Machine learning algorithms do not outperform preoperative thresholds in predicting clinically meaningful improvements after total knee arthroplasty.

PURPOSE: Patient-reported outcome measures (PROMs) are important measures of success after total knee arthroplasty (TKA) and being able to predict their improvements could enhance preoperative decision-making. Our study aims to compare the predictive performance of machine learning (ML) algorithms and preoperative PROM thresholds in predicting minimal clinically important difference (MCID) attainment at 2 years after TKA. METHODS: Prospectively collected data of 2840 primary TKA performed between 2008 and 2018 was extracted from our joint replacement registry and split into a training set (80%) and test set (20%). Using the training set, ML algorithms were developed using patient demographics, comorbidities and preoperative PROMs, whereas the optimal preoperative threshold was determined using ROC analysis. Both methods were used to predict MCID attainment for the SF-36 PCS, MCS and WOMAC at 2 years postoperatively, with predictive performance evaluated on the independent test set. RESULTS: ML algorithms and preoperative PROM models performed similarly in predicting MCID for the SF-36 PCS (AUC: 0.77 vs 0.74), MCS (AUC: 0.95 vs 0.95) and WOMAC (AUC: 0.89 vs 0.88). For each outcome, the most important predictor of MCID attainment was the patient's preoperative PROM score. ROC analysis also identified optimal preoperative threshold values of 33.6, 54.1 and 72.7 for the SF-36 PCS, MCS and WOMAC, respectively. CONCLUSION: ML algorithms did not perform significantly better than preoperative PROM thresholds in predicting MCID attainment after TKA. Future research should routinely compare the predictive ability of ML algorithms with existing methods and determine the type of clinical problems which may benefit the most from it. LEVEL OF EVIDENCE: II.

Do Patients With Diabetes Have Poorer Improvements in Patient-Reported Outcomes After Total Knee Arthroplasty?

BACKGROUND: Diabetes is one of the most common comorbidities in patients undergoing total knee arthroplasty (TKA) for osteoarthritis. However, the evidence remains unclear on how it affects patient-reported outcome measures after TKA. METHODS: We reviewed prospectively collected data of 2840 patients who underwent primary unilateral TKA between 2008 and 2018, of which 716 (25.2%) had diabetes. All patients had their HbA1c measured within 1 month before surgery, and only well-controlled diabetics (HbA1c <8.0%) were allowed to proceed with surgery. Patient demographics and comorbidities were recorded, and multiple regression was performed to evaluate the impact of diabetes on improvements in patient-reported outcome measures (Short Form 36 (SF-36), Western Ontario and McMaster University Osteoarthritis Index (WOMAC), Knee Society Score (KSS)) and knee range of motion (ROM). RESULTS: Compared with nondiabetics, patients with diabetes were more likely to possess a higher body mass index (P-value <.001), more comorbidities (P-value <.001), and poorer preoperative SF-36 Physical Component Summary (PCS) (P-value <.001), WOMAC (P-value = .002), KSS-function (P-value <.001), and knee ROM (P-value <.001). Multiple regression showed that diabetic patients experienced marginally poorer improvements in KSS-knee (-1.22 points, P-value = .025) and knee ROM (-1.67°, P-value = .013) than nondiabetics. However, there were no significant differences in improvements for SF-36 PCS (P-value = .163), Mental Component Summary (P-value = .954), WOMAC (P-value = .815), and KSS-function (P-value = .866). CONCLUSION: Patients with well-controlled diabetes (HbA1c <8.0%) can expect similar improvements in general health and osteoarthritis outcomes (SF-36 PCS and Mental Component Summary, WOMAC, and KSS-function) compared with nondiabetics after TKA. Despite having marginally poorer improvements in knee-specific outcomes (KSS-knee and knee ROM), these differences are unlikely to be clinically significant.

Short-duration chemoprophylaxis might reduce incidence of deep vein thrombosis in Asian patients undergoing total knee arthroplasty.

BACKGROUND: Venous thromboembolism (VTE) is a serious complication that may occur after total knee arthroplasty (TKA), leading to the recommendation of routine chemoprophylaxis by international guidelines. This study aims to determine if short-duration chemoprophylaxis after TKA reduces the incidence of VTE in an Asian population. METHODS: A retrospective study of 316 patients who underwent unilateral primary TKA between 1 January 2011 and 31 December 2013 was conducted. All patients received mechanical prophylaxis. One hundred seventeen patients (37%) received additional chemoprophylaxis, whereas 199 patients (63%) did not. A Doppler ultrasound (DUS) of both lower limbs was conducted for all patients within 6 days after surgery (median = 3 days) to assess for both proximal and distal DVT. Chemoprophylaxis in the form of enoxaparin (low molecular weight heparin; LMWH), aspirin, or heparin was administered until patients had a normal DUS, for a median duration of 4 days. Patients were followed up clinically for a minimum of 6 months to monitor for delayed or recurrent VTE and at least 2 years for patient-reported outcome measures. RESULTS: Overall, 24 patients (7.59%) developed deep vein thrombosis (DVT): three proximal and 21 distal DVTs. Twenty-three of the 24 patients were asymptomatic. Twenty of 199 patients (10.05%) with only mechanical prophylaxis developed DVT, whereas four of 117 patients (3.42%) with additional chemoprophylaxis developed DVT. Multivariate analysis showed that chemoprophylaxis use was associated with reduced incidence of DVT (odds ratio = 0.19, p value = 0.011). Other factors associated with increased DVT incidence include female gender (odds ratio = 5.45, p value = 0.034), positive history of cancer (odds ratio = 5.14, p value = 0.044), and increased length of stay in hospital (odds ratio = 1.19, p value < 0.001). CONCLUSIONS: Our study has shown that despite the low incidence of DVT in Asian patients undergoing TKA, short-duration chemoprophylaxis might be effective in reducing the incidence of DVT. However, most DVTs observed in our study were distal and may be of limited clinical significance. Further studies are needed to investigate the impact of chemoprophylaxis use on the incidence of PE and overall mortality rates among Asian patients.

Male orthopaedic surgeons and anaesthetists: equally good at estimating fluid volumes (and changing light bulbs) but equally poor at estimating procedure duration.

INTRODUCTION: How many orthopods does it take to change a light bulb? One - to refer to the medics for 'Darkness ?Cause'. Additionally, anaesthetists and surgeons often disagree on the estimated blood loss during surgery and the estimated procedure duration. We designed this study to compare the ability of orthopaedic surgeons and anaesthetists in: (a) estimating fluid volumes; (b) estimating procedure durations; and (c) changing light bulbs. METHODS: Participants had to either be a specialist in anaesthesia or orthopaedic surgery, or a trainee in that specialty for at least two years. Three different fluid specimens were used for volume estimation (44 mL, 88 mL and 144 mL). Two videos of different lengths (140 seconds and 170 seconds), showing the suturing of a banana skin, were used for procedure duration estimation. To determine the ability at changing light bulbs, the participants had to match eight different light sockets to their respective bulbs. RESULTS: 30 male anaesthetists and trainees and 31 male orthopaedic surgeons and trainees participated in this study. Orthopaedic surgeons underestimated the three fluid volumes by 3.9% and anaesthetists overestimated by 5.1% (p = 0.925). Anaesthetists and orthopaedic surgeons overestimated the duration of the two procedures by 21.2% and 43.1%, respectively (p = 0.006). Anaesthetists had a faster mean time in changing light bulbs (70.1 seconds vs. 74.1 seconds, p = 0.319). CONCLUSION: In an experimental environment, male orthopaedic surgeons are as good as male anaesthetists in estimating fluid volumes (in commonly seen surgical specimens) and in changing light bulbs. Both groups are poor at estimating procedure durations.

Postoperative venous thrombotic events in Asian elderly patients with surgically treated hip fractures with and without chemoprophylaxis.

BACKGROUND AND PURPOSE: Chemothromboprophylaxis for hip fracture surgery has been under-practiced in Asia. We investigate the postoperative incidence of venous thrombotic events (VTE) in elderly Asian patients with hip fractures. PATIENTS AND METHODS: Patients over 60 years old with surgically treated hip fractures were reviewed. All patients followed a hip fracture protocol, which included a post-operative ultrasound Doppler venous scan of both lower limbs five days after injury to detect any deep vein thrombosis (DVT). All patients were given mechanical prophylaxis. Patients who had additional chemoprophylaxis were included. RESULTS: From 2010 to 2012, 454 patients were reviewed. The overall incidence of DVT was 6.4% (29 patients). The incidence of PE was 1.3% (six patients). There was no significant correlation between DVT and age, gender, race, hypertension, diabetes mellitus, end stage renal failure, type of fracture and operation, preoperative haemoglobin, platelet count or urea and creatinine levels. Sixteen DVTs (55.2%) were on the ipsilateral side as the fracture, eight (27.6%) were on the contralateral side and five (17.2%) were bilateral. Chemoprophylaxis was not given to 399 patients, of which 27 (6.8%) developed DVT. Ten (2.5%) were proximal DVTs in which four (1%) developed PE. One (0.25%) patient developed PE without DVT. Fifty-five patients were on chemoprophylaxis, of which two (3.6%) developed DVT. Both were proximal DVTs with one subsequently developing PE. CONCLUSIONS: The incidence of VTE in Asian patients may be lower compared to Western populations when mechanical prophylaxis is used. Chemoprophylaxis may assist in reducing the rates of DVT but not PE. The low incidence may not justify the use of routine chemoprophylaxis.

Early versus late flap coverage for open tibial fractures.

PURPOSE: To compare early versus late flap coverage for open tibial fractures. METHODS: Medical records of 83 men and 6 women (mean age, 38 years) who underwent fixation for open tibial fractures (Gustilo grades IIIB and IIIC) followed by flap coverage within (n=30) or after (n=59) 72 hours were reviewed. All fractures were treated within 24 hours. Outcome measures included bone union, infection, flap failure, the need for secondary procedures to achieve union, and eventual amputation. The early and late flap coverage groups were compared. RESULTS: Early flap coverage was associated with shorter length of hospitalisation (31.4 vs. 55.8 days, p<0.01), lower deep infection rates (23% vs. 54%, p<0.01), and smaller number of surgical procedures (6.4 vs. 9.2, p=0.01). The 2 groups did not differ significantly in terms of the time to bone union, flap failure, amputation, and the need of secondary procedures to facilitate bone union. CONCLUSION: In severe open tibial fractures, early soft-tissue coverage (within 72 hours) was associated with more favourable outcomes in terms of length of hospitalisation and infection.

Total hip replacement using a highly crosslinked polyethylene liner in Asians with small acetabulum.

PURPOSE: To evaluate the medium-term outcomes of total hip replacement (THR) using a thin highly cross-linked polyethylene (HXLPE) liner in an Asian population. METHODS: Medical records of 20 men and 44 women aged 26 to 80 (mean, 59) years who underwent 80 THRs using a thin HXLPE liner by a single surgeon were reviewed. Indications for THR included dysplasia (n = 26), avascular necrosis (n=26), osteoarthritis (n = 22), rheumatoid arthritis (n = 4), and ankylosing spondylitis (n = 2). The surgical technique, implant used, and rehabilitation protocol were standardised. Radiographs were evaluated for cup migration and peri-implant radiolucency. Femoral head penetration was assessed at day 1 and last follow-up using the Kang modification of the Dorr and Wan method and the PowerPoint method. RESULTS: After a mean follow-up of 7.5 (range, 5.0-11.9) years, no patient had undergone acetabular revision. Two hips were revised for the femoral stem. Femoral head penetration did not correlate with sex, age at surgery, hip pathology, acetabular inclination angle, cup size, or liner thickness. The maximum femoral head penetration was 0.09 mm/year, which was below the osteolysis threshold of 0.1 mm/year. The mean ± standard deviation femoral head penetration was 0.29 ± 0.12 mm (or 0.04 ± 0.02 mm/year) by the PowerPoint method and 0.17 ± 0.22 mm (or 0.02 ± 0.03 mm/year) by the Kang modification of the Dorr and Wan method. Femoral head penetration correlated with the number of years of follow-up (p = 0.012). CONCLUSION: The use of a thin HXLPE liner in primary THR for patients with small acetabulum achieved good outcomes after a mean of 7.5 years.

Epidemiological analysis of outcomes in 323 open tibial diaphyseal fractures: a nine-year experience.

INTRODUCTION: Open fractures of the tibia pose a challenge to orthopaedic and plastic surgeons. A retrospective observational review was conducted to evaluate the epidemiological factors and fracture outcomes in the Singapore context. METHODS: A nine-year period of open tibial shaft fractures presenting to our institution was reviewed. Demographic and management data were recorded. Statistical analysis was performed on the outcomes of length of hospital stay, number of operations, time to union and infection rates. RESULTS: 323 fractures met our inclusion criteria (Gustilo [G] 1=53, G2=100, G3=170). Mean age of patients was 36.5 years, 91.3% were male and 40.9% were non-Singaporeans. 69.3% of fractures occurred from road traffic accidents and 21.7% from industrial accidents. Mean length of hospital stay was 28.7 days and number of operations was 4.29. Time to union was 10.7 months and overall infection rate was 20.7%. Infection rates were significantly higher in G3b/G3c compared to G3a (45.7% vs. 21.1%) patients. There was no significant reduction in infection rates when open tibial fractures were operated on within six hours of admission. Multiple injured patients required a longer time to union and hospital stay. There was an exponential cost increase with greater severity of fracture. CONCLUSION: High Gustilo and AO classification injuries positively correlate with high non-union and infection rates, requiring multiple operations and long hospital stay. There is no benefit in performing surgery on open tibial fractures within six hours of presentation. A significant proportion of these patients would be polytraumatised, indirectly affecting fracture union.

Hybrid surgical management of a saccular aneurysm of the internal jugular vein.

Internal jugular vein saccular aneurysms are rare. Surgical management is indicated when there is a risk of rupture. We recommend a hybrid procedure combining endovascular and open techniques. We report the case of a 77-year-old female patient with a saccular aneurysm of the internal jugular vein, which was excised with combined endovascular and vascular techniques.

Electronic software significantly improves quality of handover in a London teaching hospital.

The study compared paper-based and electronic-based medical handover with respect to quality of information transfer during hospital out-of-hours shifts, and analysed the caseload burden of the hospital night team. The participants were 1645 hospital patients transferred from the day team to the out-of-hours team over four months in 2006. Quality of information transfer was determined by clinical data handed over within pre-set fields, and caseload burden by the frequency of tasks required at handover. Handover information fields comprised patient's demographics and location, primary diagnosis, current problem, plan of action and primary care team details. Electronic handover achieved a significantly higher number of completed fields than paper-based handover. Blood collection, checking blood and X-ray results, and adjusting fluid balance and intravenous cannulation account for most of the workload during the nightshift. Electronic handover provides better continuity of care than paper-based handover, and redistribution of tasks during working hours would benefit night staff.

Effect of inhibitors of the renin-angiotensin system and other antihypertensive drugs on renal outcomes: systematic review and meta-analysis.

BACKGROUND: A consensus has emerged that angiotensin-converting-enzyme (ACE) inhibitors and angiotensin-II receptor blockers (ARBs) have specific renoprotective effects. Guidelines specify that these are the drugs of choice for the treatment of hypertension in patients with renal disease. We sought to determine to what extent this consensus is supported by the available evidence. METHODS: Electronic databases were searched up to January, 2005, for randomised trials assessing antihypertensive drugs and progression of renal disease. Effects on primary discrete endpoints (doubling of creatinine and end-stage renal disease) and secondary continuous markers of renal outcomes (creatinine, albuminuria, and glomerular filtration rate) were calculated with random-effect models. The effects of ACE inhibitors or ARBs in placebo-controlled trials were compared with the effects seen in trials that used an active comparator drug. FINDINGS: Comparisons of ACE inhibitors or ARBs with other antihypertensive drugs yielded a relative risk of 0.71 (95% CI 0.49-1.04) for doubling of creatinine and a small benefit on end-stage renal disease (relative risk 0.87, 0.75-0.99). Analyses of the results by study size showed a smaller benefit in large studies. In patients with diabetic nephropathy, no benefit was seen in comparative trials of ACE inhibitors or ARBs on the doubling of creatinine (1.09, 0.55-2.15), end-stage renal disease (0.89, 0.74-1.07), glomerular filtration rate, or creatinine amounts. Placebo-controlled trials of ACE inhibitors or ARBs showed greater benefits than comparative trials on all renal outcomes, but were accompanied by substantial reductions in blood pressure in favour of ACE inhibitors or ARBs. INTERPRETATION: The benefits of ACE inhibitors or ARBs on renal outcomes in placebo-controlled trials probably result from a blood-pressure-lowering effect. In patients with diabetes, additional renoprotective actions of these substances beyond lowering blood pressure remain unproven, and there is uncertainty about the greater renoprotection seen in non-diabetic renal disease.