Background

In the UK, approximately 60,000 patients each year access specialised prosthetic services [1]. However, true estimates of the UK population living with amputation are difficult to make, with contributions from congenital limb deficiency, trauma, peripheral arterial disease and the military. The prevalence of new lower limb amputation is estimated as 26.3/100,000 among 50 to 84 year olds with peripheral arterial disease, amounting to over 25,000 amputations over a 6-year period alone [2]. Irrespective of the exact figure, there are a large number of people living with lower limb amputation, of which the majority will use prosthetics to walk. As with any ambulant individual, they are subject to the same risks of degenerative joint disease or traumatic injury that can affect non-amputees.

Modern prosthetics benefit from ongoing developments in materials science and production technology, coupled with an ever-improving understanding of fundamental gait biomechanics and stump-socket interface [3]. However, despite these ongoing advancements, a prosthesis will still alter the transfer of energy during the gait cycle compared with a normal biological limb, leading to increased contact forces on other joints through compensatory processes [4, 5]. This has been demonstrated through the increased rates of hip and knee osteoarthritis (OA) observed among lower limb amputees, particularly affecting the non-amputated limb [6,7,8]. Norwell et al. found a significantly increased prevalence of OA among amputees (16.1%) compared with non-amputees (11.7%), in a military population [8]. However, Welke et al. found no clear difference in prevalence from the general population, though amputees were noted to develop OA at an earlier stage, suggesting accelerated disease in pre-disposed individuals [9].

For people with symptomatic hip or knee osteoarthritis, undergoing joint replacement is a proven and effective treatment [10]. However, the presence of concomitant lower limb amputation poses a unique challenge to both the patient and surgeon. From the patient’s perspective, post-operative rehabilitation is made more challenging, as an amputated limb utilising a prosthetic may be unable to provide the same level of compensation to gait as a normal limb in non-amputees. Meanwhile, stump specific complications such as swelling and wound healing may also limit ability to weight bear, delaying ambulation and rehabilitation progress. From the surgical perspective, it is technically difficult to achieve the optimum length, alignment and rotation of an implant without a contralateral limb to reference. Furthermore, when operating on the amputated limb, the length of the residual limb may restrict manoeuvrability during surgical dissection and implantation, while limiting the range of suitable implants compared to non-amputees. Although joint replacement in amputees remains uncommon compared with the total population treated, understanding the optimum surgical management and likely effectiveness of joint replacement in these patients is important, as this can guide consent discussions regarding patient expectations and complication profile for a distinct population group, while providing surgeons with valuable insight for the intra-operative management of patients they will not treat frequently.

The primary aim of this study was to identify and summarise the existing evidence base relating to hip and knee replacement in lower limb amputees, to explore whether clinical outcomes and post-operative complications are comparable to non-amputees, and to synthesise reported intra-operative techniques.

Methods

Search strategy

An electronic search of MEDLINE (PubMed), Cochrane Library, EMBASE and CINAHL was completed (date of inception to 1stApril 2023). The search terms used were developed through preliminary searches and included relevant MeSH terms. The final search strategy was refined in conjunction with an Information Specialist (see Supplementary appendices). ClinicalTrials.gov and Google Scholar index were searched for pre-print publications, and reference list searching of included studies was also performed. Non-peer reviewed articles were not explored. The protocol for this systematic sco** review was developed in conjunction with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement for sco** reviews [11], and was registered prospectively with the Open Science Framework [12].

Selection process

All articles identified by the search strategy were imported into Rayyan [13]. Relevant titles and abstracts were screened independently by two authors (TW and AC) and selected for full text retrieval, depending on conformity to eligibility criteria. This process was repeated following retrieval of relevant full text articles, to determine final inclusion. A third author (VG) was available to settle disputes if necessary.

Eligibility criteria

Inclusion criteria were: (i) English language literature, (ii) Any original research or report; case series, case report, technical tip, (iii) Adults > 18 years, (iv) Undergoing hip or knee replacement, for any indication, (v) Concomitant pre-existing lower limb amputation, for any indication, (vi) Any outcome data. Specific exclusion criteria were also applied: (i) Conference abstracts or opinion pieces, (ii) Amputation after arthroplasty procedure.

Data extraction

A data extraction form was developed, piloted and refined by TW and AC prior to expansion across the remaining included studies. Data on study characteristics, patient demographics, amputation (level, laterality, indication), operation (type, laterality, indication), outcomes, complications and technical tips were all collected and imported into Microsoft Excel (iOS Version 2.76).

Quality assessment

A sco** review methodology without quality assessment was performed, according to the framework outlined by Arksey and O’Malley (2005) [14]. In the authors’ opinion, the nature of the studies available for review meant quality assessment would not meaningfully enhance interpretation or discussion of findings.

Data synthesis

A descriptive and narrative data analysis was performed, to summarise the existing evidence.

Results

Following screening, 40 studies were included in the review (Fig. 1). The majority of included studies were case reports or case series (Table 1) [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,18], while the outcomes reported across the included studies varied widely. Therefore, it is not possible for this review to draw any definitive conclusions regarding the expected outcomes of hip or knee replacement in lower limb amputees, or to guide optimum treatment. Furthermore, as part of the primary aim of this study to broadly summarise the existing evidence base, it is important to highlight a severe lack of reporting within the literature regarding post-operative rehabilitation protocols. Given the clear challenge to ambulation faced by amputees, this is a key omission.

Total hip replacement

From the included literature, the majority focused on THR. The mean age at surgery for amputees undergoing THR was 58.7 years; this is considerably younger than the mean age for THR among the general population of the UK, which is 69 years according to registry data [10]. Osteoarthritis was the most common indication for surgery, while THR was more frequently contralateral to the side of amputation. This evidence appears to support the hypothesis from Struyf et al. that amputation places greater strain on the remaining limb, and therefore necessitates early arthroplasty [7]. However, the time to joint replacement varied widely, with one patient only requiring surgery 60 years after bilateral amputation for congenital abnormality [25], while another patient had pre-existing hip OA exacerbated by ipsilateral amputation, leading to early surgical intervention within 2 years [39]. This clearly demonstrates there are other factors not yet accounted for which dictate the time to joint replacement from amputation, requiring further investigation.

In terms of outcome, THR is often regarded as one of the most successful surgical interventions [55]. The majority of patients report excellent, very good or good results, and 97% demonstrating improvement in function [56]. From the PROMS data reported for amputees, it appears that THR is also a highly successful procedure, with similar levels of function reported post-operatively across a range of PROMS. However, while registry data suggest 58% of THR survive 25 years before requiring revision surgery [57], it remains unclear whether the same altered biomechanics that potentially predispose amputees to OA in the first instance, would also predispose to early revision surgery.

For any patient undergoing THR, dislocation and periprosthetic fracture are devastating complications, and these appear to be uncommon among amputees. However, the low numbers in this study make definitive population wide conclusions hard to make, and one study demonstrated increased risk of falls post-operatively [19], which would predispose to both dislocation and periprosthetic fracture; this requires further study. In relation to specific complications for amputees, post-operative stump swelling was identified as a key issue by one study [34], but was less common for THR patients than TKR patients. This could limit rehabilitation by restricting ambulation through an inability to wear a prosthesis. Compression wraps and bandages could be routinely utilised in amputees to prevent this occurring.

From the available data, there was no clear preponderance for a surgical approach, with both the posterior and anterolateral approaches being used equally. However, one study found functional status among those treated with the posterior approach was significantly better at 3 months, alongside a significantly lower risk of falls [19]. However, by 6 months, these differences were no longer observed [19]. This would support the hypothesis put forward by proponents of the posterior approach, that preserving the hip abductors is important for function. Among amputees, the posterior approach would appear to optimise early functional outcomes, until those receiving anterolateral approach have had time for their abductor repair to heal and recondition. The only study to directly compare THR among amputees with non-amputees similarly found a significant difference in functional status at 3 months, resolving by 6 months [18]. This could indicate that long term functional outcomes are similar, but amputees require a longer period rehabilitation to achieve this, and the surgical approach may present a factor in this.

In terms of the intra-operative techniques, the majority of technical tips centred around the control of the residual limb to facilitate dislocation and relocation of the femoral head. No comparisons were made to suggest one technique is better than another; the utility of each is likely to be specific to the patients' individual anatomy. There was an apparent preference for uncemented femoral stems among the reported implants, with some authors concerned about cement extrusion through the distal femur in transfemoral amputees with short residual limbs [20, 23, 34, 36]. However, in cases where the distal femur is flared, with associated widening of the canal and thin cortical bone, it was recognised that cementation may provide improved stem fixation in sufficient length femurs [15, 39]. Importantly, the use of templating was highlighted by several authors as key to the pre-operative planning [20, 23, 39, 42, 50], as implant size was sometimes limited by the length of the remaining femur.

Total knee replacement

For the studies exploring TKR, the mean age at operation was 61.0 years, similarly representing a much younger age of operation than the UK average for non-amputees, though notably older than amputees undergoing THR. The vast majority underwent surgery for osteoarthritis, more commonly on the contralateral side to their amputation. Similarly to THR, the time to arthroplasty varied widely. However, the slightly higher average age at time of operation potentially indicates a slower onset of degenerative disease at the knee compared to the hip. This warrants further investigation with clinical and biomechanical studies, to explore the comparative risk of OA in hip and knee among amputees.

In terms of outcomes, the functional status achieved was generally reported to be good, though there were no comparative studies to indicate whether the level of function is comparable to non-amputees. Overall, complications were uncommon. However, stump complications were reported more frequently among TKR patients, with both swelling and wound infection reported [28, 44]. This would appear to be a higher risk for ipsilateral TKR, as the incision may extend into the contact area for the prosthetic socket, while localised limb swelling from the surgery will be more consequential. For these patients, managing socket wear and return to ambulation are particularly challenging, with no clear consensus in the literature to support early vs delayed wear.

The technical tips for TKR similarly focused on the management of the residual limb intra-operatively. A key issue with performing TKR in the ipsilateral limb is the reduced length available for use with the bolster and support. The reporting of implants used was poor, and where available there was no clear preponderance for any type, with cruciate retaining, posterior stabilised, highly constrained and linked prostheses all being described.

Strengths and limitations

A key strength of this review is the thorough search strategy employed, refined through consultation with an information specialist, and conformation to PRISMA reporting guidelines. As a result, the authors are confident that the studies included in this review are fully representative of the available evidence. Furthermore, this review is comprehensive in it’s inclusion of both ipsilateral and contralateral arthroplasty, facilitating a more complete summary of the population in question.

However, the sco** nature of this study inherently limits the potential for true data synthesis and comparisons. Furthermore, a formal quality assessment was not performed, in accordance with the framework set out for sco** reviews. Although the quality of available evidence was assumed to be low by the very nature of the studies included, the lack of formal quality assessment further limits the potential for discussion regarding the strength of conclusions made. When develo** the protocol for this review, no formal patient and public involvement and engagement (PPIE) was sought. This may have helped guide the discussion and focus of the review in a more patient centred direction.

Conclusion

This study has demonstrated a paucity of high quality evidence reporting on lower limb amputees undergoing hip or knee replacement, although the available evidence appears to suggest outcomes comparable to non-amputees are achievable. There is a need for more high-quality observational studies to establish the association between amputation and subsequent need for joint replacement. Furthermore, comparative studies are needed to identify whether amputees can be expected to achieve similar functional outcomes after surgery, and if they are at higher risk of complications.