients.11–19 (3) In two of the landmark-guided studies, patients were put under general anaesthesia prior to the injection.14 ,17 Kurup and Ward14 noted a portion of patients underwent general anaesthesia, but failed to mention the number who did, or the reason(s) for this. (4) Five studies assessed the flow pattern of the contrast agent under fluoroscopy to assess accuracy.11 ,12 ,14 ,17 ,18 One study deemed the injection to be accurate if the needle was in the appropriate spot and the air arthrogram outlined the joint space under fluoroscopy.15 One study assessed the flow pattern of the contrast agent under CT to assess accuracy.16 One study assessed the flow pattern of the contrast agent under MRI arthrography to assess accuracy, where all patients were imaged within 10 min of the injection.13 One study used direct visualisation, where the injection was deemed accurate if the dye was only visualised after capsulotomy.19 All studies were viewed to have an acceptable methodology for accuracy (table 1).

Limitations related to data analysis

All studies outlined the methods used for statistical analysis of data (table 2(iv)).

Computed scores

Table 2 outlines the total score of each study and the score in each section of the critical appraisal. Of the reviewed US-guided studies, one was high quality, two were moderate quality and one was low quality. Of the reviewed landmark-guided studies, three were moderate quality and two were low quality.


There was significant variation in the proportion of successful joint injections with both techniques, Q(df)=93.92(8), p<0.001 and I2=90.8. The mixed-effects model estimated weighted mean for US and landmark guidance were 100% (95% CI 98% to 100%) and 72% (95% CI 56% to 85%), respectively. A forest plot (figure 2A, B) detailed the mean from each study and CI, and the weighted mean and CI for US and landmark guidance. The difference between weighted means for US and landmark guidance was significant, Q(df)=31.1(7), p<0.0001.

Figure 2

Forest plot graph for proportion of successful hip joint injections, (A) ultrasound-guided and (B) landmark-guided injections. RE, random effects.


This systematic review and meta-analysis showed a significantly higher accuracy for US-guided hip joint injections compared with those that are landmark guided. The CI for the weighted mean accuracy of the US-guided hip injections was small in comparison to the landmark-guided injections. This may be due in whole or in part to a ceiling effect, as the weighted mean was close to 100%. From a clinical perspective, the use of general anaesthesia in two of the landmark-guided hip injections was a concern.14 ,17 The reason(s) for its use were not clearly identified, but having to do so in clinical practice for landmark-guided hip injections would undoubtedly limit its use in an outpatient setting. In addition, when injecting an anaesthetised patient, the needle trajectory would not be altered by inadvertent patient movement, making the injection technically less difficult than that in an awake patient. For the same reason, cadaveric studies were excluded completely from this review (box 1).

A recent position statement by the American Medical Society for Sports Medicine systematically reviewed the literature for US-guided and landmark-guided injections of major (including hip joints), intermediate and small joints; and soft tissues.6 They categorised studies using the scale published by the Journal of Bone and Joint Surgery.20 The authors found four level 1 studies for US-guided hip injections with a mean accuracy of 99% and range of 97–100%; these four studies were also included in our review.11 ,13 ,16 ,18 The two studies they classified as level 4 and 5 were excluded from our paper.21 ,22 Their review also identified two level 2 studies for landmark-guided hip injections with a mean accuracy of 73% and a range of 67–73%.12 ,19 Along with including these two studies, we identified an additional three studies in the literature.14 ,15 ,17

Implications for clinical practice

Although most subspecialties have moved away from landmark-guided hip joint injections, it continues to be used or advocated for in clinical practice.12 ,14 ,15 ,17 ,19 Our study supports the results of Finnoff et al6 in the use of image guidance (US) over landmark guidance for hip joint injections. The need for image guidance may be of even higher importance in certain population subsets, as a handful of the landmark-guided injection studies noted a lower accuracy in patients with advanced arthritis or an elevated body mass index.14 ,15 ,17

Although fluoroscopy predates US guidance as an imaging modality for hip joint injections, the recent increase in use of US guidance may result in it supplanting fluoroscopy for hip joint injections.3 ,23 US offers several advantages over fluoroscopy, including decreased cost, increased availability, no radiation exposure, less pain and increased patient preference.4 ,5 ,24 The efficacy and safety of image guidance over landmark guidance, and US over fluoroscopy, have yet to be directly studied. Based on our findings, US and fluoroscopy appear to have similar accuracy, as fluoroscopy was used as the gold standard in two of the four US-guided hip joint injection studies, and the weighted mean accuracy for US was 100% (table 1).

Limitations to this systematic review

We report several limitations. One, we did not identify studies that directly compared both techniques. This undoubtedly decreased the quality of the meta-analysis data. Second, efficacy and safety data were not compared. Third, demographic data were not comparable among studies, as data were incomplete (table 2(ii)). Fourth, fluoroscopic guidance was not included as a third technique in this systematic review. However, this would have substantially weakened the overall review, as several of the included papers used fluoroscopy as a gold standard for accuracy confirmation (table 1).


In summary, our meta-analysis revealed US-guided hip injections are significantly more accurate than those that are landmark guided. Although none of the identified studies directly compared both techniques, our results support the use of US guidance over landmark guidance in clinical practice. As most subspecialties already use image guidance for hip joint injections, the need for well-designed studies examining the efficacy and safety profile of US guidance and landmark guidance for hip joint injections could be superfluous. As the use of US expands, it may be more advantageous to compare its accuracy, efficacy, safety profile, cost-effectiveness and patient satisfaction with fluoroscopic-guided hip joint injections.

What are the new findings

  • This systematic review and meta-analysis showed that ultrasound (US)-guided injections of the hip joint provide superior accuracy to those that are landmark guided.

  • There was insufficient evidence to comment on whether this increased accuracy would lead to increased efficacy. Given the large difference in accuracy, future studies examining this would find difficulty in obtaining ethics approval and recruiting patients.

  • With the widespread use of US guidance in sports medicine, it would be of benefit to directly compare it against fluoroscopy for hip joint injections looking at multiple outcome variables.

How might it impact on clinical practice in the near future

  • Landmark-guided hip injection should not be performed in clinical practice.

  • US-guided hip injections offer an accuracy that is likely equal to that of fluoroscopic-guided injections.

  • None of the literature identified patient factors that clinicians should be aware of when selecting patients for US-guided hip injections. There was a suggestion that elevated body mass index would make US-guided hip joint injections technically difficult.


View Abstract


  • Contributors SH performed the literature review and critical appraisal, and prepared the manuscript. A-RA designed the study, performed critical appraisal, and reviewed the manuscript. NA reviewed the study design and the manuscript. SR conceived and designed the study, reviewed search results, performed and reviewed critical appraisal, performed statistical analysis, prepared and reviewed the manuscript. All the authors approved the final version of the manuscript.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.