Aims. Deep surgical site infection (SSI) is common after lower limb fracture. We compared the diagnosis of deep SSI using alternative methods of data collection and examined the agreement of clinical photography and in-person clinical assessment by the Centers for Disease Control and Prevention (CDC) criteria after lower limb fracture surgery. Methods. Data from two large, UK-based multicentre randomized controlled major trauma trials investigating SSI and wound healing after surgical repair of open lower limb fractures that could not be primarily closed (UK WOLLF), and surgical incisions for fractures that were primarily closed (UK WHiST), were examined. Trial interventions were standard wound care management and negative pressure wound therapy after initial surgical debridement. Wound outcomes were collected from 30 days to six weeks. We compared the level of agreement between wound photography and clinical assessment of CDC-defined SSI. We are also assessed the level of agreement between blinded independent assessors of the photographs. Results. Rates of CDC-defined deep SSI were 7.6% (35/460) after open fracture and 6.3% (95/1519) after closed incisional repair. Photographs were obtained for 77% and 73% of WOLLF and WHiST cohorts respectively (all participants n = 1,478). Agreement between photographic-SSI and CDC-SSI was fair for open fracture wounds (83%; k = 0.27 (95% confidence interval (CI) 0.14 to 0.42)) and for closed incisional wounds (88%; k = 0.29 (95% CI 0.20 to 0.37)) although the rate of photographically detected deep SSIs was twice as high as CDC-SSI (12% vs 6%). Agreement between different assessors for photographic-SSI (WOLLF 88%, k = 0.63 (95% CI 0.52 to 0.72); WHiST 89%; k = 0.61 (95% CI 0.54 to 0.69)); and wound healing was good (WOLLF 90%; k = 0.80 (95% CI 0.73 to 0.86); WHiST 87%; k = 0.57 (95% CI 0.50 to 0.64)). Conclusion. Although wound photography was feasible within the research context and inter-rater assessor agreement substantial, digital photographs used in isolation overestimated deep SSI rates, when compared to CDC criteria. Wound photography should not replace clinical assessment in pragmatic trials but may be useful for screening purposes where surgical infection outcomes are paramount. Cite this article: Bone Joint J 2021;103-B(12):1802–1808

Satisfactory primary wound healing following total joint replacement is essential. Wound healing problems can have devastating consequences for patients. Assessment of their healing capacity is useful in predicting complications. Local factors that influence wound healing include multiple previous incisions, extensive scarring, lymphoedema, and poor vascular perfusion. Systemic factors include diabetes mellitus, inflammatory arthropathy, renal or liver disease, immune compromise, corticosteroid therapy, smoking, and poor nutrition. Modifications in the surgical technique are necessary in selected cases to minimise potential wound complications. Prompt and systematic intervention is necessary to address any wound healing problems to reduce the risks of infection and other potential complications. Cite this article: Bone Joint J 2013;95-B, Supple A:144–7

Aims. Local antibiotics are used in the surgical management of foot infection in diabetic patients. This systematic review analyzes the available evidence of the use of local antibiotic delivery systems as an adjunct to surgery. Materials and Methods. Databases were searched to identify eligible studies and 13 were identified for inclusion. Results. Overall, the quality of the studies was poor. A single trial suggested that wound healing is quicker when a gentamicin-impregnated collagen sponge was implanted at time of surgery, with no difference in length of stay or rate of amputation. Results from studies with high risk of bias indicated no change in wound healing when a gentamicin-impregnated sponge was implanted during transmetatarsal amputation, but a reduction in the incidence of wound breakdown (8% vs 25%, not statistically significant) was identified. A significant cost reduction was identified when using an antimicrobial gel to deliver antibiotics and anti-biofilm agents (quorum-sensing inhibitors) compared with routine dressings and systemic antibiotics. Analyses of case series identified 485 patients who were treated using local antibiotic delivery devices. The rates of wound healing, re-operation, and mortality were comparable to those that have been previously reported for the routine management of these infections. Conclusion. There is a lack of good-quality evidence to support the use of local antibiotic delivery devices in the treatment of foot infections in patients with diabetes. Cite this article: Bone Joint J 2018;100-B:1409–15

In 42 amputations (28 below-knee and 14 above-knee), healing of the stump was found to correlate with the pre-operative transcutaneous oxygen pressure (tcPO2) measured at the planned site of incision. Of 31 patients with a tcPO2 of 30 mmHg or more, 27 healed primarily, three had delayed healing and one failed to heal; of seven patients with values between 20 and 30 mmHg three healed primarily and four had delayed healing; and all four patients with a value below 20 mmHg had stumps which failed to heal and needed reamputation. It is concluded that a tcPO2 of 20 mmHg or more at the intended site of amputation indicates that the wound will heal

We performed a prospective, blind, controlled study on wound infection after implant surgery involving 41 procedures in patients infected with the human immunodeficiency virus (HIV) and 141 in HIV-negative patients. The patients were staged clinically and the CD4 cell count determined. Wound infection was assessed using the asepsis wound score. A risk category was allocated to account for presurgical contamination.

In HIV-positive patients, with no preoperative contamination, the incidence of wound infection (3.5%) was comparable with that of the HIV-negative group (5%; p = 0.396). The CD4 cell count did not affect the incidence of infection (r = 0.16). When there was preoperative contamination, the incidence of infection in HIV-positive patients increased markedly (42%) compared with that in HIV-negative patients (11%; p = 0.084).

Our results show that when no contamination has occurred implant surgery may be undertaken safely in HIV-positive patients.

There are 33 million people worldwide currently infected with human immunodeficiency virus (HIV). This complex disease affects many of the processes involved in wound and fracture healing, and there is little evidence available to guide the management of open fractures in these patients. Fears of acute and delayed infection often inhibit the use of fixation, which may be the most effective way of achieving union.

This study compared fixation of open fractures in HIV-positive and -negative patients in South Africa, a country with very high rates of both HIV and high-energy trauma. A total of 133 patients (33 HIV-positive) with 135 open fractures fulfilled the inclusion criteria. This cohort is three times larger than in any similar previously published study.

The results suggest that HIV is not a contraindication to internal or external fixation of open fractures in this population, as HIV is not a significant risk factor for acute wound/implant infection. However, subgroup analysis of grade I open fractures in patients with advanced HIV and a low CD4 count (< 350) showed an increased risk of infection; we suggest that grade I open fractures in patients with advanced HIV should be treated by early debridement followed by fixation at an appropriate time.

We compared early post-operative rates of wound infection in HIV-positive and -negative patients presenting with open tibial fractures managed with surgical fixation.

The wounds of 84 patients (85 fractures), 28 of whom were HIV positive and 56 were HIV negative, were assessed for signs of infection using the ASEPIS wound score. There were 19 women and 65 men with a mean age of 34.8 years. A total of 57 fractures (17 HIV-positive, 40 HIV-negative) treated with external fixation were also assessed using the Checkett score for pin-site infection. The remaining 28 fractures were treated with internal fixation. No significant difference in early post-operative wound infection between the two groups of patients was found (10.7% (n = 3) vs 19.6% (n = 11); relative risk (RR) 0.55 (95% confidence interval (CI) 0.17 to 1.8); p = 0.32). There was also no significant difference in pin-site infection rates (17.6% (n = 3) vs 12.5% (n = 5); RR 1.62 (95% CI 0.44 to 6.07); p = 0.47).

The study does not support the hypothesis that HIV significantly increases the rate of early wound or pin-site infection in open tibial fractures. We would therefore suggest that a patient’s HIV status should not alter the management of open tibial fractures in patients who have a CD4 count > 350 cells/μl.

Cite this article: Bone Joint J 2013;95-B:1703–7.

Aims. The aim of this study was to explore risk factors for complications associated with dural tear (DT), including the types of DT, and the intra- and postoperative management of DT. Patients and Methods. Between 2012 and 2017, 12 171 patients with degenerative lumbar diseases underwent primary lumbar spine surgery. We investigated five categories of potential predictors: patient factors (sex, age, body mass index, and primary disease), surgical factors (surgical procedures, operative time, and estimated blood loss), types of DT (inaccessible for suturing/clipping and the presence of cauda equina/nerve root herniation), repair techniques (suturing, clipping, fibrin glue, polyethylene glycol (PEG) hydrogel, and polyglycolic acid sheet), and postoperative management (drainage duration). Postoperative complications were evaluated in terms of dural leak, prolonged bed rest, headache, nausea/vomiting, delayed wound healing, postoperative neurological deficit, surgical site infection (SSI), and reoperation for DT. We performed multivariable regression analyses to evaluate the predictors of postoperative complications associated with DT. Results. In total, 429/12 171 patients (3.5%) had a DT. Multivariable analysis revealed that PEG hydrogel significantly reduced the incidence of dural leak and prolonged bed rest, and that patients treated with sealants (fibrin glue and PEG hydrogel) significantly less frequently suffered from headache. A longer drainage duration significantly increased the incidence of headache, nausea/vomiting, and delayed wound healing. Headache and nausea/vomiting were significantly more prevalent in younger female patients. Postoperative neurological deficit and reoperation for DT significantly depended on the presence of cauda equina/nerve root herniation. A longer operative time was the sole independent risk factor for SSI and was also a risk factor for dural leak, prolonged bed rest, and nausea/vomiting. Conclusion. Sealants, particularly PEG hydrogel, may be useful in reducing symptoms related to cerebrospinal fluid leakage, whereas prolonged drainage may be unnecessary. Younger female patients should be carefully treated when DT occurs. Cite this article: Bone Joint J 2019;101-B:1115–1121

Aims. We report the medium-term outcomes of a consecutive series of 118 Zenith total ankle arthroplasties (TAAs) from a single, non-designer centre. Methods. Between December 2010 and May 2016, 118 consecutive Zenith prostheses were implanted in 114 patients. Demographic, clinical, and patient-reported outcome measures (PROMs) data were collected. The endpoint of the study was failure of the implant requiring revision of one or all of the components. Kaplan-Meier survival curves were generated with 95% confidence intervals (CIs) and the rate of failure calculated for each year. Results. Eight patients (ten ankles) died during follow-up, but none required revision. Of the surviving 106 patients (108 ankles: rheumatoid arthritis (RA), n = 15; osteoarthritis (OA), n = 93), 38 were women and 68 were men, with a mean age of 68.2 years (48 to 86) at the time of surgery. Mean follow-up was 5.1 years (2.1 to 9.0). A total of ten implants failed (8.5%), thus requiring revision. The implant survival at seven years, using revision as an endpoint, was 88.2% (95% CI 100% to 72.9%). There was a mean improvement in Manchester-Oxford Foot and Ankle Questionnaire (MOXFQ) from 85.0 to 32.8 and visual analogue scale (VAS) scores from 7.0 to 3.2, and overall satisfaction was 89%. The three commonest complications were malleolar fracture (14.4%, n = 17), wound healing (13.6%, n = 16), and superficial infection (12.7%, n = 15). The commonest reason for revision was aseptic loosening. No patients in our study were revised for deep infection. Conclusion. Our results show that Zenith survival rates are comparable with those in the literature for other implants and in the National Joint Registry (NJR). Overall patient satisfaction was high as were functional outcomes. However, the data highlight potential complications associated with this surgery. The authors believe that these figures support ankle arthroplasty as an option in the treatment of ankle arthritis. Cite this article: Bone Joint J 2021;103-B(4):696–703

We prospectively randomised 78 patients into two groups, ‘drains’ or ‘no drains’ to assess the effectiveness of suction drains in reducing haematoma and effusion in the joint and its effect on wound healing after total knee replacement. Ultrasound was used to measure the formation of haematoma and effusion on the fourth post-operative day. This was a semi-quantitative assessment of volume estimation. There was no difference in the mean effusion between the groups (5.91 mm in the drain group versus 6.08 mm in the no-drain, p = 0.82). The mean amount of haematoma in the no-drain group was greater (11.07 mm versus 8.41 mm, p = 0.03). However, this was not clinically significant judged by the lack of difference in the mean reduction in the post-operative haemoglobin between the groups (drain group 3.4 g/dl; no-drain group 3.0 g/dl, p = 0.38). There were no cases of wound infection or problems with wound healing at six weeks in any patient. Our findings indicate that drains do not reduce joint effusion but do reduce haematoma formation. They have no effect on wound healing

1. The incidence of failure of wound healing by primary intention due to infection, haematoma and gaping of wound edges, has been compared in 100 patients with rheumatoid arthritis and in 100 matched controls following a variety of orthopaedic operations. 2. There were thirty-one cases of failure of wound healing in the patients with rheumatoid arthritis and sixteen in the controls (P< ·02). There was no significant difference in the numbers of infected wounds or in those in which a haematoma developed. However, there was a significant excess of wound edge separation in the rheumatoid patients (thirteen) compared with the controls (four) (P< ·05). 3. The problems of wound healing were minor since there was no difference in the mean number of days ±1 standard deviation to wound healing between the patients: l6·6±7·5 days; and the controls: 15·2±7·9 days. 4. There was no correlation between duration or severity of rheumatoid disease and wound healing. 5. There was no difference in wound healing between patients with sero-positive and seronegative disease. 6. Forty-nine patients received corticosteroid therapy in small dosage. This was associated with an increased incidence of wound infection. Treatment for more than three years was associated with a significant increase in the mean number of days to wound healing. 7. The results are discussed in the light of the increased incidence of infection in several sites in patients with rheumatoid arthritis and of the effect of corticosteroid therapy on wound healing in man and experimental animals. It is suggested that more marked abnormalities might have been expected, and that these findings may need to be considered in the future surgical management of such patients

We describe a retrospective review of the clinical and radiological parameters of 32 feet in 30 patients (10 men and 20 women) who underwent correction for malalignment of the hindfoot with a modified double arthrodesis through a medial approach. The mean follow-up was 21 months (13 to 37). Fusion was achieved in all feet at a mean of 13 weeks (6 to 30). Apart from the calcaneal pitch angle, all angular measurements improved significantly after surgery. Primary wound healing occurred without complications. The isolated medial approach to the subtalar and talonavicular joints allows good visualisation which facilitated the reduction and positioning of the joints. It was also associated with fewer problems with wound healing than the standard lateral approach

Aims. The aim of this study was to describe the use of 3D-printed sacral endoprostheses to reconstruct the pelvic ring and re-establish spinopelvic stability after total en bloc sacrectomy (TES) and to review its outcome. Patients and Methods. We retrospectively reviewed 32 patients who underwent TES in our hospital between January 2015 and December 2017. We divided the patients into three groups on the basis of the method of reconstruction: an endoprosthesis group (n = 10); a combined reconstruction group (n = 14), who underwent non-endoprosthetic combined reconstruction, including anterior spinal column fixation; and a spinopelvic fixation (SPF) group (n = 8), who underwent only SPF. Spinopelvic stability, implant survival (IS), intraoperative haemorrhage rate, and perioperative complication rate in the endoprosthesis group were documented and compared with those of other two groups. Results. The mean overall follow-up was 22.1 months (9 to 44). In the endoprosthesis group, the mean intraoperative hemorrhage was 3530 ml (1600 to 8100). Perioperative complications occurred in two patients; both had problems with wound healing. After a mean follow-up of 17.7 months (12 to 38), 9/10 patients could walk without aids and 8/10 patients were not using analgesics. Imaging evidence of implant failure was found in three patients, all of whom had breakage of screws and/or rods. Only one of these, who had a local recurrence, underwent re-operation, at which solid bone-endoprosthetic osseointegration was found. The mean IS using re-operation as the endpoint was 32.5 months (95% confidence interval 23.2 to 41.8). Compared with the other two groups, the endoprosthesis group had significantly better spinopelvic stability and IS with no greater intraoperative haemorrhage or perioperative complications. Conclusion. The use of 3D-printed endoprostheses for reconstruction after TES provides reliable spinopelvic stability and IS by facilitating osseointegration at the bone-implant interfaces, with acceptable levels of haemorrhage and complications. Cite this article: Bone Joint J 2019;101-B:880–888

Aims. Survival rates and local control after resection of a sarcoma of the pelvis compare poorly to those of the limbs and have a high incidence of complications. The outcome for patients who need a hindquarter amputation (HQA) to treat a pelvic sarcoma is poor. Our aim was to evaluate the patient, tumour, and reconstructive factors that affect the survival of the patients who undergo HQA for primary or recurrent pelvic sarcoma. Methods. We carried out a retrospective review of all sarcoma patients who had undergone a HQA in a supraregional sarcoma unit between 1996 and 2018. Outcomes included oncological, surgical, and survival characteristics. Results. A total of 136 patients, with a mean age of 51 (12 to 83) underwent HQA, 91 for a bone sarcoma and 45 for a soft tissue sarcoma. The overall survival (OS) after primary HQA for a bone sarcoma was 90.7 months (95% confidence interval (CI) 64.1 to 117.2). In patients undergoing a secondary salvage HQA it was 90.3 months (95% CI 58.1 to 122.5) (p = 0.727). For those treated for a soft tissue sarcoma (STS), the mean OS was 59.3 months (95% CI 31.1 to 88.6) for patients with a primary HQA, and 12.5 months (95% CI 9.4 to 15.5) for those undergoing a secondary salvage HQA (p = 0.038). On multivariate analysis, high histological grade (hazard ratio (HR) 2.033, 95% CI 1.127 to 3.676; p = 0.018) and a diagnosis of STS (HR 1.653, 95% CI 1.027 to 2.660; p = 0.039) were associated with a poor prognosis. The 30-day mortality for patients with curative intent was 0.8% (1/128). For those in whom surgery was carried out with palliative intent it was 33.3% (2/6) (p = 0.001). In total, 53.7% (n = 73) of patients had at least one complication with 23.5% (n = 32) requiring at least one further operation. Direct closure was inferior to flap reconstruction in terms of complete primary wound healing (60.0% (3/5) vs 82.0% (82/100); p = 0.023). Conclusion. In carefully selected patients HQA is associated with satisfactory overall survival, with a low risk of perioperative mortality, but considerable morbidity. However, caution must be exercised when considering the procedure for palliation due to the high incidence of early postoperative mortality. Cite this article: Bone Joint J 2020;102-B(6):788–794

Aims. This study reports the outcomes of a technique of soft-tissue coverage and Chopart amputation for severe crush injuries of the forefoot. Patients and Methods. Between January 2012 to December 2016, 12 patients (nine male; three female, mean age 38.58 years; 26 to 55) with severe foot crush injury underwent treatment in our institute. All patients were followed-up for at least one year. Their medical records, imaging, visual analogue scale score, walking ability, complications, and functional outcomes one year postoperatively based on the American Orthopedic Foot and Ankle Society (AOFAS) and 36-Item Short-Form Health Survey (SF-36) scores were reviewed. Results. The mean length of follow-up was 18.6 months (13 to 28). Two patients had a local infection, flap necrosis was seen in one patient, and one patient experienced a skin graft wound healing delay. Of the 12 patients, one had persistent infection and eventually required below-knee amputation, but pain-free walking was achieved in all the other patients. The mean one-year postoperative AOFAS and SF-36 scores were 75.6 (68 to 80) and 82 (74 to 88), respectively. Conclusion. Although our sample size was small, we believe that this treatment method may be a valuable alternative for treating severe foot crush injuries. Cite this article: Bone Joint J 2018;100-B:1359–63

Aims. The classical longitudinal incision used for the direct anterior approach (DAA) to the hip does not follow the tension lines of the skin and can lead to impaired wound healing and poor cosmesis. The purpose of this retrospective study was to determine the satisfaction with the scar, and functional and radiographic outcomes comparing the classic longitudinal incision with a modified skin crease ‘bikini’ when the DAA is used for total hip arthroplasty (THA). Patients and Methods. A total of 964 patients (51% female; 59% longitudinal, 41% ‘bikini’) completed a follow-up questionnaire between two and four years postoperatively, including the Oxford Hip Score (OHS), the University of North Carolina ‘4P’ scar scale (UNC4P) and two items for assessing the aesthetic appearance of the scar and symptoms of numbness. The positioning of the components, rates of heterotopic ossification (HO) and rates of revision were assessed. Results. The mean OHS was similar in both groups (p = 0.41). The mean UNC4P total score was slightly better (p = 0.01) and the proportion of patients who were very satisfied with the cosmetic aspects of the scar was higher in the ‘bikini’ group (p < 0.001). The proportion of patients reporting numbness in the scar was higher in the longitudinal group (14.5% vs 7.5%, respectively, p < 0.001). The abduction angle of the acetabular component, the position of the stem and rates of HO did not differ between the groups. There were no differences in the revision rates of both groups, being 2.3% in the longitudinal and 1.5% in the ‘bikini’ group (p = 0.911). Conclusion. We found that a short oblique ‘bikini’ skin crease incision is safe when used for the DAA at THA, without compromising the positioning of the components or increasing the rate of lateral femoral cutaneous nerve dysaesthesia. Although it leads to a superior scar satisfaction, as it is less extensile, it should be used after having gained experience with the classic longitudinal incision. Cite this article: Bone Joint J 2018;100-B:853–61

Aims. To establish whether there was a consensus among the members of the Hip Society (HS) on the role of direct anterior approach (DAA) contemporary primary total hip arthroplasty (THA). Methods. An online survey was sent to all 112 active and senior members of the HS, to which 71 members responded. The survey was constructed to determine whether they believed that evidence-based medicine proves, in modern clinical practice, that the DAA has significant benefits compared to risks when contrasted with other approaches. In addition, they were asked if they currently used the DAA. Results. While only 16.9% (12/71) of respondents had been trained in a generic anterior approach during residency, 49.3% (35/71) had used the DAA in their clinical practice in the past or were using it at the present time. Unexpectedly, 42.9% (15/35) of respondents who had used the DAA in the past had abandoned it by the time of this survey. Only 22.5% (16/71) of all respondents believed that evidence-based medicine proves that the DAA has significant benefits compared to risks in contrast to other approaches. Conclusion. A comprehensive literature review found only three prospective randomized clinical trials (RCT) comparing the DAA with another approach with greater than one-year follow-up. Two showed minor benefits within the early postoperative period only, and one of those showed poorer mid-term results. Most of the published comparison studies with short follow-up show longer surgical times and greater blood loss for the DAA, and many three-month comparison studies show higher complication rates for the DAA using a proprietary traction table. The complications included problems with wound healing, lateral femoral cutaneous nerve injury, femoral component loosening, and femoral fractures. Because of the lack of evidence from RCTs showing superiority of the DAA over other approaches and reports of higher complications, the opinion of a large majority (77.5%; 55/71) of HS surgeons was that the DAA lacks sufficient evidence to warrant its use. Cite this article: Bone Joint J 2020;102-B(7 Supple B):57–61

1. Eighty consecutive open fractures of the tibial shaft were treated by primary internal fixation and wound closure. Wound healing was complicated by deep infection in eight patients (10 per cent) and by skin loss of varying degree in ten (l2·5 per cent). 2. Careful selection of patients on the basis of associated soft-tissue injuries is urged. A simple method of grading open fractures by the appearance of the wound and adjacent skin and the effectiveness of wound closure is suggested. If internal fixation is indicated on mechanical grounds, the nature of the soft-tissue injury should be the deciding factor in the choice of the method of treatment. In the less severe (Grade 1) fractures internal fixation and wound closure may be safely employed. In the severe (Grade 3) injuries, primary wound closure with or without internal fixation should be avoided. Moderately severe (Grade 2) fractures should be carefully assessed and treated by internal fixation and wound closure only if primary wound healing is confidently expected. 3. Wound healing by first intention requires, in addition to adequate debridement of the deep layers of the wound, careful approximation of healthy wound edges without excessive tension. An adequate knowledge of skin-plastic procedures is essential to achieve this. 4. A combination of systemic penicillin and streptomycin in adequate doses is a safe and effective prophylactic antibiotic for use in the treatment of open fractures

In this paper, we consider wound healing after total knee arthroplasty

Aims. Infected and deformed neuropathic feet and ankles are serious challenges for surgical management. In this study we present our experience in performing ankle arthrodesis in a closed manner, without surgical preparation of the joint surfaces by cartilaginous debridement, but instead using an Ilizarov ring fixator (IRF) for deformity correction and facilitating fusion, in arthritic neuropathic ankles with associated osteomyelitis. Methods. We retrospectively reviewed all the patients who underwent closed ankle arthrodesis (CAA) in Ilizarov Scientific Centre from 2013 to 2018 (Group A) and compared them with a similar group of patients (Group B) who underwent open ankle arthrodesis (OAA). We then divided the neuropathic patients into three arthritic subgroups: Charcot joint, Charcot-Maire-Tooth disease, and post-traumatic arthritis. All arthrodeses were performed by using an Ilizarov ring fixator. All patients were followed up clinically and radiologically for a minimum of 12 months to assess union and function. Results. The union rate for Group A was 81% (17/21) while it was 84.6% (33/39) for Group B. All the nonunions in Group A underwent revision with an open technique and achieved 100% union. Mean duration of IRF was 71.5 days (59 to 82) in Group A and 69 days (64.8 to 77.7) in Group B. The American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot score was similar in both groups. The postoperative hospital stay was shorter in Group A (21 days (SD 8)) than Group B (28 days (SD 9)). In the latter Group there were more problems with wound healing and greater requirement for antibiotic treatment. The mean operating time was 40 minutes (SD 9) in Group A compared to 80 minutes (SD 13) in Group B. Recurrence of infection occurred in 19% (4/21) and 15.5% (6/39) for Group A and Group B respectively. Conclusion. We found CAA using an IRF to be an effective method for ankle arthrodesis in infected neuropathic foot and ankle cases and afforded comparable results to open methods. Due to its great advantages, Ilizarov method of CAA should always be considered for neuropathic ankles in suitable patients. Cite this article: Bone Joint J 2020;102-B(4):470–477