Open access Protocol Rationale and design of the HIP fracture Accelerated surgical TreaTment And Care tracK (HIP ATTACK) Trial: a protocol for an international randomised controlled trial evaluating early surgery for hip fracture patients Flavia K Borges, 1 Mohit Bhandari,2 Ameen Patel,3 Victoria Avram,2 Ernesto Guerra-Farfán,4 Alben Sigamani,5 Masood Umer,6 Maria Tiboni,3 Anthony Adili,2 John Neary,3 Vikas Tandon,3 Parag K Sancheti,7 AbdelRahman Lawendy,8 Richard Jenkinson,9,10 Mmampapatla Ramokgopa,11,12 Bruce M Biccard,13 Wojciech Szczeklik,14 Chew Yin Wang,15 Giovanni Landoni,16,17 Patrice Forget,18 Ekaterine Popova, 19 Gavin Wood,20 Aamer Nabi Nur,21 Bobby John,22 Paweł Ślęczka,23 Robert J Feibel,24 Mariano Balaguer-Castro, 25 Benjamin Deheshi,26 Mitchell Winemaker,2 Justin de Beer,2 Richard Kolesar,27 Jordi Teixidor-Serra,4 Jordi Tomas-Hernandez,4 Michael McGillion,28 Harsha Shanthanna,27 Iain Moppett,29 Jessica Vincent,1 Shirley Pettit,1 To cite: Borges FK, Valerie Harvey,1 Leslie Gauthier,30 Kim Alvarado,28 P J Devereaux1,31 Bhandari M, Patel A, et al. Rationale and design of the HIP fracture Accelerated surgical TreaTment And Care Abstract tracK (HIP ATTACK) Trial: a Strengths and limitations of this study Introduction Annually, millions of adults suffer hip protocol for an international fractures. The mortality rate post a hip fracture is 7%– randomised controlled trial ►► HIP fracture Accelerated surgical TreaTment And evaluating early surgery for hip 10% at 30 days and 10%–20% at 90 days. Observational Care tracK (HIP ATTACK) is the first large randomised fracture patients. BMJ Open data suggest that early surgery can improve these controlled trial powered to determine the effects of 2019;9:e028537. doi:10.1136/ outcomes in hip fracture patients. We designed a clinical accelerated surgery compared with the standard of bmjopen-2018-028537 trial—HIP fracture Accelerated surgical TreaTment care in hip fracture patients. And Care tracK (HIP ATTACK) to determine the effect of ►► Prepublication history and ►► HIP ATTACK trial implemented patient engagement accelerated surgery compared with standard care on the additional material for this strategies, including research governance, trial out- paper are available online. To 90-day risk of all-cause mortality and major perioperative come evaluation and knowledge translation. view these files, please visit complications. ►► Patients, healthcare providers and study personnel the journal online (http://​dx.​doi.​ Methods and analysis HIP ATTACK is a multicentre, are unblinded to patient treatment allocation; how- org/​10.​1136/​bmjopen-​2018-​ international, parallel group randomised controlled ever, outcome adjudicators are blinded to treatment 028537). trial (RCT) that will include patients ≥45 years of age allocation. and diagnosed with a hip fracture from a low-energy Received 12 December 2018 ►► HIP ATTACK will only inform the effect of accelerated mechanism requiring surgery. Patients are randomised to Revised 11 February 2019 surgery versus standard care during hospital work- accelerated medical assessment and surgical repair (goal Accepted 21 February 2019 ing hours and does not inform the effects outside of within 6 h) or standard care. The co-primary outcomes working hours. are (1) all-cause mortality and (2) a composite of major perioperative complications (ie, mortality and non-fatal © Author(s) (or their myocardial infarction, pulmonary embolism, pneumonia, employer(s)) 2019. Re-use sepsis, stroke, and life-threatening and major bleeding) at journal, conference presentations, engagement of permitted under CC BY-NC. No 90 days after randomisation. All patients will be followed influential medical organisations, and providing public commercial re-use. See rights up for a period of 1 year. We will enrol 3000 patients. awareness through multimedia resources. and permissions. Published by Trial registration number NCT02027896; Pre-results. BMJ. Ethics and dissemination All centres had ethics approval before randomising patients. Written informed consent is For numbered affiliations see required for all patients before randomisation. HIP ATTACK end of article. Introduction is the first large international trial designed to examine Correspondence to whether accelerated surgery can improve outcomes Worldwide, millions of adults suffer a hip Dr P J Devereaux; in patients with a hip fracture. The dissemination plan fracture annualy.1 A hip fracture results in ​philipj@​mcmaster.​ca includes publishing the results in a policy-influencing trauma, pain, bleeding and immobility. These Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 1 Open access factors may trigger inflammation, hypercoagulability, Box 1 Exclusion criteria of the HIP ATTACK trial catabolism and stress,2–5 which can precipitate periopera- tive complications. The most commonly reported causes Patients fulfilling any of the following criteria are excluded: of short-term mortality after a hip fracture are coronary ►► Requiring emergent surgery or emergent interventions for anoth- heart disease, stroke, pneumonia, sepsis and pulmonary er reason (eg, subdural hematoma, abdominal pathology requiring embolism.6 7 The mortality rate post a hip fracture is 7% urgent laparotomy, acute limb ischemia, other fractures or trauma to 10% at 30 days and 10% to 20% at 90 days.8–13 requiring emergent surgery, necrotising fasciitis, coronary revascu- The impact of early surgery on the risk of periopera- larisation, pacemaker-implantation). tive complications and mortality in hip fracture patients ►► Open hip fracture. was evaluated in a systematic review and meta-analysis of ►► Bilateral hip fractures. ►► Peri-prosthetic fracture. observational studies.14 Earlier surgery was associated with ►► Therapeutic anticoagulation not induced by a vitamin K antagonist, a significant reduction in mortality (relative risk [RR], unfractionated heparin (eg, any administration of therapeutic low 0.81; 95% CI 0.68 to 0.96; p=0.01) in five studies (4208 molecular weight heparin [>6000 u/24 hours] in the 24 hours prior patients,721 deaths). Earlier surgery was also associated to enrolment) or intake of any other oral anticoagulant(s) for which with reduced risk of pressure sores (RR, 0.48; 95% CI 0.34 there is no reversing agent available. to 0.69; p<0.001) and in-hospital pneumonia (RR, 0.59; ►► Taking a therapeutic vitamin K antagonist with a history of hepa- 95% CI 0.37 to 0.93; p=0.02).14 rin-induced thrombocytopaenia. The HIP fracture Accelerated surgical TreaTment And ►► Refusing participation. Care tracK (HIP ATTACK) Pilot Trial included 60 patients ►► Previously enrolled in the trial. and established the feasibility of a trial of accelerated HIP ATTACK , HIP fracture Accelerated surgical TreaTment And Care tracK. surgery in patients with a hip fracture. Among patients randomised to accelerated surgery, 30% had a major perioperative complication (ie, mortality and non-fatal that have >80% of their hip fracture patients undergoing preoperative myocardial infarction [MI], myocardial surgery within 48 hours. injury after noncardiac surgery, pulmonary embolism, pneumonia, stroke, and life-threatening and major Patient recruitment bleeding) within 30 days of randomisation as compared Emergency department physicians and nurses receive with 47% of the patients allocated to standard care (HR , a trial in-service, during which we encourage them to 0.60; 95% CI 0.26 to 1.39; p=0.23).15 triage patients with a potential hip fracture for rapid We designed the HIP ATTACK Trial to determine the assessment during working hours, similar to how patients effect of accelerated medical clearance and accelerated with a potential MI or stroke are rapidly assessed. The surgery compared with standard care on the 90-day radiology department expedites imaging of all potential risk of the following two co-primary outcomes: all-cause hip fractures during working hours. Immediately on diag- mortality and major perioperative complications. nosing a hip fracture, the emergency department physi- cian consults the orthopaedic team on call and informs the HIP ATTACK research team about the patient. After reviewing the films and confirming a hip fracture Methods and analysis requiring surgical intervention, the orthopaedic surgeon Trial design immediately informs the study personnel. Research The HIP ATTACK Trial is a multicentre international, personnel approach all eligible patients to participate in parallel group randomised controlled trial (RCT) of 3000 the trial. patients with a hip fracture that requires a surgical inter- vention. Patients are randomised to accelerated medical Randomisation and blinding assessment and surgical repair (ie, goal of surgery within Randomisation occurs immediately after a patient 6 hours of hip fracture diagnosis) or standard care. is deemed eligible and written informed consent is obtained. Research personnel randomise the patients via Trial population an Interactive Web Randomisation System (IWRS). The We include patients ≥45 years of age who were diagnosed IWRS is a 24 hours computerised randomisation internet with a hip fracture during working hours, due to a low-en- system maintained by the coordinating centre at the ergy mechanism, and requiring surgery. All centres are Population Health Research Institute (PHRI), which is able to define their own study working hours according to part of Hamilton Health Sciences and McMaster Univer- the feasibility of randomising patients to the accelerated sity in Hamilton, Ontario, Canada. surgery within 6 hours from diagnosis. Box 1 reports the The randomisation process uses block randomisation exclusion criteria. stratified by the centre and by the type of planned surgery Currently, across Canada, 80%–90% of patients with (open reduction and internal fixation; or arthroplasty). a hip fracture undergo hip surgery within 48 hours after We use randomly varying block sizes; the study personnel the diagnosis.16 To minimise the variation in the timing of and investigators are not aware of the exact sizes. We surgery between centres, we have only included centres randomise patients in a 1:1 fashion to receive accelerated 2 Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 Open access Figure 1 The HIP ATTACK RCT flow chart. ECG, electrocardiogram; HIP ATTACK, HIP fracture Accelerated surgical TreaTment And Care tracK; NPO, nil per os; OR, operating room; RCT, randomised controlled trial. medical clearance and accelerated surgery versus standard randomised to accelerated care (ie, medical clearance care (figure 1). The randomisation procedure ensures and surgery), who are therapeutically anticoagulated with concealment for the purpose of minimising bias. Due to a vitamin K antagonist, receive prothrombin complex the nature of the trial, it is not possible to blind research concentrate to target an International Normalised Ratio personnel, participants or care providers involved in (INR) <1.5. a patient’s care. Outcome assessors are blinded to the Patients randomised to accelerated care, after obtaining trial intervention. medical clearance, move into the next orthopaedic trauma room or elective operating room slot depending Trial intervention on availability (ie, they are prioritised over scheduled Patients randomised to accelerated care undergo medical elective cases). In centres with a dedicated trauma room, clearance by a medical specialist (ie, internist, geriatri- there is minimal impact to the workflow with case prior- cian, cardiologist or anaesthesiologist), who is available to quickly arrive in the emergency department for the ities being adjusted to accommodate the HIP ATTACK assessment. This specialist uses his/her own individual case booking. In addition, on evenings or weekends, HIP judgement regarding management when considering ATTACK patients are prioritised over other non-urgent any medical conditions identified, and weighs the poten- emergency cases. Immediately after medical clearance tial benefits of delaying surgery for medical management is obtained, research personnel inform all the relevant versus the potential negative consequences of protracted stakeholders (ie, surgical booking clerk, orthopaedic exposure to the inflammatory, hypercoagulable, stress surgeon and anaesthesiologist) to facilitate the exchange and catabolic states associated with a hip fracture. The of the elective and the accelerated hip fracture case. The medical specialist is aware of all the conditions that the scheduled elective cases shift a slot forward, and there- trial consensus group believe are likely to benefit from fore they occur a few hours later than originally planned. medical optimisation before surgery (box 2). Patients randomised to standard care undergo medical Following medical clearance, the orthopaedic surgeon clearance based on local standard practices. After the and anaesthesiologist need to agree that the patient is patient is medically cleared, he/she is waitlisted for appropriate for surgery for the case to proceed. Patients surgery according to local standard practices. Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 3 Open access that they have experienced an outcome, the study team Box 2 Conditions likely to benefit from medical obtains the appropriate documentation. optimisation prior to surgery To accurately capture perioperative MI, we obtain daily ►► Acute myocardial infarction associated with a mechanical complica- troponin measurements until day 7 after randomisation. tion or ST-elevation myocardial infarction. Research personnel screen all patients for postopera- ►► Cardiac arrest. tive delirium applying the confusion assessment method ►► Cardiogenic shock or frank pulmonary oedema. (CAM)17 18 daily from day 1 to 7 after randomisation. ►► Respiratory failure requiring mechanical ventilation. Study personnel administer the short form quality of life ►► Known pulmonary artery hypertension (>80 mmHg). (SF-36) questionnaire19 to address patients’ quality of life ►► Home oxygen therapy with concomitant clopidogrel. at baseline, 30 days, and 1 year after randomisation. Func- ►► Presumptive bacteremia. tional independence measure (FIM) motor domain is ►► Hereditary or acquired coagulopathy that cannot be corrected within determined at 30 days and 1 year, as validated in hip frac- 2 hours to an INR <1.5. 9 ture patients.20 The phone administration of the SF-36 ►► Thrombocytopaenia (platelets <75 ×10 /L) of unknown origin that cannot be corrected within 2 hours or in case of known chronic questionnaire has also been validated in hip arthroplasty thrombocytopaenia with Platelets <50 ×109/L. patients.21 Research personnel are trained in admin- ►► Deep venous thrombosis in the last month requiring implantation of istering the CAM, FIM and SF-36. Research personnel vena cava filter. record all the trial data on case report forms with infor- ►► Acute stroke within 7 days of fracture. mation entered directly into an electronic data capture ►► Subarachnoid haemorrhage within 1 month of fracture. programme (iDataFax). ►► Impaired consciousness of unknown origin (GCS<12). ►► Fractures during seizure without known history of epilepsy. Trial outcomes ►► Na<120 mmol/L or >155 mmol/L; or Na<125 mmol/L There are two primary outcomes: (1) all-cause mortality or >150 mmol/L with neurological symptoms. at 90 days after randomisation and (2) composite of major ►► K>5.5 mmol/L with QRS-complex >120 ms in patients without perioperative complications (ie, mortality, and non-fatal known previous QRS-complex >120 ms or K<2.8 mmol/L not amenable to correction within 2 hours. MI, pulmonary embolism, pneumonia, sepsis, stroke, ►► pH<7.15, not amenable to correction within 2 hours. and life-threatening and major bleeding) at 90 days after ►► Indication for acute dialysis. randomisation. Individual secondary outcomes at 90 days after rando- GCS, Glasgow Coma Scale; INR, International normalised ratio. misation include all-cause mortality, vascular mortality, non-vascular mortality, MI, myocardial injury after rando- misation not meeting the third universal definition of The central data management team monitors data MI),22–24 cardiac revascularisation procedure (ie, percu- quality, adherence to the trial intervention and provides taneous coronary intervention or coronary artery bypass a feedback to local investigators to ensure adherence to grafting surgery), congestive heart failure, new clinically the protocol. important atrial fibrillation, non-fatal cardiac arrest, Co-interventions stroke, peripheral arterial thrombosis, pulmonary embo- For patients undergoing arthroplasty, the choice of lism, deep venous thrombosis, pneumonia, sepsis, infec- the surgical implant is left to the surgeon’s discretion tion, life-threatening bleeding, major bleeding, acute in both accelerated and standard care groups. All other kidney injury, new acute renal failure resulting in dial- perioperative management (monitoring, fluids, type of ysis, peri-prosthetic fracture, prosthetic hip dislocation, anaesthesia, analgesia and transfusions) and postopera- implant failure, hip re-operation, time to first mobilisa- tive care are at the discretion of the attending anaesthesi- tion, length of hospital stay, length of critical care stay, ologist, surgeon and medical specialist. Study personnel length of rehabilitation stay, new residence in a nursing record data on co-interventions. Study investigators home, new pressure ulcers and persistent post-surgical strongly encourage appropriate venous thromboembo- pain. Online supplementary appendix 1 describes all lism prophylaxis in all randomised patients. We also advo- outcome definitions. cate early mobilisation within 12 hours of hip surgery in The FIM motor domain and its mobility and locomo- all randomised patients, unless medically or surgically tion subscores, and the SF-36 score are assessed at 30 contraindicated. days after randomisation.19 25 An additional secondary outcome is delirium up to 7 days after randomisation. We Follow-up determine the presence of delirium using CAM, which is All trial patients receive the same structured follow-up a validated tool for the detection of delirium in elderly assessment. Research personnel follow patients hospitalised patients.17 18 For the diagnosis of delirium, throughout their time in hospital evaluating them, the CAM requires acute fluctuating changes in mental reviewing their medical records, ensuring trial orders status, including inattention, incoherent thoughts and are followed and noting any outcomes. The research alterations in the consciousness level.17 personnel contact the patients by telephone at 30 days, 90 The motor domain of FIM consists of 13 items each days and 1 year after randomisation. If patients indicate scored 1 to 7.26 The motor domain scores range from 13 4 Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 Open access to 91, with high scores indicating higher function. The Table 1 Sample size calculations: sample size fixed at SF-36 measures health-related quality of life by scoring 3000 and α for the first co-primary outcome (mortality) fixed eight domains (physical function, role limitation due at 0.04 and the second co-primary outcome calculated at to physical health, pain, general health perception, 0.012 vitality, social function, role limitations due to emotional Power calculation health and mental health) from 0 to 100. High scores Control event rate indicate good quality of life. The validity of the SF-36 in patients following a hip fracture is established, as well as Sample size fixed at 3000 and α for the first co-primary responsiveness to changes in the SF-36 physical function outcome (mortality) fixed at 0.04 domain.19 25  HR 12.5% 13.0% 13.5 %  0.73 0.785 0.801 0.816 Outcomes at 1 year  0.70 0.873 0.886 0.897 For long-term follow-up, the primary outcome is a  0.65 0.96 0.966 0.971 composite at 1 year after randomisation of all-cause mortality and non-fatal MI, pulmonary embolism, pneu- Sample size fixed at 3000 and α for the second co-primary monia, sepsis and stroke. Individual secondary 1-year outcome* at 0.012 (calculated) follow-up outcomes include all-cause mortality, vascular  HR 27.5% 30% 32.5% mortality, non-vascular mortality, MI, congestive heart  0.75 0.916 0.941 0.958 failure, non-fatal cardiac arrest, coronary revascularisa-  0.70 0.988 0.993 0.996 tion, stroke, peripheral arterial thrombosis, pulmonary  0.65 0.9992 0.9997 0.9999 embolism, deep venous thrombosis, pneumonia, sepsis, *Composite outcome of major perioperative complications (ie, new acute renal failure requiring dialysis, peri-prosthetic mortality, non-fatal myocardial infarction, pulmonary embolism, fracture, prosthetic hip dislocation, implant failure, hip pneumonia, sepsis, stroke, and life-threatening and major re-operation, new residence in a nursing home, hospital bleeding). readmission, persistent postsurgical pain, FIM motor domain and its mobility and locomotion subscores, and the SF-36 score. reduction (RRR) of 30% (ie, a HR of 0.70) for the first co-primary outcome (90-day all-cause mortality) with a Adjudication of outcomes two-sided α=0.04, assuming an event rate of 13.0% in the The Event Adjudication Committee is a committee of control group. Even with an observed RRR of 27% (ie, a clinicians with expertise in perioperative outcomes. HR of 0.73), the trial would have 80% power for the first These individuals are blinded to treatment allocation co-primary outcome. The trial will also have 99% power and adjudicate the following outcomes: myocardial injury to detect a 30% RRR (ie, a HR of 0.70) for the second after randomisation, MI, non-fatal cardiac arrest, stroke, co-primary outcome assuming a standard care event pulmonary embolism, deep vein thrombosis, new conges- rate of 30% with an α=0.012 (two-sided). Even with an tive heart failure, pneumonia, sepsis and bleeding. All observed RRR of 25% (ie, HR of 0.75), there would be adjudicators are trained before commencing trial adjudi- 91% power for the second co-primary outcome assuming cation. We will use the decisions of the outcome adjudica- an event rate of 27.5% in the control group. tors for all statistical analyses of these events. Main analysis Statistical considerations We will analyse patients in the treatment group to which Sample size they are allocated, according to the intention-to-treat The overall type 1 error rate for the two co-primary principle. We will include all patients randomised in outcomes will be 5% (0.05) and this will be partitioned these analyses, regardless of the timing of surgery. We will between the two co-primary outcomes, taking into account compare patients allocated to accelerated medical clear- the overlap between the outcomes (ie, all-cause mortality ance and surgery with patients allocated to standard care. is a subset of the composite). Assuming a 20% overlap, We will present the binary analyses using the Kaplan- with the pre-specified α of 0.04 for the first co-primary Meier estimator. We will use log-rank tests to compare the outcome (all-cause mortality at 90 days), the α of 0.012 for rate of occurrence of the primary outcome between the the second co-primary outcome (composite) was calcu- accelerated care group and the standard care group. We lated via simulation. will use Cox proportional hazards models to estimate the The sample size calculations were performed using a effect of accelerated care on the HR for the primary and time-to-event analysis (Cox proportional hazards model dichotomous secondary outcomes including the 1-year comparison with two equal groups), two-sided α=0.05 outcomes. We will calculate the HRs and their associ- using Power Analysis and Sample Size software V.13 ated 95% CIs. We will estimate the effect of accelerated (2014) (see table 1). care versus standard care on SF-36 and FIM scores with a With a sample size of 3000 patients, the HIP ATTACK generalised linear model. We will infer statistical signifi- Trial will have 88% power to detect a relative risk cance if the computed two-sided p value is <0.05. Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 5 Open access Subgroup analysis patient-oriented research patient engagement framework Cox proportional hazards model assessing the primary spanning research governance, strategy, and methods.27 outcome will provide the basis for evaluating our single Examples include (a) governance auditing—-en- planned subgroup analysis (ie, patients who present to gaged patient representatives maintain an audit trail of the hospital ≥4 hours after their hip fracture). We expect strategy—and execution—related decisions in order to a larger treatment effect in patients who present within guide ongoing activities; (b) ‘Word on the Street’ videos— 4 hours of their hip fracture. We will infer a subgroup brief (40 to 60 s) commentaries target personal experi- effect if the interaction term of treatment and subgroup ence of our research and are shared via Twitter using Bitly is statistically significant at p<0.05. software and (c) Outcome evaluation—we use an interac- tive audience response system so patient partners can tell Interim analysis us, from their perspective, which trial outcomes matter We will perform two interim efficacy analyses based on the most. We use this information to inform our future trial co-primary outcomes when 50% and 75% of the patients communications strategy. have been followed for 90 days. The independent Trial Monitoring Committee (TMC) will employ the modified Haybittle-Peto rule of 4 SDs (α=0.0001) for analyses in the Ethics and dissemination first half of the trial (including the first planned interim We require documentation of Research Ethics Committee analysis) and three SDs (α=0.00047) for all analyses in the or Institutional Review Board (REC/IRB) approvals second half. For a finding to be considered significant for before sites are activated to enrol patients. All commit- either co-primary outcome, these predefined boundaries tees are described in detail in the Supplement File under will have to be exceeded in at least two consecutive anal- the Supplemental Trial Groups and Investigators section. yses, three or more months apart. If either co-primary Investigators are informed of any protocol amendments, outcome fulfils these criteria, the TMC will consider initi- and REC/IRBs are requested to approve them. Research ating discussion with the Project Ooffice Ooperations personnel or good clinical practice trained healthcare Ccommittee about potentially terminating the trial. professionals participating in the study obtain written The α-level for the final analysis will remain the conven- informed consent (online supplementary appendix 2) tional α=0.05, given the infrequent interim analyses and for each patient before randomisation. All data are stored associated low α levels, as well as the requirement for on a centrally encrypted, high-security computer system confirmation with subsequent analyses. We will appor- and kept strictly confidential. The online supplementary tion the α between the two co-primary outcomes in the appendix 3, presents the list of the HIP ATTACK trial final analysis. We will split the α with the first co-primary participant sites and countries. outcome (all-cause mortality at 90 days) at 0.04 and the second co-primary outcome (composite) at 0.012, due to Dissemination policy overlap. The knowledge dissemination plan includes traditional At any time during the trial, if safety concerns arise, the modes of dissemination (ie, publication in a policy-driving TMC chairperson will assemble a formal meeting of the journal, national/international conference presenta- full committee. The TMC will make their recommenda- tions) as well as the engagement of influential medical tions to the Project Office Operations Committee after organisations (ie, emergency medicine and orthopaedic considering all the available data and any external data surgery organisations). Broader dissemination will be from relevant studies. If a recommendation for termina- performed by the HIP ATTACK public website (http:// tion is being considered the TMC will invite the Interna- www.​ hipattacktrial.​ com), Twitter account (@HIPAT- tional Operations Committee to explore all possibilities TACKTrial), Facebook page and LinkedIn Profile. The before a decision is made. Reducing Global Perioperative Risk Multimedia Resource Centre, linked to Elsevier’s entire online global reader- Trial organisation ship, will disseminate slide and audioinstructional videos, PHRI is the coordinating centre for this trial worldwide full-text articles, links to abstracts and data summaries. and is primarily responsible for the organisation of the trial, development of the randomisation scheme, the study database, data consistency checks, data analysis Discussion and coordination of the study centres. The trial structure Hip fractures are a worldwide health concern due to includes the following groups: the Project Office Oper- their high incidence, poor outcomes and high health ations Committee, International Operations Committee, economic costs. Population ageing will probably worsen Steering Committee, National Coordinators, Investiga- this scenario in the near future. Age, male gender, clin- tors, Coordinating Centre, and Adjudication Committee. ical comorbidities, dementia, nursing home residency and surgical delay are associated with an increased risk Patient and public involvement of mortality after a hip fracture.28 Surgical timing is the Our approach to patient engagement is guided by the only potential modifiable risk factor for postoperative Canadian Institutes of Health Research strategy for mortality and major complications. 6 Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 Open access Evidence from several observational studies, including a target time of <6 hours, which was demonstrated as systematic reviews and meta-analyses, demonstrates that feasible in the HIP ATTACK pilot.15 HIP ATTACK is a early surgery is associated with better outcomes and large international trial powered to inform the effect of with decreased mortality in patients suffering a hip frac- accelerated surgery on patient-important outcomes. ture.14 29–31 Uzoigwe and colleagues published prospec- tive data on 2056 patients in UK. Patients who had Author affiliations 1 surgery more than 12 hours after hip fracture diagnosis Department of Perioperative Medicine, Population Health Research Institute, Hamilton, Ontario, Canada had adjusted OR of 3.8 (95% CI 1.03–14.50; p=0.046) 2 Department of Surgery, McMaster University, Hamilton, Ontario, Canada for in-hospital mortality compared with those who had 3 Department of Medicine, McMaster University, Hamilton, Ontario, Canada surgery within 12 hours.32 4 Department of Orthopaedic Surgery and Traumatology, Vall d'Hebron University Results from an intervention study in Canada demon- Hospital, Barcelona, Spain 5 strated that coordinated, region-wide efforts directed at Department of Clinical Research, Narayana Health, Bangalore, India 6 Department of Orthopaedic Surgery, Aga Khan University, Karachi, Pakistan meeting a 48 hours benchmark for hip fracture surgery 7 Department of Orthopaedic Surgery, Sancheti Institute for Orthopaedics & was successful in reducing time to surgery, length of Rehabilitation, Pune, India stay, adjusted in-hospital and 1-year mortality.33 There 8 Department of Surgery, London Health Sciences Centre, London, Ontario, Canada 9 were 3525 preintervention and 3007 postinterven- Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada 10 tion patients ≥50 years of age. Surgery within 48 hours Department of Surgery and Institute of Health Policy Management and Evaluation, University of Toronto, Toronto, Ontario, Canada increased from 66.8% to 84.6%, length of stay decreased 11 Department of Orthopaedic Surgery, Chris Hani Baragwanath Academic Hospital, from 13.5 to 9.7 median days and in-hospital mortality Johannesburg, South Africa decreased from 9.6% to 6.8% (all p<0.001). In-hospital 12 University of the Witwatersrand, Johannesburg, South Africa mortality (HR, 0.68; 95% CI 0.57 to 0.81) and mortality 13 Department of Anaesthesia and Perioperative Medicine, University of Cape Town, at 1-year follow-up (HR, 0.87; 95% CI 0.79 to 0.96) were Cape Town, South Africa 14 Department of Intensive Care and Perioperative Medicine, Jagiellonian University reduced in adjusted analyses. Medical College, Kraków, Poland Hip fracture patients who undergo surgery have worse 15 Department of Anaesthesiology, University of Malaya, Kuala Lumpur, Malaysia outcomes compared with matched patients who undergo 16 Vita-Salute San Raffaele University, Milano, Italy elective hip surgery.34 This suggests that a hip fracture 17 Department of Anaesthesia and Intensive Care, IRCCS San Raffaele Scientific initiates processes that increase patients’ risk indepen- Institute, Milan, Italy 18 dent of surgery. A hip fracture causes pain, immobilisation Department of Anesthesiology and Perioperative Medicine, Universitair Ziekenhuis Brussel, Brussels, Belgium and bleeding, which trigger a cascade of inflammation, 19 Biomedical Research Institute (IIB – SANT PAU), Barcelona, Spain sympathetic activation, hypercoagulability and catabolism 20 Department of Surgery, Queen’s University, Kingston, Ontario, Canada that can ultimately cause acute clinical complications (eg, 21 Shifa International Hospital, Islamabad, Pakistan 22 thromboembolism, acute MI, infection and death). Delay 23 Department of Orthopaedic Surgery, Christian Medical College, Ludhiana, India in repairing a hip fracture will increase the duration of Orthopaedics, SPZOZ, Myslenice, Poland 24 Department of Surgery, University of Ottawa, Ottawa, Ontario, Canada time a patient is exposed to these negative physiological 25 Department of Orthopaedic Surgery and Traumatology, Parc Taulí Hospital stressors. It is possible that urgent surgical treatment Universitari, Barcelona, Spain of a hip fracture will yield benefit, similar to how rapid 26 Department of Orthopaedics, Sharif Surgical Oncology, Fort Worth, Texas, USA 27 treatment of an acute MI and stroke have yielded benefit Department of Anaesthesia, McMaster University, Hamilton, Ontario, Canada 28 from rapid reversal of the underlying physiological School of Nursing, McMaster University, Hamilton, Canada 29 Department of Anaesthesia & Critical Care, University of Nottingham, Nottingham, abnormalities. UK Although previous hip fracture studies provide insights 30 Hamilton Health Sciences, Hamilton, Ontario, Canada into this issue, there are many examples of risk-adjusted 31 Departments of Health Research Methods, Evidence, and Impact (HEI), McMaster observational studies reporting misleading results. For University, Hamilton, Canada example, observational studies suggested harm with transfusion of older blood; however, subsequent large Contributors FKB, MB and PJD wrote the first draft of the protocol manuscript. FKB, MB, AS, BMB, WS, MM, JV, SP, VH and PJD planned the conceptualization and RCTs showed older blood was safe.35 Currently, evidence the design of the study, and the protocol. AP, VA, EG-F, MU, MT, AA, JN, VT, PKS, ARL, on best timing to perform hip fracture surgery is based RJ, MR, CYW, GL, PF, EP, GW, ANN, BJ, PŚ, RJF, MB-C, BD, MW, JdB, RK, JT-S, JT-H, mostly on observational studies, which are at risk of HS, IM, LG, and KA contributed to the design and implementation of the protocol. residual confounding. These studies may have overesti- All authors provided critical revisions to the manuscript before approving the final version. mated the effects of early surgery because sicker patients likely went to surgery later than less ill patients, due to Funding This work was supported by the following grants: Canadian Institute of Health and Research (CIHR) Foundation Award, CIHR’s Strategy for Patient Oriented clinical optimisation before surgery. On the other hand, Research (SPOR), through the Ontario SPOR Support Unit (OSSU) as well as the observational data could underestimate the effects Ontario Ministry of Health and Long-Term Care, and a grant from Smith & Nephew of ultra-early surgery, such as surgery within 6 hours, to recruit 300 patients in Spain. which has not been evaluated in clinical studies. Only Disclaimer The funders had no role in the study design, data collection, large, high-quality RCTs can minimise bias and provide a management, analysis, writing of the report, the decision to submit the report for valid estimate of treatment effects. publication, and they will not have ultimate authority over any of these activities. In HIP ATTACK, the goal in the accelerated surgery Competing interests None declared. arm is to operate on patients as soon as possible, with Patient consent for publication Not required. Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537 7 Open access Provenance and peer review Not commissioned; externally peer reviewed. 18. Wei LA, Fearing MA, Sternberg EJ, et al. The Confusion Assessment Method: a systematic review of current usage. J Am Geriatr Soc Open access This is an open access article distributed in accordance with the 2008;56:823–30. Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which 19. Latham NK, Mehta V, Nguyen AM, et al. Performance-based or permits others to distribute, remix, adapt, build upon this work non-commercially, self-report measures of physical function: which should be used and license their derivative works on different terms, provided the original work is in clinical trials of hip fracture patients? Arch Phys Med Rehabil properly cited, appropriate credit is given, any changes made indicated, and the use 2008;89:2146–55. is non-commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/. 20. Petrella RJ, Overend T, Chesworth B. FIM after hip fracture: is telephone administration valid and sensitive to change? Am J Phys Med Rehabil 2002;81:639–44. 21. McGrory BJ, Shinar AA, Freiberg AA, et al. Enhancement of the value of hip questionnaires by telephone follow-up evaluation. J References Arthroplasty 1997;12:340–3. 1. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. 22. Botto F, Alonso-Coello P, Chan MT, et al. Myocardial injury after Bull World Health Organ 2003;81:646–56. noncardiac surgery: a large, international, prospective cohort study 2. Beloosesky Y, Hendel D, Weiss A, et al. Cytokines and C-reactive establishing diagnostic criteria, characteristics, predictors, and 30- protein production in hip-fracture-operated elderly patients. J day outcomes. Anesthesiology 2014;120:564–78. Gerontol A Biol Sci Med Sci 2007;62:420–6. 23. Devereaux PJ, Biccard BM, Sigamani A, et al. Association of 3. Miller RR, Shardell MD, Hicks GE, et al. Association between Postoperative High-Sensitivity Troponin Levels With Myocardial interleukin-6 and lower extremity function after hip fracture--the role Injury and 30-Day Mortality Among Patients Undergoing Noncardiac of muscle mass and strength. J Am Geriatr Soc 2008;56:1050–6. Surgery. JAMA 2017;317:1642–51. 4. Svensén CH. Vascular endothelial growth factor (VEGF) in plasma 24. Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of increases after hip surgery. J Clin Anesth 2004;16:435–9. myocardial infarction. J Am Coll Cardiol 2012;60:1581–98. 5. Bessissow A, Chaudhry H, Bhandari M, et al. Accelerated versus 25. Mangione KK, Craik RL, Palombaro KM, et al. Home-based leg- standard care in hip fracture patients: does speed save lives? J strengthening exercise improves function 1 year after hip fracture: a Comp Eff Res 2014;3:115–8. randomized controlled study. J Am Geriatr Soc 2010;58:1911–7. 6. LeBlanc ES, Hillier TA, Pedula KL, et al. Hip fracture and increased 26. Stineman MG, Shea JA, Jette A, et al. The Functional Independence short-term but not long-term mortality in healthy older women. Arch Measure: tests of scaling assumptions, structure, and reliability Intern Med 2011;171:1831–7. 7. Prevention of pulmonary embolism and deep vein thrombosis with across 20 diverse impairment categories. Arch Phys Med Rehabil low dose aspirin: Pulmonary Embolism Prevention (PEP) trial. Lancet 1996;77:1101–8. 2000;355:1295–302. 27. McGillion MH, Lin-Rogano L, Borges FK. Patient engagement in 8. Brauer CA, Coca-Perraillon M, Cutler DM, et al. Incidence research related to accelerated surgical care and treatment for hip and mortality of hip fractures in the United States. JAMA fracture. CMAJ 2018;190(Suppl):S38–9. 2009;302:1573–9. 28. Smith T, Pelpola K, Ball M, et al. Pre-operative indicators for mortality 9. Maxwell MJ, Moran CG, Moppett IK. Development and validation of following hip fracture surgery: a systematic review and meta- a preoperative scoring system to predict 30 day mortality in patients analysis. Age Ageing 2014;43:464–71. undergoing hip fracture surgery. Br J Anaesth 2008;101:511–7. 29. Shiga T, Wajima Z, Ohe Y. Is operative delay associated 10. White SM, Griffiths R, Holloway J, et al. Anaesthesia for proximal with increased mortality of hip fracture patients? Systematic femoral fracture in the UK: first report from the NHS Hip Fracture review, meta-analysis, and meta-regression. Can J Anaesth Anaesthesia Network. Anaesthesia 2010;65:243–8. 2008;55:146–54. 11. Roberts SE, Goldacre MJ. Time trends and demography of mortality 30. Moja L, Piatti A, Pecoraro V, et al. Timing matters in hip fracture after fractured neck of femur in an English population, 1968-98: surgery: patients operated within 48 hours have better outcomes. A database study. BMJ 2003;327:771–5. meta-analysis and meta-regression of over 190,000 patients. PLoS 12. SooHoo NF, Farng E, Chambers L, et al. Comparison of complication One 2012;7:e46175. rates between hemiarthroplasty and total hip arthroplasty for 31. Maheshwari K, Planchard J, You J, et al. Early Surgery Confers intracapsular hip fractures. Orthopedics 2013;36:e384–9. 1-Year Mortality Benefit in Hip-Fracture Patients. J Orthop Trauma 13. Nichols CI, Vose JG, Nunley RM. Clinical Outcomes and 90-Day 2018;32:105–10. Costs Following Hemiarthroplasty or Total Hip Arthroplasty for Hip 32. Uzoigwe CE, Burnand HG, Cheesman CL, et al. Early and ultra- Fracture. J Arthroplasty 2017;32(9S):S128–34. early surgery in hip fracture patients improves survival. Injury 14. Simunovic N, Devereaux PJ, Sprague S, et al. Effect of early surgery 2013;44:726–9. after hip fracture on mortality and complications: systematic review 33. Bohm E, Loucks L, Wittmeier K, et al. Reduced time to surgery and meta-analysis. CMAJ 2010;182:1609–16. improves mortality and length of stay following hip fracture: results 15. Hip Fracture Accelerated Surgical Treatment and Care Track (HIP from an intervention study in a Canadian health authority. Can J Surg ATTACK) Investigators. Accelerated care versus standard care 2015;58:257–63. among patients with hip fracture: the HIP ATTACK pilot trial. CMAJ 34. Le Manach Y, Collins G, Bhandari M, et al. Outcomes After Hip 2014;186:E52–60. Fracture Surgery Compared With Elective Total Hip Replacement. 16. Canada Health Act - Annual Report 2011–2012, 2012. JAMA 2015;314:1159–66. 17. Inouye SK, van Dyck CH, Alessi CA, et al. Clarifying confusion: 35. Heddle NM, Cook RJ, Arnold DM, et al. Effect of Short-Term vs. the confusion assessment method. A new method for detection of Long-Term Blood Storage on Mortality after Transfusion. N Engl J delirium. Ann Intern Med 1990;113:941–8. Med 2016;375:1937–45. 8 Borges FK, et al. BMJ Open 2019;9:e028537. doi:10.1136/bmjopen-2018-028537