Before the 1990s, the main indication for elbow prosthesis was rheumatoid arthritis [4, 15,16,17,18,19], an indication which has undergone a progressive reduction especially in the last decade [18, 20,21,22].

Young et al [20]. in 2018 studied the correlation in RA patients between the use of DMARDs (“disease modifying antirheumatic drugs”) [20, 23, 24] and the reduction in the number of patients undergoing arthroplasty surgery in any joint district; the study highlighted a real decline in the number of elbow arthroplasty implants, greater than in any other joint district (from 24% in 2002 to 12% in 2012).

The results we obtained are in line with what has been published, with a significant decrease (87.5%) in cases of RA patients treated with CM prostheses: this may be correlated to the diffusion of these drugs has revolutionized the trend and the treatment of the pathology.

There has been a notable rise (44.4%) in the utilization of CM prostheses for the management of complex acute fractures of the distal humerus [25, 26].

This finding aligns with a recent systematic review conducted by Macken et al. [27]. in 2020.

This shift is mainly attributed to the increasing surgical confidence in applying a well-established technique. Additionally, elbow surgeons are increasingly seeking simpler and earlier patient rehabilitation without compromising the extensor mechanism, resulting in more reproducible outcomes compared to synthesis surgery, which can be challenging, particularly in the elderly, osteoporotic, and less compliant patients [27, 28].

Although several studies have analyzed the results of CM prostheses [8, 15, 16, 29,30,31], there is no systematic comparison of long-term results based on individual implant indications in the literature.

Revision surgery-free survival in our study was 90% at 5years (60 months), 85% at 10years (120 months) and 83% at 15years (180months); this finding is in line with the results of other studies in the literature [31, 32].

Only recently has a comparison been initiated between the implant indication and its longevity, revealing a generally higher failure rate in the AR group compared to FX group [21, 31, 32].

This study’s analysis included a comparison of patients with significant elbow instability, osteoarthritis, or ankylosis. The FX group exhibited a lower failure rate compared to the OA group, a finding corroborated by previous studies [21, 33], as well as INS, RA, and ANK groups. A high survival of the implant can therefore be more easily achieved in the acute treatment of a complex fracture of the distal humerus rather than in the subsequent execution of a prosthetic implant on a stiff, unstable or arthritic elbow (possible complications or outcomes of a previous trauma treated in another way); the data could, however, be explained by the higher age (average of 74.4 versus 62.5years) and the lower average monitoring (average of 84.6 versus 138.7months) of FX group compared to the others.

The complications and causes leading to revision after CM prosthesis implantation, as documented in our study, align with findings from comparable analyses in terms of patient cohorts or duration of clinical monitoring. Damage to the ulnar nerve is an often underestimated eventuality (between 2 and 26%) which, although in the majority of cases presents with minor paraesthesia and rarely with an involvement of the motor component (5.7%), constitutes a relatively frequent complication that should be made explicit in the collection of informed consent [4, 34].

The occurrence of wear in the high molecular weight polyethylene (UHMWPE) component of the prosthesis bushing is rare but feasible. However, aseptic loosening remains the predominant cause of revision surgery, constituting 5.7%, consistent with the literature’s range of 5–15% [18, 32, 35].

Conversely, septic mobilization exhibits a higher incidence (3.3%) compared to other joint areas like the hip and knee (0.5-2%) [36]. This disparity may be attributed in part to a significant number of patients with rheumatoid arthritis and concurrent use of immunosuppressive therapy. It is no coincidence that the studies in the literature that estimate the highest rates of periprosthetic infection (up to 9%) [36, 37] predate the diffusion of DMARDs.

The literature is full of studies that evaluate the functional results of the CM prosthesis, but there is a lack of data in this aspect that compare the various indications for the implant.

Mansat et al. and Hildebrand et al. [29, 31]. have presented generally better results in the group of patients suffering from rheumatoid arthritis when compared with those treated for traumatic pathologies, either in acute or for chronic complications.

Our work stands out in considering separately the implants performed on FX, OA, INS, RA, and ANK: this substantial difference in method justifies the profound differences in our results.

Using the MEPS, superior outcomes were observed in the FX, INS, and OA groups compared to the RA and ANK groups, although the latter exhibited generally satisfactory evaluations. Similarly, the greatest amplitude in the range of motion was achieved by FX, INS and OA, while a more limited mobilization was detected in ANK and RA patients, leading overall to a restoration of the functional range of motion (30–130°, so as defined by Morrey) in 64.7% of patients. The brightest results in particular were obtained in the OA group (p < 0.005).

The functionality of this joint clearly depends not only on the restoration of the joint surfaces but also on the quality of the surrounding soft tissues: greater retraction and reduced elasticity of these, present in patients with greater preoperative joint stiffness such as those affected by RA and by definition in those of the ANK group, justify the poor results; on the contrary, in the presence of a pathological picture that affects the articular surfaces only (OA group) with greater selectivity, greater recovery can be achieved.

The QuickDASH assesses residual disability across the entire upper extremity. Conditions with potential multi-district involvement, such as rheumatoid arthritis and osteoarthritis, may consequently not result in complete restoration of limb function during examination. Notably, RA and OA exhibit the least favourable outcomes among the examined groups.

The pain assessment using the VAS scale led across the board to very satisfactory results: 72.5% did not report any painful symptoms and no patient complained of moderate-severe pain (VAS ≥ 6).

Finally, we deemed it valuable to assess the subjective satisfaction level of the patients. Despite the heterogeneity observed across the examined cohorts (with ages ranging from 39 to 88years) and diverse initial pathological conditions, the proposed surgical intervention consistently met the needs and expectations of the patients across the board.