Fractures of bone are very common, affecting 2% of the population per annum. Whilst the majority heal uneventfully, 10-15% exhibit delayed or non-union. These complications tend to occur in patients who have sustained high-energy open fractures, which are limb-threatening injuries, or low-energy osteoporotic fractures, which are associated with high morbidity and mortality rates. Enhancement or acceleration of fracture repair would confer significant benefit to these patients as well as reduce the public health burden. Inflammation represents the earliest response following trauma and initiates a cascade of downstream events crucial for wound healing. However, the mechanism by which this occurs remains poorly defined. A detailed understanding of how these upstream events initiate fracture healing is a necessary step in the development of therapeutics to enhance this process. Our group previously reported that addition of low dose recombinant human TNF (rhTNF) at the fracture site accelerated fracture repair in a murine tibial fracture model. Here I show that local rhTNF treatment is only effective when administered within 24 hours of injury, when neutrophils represent the major inflammatory cell infiltrate. Endogenous TNF was expressed at the fracture site initially by neutrophils and after 3 days by monocytes/macrophages. Systemic administration of anti-TNF resulted in impaired fracture healing. The addition of rhTNF to the fracture environment in an air pouch model enhanced neutrophil recruitment, and promoted the recruitment of monocytes through CCL2 production. Conversely, inhibition of either neutrophils or the chemokine receptor CCR2 resulted in significantly impaired fracture healing. Fragility fractures represent a major unmet medical need and they are associated with permanent disability and premature death. Using a murine model of fragility fractures, rhTNF treatment improved fracture healing during the early phase of repair. Translated clinically, accelerated healing would permit earlier load bearing and reduce the morbidity and mortality associated with delayed patient mobilisation.

Introduction: Trauma is the main cause of death in working-age adults in the UK. Multiple organ failure (MOF) is associated with a high proportion of late trauma deaths, and MOF survivors have poor long-term outcomes. Early prediction of patients at risk of MOF would assist treatment decisions and allow targeted interventions. Methods: A cohort of major trauma patients requiring intensive care unit (ICU) treatment at the John Radcliffe Hospital was identified. Data were obtained from the two national databases of the Trauma Audit Research Network and the Intensive Care National Audit and Research Centre, and from a local ICU database with hourly data recording. Literature review and questionnaire analysis of trauma clinicians identified candidate predictors of MOF, grouped into patient, injury, physiological, laboratory and management variables. MOF scoring systems were reviewed to determine the most appropriate for use in trauma patients. Prediction models of post-trauma MOF were developed using logistic regression at a range of times from 0 to 48 hours after injury. Models were internally validated using bootstrapping. Results: 517 adult trauma patients were identified from 2003-2011. Overall mortality was 14.9%, with 491 patients surviving more than 48 hours, and therefore being at risk of MOF development. For these 491 patients, MOF incidence depended on the definition, and ranged from 23% (Denver score) to 58% (SOFA score). MOF was associated with mortality, time to ICU admission, and length of ICU and hospital stay. MOF could be predicted with an accuracy of up to 81.3% at 2 hours post-injury, and 84.2% at 12 hours post-injury using small numbers of clinical variables. Age, head injury, abdominal injury, maximum heart rate and the need for vasopressors were strong predictors of all definitions of MOF. Conclusions: Post-trauma MOF can be predicted early after injury using combinations of clinical variables. Further validation of the identified variables on external populations would allow development of a clinical score to assist clinicians in trauma management.

Knee osteoarthritis (OA) is the commonest form of lower limb OA with a lifetime risk of over 40%. It is a disease characterised by symptoms such as pain and loss of function. In addition there are typical structural features on both radiographs and MRI. Knee OA represents a spectrum of disease, ranging from early preclinical cartilage change to established full thickness disease. Anteromedial knee OA is a particular phenotype of knee OA where disease is confined to the medial compartment. Whilst end stage arthritis is treated reliably with joint arthroplasty, those with early stage disease are treated with a variety of non- surgical interventions with varying success. This thesis is concerned with understanding the disease of patients that have early radiographic changes but symptoms not controlled by conservative measures. Up to 150 of these patients a year present to the Nuffield Orthopaedic Centre, Oxford. They have been described as being in the “Treatment Gap”. A series of validation studies were performed to determine the optimal method for diagnosing cartilage defects within the knee. The three commonest diagnostic methods were examined for their validity. Arthroscopic assessments of cartilage lesions demonstrated a moderate level of intra and inter observer reliability. In contrast, radiographs and MRI demonstrated high levels of reliability. When using MRI as a criterion standard, both radiographs and arthroscopic assessment were found to have poor accuracy. Based on the work in this thesis a formal definition of the cartilage changes exhibited in early knee OA was proposed. A cross sectional cohort of 100 patients with the symptoms and radiological features of early knee OA were identified. Their pain and function profile was compared to two comparison groups of patients at the end stage of knee OA (defined by the need for partial or total arthroplasty). In up to 78% of individual cases those with early OA had pain and function profiles as bad as those with end stage disease. The cross sectional symptoms of early knee OA demonstrate a marked discordance with their mild radiographic changes. The same cohort was extended to 125 patients. They were followed over one year with monthly PROM assessments to determine how symptoms change over time. 43% of patients experience a clinical improvement over 12 months, 31% experience a clinical deterioration and 26% remain unchanged. The range in OKS variation over 12 months was on average 12 points, with clinically relevant variation occurring on 45% of monthly measurements. Patients with early knee OA can expect to experience considerable variation in their symptoms over 12 months and this must be considered when planning interventions. A number of patients with early knee OA were noticed to demonstrate medial meniscal extrusion. Using data from the Osteo Arthritis Initiative (OAI) a nested case control study was designed to determine how the presence of meniscal extrusion in an otherwise normal knee affects the risk of developing knee OA over the next 48 months. This demonstrated an Odds Ratio of 3.5, suggesting that meniscal extrusion is a considerable risk factor for the development of OA. The presence of a knee injury or operative intervention to the index meniscus was shown to increase this risk. Many phenotypes of OA are known to demonstrate familial aggregation. In an attempt to determine where the earliest structural changes occur in medial compartment knee OA, a cohort of patients selected only for their family history of the disease were developed. This cohort was compared to spouse controls for the presence of knee OA, as well as meniscal extrusion and long leg alignment. In addition, a functional analysis of their cartilage was performed. This cohort was not shown to be at increased risk of disease compared to controls. Discussion of the possible reasons for this finding is presented. Early knee osteoarthritis is a considerable clinical problem. This thesis has aided the understanding of the condition by firstly defining the radiological description of these patients. Secondly, their cross sectional and longitudinal symptom profile have been described for the first time. In addition, the presence of an extruded meniscus has been demonstrated as a substantial risk factor for the disease. Finally, family history has not been demonstrated as a risk factor for the disease within the limits of the study described here. Future work has been proposed.

Bakgrund: Djup ventrombos (DVT) är en allvarlig komplikation efter kirurgiska och ortopediska ingrepp. Detta leder till ökat lidande för patienten, längre vårdtider och därav ökade kostnader för samhället. För att förebygga DVT krävs kunskap om vilka preventiva omvårdnadsåtgärder som har effekt och hur de ska användas. Syfte: Syftet med studien var att genom en systematisk litteraturstudie se vad kunskapsläget säger om preventiva omvårdnadsåtgärder mot DVT samt vad som ser mest effektivt ut. Syftet var också att genom en enkät ta reda på hur sjuksköterskor arbetar kliniskt för att förebygga DVT. Metod: Litteraturstudie som baserades på tio orginalartiklar. Som komplement till litteraturstudien gjordes även en enkätstudie som inkluderade tio sjuksköterskor. Resultat: Tidig mobilisering tillsammans med mekanisk profylax ses som effektivt för att förebygga DVT. I många av studierna som granskats gavs patienterna lågmolekylärt heparin (LMH). Det var dock ingen lägre frekvens av DVT hos dessa patienter jämfört med de som ej fick LMH. Det var vid införandet av profylaktiska omvårdnadsåtgärder som tidig mobilisering, kompressionsstrumpor och kompressionspumpar som frekvensen av DVT sjönk. Sjuksköterskorna arbetar aktivt med förebyggande omvårdnadsåtgärder men det finns dock inga specifika riktlinjer för dessa omvårdnadsåtgärder i den kliniska verksamheten. Slutsats:  Studien visar att ett behov finns för utvecklandet av att arbeta med individanpassad profylax då patienterna utifrån olika riskgrupper har olika behov av profylaxtyper. Stor vikt bör läggas på att utveckla och implementera riktlinjer för tidig mobilisering då detta utifrån denna studies resultat har god effekt för att förebygga uppkomsten av DVT. / Background: Deep vein thrombosis is a serious complication during surgical and orthopaedic procedures. This leads to increased patient suffering, longer hospital stays and thereby also higher cost for society. To prevent DVT requires knowledge about preventive care measures that are effective, and how they will be used. Aim: The aim of this study was to examine the current state of knowledge about preventive care measures against DVT and how effective they are. The aim was also, by conducting a survey, to find out how nurses work clinically for the prevention of DVT. Study design: A literature study was conducted. The study was based on ten original articles. As a complement to the literature study a questionnaire was created and handed out to ten surgical and orthopaedic nurses. Results: Early mobilization in combination with mechanical prophylactics is seen as effective for preventing DVT. In many of the included studies patients were given low molecular weight heparin (LMH). However the frequency of DVT in these patients was not lower compared to those who did not get LMH. Instead it was at the introduction of prophylactic care measures such as early mobilization, compression stockings and compression pumps as the frequency of DVT fell. The questionnaires showed that clinical nurses worked actively with care measures. However, there were no specific guidelines for the measures in the organisation. Conclusion: The study shows that there is a need for developing methods for individual prophylactic care as patients from different risk groups have different needs. There should also be focus on developing and implementing guidelines for early mobilization as this study has shown it is effective in preventing DVT.

Thesis questions: • Is the glutaminergic system altered in rotator cuff tendinopathy? • Is the glutaminergic system altered by common treatments? • Are glutaminergic changes related to pain symptoms? • What are the effects of glutamate and glutamate receptor modulation on tendon derived cells? Summary answers: • The glutaminergic system is altered in rotator cuff tendinopathy • Changes within this system are seen after common treatments • Specific glutaminergic changes are associated with the resolution of pain following shoulder surgery but do not predict the severity of pain symptoms • Glutamate has significant effects on tendon derived cells. What is known: It is known that extracellular glutamate concentrations are increased in both Achilles and patellar tendinopathy. It has also been previously shown that the glutamate receptors NMDAR1 and mGluR5 are upregulated in patellar tendinopathy. What this thesis adds: This thesis has shown for the first time that glutamate and NMDAR1 are increased in rotator cuff tendinopathy. Increases in cell proliferation, vascularity and HIF1α are seen after surgical rotator cuff repair and these features are not seen after glucocorticoid injection. There are significant differences between painful and pain-free rotator cuff tendons in terms of glutamate receptor expression (KA1, mGluR7 and mGluR2) and inflammatory cell numbers (CD45 and CD206). Exposure to 1.875mM glutamate for 72 hours results in reduced cell viability, decreased collagen (COL1A1 and COL3A1) and increased aggrecan gene expression; NMDAR antagonism with MK-801 attenuates the deleterious effect on cell viability but had no effect on the changes in matrix gene expression. Bias, confounding and other reasons for caution: The observational histological work was limited by the control tissue. Some control tissue was not age matched, while some of the pain-free control tendons were post-surgical intervention. Confounding factors include tendon structure, length of symptoms and previous treatments. Caution must be applied when discussing the in vivo implications of the in vitro work.

Back pain is a massive global public health problem with multiple contributing factors including obesity. Obesity is thought to be linked to back pain through mechanical factors. However, obesity also causes a systemic low-grade inflammatory milieu. This would suggest a possible biochemical link between obesity, intervertebral disc degeneration, and back pain. Furthermore, the relationship between obesity and the clinical presentation of spine patients is unclear. This thesis aims to examine the effect of and relationship between obesity, the intervertebral discs, and back pain from biochemical, clinical, and epidemiological perspectives. In this thesis, an in vitro study assessed the effect of leptin, a fat-specific cytokine, upon the intervertebral disc. The bovine intervertebral disc was used as a model in a cell culture system. An ex vivo study examined leptin and pro-inflammatory cytokines produced by paraspinal adipose tissue taken during routine surgical procedures from spinal patients. Plasma taken from patients presenting with low back pain was analysed by mass spectrometry and multiplex immunoassay to identify possible protein biomarkers. At an epidemiological level, statistical modelling of the Genodisc patient population was conducted. This was a pan-European study of 2636 patients presenting to tertiary spinal units. Analyses were performed to examine relationships between obesity, quantified by body mass index (BMI), and pain, clinical diagnosis, and spinal degeneration identified on magnetic resonance imaging (MRI). Leptin was shown to increase the production of and expression of degradative and pain-generating molecules by disc cells. A pro-inflammatory environment, especially IL-6, potentiated this response. Leptin and pro-inflammatory cytokines produced by paraspinal fat were unrelated to clinical symptoms. However, levels of the pro-inflammatory cytokines, TNF-α and IL-6, were raised in the plasma of patients with greater pain or those with spinal stenosis. Furthermore, clusterin and complement were identified, by mass spectrometry, as potential biomarkers for spine patients. Epidemiological analyses revealed that obesity was associated with greater back pain, although the magnitude of this association was small. Similarly, obesity was associated with a diagnosis of spinal stenosis. Finally, increased BMI was found to be an independent predictor of disc degeneration, spinal stenosis, and disc herniation on MRI. In summary, this thesis has furthered the clinical understanding of lumbar spine pathology and back pain. It will provide clinicians with a better framework to assess spine patients. These results show that obesity is associated with lumbar spine degeneration and pain. Leptin could be a factor mediating this relationship. Further studies should concentrate on clarifying the mechanism of action of leptin upon the intervertebral disc and assessing the longitudinal effect of obesity upon the lumbar spine. In this thesis, an in vitro study assessed the effect of leptin, a fat-specific cytokine, upon the intervertebral disc. The bovine intervertebral disc was used as a model in a cell culture system. An ex vivo study examined leptin and pro-inflammatory cytokines produced by paraspinal adipose tissue taken during routine surgical procedures from spinal patients. Plasma taken from patients presenting with low back pain was analysed by mass spectrometry and multiplex immunoassay to identify possible protein biomarkers. At an epidemiological level, statistical modelling of the Genodisc patient population was conducted. This was a pan-European study of 2636 patients presenting to tertiary spinal units. Analyses were performed to examine relationships between obesity, quantified by body mass index (BMI), and pain, clinical diagnosis, and spinal degeneration identified on magnetic resonance imaging (MRI). Leptin was shown to increase the production of and expression of degradative and pain-generating molecules by disc cells. A pro-inflammatory environment, especially IL-6, potentiated this response. Leptin and pro-inflammatory cytokines produced by paraspinal fat were unrelated to clinical symptoms. However, levels of the pro-inflammatory cytokines, TNF-α and IL-6, were raised in the plasma of patients with greater pain or those with spinal stenosis. Furthermore, clusterin and complement were identified, by mass spectrometry, as potential biomarkers for spine patients. Epidemiological analyses revealed that obesity was associated with greater back pain, although the magnitude of this association was small. Similarly, obesity was associated with a diagnosis of spinal stenosis. Finally, increased BMI was found to be an independent predictor of disc degeneration, spinal stenosis, and disc herniation on MRI. In summary, this thesis has furthered the clinical understanding of lumbar spine pathology and back pain. It will provide clinicians with a better framework to assess spine patients. These results show that obesity is associated with lumbar spine degeneration and pain. Leptin could be a factor mediating this relationship. Further studies should concentrate on clarifying the mechanism of action of leptin upon the intervertebral disc and assessing the longitudinal effect of obesity upon the lumbar spine.

Patellofemoral pain syndrome (PFPS) is the most common injury in runners. This thesis investigates the theory that prolonged eversion at the rear-foot causes prolonged tibial internal rotation and excessive femoral internal rotation, and predisposes female, distance runners to PFPS. Since eversion occurs at the subtalar joint, the morphology of the talus and calcaneus were also assessed. The study was a case-control investigation between female runners with a history of PFPS (n = 9) and normal controls (n = 10). Gait analysis was used to measure lower-limb joint angles during barefoot, treadmill running. It was hypothesised that runners with PFPS would demonstrate prolonged rearfoot eversion and tibial internal rotation, and increased hip internal rotation compared to normal controls. Computed tomography was used to image the foot and ankle in simulated weight-bearing using a custom-built loading rig. Three-dimensional models of the talus and calcaneus were generated and their shapes were quantified using principal axis lengths and orientations. The results did not support the theory that prolonged eversion and rear-foot structure predispose to PFPS during running, and attributing PFPS to these factors should be done with discretion. However, runners with a history of PFPS exhibited increased rear-foot eversion, reduced rear-foot dorsiflexion and increased knee internal rotation compared to normal controls during running, walking and squatting. Subjects with PFPS also demonstrated increased dorsiflexion at the mid-foot. It was proposed that increased eversion was secondary to reduced rear-foot dorsiflexion as this enabled compensatory dorsiflexion at the mid-tarsal complex. Due to the tight articulation of the ankle mortise, increased knee internal rotation corresponds well with excessive rear-foot eversion. A prospective study is required to establish whether these kinematic alterations are a cause or an effect of PFPS.

Conventional anterior cruciate ligament (ACL) reconstruction grafts have not been able to replicate the mechanical behaviour of the native ACL, reproduce normal knee mechanics and kinematics, or prevent degenerative disease progression of the knee. The aim of this thesis was to investigate a novel ACL design to more closely mimic the normal mechanical behaviour of the ACL, reconstruct the isometric ACL fibre and potentially reproduce the normal kinematics and mechanics of the knee. The designed artificial ACL reconstruction (ACLR) system could be used as a stand-alone device or in conjunction with a total knee replacement (TKR). The nominated design option for the ACLR system consisted of a connecting cord made of ultra-high molecular weight polyethylene (UHMWPE) fibres and an elastic system made of cobalt-chrome-molybdenum (CoCrMo) alloy with similar load-elongation characteristics to the native ACL. The design requirements were defined based on the mechanical properties of the native ACL, size constraints from the bony geometry and TKR components, and the location of the isometric fibres of the native ACL. The in vitro mechanical tests performed in this project on the designed cord showed a 2-3 times greater ultimate tensile load compared to the ACL in young human cadavers. The decreasing creep modulus of the UHMWPE cord under fatigue loading in simulated body conditions (3118 MPa at 6.5×10⁶ cycle) indicated nominal creep and stabilised mechanical properties by the 3000^th loading cycle. To replicate the non-linear stiffness of the ACL with ~38 N mm^-1 toe and ~100 N mm^-1 linear regions, the artificial ACLR device consisted of a femoral spring (~60 N mm^-1) in series with a tibial spring (~100 N mm^-1) and a connecting cord (~2000 N mm^-1). Two helical springs in series were used for the stand-alone ACLR, whereas a helical spring in series with a spiral spring was designed for the ACLR-TKR. As both the helical and spiral springs had a constant stiffness, stop mechanisms were added to the springs to create a non-linear stiffness and control the maximum safe deformation limit of each spring. To understand the mechanical behaviour of the reconstructed isometric fibre of the ACL, passive and loaded motions were simulated in 18 sets of segmented MRI models of healthy human knees. Constant load and elongation was observed throughout flexion during the passive movements, whereas maximal load and elongation in the reconstructed ACL was identified at 50 º of flexion during loaded motion. An ACL attachment placement sensitivity study, conducted in this project to assess the effect of surgical implantation error on the behaviour of the reconstructed ACL, revealed that misplacement of the femoral attachment would significantly influence the load-elongation of the reconstructed ACL. Finite element (FE) models of the designed ACLR devices enabled their behaviour under simulated axial loading, squatting and the Lachman test to be assessed. Both ACLR devices successfully reproduced stiffness of the native ACL with a multi-linear stiffness curve, however, elongation greater than 3.1 mm could not be achieved. It can be concluded that the designed artificial ACLR devices were able to mimic the mechanical behaviour of the ACL provided it was positioned at the isometric attachment points; potentially enabling achievement of more natural kinematics and mechanics of the reconstructed knee. However, ACL placement was shown to have a significant impact on the behaviour of the reconstructed ACL, therefore, placement error may over-constrain the joint. For this reason, a more forgiving design with a lower stiffness and a larger deformation limit would be advised.

Unicompartmental knee replacement (UKR) is the principal alternative to total knee replacement (TKR) in the treatment of end-stage knee osteoarthritis. It involves less tissue resection, resulting in lower rates of morbidity and faster recoveries compared to TKR. However, UKR has a significantly higher revision rate compared to TKR. As a result, whilst over a third of patients are eligible for UKR, only around 8% receive it. A comprehensive comparison of matched patients undergoing TKR and UKR was undertaken using a large dataset from the National Joint Registry for England and Wales (NJR). Failure rates (revision, reoperation, complications and mortality), length of stay and patient-reported outcomes (PROMs) were studied. Whilst patients undergoing TKR had lower reoperation and revision rates, they had higher rates of morbidity and mortality, longer hospital stays, and inferior PROMs compared to UKR. The main reason for revision in UKR was loosening. In view of the high revision rate in UKR, NJR data was studied to identify modifiable risk factors for failure in UKR. Important patient factors were identified including age, gender and pre-operative function. Surgeons with a higher UKR caseload had significantly lower revision rates and superior patient-reported outcomes. Increasing usage (offering UKR to a greater proportion of knee replacement patients) appears to be a viable method of increasing caseload and therefore of improving results. Surgeons with optimal usage (around 50% of patients, using appropriate implants) achieved revision/reoperation rates similar to matched patients undergoing TKR. Two clinical studies were conducted to establish whether the use of cementless fixation would improve fixation and reduce the revision rate of UKR. Cementless UKR was demonstrated to be safe and reliable, with PROMs similar or superior to those demonstrated in cemented UKR. Patients with suboptimal cementless fixation were examined and pre-disposing technical factors were identified. Finally, using NJR data, the effect of the introduction of cementless UKR on overall outcomes was examined. The number of cementless cases was small, and no significant effect on implant survival was demonstrated. However, patients undergoing cementless UKR demonstrated superior PROMs. These studies demonstrate that UKR has numerous advantages over TKR in terms of morbidity, mortality and PROMs. If surgeons perform high volumes of UKR (achievable by increasing their UKR usage), these advantages can be attained without the large difference in revision rates previously demonstrated. Cementless UKR is safe and provides superior fixation and outcomes in the hands of high-volume surgeons. Further work is needed to quantify the revision rate of cementless UKR, and to assess its results in the hands of less experienced surgeons.

The Oxford Knee Score (OKS) is a popular single summary questionnaire developed to measure the effect of knee replacement surgery from the patients' perspective. There has been a recent interest in the use of the OKS in populations of patients and in roles it has not been originally developed for. To date, no evidence has been provided about the measurement properties of the OKS when it is used outside the context or purpose for which it was originally designed. The general aim of this thesis is to investigate the measurement properties of the OKS when used in extended roles: a) within the population for which the OKS was originally intended and developed for (knee replacement) and, b) when applied on a different population, viz. patients undergoing non-operative treatment for knee osteoarthritis (OA). Four existing large-scale databases of patients undergoing knee replacement surgery and a database obtained from a prospective study on patients undergoing non-surgical management for knee OA were analyzed. The results demonstrate that: 1) it is possible to extract separate information on pain and functional disability from the OKS in a meaningful way (in the form of subscales). 2) For the first time, anchor-based Minimal Important Change (MIC) of 9 points and Minimal Important Difference (MID) of 5 points were established for joint replacement surgery. 3) The OKS demonstrated satisfactory evidence reliability, validity, responsiveness, and interpretability, when used in patients who are undergoing non-operative management for their knee OA. 4) Further evidence of validity was demonstrated by fitting the OKS to the Rasch model. 5) Lastly, it was demonstrated that thresholds can be applied on the OKS to distinguish between patients who consider their knee problem to be severe enough to warrant joint replacement surgery versus patients who do not. This supports the potential use of the OKS in decision making aids for secondary care referral. Overall the thesis provides critical evidence, not previously existing, to support the continued use, and extended use, of the OKS in orthopaedic medicine.