Safety concerns surrounding osseointegration are a significant barrier to replacing socket prosthesis as the standard of care following limb amputation. While implanted osseointegrated prostheses traditionally occur in two stages, a one-stage approach has emerged. Currently, there is no existing comparison of the outcomes of these different approaches. To address safety concerns, this study sought to determine whether a one-stage osseointegration procedure is associated with fewer adverse events than the two-staged approach.
A comprehensive electronic search and quantitative data analysis from eligible studies were performed. Inclusion criteria were adults with a limb amputation managed with a one- or two-stage osseointegration procedure with follow-up reporting of complications.
A total of 19 studies were included: four one-stage, 14 two-stage, and one article with both one- and two-stage groups. Superficial infection was the most common complication (one-stage: 38% vs two-stage: 52%). There was a notable difference in the incidence of osteomyelitis (one-stage: nil vs two-stage: 10%) and implant failure (one-stage: 1% vs two-stage: 9%). Fracture incidence was equivocal (one-stage: 13% vs two-stage: 12%), and comparison of soft-tissue, stoma, and mechanical related complications was not possible.
This review suggests that the one-stage approach is favourable compared to the two-stage, because the incidence of complications was slightly lower in the one-stage cohort, with a pertinent difference in the incidence of osteomyelitis and implant failure.
Cite this article: Bone Jt Open 2023;4(7):539–550.
Take home message
One-stage approach of an osseointegrated prosthesis is favourable compared to a two-stage approach.
Incidence of complications was found to be slightly lower in the one-stage cohort.
Of clinical importance is the reduction of osteomyelitis incidence found in the one-stage cohort when compared to the two-stage cohort.
There is significant morbidity associated with limb amputation and its prevalence is expected to increase.1-3 Common indications for limb amputations include trauma, tumour, infection, and peripheral vascular disease (PVD).1,4 The burden associated with limb amputation and increasing prevalence means that research into improved management of these patients is imperative.1-5
Conventionally, patients with limb amputation are treated with a socket prosthesis, where the stump sits inside the (socket of a) prosthetic device.6 However, patients report low satisfaction with prosthetic function and fit, or experience complications such as pain, fracture, and skin breakdown.4,6,7 An alternative to socket prosthesis is osseointegrated, or “bone-anchored prosthesis”.1,4,5,8-12 This involves direct anchorage of a prosthetic implant into residual bone via an intramedullary implant, depicted in Figure 1.
Initially described in the early 1950s by Brånemark et al,3-5 osseointegrated prosthesis (OIP) has become a clinically viable procedure over the last 30 years. Currently, there are multiple osseointegration systems for the treatment of amputees.3-5,9,10,12 Implantation of osseointegrated prosthesis traditionally occurs in two stages: implantation of the intramedullary component (S1) and the creation of a percutaneous opening for the attachment site of the prosthesis (S2).5,13-15 Alternatively, a one-stage approach has been developed, which involves inserting the intramedullary implant and fashioning the stoma in one procedure.13
However, concerns regarding the safety of osseointegration are a considerable barrier to this becoming the standard of care following limb amputation.8,9,16 Successful OIP relies on the host bone, the implant, and the skin-implant interface. If any of these three elements are compromised, complications can occur.10,17 The incidence of complications due to OIP is well reported.8-10 While serious adverse events, such as osteomyelitis, fracture, and implant failure, are rare, they are clinically important, as these are associated with significant morbidity.1,8,9,16,18
The most-reported complication of OIP is infection,5,8-11,16,18 and Hoellwarth et al5 suggested that the risk of infection is decreasing with improved management of soft-tissues. The interface between the soft-tissue and the bone-anchored implant is important in bacterial infection of the OIP due to the ‘race’ to colonize the surface of the OIP between epithelial tissue, bone, and bacteria.5,10,17,19,20 Since infection is unavoidable if bacteria colonize the implant prior to tissue integration,20 good closure of the implant-soft-tissue interface is required to prevent infectious complications.10,17,21 Furthermore, Hoellwarth et al5 noted that the risk of infection, including the risk of implant removal secondary to an infection, was reduced with the one-stage procedure because of the improved management of soft-tissues.
Interestingly, there is no literature comparing the outcomes of the one- versus two-stage approach to implantation of an OIP. Therefore, the current study contributed to the literature by determining whether there is evidence that a one-stage procedure is associated with lower infection rates compared to the two-stage approach. Furthermore, because of the significant burden associated with other osseointegration complications, such as fracture and implant failure, the scope of this review included all adverse outcomes. By identifying and comparing the incidence of adverse events after one-stage and two-stage OIPs, this study sought to determine which procedure has a favourable complication profile.
This systematic review followed the PRISMA guidelines.22 Relevant studies published before 29 December 2020 (date last searched) were identified using OVID to concurrently search MEDLINE ALL (1946 to 23 December 2020), Ovid Emcare (1995 to 2020, week 51), and Embase Classic + Embase (1947 to 24 December 2020). The electronic search strategy used a combination of MeSH and free-text keywords related to the population (e.g. amput*, artificial lim*), intervention (e.g. osseointegrat*, bone-anchor *), and outcomes (e.g. safety, failure, complicat*). The search was limited to humans, and the OVID deduplicate function was used (n = 271 to n = 143) to remove duplicated papers automatically. The full search string is provided in Figure 2. Additional relevant studies were retrieved by manually scanning the reference lists of articles identified by this search (systematic reviews) and were assessed using the same eligibility criteria.
The inclusion criteria were adults with an upper and/or lower limb amputation managed with a one- or two-stage osseointegration procedure, and had follow-up reporting of complications or adverse events associated with their bone-anchored prosthesis. Eligible studies were observational studies published before 29 December 2020 (i.e. date last searched).
Articles were excluded if they did not include follow-up reporting of the incidence and types of adverse events, were a conference abstract or case report, presented non-original or duplicate data, or were not in the English language.
The electronic search results were imported into Microsoft Excel via Endnote, and study selection was then conducted in two phases. The first phase was screening the titles and abstracts to identify studies that potentially met the inclusion criteria. Additional studies were identified by manually searching the included studies’ reference lists. All studies included were evaluated in the same manner. A full-test evaluation of potentially eligible studies was conducted in phase two, and exclusion reasons were coded, as seen in Figure 3. If patient cohorts completely overlapped, the study with the most relevant data was selected.
Data were extracted from eligible studies and organized into a second Microsoft Excel (USA) document, based on surgery type (i.e. one- vs two-stage). Data included were the study design, follow-up period, and all reported complications. This organization strategy revealed significant data overlap in patient cohorts, data collection periods, and follow-up duration. Because of this, quantitative pooling of data was not possible due to the extensive heterogeneity of implant design, methodology, follow-up duration, and reported complications.
The Journal of Bone & Joint Surgery level of evidence (LOE)23 rating system assesses the clinical application of research findings by study type (diagnostic, prognostic, therapeutic, economic). This hierarchical system was used to assess the study quality, with therapeutic Level I as a randomized control trial, down to Level V (mechanism-based reasoning).23 The Newcastle-Ottawa Scale (NOS)24 is a scoring system to evaluate non-randomized studies based on participant selection, comparability, and outcome determination, and was used to evaluate the risk of bias. Based on published evidence,8 studies scoring nine points were assessed as having a low risk of bias, seven or eight points as medium risk, and a score of six points or less was judged high-risk.
Quantitative data were extracted from eligible studies, processed using Microsoft Excel to calculate mean and standard deviation. SPC for Excel was used to determine upper and lower confidence limits for outcomes of interest and generate datasets for graphs.
Table I provides a summary of study characteristics. There were 19 included studies: four reported results for the one-stage approach,17,25-27 14 reported outcomes for the two-stage approach,3,14,28-39 and one article with both one- and two-stage groups.40 Osseointegration is a relatively new procedure, and included studies were published between 2010 and 2020. They were conducted across Australia, the UK, Europe (Sweden, Germany, and the Netherlands), and the USA. Included studies were Therapeutic LOE II-IV23 and had a high risk of bias (NOS 5 to 6).24
|Author||Location||Procedure type||Implantation period, yrs||Follow-up period||Study design||Level of evidence||NOS quality score|
|Muderis et al30||Australia||Two-stage||Not recorded||Mean: 21.5 months after S1||Prospective cohort||II||5|
|Al Muderis et al28||Australia and the Netherlands||Two-stage||2009 to 2013||Median: 34 months
Range: 24 to 71 months
|Al Muderis25||Australia||One-stage||2013 to 2014||Mean: 14 months
Range: 10 to 30 months
|Muderis et al29||Australia||Two-stage||Not recorded||Mean: 36.4 months
Range: 24 to 60 months
|Prospective case series||IV||5|
|Aschoff et al14||Germany||Two-stage||1999 to 2009||NR||Retrospective cohort study||III||5|
|Attallah et al26||Australia||One-stage||2015 to 2018||12 months||Multicentre case series||IV||N/A|
|Branemark et al31||Sweden||Two-stage||1999 to 2007||Two years||Prospective cohort||II||5|
|Branemark et al3||Sweden||Two-stage||1999 to 2007||Five years||Prospective cohort||II||5|
|Hagberg32||Sweden||Two-stage||1990 to 2015||Median: 7 years
Range: 1 to 20 years
|Hagberg et al33||Sweden||Two-stage||1999 to 2017||15 years||Prospective cohort||II||5|
|Juhnke et al34||Germany||Two-stage||1999 to 2013||Range: 1 to 144 months||Retrospective comparison||III||6|
|Marano et al17||USA||One-stage||2017 to 2019||Mean: 28 weeks
Range: 10 to 73 weeks
|Matthews et al35||UK||Two-stage||1997 to 2008||Range: 1.8 to 15.9 years||Prospective cohort||II||6|
|McGough et al40||USA||One- and Two-stage||2012 to publication (2017)||NR||Prospective cohort||II||6|
|Reetz et al36||The Netherlands||Two-stage||2009 to 2013||5 years||Retrospective cohort||III||5|
|Tillander et al38||Sweden||Two-stage||2005||Mean: 56 months
Range: 3 to 132 months
|Tillander et al37||Sweden||Two-stage||1990 to 2010||Mean: 7.9 years
Range: 1.5 to 19.6 years
|Tsikandylakis et al39||Sweden||Two-stage||1995 to 2010||Median: 8 years
Range: 2 to 19 years
|Wood et al27||UK||One-stage||2015 to 2017||Up to 3 years||Case-series||IV||N/A|
NOS, Newcastle-Ottawa Scale; NR, not reported.
Patient characteristics are presented in Table II. Over the five papers reporting one-stage outcomes,17,25-27,40 the cohort size ranged from four to 22 patients (mean: 11 patients). Of the 15 studies of the two-stage approach,3,14,28-40 the cohort ranged from five to 111 patients (mean: 35 patients).
|Number of patients||Amputation site||Indication for amputation|
|One-stage||Al Muderis et al25||22 patients||Unilateral TFA||Trauma, neoplasia, and infection|
|Attallah et al26||4 patients||Unilateral TTA||Salvage knee joint alternative to above-knee amputation, excessive phantom limb pain, and socket-interface problems|
|Marano et al17||14 patients||Lower limb - unilateral
12 × TFA
2 × TTA
|McGough et al40||6 patients||Unilateral TFA||Oncologic and traumatic|
|Wood et al27||7 patients||6 × bilateral TFA, 1 × unilateral TFA (bilateral amputee)||Trauma (military - complex ballistic injuries)|
|Two-stage||Muderis et al30||50 patients||Unilateral TFA||Trauma, blast injury, infection, oncology, congenital|
|Al Muderis et al28||86 (91 implants)
44 in Australia
42 in Norway
|Unilateral TFA||Trauma, tumour, infection, congenital, other|
|Muderis et al29||37 patients||Unilateral TFA||Not recorded|
|Aschoff et al14||37 (39 implants)||37 × unilateral TFA, 2 × bilateral TFA||Trauma, tumour, other|
|Branemark et al31||48 patients, 52 implants||TFA: 45 × unilateral, 6 × bilateral||Trauma, tumour, other|
|Branemark et al3||40 patients||TFA (majority unilateral)||Trauma, tumour, other|
|Hagberg32||12 patients||10 × bilateral TFA
2 × unilateral TFA
|Hagberg et al33||111 patients||Unilateral TFA||Trauma, tumour, emboli, infection|
|Juhnke et al34||69 patients||65 × unilateral TFA
4 × bilateral TFA
|Trauma, tumour, infection, fourth-degree burn, other|
|Matthews et al35||18 patients||Unilateral TFA||Trauma|
|McGough et al40||5 patients||4 × TFA, 1 × THA||Oncological, traumatic, and infection|
|Reetz et al36||39 patients||38 × unilateral TFA
1 × bilateral TFA
|Trauma, tumour, infection, other (compartment syndrome)|
|Tillander et al38||39 patients, 45 implants||45 implants
33 × TFA, 1 × TTA, 4 × ulnar, 4 × radial, 3 × THA
|Trauma or neoplasia|
|Tillander et al37||96 patients||90 × unilateral TFA
6 × bilateral TFA
|Tumour, trauma, ischaemic event, primary deep-seated infection|
|Tsikandylakis et al39||18 patients||Unilateral THA||Trauma, tumour|
TFA, transfemoral amputation; TTA, transtibial amputation.
Common indications for amputation were trauma, tumour, and infection. Patients had a variety of amputation sites: most patients had unilateral lower limb amputations (transfemoral amputation (TFA), trans-tibial amputation (TTA)), and other sites included upper limb (transhumeral amputation, bilateral, or mixed amputation sites. Common inclusion criteria were complications with socket prosthesis, skeletal maturity, ability to comply with rehabilitation protocol, and overall good health with no ongoing chemotherapy.3,14,17,25-40 All studies excluded patients with peripheral artery disease from receiving osseointegration surgery, except a single one-stage case series.26
A tabulated summary of the incidence of complications is provided in Table III. Nine articles reported the incidence of one or more complications.26-31,34,36,40 Total complications of the one-stage procedure were reported in three papers,26,27,40 with a mean incidence of 51% (17% to 86%, SD 35%). Seven papers reported the total complications for the two-stage method,28-31,34,36,40 with a mean incidence of 59% (40% to 96%, SD 21%).
|Procedure type||Author||1 or more complication||Infection||Fracture||Soft-tissue and stoma complications||Implant failure||Mechanical complications|
|One-stage||Al Muderis et al25||NR*||12/22 to 55%||Nil||Nil||6/22 to 27% elective soft-tissue refashioning||Nil||NR*|
|Atallah et al26||2/4 to 50%||2/4 to 50%||Nil||Nil||Nil||Nil||Nil|
|Marano et al17||NR*||2/14 to 14%||Nil||1/14 to 7%||Nil||1/14 to 7%||Not adequately reported†|
|McGough et al40||1/6 to 17%||Nil||Nil||1/6 to 17%||Nil||Nil||Nil|
|Wood et al27||6/7 to 86%||5/7 to 71%||Nil||3/7 to 43%||3/7 to 43% required soft-tissue refashioning||Nil||Nil|
|Two-stage||Muderis et al30||27 / 50 to 54%||21/50 to 42%||Nil||4/50 to 8%||10/50 to 20% required soft-tissue refashioning||2/50 to 40%||NR*|
|Al Muderis et al28||55/86 to 64%||29/86 to 34%||Nil||3/86 to 3%||14/86 to 16% with issues related to soft-tissue||17/86 to 20% with stoma hyper granulation||3/86 to 3%||25/86 to 29%|
|Muderis et al29||16/37 to 43%||16/37 to 43%||Nil||1/37 to 3%||6/27 to 16% elective soft-tissue refashioning||Nil||NR*|
|Aschoff et al14||NR*||NR*||1/37 to 3%||2/37 to 5%||14/37 to 38% revision due to stoma issues||4/37 to 11%||NR*|
|Brånemark et al31||46/48 to 96%||28/48 to 58%||4/48 to 8%||4/48 to 8%||NR*||4/48 to 8%||4/48 to 8%|
|Brånemark et al3||NR*||34/40 to 85%||11/40 to 28%||NR*||NR*||4/40 to 10%||15/40 to 10%|
|Hagberg32||NR*||10/12 to 83%||1/12 to 8%||2/12 to 17%||NR*||1/12 to 8%||8/12 to 67%|
|Hagberg et al33||NR*||NR*||NR*||5/111 to 5%||NR*||18/111 to 16%||61/111 to 55%|
|Juhnke et al34||29/69 to 42%||23/69 to 33%||1/69 to 1%||5/69 to 7%||24/69 to 35% with intervention for soft-tissue problems/ problems at stoma||4/69 to 6%||1/69 to 1%|
|Matthews et al35||NR*||11/18 to 61%||5/18 to 28%||3/18 to 17%||NR*||5/18 to 28%||12/18 to 67%|
|McGough et al40||2/5 to 40%||Nil||Nil||1/5 to 20%||1/5 to 20% with taper mismatch||Nil||Nil|
|Reetz et al36||30/39 to 77%||30/39 to 77%||4/39 to 10%||NR*||14/39 to 36% required soft-tissue refashioning||8/39 to 21% with stoma hyper granulation||5/39 to 13%||Not adequately reported†|
|Tillander et al38||NR*||NR*||7/39 to 18%||NR*||NR*||Not adequately reported†||NR*|
|Tillander et al37||NR*||NR*||16/96 to 17%||NR*||NR*||Not adequately reported†||NR*|
|Tsikandylakis et al39||NR*||5/18 to 28%||1/18 to 6%||8/18 to 44%||8/18 to 44% with skin irritation||3/18 to 17%||NR*|
NR = endpoint not reported.
Not adequately reported – did not detail the number of patients with events, or total incidence of outcome (e.g. only reported septic implant failure).
Superficial infection was the most common complication, and the incidence of one or more infections was reported in 14 articles. The incidence of superficial infections from the one-stage procedure was reported in five papers,17,25-27,40 with a mean incidence of 38% (0% to 71%, SD 30%). The incidence of superficial infections for a two-stage approach was reported in 11 papers,3,28-32,34-36,39,40 and the mean incidence was 52% and ranged from 0% to 85% (SD 27%).
Osteomyelitis (deep infection)
Overall, 17 articles reported the incidence of osteomyelitis, depicted in Figure 4. There were no cases of osteomyelitis across five papers reporting outcomes in the one-stage cohort.17,25-27,40 Across the 14 articles reporting outcomes of the two-stage procedure,3,28-32,34-36,39,40 the mean incidence of osteomyelitis was 9%, ranging from 0% to 28% (SD 10%).
The incidence of implant failure was reported in 17 articles, depicted in Figure 5. In the five papers reporting outcomes in the one-stage cohort,17,25-27,40 there was one reported case of implant failure (mean 1%; SD 3%). In the 13 studies reporting outcomes of the two-stage approach,3,14,28-36,39,40 the mean incidence of implant failure was 9%, ranging from 0% to 28% (SD 8%).
Fracture incidence (one or more fractures, intraoperative or postoperative) was reported in 16 articles. The mean incidence of fractures due to the one-stage procedure was 13% (0% to 43%; SD 18%) reported in five papers (13, 18, 26, 27, 40). In the 12 articles (3, 15, 28 to 35, 39, 40) reporting fracture incidence from the two-stage approach, the mean incidence was 12% (0% to 50%; SD 14%). There was one case of intraoperative fracture in the one-stage cohort,27 and eight reported cases of intraoperative fracture in the first stage of the two-stage approach.39
Soft-tissue and stoma-related complications were reported in both one- and two-stage cohorts, as seen in Table III. Reporting was infrequent and inconsistent, which prevented quantitative analysis.
There was limited reporting of mechanical complications in both the one- and two-stage groups. Evaluation of the incidence of mechanical complications was not possible due to inconsistent reporting of the mechanism and classification of these complications.
Safety concerns are a considerable barrier to OIP becoming the standard of care for patients after limb amputation.8,9,16 Adverse events following OIP range from minor (e.g. soft-tissue infections and complications) to severe (e.g. implant infection, implant failure),9,18,39 and there is no literature comparing the incidence of complications of the one-stage versus two-stage approach.
Complications were common in patients treated with OIPs regardless of procedure type. In articles that reported the incidence of any complication,26-31,34,36,40 more than half the patients in both the one- and two-stage cohorts experienced an adverse event. Furthermore, some patients experienced more than one complication, with either several episodes of the same event, or separate complications. Thus, concerns regarding the safety of OIPs are warranted; however, socket prostheses are also associated with notable complications.6,41
Infection remains an important concern for patients treated with an OIP,8,11,16 and the primary focus of this review was to determine if there was a difference in the incidence of infection between the one- and two-stage approaches. The Al Muderis et al28 classification of infection related to the osseointegrated implants categorizes infection as superficial (grade 1 or 2) or deep (bone infection: grade 3, or implant failure: grade 4). This classification is important because superficial and deep infection are associated with different disease processes, treatments, and sequelae.10,28,38
The literature suggests that the risk of infection is decreasing with ‘improved surgical technique’ and management of the soft-tissue bone-anchored implant interface.5,17,18 Hoellwarth et al5 noted that the risk of infection was reduced with the one-stage procedure, and suggested that lower infection rates were a result of improved management of soft-tissue. Thus, because soft-tissue management is crucial to preventing infection in osseointegration procedures, 10 and soft-tissue optimization is the focus of the one-stage approach,13,17 we hypothesized that the one-stage approach enables superior soft-tissue management and subsequently results in lower infection rates.
As expected, superficial infection was the most reported complication in both the one- and two-stage cohorts.10,18 There was a slight difference in the incidence between the one- (37%) and two-stage (52%) approaches, favouring the one-stage procedure. This finding supports our hypothesis and suggests that the one-stage approach provides a superior soft-tissue seal, which accounts for the improved outcomes.17,21 However, inconsistent reporting of the number of events prohibited the comparison of event frequency between the one- and two-stage procedures.
The difference in the incidence of osteomyelitis (deep infection/grade 3 infectious complication) is the most compelling outcome of this review. Osteomyelitis is bone inflammation secondary to infection leading to bone destruction,10,42 and is clinically significant due to its high patient morbidity, mortality, and economic burden.9,42 Ideally, the surface of the OIP is colonized by bone and epithelial tissue, not bacteria,10,17,21 which is facilitated by tight closure of the soft-tissue bone-anchored implant interface.5,17 The hypothesis that the one-stage approach leads to superior stump closure, with a tight soft-tissue seal, is further supported by the fact that there were no cases of osteomyelitis in the one-stage cohort, compared to an average of 10% in the two-stage cohort. However, the risk of deep infection continues with time,3,16,37 and the maximum follow-up in the one-stage cohort was three years,27 compared to 20 years in the two-stage cohort.32 Furthermore, the risk of osteomyelitis may also be related to implant design.18,37 Overall, this review found that the one-stage procedure was associated with a lower incidence of osteomyelitis and implies that this is the more preferable approach. Because osteomyelitis is a major complication of an OIP, this finding has the potential to inform operative technique.9,18
Implant failure is another major complication of OIP.9,16,18 This is defined as implant loosening or explantation, and may be secondary to infection (septic loosening/grade 4 infection) or other processes such as failed osseointegration, implant breakage, and fatigue failure (aseptic loosening).4,10,16 This review found a notably lower incidence of implant failure (septic and aseptic) in the one-stage cohort, suggesting that the one-stage approach is favourable compared to the two-stage procedure. However, other factors affect osseointegration and osteoblast adhesion beyond surgical technique, such as implant design and quality of host bone.4,10,43
Fractures are another rare but serious adverse event associated with OIP.9,16 The overall definition of ‘fracture’ included periprosthetic fractures,14,27,28,30,31,34,40 fractures in secondary sites such as the vertebrae,3 fractures secondary to falls,17,27,32,35 and fractures secondary to septic loosening. 35 The average incidence of overall fractures was equivocal between the one- and two-stage approaches, as they are more likely correlated with the quality of bone and implant stability,7,16,44 or the patient’s return to activity.27 Furthermore, Hoellwarth et al7 suggested that the risks and complications associated with a fracture should not deter patients and clinicians, because most patients who sustained a fracture continue to wear their OIP.
Quantitative analysis and comparison of soft-tissue-, stoma-, and mechanical-related complications were not possible because of inconsistent reporting. These are areas of interest because when reported, they were not uncommon, and patients often experienced several events. Requirement for revision surgery was the most common reporting tool for soft-tissue/stoma complications;14,25,27,29,30,34 infrequent and inconsistent reporting of the origin (primarily stoma-related vs soft-tissue-related) prevented quantitative analysis. This is regrettable, since the evaluation of soft-tissue- and stoma-related complications between the one- and two-stage approach is an important part of assessing whether there is a difference in the quality of the soft-tissue management between these procedures. Difficulty in evaluating mechanical complications stems from inconsistent reporting (e.g. “extramedullary breakage”28,31 vs “dual-cone adaptor breakage”36 vs “mechanical complications”15,17,33) and the variety of osseointegrated implants used, which may have confounded the results.4,10
While these findings imply that the one-stage approach is preferable to the two-stage, they should be interpreted with caution, as this systematic review has several limitations. First, there was significant heterogeneity of outcome reporting, implant type, patient factors (i.e. peripheral vascular disease status, amputation site, age), rehabilitation protocols, and follow-up period, which may have confounded results. This heterogeneity and lack of clearly reported information regarding covariants, such as implant type, surgeon, and length of implant time in situ, further hinders the ability to provide robust statistical analysis to assess the optimal surgical approach. Furthermore, this limits the ability to evaluate the incidence of adverse events as a function of implant date and consider the effects of improving implant design and manufacture on the results. Second, data availability may have contributed to the distortion of the review outcomes because it was limited to published articles, with a notable overlap of patient cohorts despite efforts to prevent this in the search strategy. Third, the review was limited by the data quality because the available data had a high risk of bias, lower level of evidence (i.e. no therapeutic level I studies), generally small sample size (especially one-stage cohorts), variable follow-up periods, and inconsistent outcome reporting. Finally, the variability and low quality of the data prevented meta-analysis, limiting the analysis to a direct comparison of basic statistics.
Future research is needed to improve our understanding of the specific comparisons between one- and two-step procedures. This may include a randomized control trial that would enable control of patient and implant confounders, and a repeat review is indicated because subgroup analysis43 and a prospective trial13 of one-stage osseointegration is currently underway. Furthermore, subgroup analysis to investigate amputation site and patient factors (i.e. comorbid disease, indication for amputation), economic evaluation, and future meta-analysis of the one- versus two-stage approach are additional research areas.
The evidence analyzed by this systematic review indicates that a one-stage approach of an OIP is favourable compared to a two-stage approach. The incidence of complications was slightly lower in the one-stage cohort, especially the incidence of osteomyelitis, a clinically important complication. However, adverse events still frequently occurred in patients with OIPs treated with either approach. This review has contributed to the gap in knowledge surrounding complications and adverse events for one- and two-stage osseointegrated procedures; however, further research into soft-tissue and mechanical complications is required to appreciate the outcomes of each surgery more completely.
2. Ziegler-Graham K , MacKenzie EJ , Ephraim PL , Travison TG , Brookmeyer R . Estimating the prevalence of limb loss in the United States: 2005 to 2050 . Arch Phys Med Rehabil . 2008 ; 89 ( 3 ): 422 – 429 . Crossref PubMed Google Scholar
3. Brånemark RP , Hagberg K , Kulbacka-Ortiz K , Berlin Ö , Rydevik B . Osseointegrated percutaneous prosthetic system for the treatment of patients with transfemoral amputation: A prospective five-year follow-up of patient-reported outcomes and complications . J Am Acad Orthop Surg . 2019 ; 27 ( 16 ): e743 – e751 . Crossref PubMed Google Scholar
5. Hoellwarth JS , Tetsworth K , Rozbruch SR , Handal MB , Coughlan A , Al Muderis M . Osseointegration for amputees: Current implants, techniques, and future directions . JBJS Rev . 2020 ; 8 ( 3 ): e0043 . Crossref PubMed Google Scholar
6. Paterno L , Ibrahimi M , Gruppioni E , Menciassi A , Ricotti L . Sockets for limb prostheses: A review of existing technologies and open challenges . IEEE Trans Biomed Eng . 2018 ; 65 ( 9 ): 1996 – 2010 . Crossref PubMed Google Scholar
7. Hoellwarth JS , Tetsworth K , Kendrew J , et al. Periprosthetic osseointegration fractures are infrequent and management is familiar . Bone Joint J . 2020 ; 102-B ( 2 ): 162 – 169 . Crossref PubMed Google Scholar
8. Kunutsor SK , Gillatt D , Blom AW . Systematic review of the safety and efficacy of osseointegration prosthesis after limb amputation . Br J Surg . 2018 ; 105 ( 13 ): 1731 – 1741 . Crossref PubMed Google Scholar
9. Safari R . Lower limb prosthetic interfaces: Clinical and technological advancement and potential future direction . Prosthet Orthot Int . 2020 ; 44 ( 6 ): 384 – 401 . Crossref PubMed Google Scholar
10. Overmann AL , Clark DM , Tsagkozis P , Wedin R , Forsberg JA . Validation of PATHFx 2.0: An open-source tool for estimating survival in patients undergoing pathologic fracture fixation . J Orthop Res . 2020 ; 38 ( 10 ): 2149 – 2156 . Crossref PubMed Google Scholar
11. Gerzina C , Potter E , Haleem AM , Dabash S . The future of the amputees with osseointegration: A systematic review of literature . J Clin Orthop Trauma . 2020 ; 11 ( Suppl 1 ): S142 – S148 . Crossref PubMed Google Scholar
12. Brånemark R , Brånemark PI , Rydevik B , Myers RR . Osseointegration in skeletal reconstruction and rehabilitation: a review . J Rehabil Res Dev . 2001 ; 38 ( 2 ): 175 – 181 . PubMed Google Scholar
13. Al Muderis M , Lu W , Tetsworth K , Bosley B , Li JJ . Single-stage osseointegrated reconstruction and rehabilitation of lower limb amputees: the Osseointegration Group of Australia Accelerated Protocol-2 (OGAAP-2) for a prospective cohort study . BMJ Open . 2017 ; 7 ( 3 ): e013508 . Crossref PubMed Google Scholar
14. Aschoff HH , Kennon RE , Keggi JM , Rubin LE . Transcutaneous, distal femoral, intramedullary attachment for above-the-knee prostheses: an endo-exo device . J Bone Joint Surg Am . 2010 ; 92 Suppl 2 : 180 – 186 . Crossref PubMed Google Scholar
15. Hagberg K , Brånemark R . One hundred patients treated with osseointegrated transfemoral amputation prostheses--rehabilitation perspective . J Rehabil Res Dev . 2009 ; 46 ( 3 ): 331 – 344 . PubMed Google Scholar
16. Ontario Health (Quality) . Osseointegrated prosthetic implants for people with lower-limb amputation: A health technology assessment . Ont Health Technol Assess Ser . 2019 ; 19 ( 7 ): 1 – 126 . PubMed Google Scholar
17. Marano AA , Modiri O , Rozbruch SR , Otterburn DM . Soft tissue contouring at the time of osseointegrated implant reconstruction for lower extremity amputation . Ann Plast Surg . 2020 ; 85 ( S1 Suppl 1 ): S33 – S36 . Crossref PubMed Google Scholar
18. Atallah R , Leijendekkers RA , Hoogeboom TJ , Frölke JP . Complications of bone-anchored prostheses for individuals with an extremity amputation: A systematic review . PLoS One . 2018 ; 13 ( 8 ): e0201821 . Crossref PubMed Google Scholar
20. Hoellwarth JS , Tetsworth K , Akhtar MA , Al Muderis M . Transcutaneous osseointegration for amputees: lessons from the past of relevance to the future . Bone Joint Res . 2021 ; 10 ( 10 ): 690 – 692 . Crossref PubMed Google Scholar
21. Pendegrass CJ , Goodship AE , Blunn GW . Development of a soft tissue seal around bone-anchored transcutaneous amputation prostheses . Biomaterials . 2006 ; 27 ( 23 ): 4183 – 4191 . Crossref PubMed Google Scholar
22. Moher D , Liberati A , Tetzlaff J , Altman DG , PRISMA Group . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med . 2009 ; 6 ( 7 ): e1000097 . Crossref PubMed Google Scholar
24. Wells GA , Shea B , O’Connell D , et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses . http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp ( date last accessed 17 July 2023 ). Google Scholar
25. Al Muderis M , Lu W , Li JJ . Osseointegrated prosthetic limb for the treatment of lower limb amputations: Experience and outcomes . Unfallchirurg . 2017 ; 120 ( 4 ): 306 – 311 . Crossref PubMed Google Scholar
26. Atallah R , Li JJ , Lu W , Leijendekkers R , Frölke JP , Al Muderis M . Osseointegrated transtibial implants in patients with peripheral vascular disease: A multicenter case series of 5 patients with 1-year follow-up . J Bone Joint Surg Am . 2017 ; 99-A ( 18 ): 1516 – 1523 . Crossref PubMed Google Scholar
27. Wood P , Small C , Mahoney P . Perioperative and early rehabilitation outcomes following osseointegration in UK military amputees . BMJ Mil Health . 2020 ; 166 ( 5 ): 294 – 301 . Crossref PubMed Google Scholar
28. Al Muderis M , Khemka A , Lord SJ , Van de Meent H , Frölke JPM . Safety of osseointegrated implants for transfemoral amputees: A two-center prospective cohort study . J Bone Joint Surg Am . 2016 ; 98-A ( 11 ): 900 – 909 . Crossref PubMed Google Scholar
29. Muderis MA , Lu W , Glatt V , Tetsworth K . Two-stage osseointegrated reconstruction of post-traumatic unilateral transfemoral amputees . Mil Med . 2018 ; 183 ( suppl_1 ): 496 – 502 . Crossref PubMed Google Scholar
30. Muderis MA , Tetsworth K , Khemka A , et al. The Osseointegration Group of Australia Accelerated Protocol (OGAAP-1) for two-stage osseointegrated reconstruction of amputated limbs . Bone Joint J . 2016 ; 98-B ( 7 ): 952 – 960 . Crossref PubMed Google Scholar
31. Brånemark R , Berlin O , Hagberg K , Bergh P , Gunterberg B , Rydevik B . A novel osseointegrated percutaneous prosthetic system for the treatment of patients with transfemoral amputation: A prospective study of 51 patients . Bone Joint J . 2014 ; 96-B ( 1 ): 106 – 113 . Crossref PubMed Google Scholar
32. Hagberg K . Bone-anchored prostheses in patients with traumatic bilateral transfemoral amputations: rehabilitation description and outcome in 12 cases treated with the OPRA implant system . Disabil Rehabil Assist Technol . 2019 ; 14 ( 4 ): 346 – 353 . Crossref PubMed Google Scholar
33. Hagberg K , Ghassemi Jahani SA , Kulbacka-Ortiz K , Thomsen P , Malchau H , Reinholdt C . 15-year follow-up of transfemoral amputees with bone-anchored transcutaneous prostheses: mechanical complications and patient-reported outcomes . Bone Joint J . 2020 ; 102-B ( 1 ): 55 – 63 . PubMed Google Scholar
34. Juhnke DL , Beck JP , Jeyapalina S , Aschoff HH . Fifteen years of experience with Integral-Leg-Prosthesis: Cohort study of artificial limb attachment system . J Rehabil Res Dev . 2015 ; 52 ( 4 ): 407 – 420 . Crossref PubMed Google Scholar
35. Matthews DJ , Arastu M , Uden M , et al. UK trial of the Osseointegrated Prosthesis for the Rehabilitation for Amputees: 1995-2018 . Prosthet Orthot Int . 2019 ; 43 ( 1 ): 112 – 122 . Crossref PubMed Google Scholar
36. Reetz D , Atallah R , Mohamed J , van de Meent H , Frölke JPM , Leijendekkers R . Safety and performance of bone-anchored prostheses in persons with a transfemoral amputation: A 5-year follow-up study . J Bone Joint Surg Am . 2020 ; 102-A ( 15 ): 1329 – 1335 . Crossref PubMed Google Scholar
37. Tillander J , Hagberg K , Berlin Ö , Hagberg L , Brånemark R . Osteomyelitis risk in patients with transfemoral amputations treated with osseointegration prostheses . Clin Orthop Relat Res . 2017 ; 475 ( 12 ): 3100 – 3108 . Crossref PubMed Google Scholar
38. Tillander J , Hagberg K , Hagberg L , Brånemark R . Osseointegrated titanium implants for limb prostheses attachments: infectious complications . Clin Orthop Relat Res . 2010 ; 468 ( 10 ): 2781 – 2788 . Crossref PubMed Google Scholar
39. Tsikandylakis G , Berlin Ö , Brånemark R . Implant survival, adverse events, and bone remodeling of osseointegrated percutaneous implants for transhumeral amputees . Clin Orthop Relat Res . 2014 ; 472 ( 10 ): 2947 – 2956 . Crossref PubMed Google Scholar
40. McGough RL , Goodman MA , Randall RL , Forsberg JA , Potter BK , Lindsey B . The Compress® transcutaneous implant for rehabilitation following limb amputation . Unfallchirurg . 2017 ; 120 ( 4 ): 300 – 305 . Crossref PubMed Google Scholar
41. Aydın A , Çağlar Okur S . Effects of test socket on pain, prosthesis satisfaction, and functionality in patients with transfemoral and transtibial amputations . Med Sci Monit . 2018 ; 24 : 4031 – 4037 . Crossref PubMed Google Scholar
42. Masters EA , Trombetta RP , de Mesy Bentley KL , et al. Evolving concepts in bone infection: redefining “biofilm”, “acute vs. chronic osteomyelitis”, “the immune proteome” and “local antibiotic therapy.” Bone Res . 2019 ; 7 ( 1 ): 20 . Crossref PubMed Google Scholar
43. Atallah R , van de Meent H , Verhamme L , Frölke JP , Leijendekkers RA . Safety, prosthesis wearing time and health-related quality of life of lower extremity bone-anchored prostheses using a press-fit titanium osseointegration implant: A prospective one-year follow-up cohort study . PLOS ONE . 2020 ; 15 ( 3 ): e0230027 . Crossref PubMed Google Scholar
44. Nauth A , Ristevski B , Bégué T , Schemitsch EH . Periprosthetic distal femur fractures: current concepts . J Orthop Trauma . 2011 ; 25 Suppl 2 ( Supplement 2 ): S82 – 5 . Crossref PubMed Google Scholar
E. Banducci: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing.
M. Al Muderis: Writing – review & editing.
W. Lu: Writing – review & editing.
S. R. Bested: Conceptualization, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing, Supervision, Validation.
The authors received no financial or material support for the research, authorship, and/or publication of this article.
ICMJE COI statement
M. Al Muderis reports royalties from and shares in Osseointegration International Pty Ltd, unrelated to this study. W. Lu is employed by Osseointegration International Pty Ltd.
All data generated or analyzed during this study are included in the published article and/or in the supplementary material.
Open access funding
The authors confirm that the open access fee for this study was self-funded.
© 2023 Author(s) et al. Open Access This article is distributed under the terms of the Creative Commons Attributions (CC BY 4.0) licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original author and source are credited.