This is the first survey to explore the use and practice of intra-articular corticosteroid injections for symptomatic first MTPJ OA by UK health professionals. Intra-articular steroids are administered into the first MTPJ by a range of professionals across multiple NHS settings. We found large variation in practice across all stages of the procedure, including use of different drugs, doses, equipment, use of imaging, and follow-up care.

Treatment of first MPTJ OA is primarily driven by evidence from other joints which informs management guidelines. As a result, clinical practice varies, and ranges from advice, information and support, use of insoles, injections, through to surgery [17]. A 2022 systematic review of international clinical guidelines found considerable variation around their endorsement of intra-articular corticosteroids [25]. NICE clinical guidelines emphasise non-pharmacological interventions in the management of OA [26]. Recent changes now recommend that intra-articular corticosteroid injections are considered as a short-term adjunct to support therapeutic exercise or when other pharmacological treatments have failed. NICE also acknowledge the lack of evidence for corticosteroid injections for joints other than the knee [18]. There is a lack of evidence on the effectiveness of corticosteroid injections for first MTPJ OA.

Joint injections can be undertaken using anatomical (palpation) guidance or image-guidance. Our data show that use of imaging guidance is common when injecting the first MTPJ, and this included both ultrasound and x-ray/fluoroscopy. Ultrasound can be used in outpatient clinics, but x-ray/fluoroscopy is typically used in theatres so is likely to have considerable resource implications and could require patients to incur a period of time on waiting lists. In order to justify this additional cost and potential treatment delay, such imaging would need to improve clinical outcomes.

Inaccurate placement of joint injections has long been recognised in the literature with concerns that extra-articular placement may contribute to local tissue damage (soft tissue and fat atrophy) [27]. Whilst there is little evidence from the foot and ankle, image guidance has been shown to improve accuracy of injection placement in other joints but it is less clear whether it improves clinical outcome [28].

Within the first MTPJ, Heidari et al. (2013) [29] reported low rates of unintentional periarticular injection (i.e., missing the joint capsule) when using anatomical guidance to inject methylene blue into 106 cadaveric first MTPJs (10/106 joints; 9%). More recently, Reilly et al. (2022) [30] questioned the accuracy of this technique when injecting radio-opaque contrast in a cadaveric study (n = 6 feet). However, Razavi et al. (2021) [31] found no clinical benefit when using ultrasound guidance in a small trial of 50 people with first MTPJ OA randomised to landmark guidance or ultrasound guidance.

Similarly, Ekeberg et al. (2009) [32] found no difference in clinical outcomes when the same dose of triamcinolone acetonide (20 mg) was administered using either ultrasound guidance into subacromial bursa or an intramuscular injection into the gluteal region of people with rotator cuff disease. This suggests that accurate placement may not be important for symptom relief and raises questions on whether the additional cost of imaging is warranted. Our survey found that almost half of respondents used imaging guidance when injecting the first MTPJ with ultrasound and x-ray being the most common modalities. The use of imaging warrants further investigation.

Reilly (2021) recently highlighted the lack of a standardised protocol for injecting the first MTPJ, and proposed a useful framework for palpation-guided injections [33]. This protocol recommended using 23/25 gauge needles for injecting steroids and these were the most common sizes reported in our survey. Although a dose of 20 mg triamcinolone acetonide was recommended in the Reilly protocol, this contrasts with our findings, where more than twice as many respondents used methylprednisolone acetate compared to triamcinolone acetonide. Triamcinolone acetonide use was more common amongst medical prescribers than non-medical prescribers, and the most common dose was 40 mg.

Another key finding from our survey was the large variation in dose of each steroid injected. Generally, we found medical prescribers injected higher doses of steroid than non-medical prescribers but the variation between and within groups highlights the lack of evidence upon which to base clinical decisions. Methylprednisolone acetate was the most common steroid used, and the dose suggested in the Summary of Product Characteristics is 4-10 mg for small joints such as the metacarpophalangeal joints [34]. Although most respondents (36/42 86%) using the premixed solution of methylprednisolone acetate and lidocaine hydrochloride, (all non-medical prescribers), adhered to this suggested dose, only one respondent (also a non-medical prescriber) used this dose when using separate non-mixed steroid. Exploring justification for clinical practice was beyond the scope of this survey, current legislation prevents non-medical prescribers from combining drugs within a syringe prior to administration unless they have accreditation as an independent non-medical prescriber.

The most common dose of 40 mg methylprednisolone acetate is equivalent to 1ml, and most respondents administered 1ml or less of steroid. However, MRI data suggests the volume of synovial fluid within a healthy first MTPJ is much smaller than this, with a median (IQR) 0.15ml (0.073 to 0.21) [35]. Therefore, as well as the anti-inflammatory action of the steroid, the introduction of relatively large volumes of injectate will have a mechanical effect, potentially distending the joint capsule and distracting the joint. Which of these actions is more important in relieving symptoms may be worthy of further investigation.

A quarter of our survey respondents offered an open appointment for post-injection review or did not typically offer review, although one third failed to report their practice. Post injection review is essential to for monitoring of treatment response, which can guide future management, and identify any side effects which may require treatment. Effects of corticosteroids are widespread and although injection-related complications are rare, these can include post-injection flares, facial flushing, tendon and ligament rupture, subcutaneous fat atrophy, glucose tolerance impairment, osteonecrosis, osteoporosis, menstrual cycle irregularities, and skin pigmentation changes [36]. Few studies have reported adverse events after first MTPJ injections but adverse events may be acute or chronic with delayed onset, thus true incidence of complications is challenging to monitor [36]. There is uncertainty regarding the optimal method, timing, and clinical value of post-injection review.

In 2011, the American Orthopaedic Foot and Ankle Society (AOFAS) surveyed 197 of 870 registered members (23%) about their use of corticosteroid injections for a wider range of clinical conditions, but provided few data specific to first MTPJ OA [37]. With an average of 4.1 steroid injections per year, this suggests a much lower use of corticosteroid injections than our sample, but they also found methylprednisolone acetate and lidocaine hydrochloride were the most common injectate. Data from our survey is comparable to a survey of current practice in the care of carpometacarpal OA across 32 UK centres, showing a lack of clear guidance on the use of intraarticular steroid injection and uncertainty about their clinical effectiveness [38]. Injections were also administered into the carpometacarpal joint by a range of health professionals, using a mixture of anatomical and image guidance, with no standardisation in the threshold for injection. Similarly, no centre offered an injection at the first appointment, but the most commonly administered steroid and local anaesthetic was triamcinolone acetonide and lidocaine hydrochloride.

There are limitations to this survey, and perhaps most notable is the lack of sampling frame, thus it is unclear how representative these data are of UK practice. It is possible that orthopaedic surgeons are under-represented in our sample and we note that our respondents did not include radiologists. Another limitation is that due to the anonymous nature of the survey, we were unable to clarify extreme or unusual responses (e.g., use of Tc99m bone scan to guide injection placement). We did not explore respondent characteristics (gender, age, ethnicity, years of experience, qualifications), care pathways, nor use of concomitant therapies such as joint protection, exercise therapy, and manipulation under anaesthetic. However, this must be balanced against survey burden; we aimed to maximise response rates and minimise data missingness. Additional strengths include the piloting of our survey prior to use, reporting our findings in line with current recommendations, and wide distribution to encourage a spread of responses from different professionals across a range of healthcare settings.