Search

Objective: To develop evidence‐based consensus recommendations for the optimal timing of hip and knee arthroplasty to improve patient‐important outcomes including, but not limited to, pain, function, infection, hospitalization, and death at 1 year for patients with symptomatic and radiographic moderate‐to‐severe osteoarthritis or advanced symptomatic osteonecrosis with secondary arthritis of the hip or knee who have previously attempted nonoperative therapy, and for whom nonoperative therapy was ineffective, and who have chosen to undergo elective hip or knee arthroplasty (collectively referred to as TJA). Methods: We developed 13 clinically relevant population, intervention, comparator, outcomes (PICO) questions. After a systematic literature review, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used to rate the quality of evidence (high, moderate, low, or very low), and evidence tables were created. A Voting Panel, including 13 physicians and patients, discussed the PICO questions until consensus was achieved on the direction (for/against) and strength (strong/conditional) of the recommendations. Results: The panel conditionally recommended against delaying TJA to pursue additional nonoperative treatment including physical therapy, nonsteroidal antiinflammatory drugs, ambulatory aids, and intraarticular injections. It conditionally recommended delaying TJA for nicotine reduction or cessation. The panel conditionally recommended delay for better glycemic control for patients who have diabetes mellitus, although no specific measure or level was identified. There was consensus that obesity by itself was not a reason for delay, but that weight loss should be strongly encouraged, and the increase in operative risk should be discussed. The panel conditionally recommended against delay in patients who have severe deformity or bone loss, or in patients who have a neuropathic joint. Evidence for all recommendations was graded as low or very low quality. Conclusion: This guideline provides evidence‐based recommendations regarding the optimal timing of TJA in patients who have symptomatic and radiographic moderate‐to‐severe osteoarthritis or advanced symptomatic osteonecrosis with secondary arthritis for whom nonoperative therapy was ineffective to improve patient‐important outcomes, including pain, function, infection, hospitalization, and death at 1 year. We acknowledge that the evidence is of low quality primarily due to indirectness and hope future research will allow for further refinement of the recommendations. [ABSTRACT FROM AUTHOR]

Background: The routine use of traditional chemistry-7 (chem-7) laboratory tests following total joint arthroplasty (TJA) has been called into question with the advent of short-stay procedures. Our objective was to determine the incidence, risk factors, and clinical interventions associated with inpatient abnormal routine postoperative chem-7 panels.Methods: From 2015 to 2017, 3,162 patients underwent a total of 3,721 TJA procedures, including primary total hip arthroplasty (THA) (n = 1,939; 52.1%) or primary total knee arthroplasty (TKA) (n = 1,782; 47.9%). Patients underwent routine preoperative and postoperative chem-7 testing. Clinical interventions were identified. With use of mixed-effects multivariate logistic regression, potential risk factors for abnormal chemistry panel values (including preoperative chem-7 results, type of surgery, age, sex, race, comorbidities, American Society of Anesthesiologists [ASA] score, and medications) were analyzed.Results: The rates of abnormal preoperative laboratory results were 3.4% for sodium (Na+), 7.4% for potassium (K+), 15.8% for blood urea nitrogen (BUN), and 26.4% for creatinine (Cr). The incidence of abnormal postoperative results was low for K+ (9.7%) and higher for Na+ (25.6%), BUN (55.6%), and Cr (27.9%). Preoperative abnormal laboratory results were a significant predictor of a postoperative abnormality for Na+ (odds ratio [OR] = 2.15; 95% confidence interval [CI] = 1.82 to 2.54), K+ (OR = 4.22; 95% CI = 3.03 to 5.88), and Cr (OR = 3.00; 95% CI = 2.45 to 3.68). Bilateral TJA was associated with increased odds of abnormal postoperative Na+ (OR = 1.56; 95% CI = 1.44 to 1.68). Renal disease was associated with increased odds of abnormal postoperative Cr (OR = 15.21; 95% CI = 5.67 to 40.77). Patients taking loop diuretics had increased odds of abnormal postoperative K+ (OR = 2.10; 95% CI = 1.42 to 3.11) and Cr (OR = 2.28; 95% CI = 1.56 to 3.33). Regarding intervention, 6.7% of hypokalemic patients received potassium chloride (KCl) fluid/tablets. Forty percent of hyponatremic patients received sodium chloride (NaCl) fluid/tablets. The electrolyte-related medicine consultation rate was 0.3% (13 of 3,721).Conclusions: On the basis of our findings, we recommend postoperative chem-7 testing for patients with an abnormal preoperative laboratory result (Na+, K+, BUN, Cr), preexisting renal disease, bilateral TJA, and prescribed angiotensin-converting enzyme inhibitors (ACE), angiotensin II receptor blockers (ARB), and diuretics.Level Of Evidence: Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]

Background: Recently, the Centers for Medicare and Medicaid Services (CMS) has labeled the procedural codes for total hip arthroplasty (THA) and total knee arthroplasty (TKA) as potentially misvalued and has asked the American Medical Association (AMA) and its Relative Value Scale Update Committee (RUC) to review this. To assess the validity of this claim, we aimed to catalog the specific service tasks and duration of time required for each task associated with the perioperative care of the patient who underwent primary THA and TKA.Methods: We prospectively timed preservice and immediate postservice activities performed outside of the operating room (OR) by 7 arthroplasty surgeons over a four-week period. Specific timing data for preservice activities performed in the OR were obtained retrospectively from our institutional electronic medical record for 500 patients undergoing THA and 500 undergoing TKA. Results were compared with the current approved values reviewed by the RUC in 2013 and converted to work relative value units (wRVUs) based on the intensity coefficients used by the RUC.Results: The average total preservice evaluation time was 42.2 minutes. The average time from the patient entering the OR to incision was 40.8 ± 25.4 minute. Immediate postservice tasks took 30.0 minutes. Compared with the 2013 RUC weighted wRVU value of 1.394 for preservice and 0.560 for immediate postservice activities, we found that surgeons actually perform 1.567 wRVUs of preservice and 0.672 of immediate postservice activities.Conclusion: Policymakers should consider these findings when reviewing the time and intensity spent on perioperative care for patients undergoing THA and TKA. [ABSTRACT FROM AUTHOR]

Background: Recently, self-directed physical therapy (SDPT) programs have gained popularity following total knee arthroplasty (TKA). This study evaluated the safety and efficacy of the routine use of an SDPT program in a nonselect patient population.Methods: This is a single-surgeon, retrospective study of 296 consecutive patients from August 2016 to October 2017 discharged home after primary, unilateral TKA and enrolled in a web-based SDPT program. Patients were seen 2 weeks after surgery and outpatient physical therapy (OPPT) was prescribed if flexion was less than 90°, upon patient request, or inability to use the web-based platform.Results: Overall, 195 of 296 (65.9%) patients did not require OPPT (SDPT-only) while 101 of 296 were prescribed OPPT (34.1%, SDPT + OPPT). In SDPT + OPPT, 66.3% were for flexion <90°, 27.7% by patient request, 5.0% received a prescription but did not attend OPPT, and 1.0% due to inability to use the web-based platform. The rate of manipulation under anesthesia was 2.36% overall (SDPT + OPPT, 6.93%; SDPT-only, 0.0%). Multivariate analysis identified elevated Charlson comorbidity index, elevated body mass index, higher preoperative SF12 mental score, and loss of flexion at 2 weeks as independent predictors associated with the need for OPPT.Conclusion: Web-based SDPT is safe and effective for most patients eligible for home discharge after TKA. It is difficult to preoperatively predict those patients who will require OPPT; therefore, we recommend close follow-up. It is critical to preserve these services for patients who require them after TKA as up to a third of patients required OPPT. [ABSTRACT FROM AUTHOR]

Background: Recent data suggested that unsupervised, self-directed physical therapy (SDPT) is both safe and efficacious for patients undergoing total hip arthroplasty (THA) and that formal outpatient physical therapy (OPPT) may not be routinely required. The purpose of this study was to evaluate the routine use of an SDPT program in a nonselect patient population.Methods: This is a multi-surgeon, single-institution, retrospective study of 941 consecutive patients discharged home, from January 2016 to December 2016, after primary, unilateral THA and enrolled in a web-based SDPT program. Patients were seen 4 weeks after surgery and OPPT was prescribed for perceived need, patient request, or if unable to use the web-based program. Patient-reported outcomes, medical comorbidities, and assessment of home environment were prospectively recorded.Results: Overall, 646 of 941 patients (68.7%) were not prescribed OPPT (SDPT-only group) while 295 of 941 patients (31.3%) were prescribed OPPT (SDPT + OPPT group). In the SDPT + OPPT group, 88.2% were for perceived need, 10.8% for patient request, and 1.0% due to inability to use the web-based platform. Multivariate analysis identified male sex (odds ratio, 0.64; 0.45-0.90; P = .012) and a higher preoperative Short Form-12 physical component (odds ratio, 0.98; 0.96-0.99; P = .036) as independent variables protective against requiring OPPT. At a minimum 6-month follow-up, the SDPT-only group had statistically higher hip disability and osteoarthritis outcome score junior compared to the SDPT + OPPT cohort (85.0 vs 80.9; P = .012).Conclusion: Web-based SDPT is safe and effective for most, but not all, patients eligible for home discharge after THA. It is critical to preserve OPPT services for the one-third of patients who require them.Level Of Evidence: Level III. [ABSTRACT FROM AUTHOR]

Background: The use of perioperative dexamethasone has been proven to reduce pain and shorten recovery for patients undergoing total hip and knee arthroplasty. However, the effect of these medications on 90-day readmissions and the rates of clinically significant venous thromboembolic events (VTE) after total joint arthroplasty (TJA) remains unknown.Methods: Patients undergoing unilateral, primary total joint arthroplasty between 2009 and 2016 in a single institution were identified. There were 6617 patients who did not receive dexamethasone intraoperatively compared to 1293 patients who received a single, intraoperative, intravenous dose of dexamethasone (8-10 mg). The primary outcomes were the rate of clinically significant VTE and 90-day readmission. Secondary outcomes included wound complications, periprosthetic joint infection, and 90-day mortality.Results: While the overall rate of clinically symptomatic VTE was lower in the dexamethasone group, this did not reach significance in a univariate analysis (0.1% vs 0.2%, P = .353). Only body mass index (odds ratio [OR], 1.13; 95% confidence interval [CI], 1.03-1.24; P = .006) and longer length of stay (OR, 1.16; 95% CI, 1.06-1.28; P = .001) were associated with VTE in the multivariate analysis (OR, 0.31; 95% CI, 0.04-2.36; P = .319). However, the use of intravenous dexamethasone was independently associated with a reduction in 90-day readmission rate (1.6% vs 2.5%; OR, 0.57; 95% CI, 0.36-0.90; P = .016). There was no difference in the rate of periprosthetic joint infection or mortality.Conclusion: A single, intraoperative, low dose of dexamethasone is not associated with a reduction in clinically significant VTE but may be a safe and effective adjunct medication to lower 90-day readmission rates.Level Of Evidence: Level III. [ABSTRACT FROM AUTHOR]

Background: Aspirin is a safe and effective prophylaxis for the prevention of venous thromboembolism following total joint arthroplasty. The optimal dose of aspirin prophylaxis is unknown. Our hypothesis was that lower-dose aspirin is as effective as higher-dose aspirin for the prevention of venous thromboembolism and is associated with fewer gastrointestinal side effects.Methods: In a prospective, crossover study, we analyzed 4,651 primary total joint arthroplasty cases performed from July 2013 to June 2015. For 4 weeks, 3,192 patients received enteric-coated 325-mg aspirin twice daily (the 325-mg aspirin group) and 1,459 patients received 81-mg aspirin twice daily (the 81-mg aspirin group). There were no significant differences (p > 0.05) in sex, body mass index, or Charlson Comorbidity Index between the two patient populations. Recorded complications occurring within 90 days postoperatively included symptomatic venous thromboembolism (deep venous thrombosis and pulmonary embolism), gastrointestinal complications, acute periprosthetic joint infection, and death.Results: The incidence of venous thromboembolism of 0.1% (95% confidence interval [CI], 0% to 0.3%) in the 81-mg aspirin group (1 with deep venous thrombosis and 1 with pulmonary embolism) was not significantly different (p = 0.345) from 0.3% (95% CI, 0.1% to 0.6%) in the 325-mg aspirin group (7 with deep venous thrombosis and 5 with pulmonary embolism). The incidence of gastrointestinal bleeding or ulceration of 0.3% (95% CI, 0% to 0.5%) in the 81-mg aspirin group was slightly, but not significantly (p = 0.66), lower than the 0.4% (95% CI, 0.2% to 0.6%) in the 325-mg aspirin group. The incidence of acute periprosthetic joint infection was 0.2% (95% CI, 0% to 0.4%) in the 81-mg aspirin group compared with 0.5% (95% CI, 0.2% to 0.7%) in the 325-mg aspirin group (p = 0.28). The 90-day mortality rate was similar in both groups at 0.1% (95% CI, 0% to 0.2%) in the 81-mg aspirin group and 0.1% (95% CI, 0% to 0.2%) in the 325-mg aspirin group (p = 0.78).Conclusions: Our study demonstrates that low-dose aspirin is not inferior to high-dose aspirin for venous thromboembolism prophylaxis following total joint arthroplasty. This is not unexpected, as the available literature demonstrates that low-dose aspirin is as effective as higher-dose aspirin in the prevention of acute coronary syndrome and cerebrovascular events.Level Of Evidence: Therapeutic Level II. See Instructions for Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]

In comparison to primary total knee arthroplasty, surgical time was 1.8 times greater for all knee revisions and 2.4 times greater for complex knee revisions. Knee revisions had an 8.5% higher rate of 90-day repeat procedures. In comparison to primary total hip arthroplasty, surgical time was 1.8 greater for all hip revisions and 2.6 fold greater for complex hip revisions. Hip revisions had a 3.4% higher rate of 90-day repeat procedures. Practices based on revisions or complex revisions alone would see a 32% and 50% decrease in reimbursement respectively compared to the ones based on primary arthroplasty. The projected future increase in primary arthroplasties and the relative incentive to perform primary arthroplasty may soon put patient access to physicians willing to perform revision arthroplasty at risk. [ABSTRACT FROM AUTHOR]

Background: The Coronavirus Disease 2019 (COVID-19) pandemic has caused a substantial number of patients to have their elective arthroplasty surgeries rescheduled. While it is established that patients with COVID-19 who are undergoing surgery have a significantly higher risk of experiencing postoperative complications and mortality, it is not well-known at what time after testing positive the risk of postoperative complications or mortality returns to normal.Methods: PubMed (MEDLINE), Excerpta Medica dataBASE, and professional society websites were systematically reviewed on March 7, 2022 to identify studies and guidelines on the optimal timeframe to reschedule patients for elective surgery after preoperatively testing positive for COVID-19. Outcomes included postoperative complications such as mortality, pneumonia, acute respiratory distress syndrome, septic shock, and pulmonary embolism.Results: A total of 14 studies and professional society guidelines met the inclusion criteria for this systematic review. Patients with asymptomatic COVID-19 should be rescheduled 4-8 weeks after testing positive (as long as they do not develop symptoms in the interim), patients with mild/moderate COVID-19 should be rescheduled 6-8 weeks after testing positive (with complete resolution of symptoms), and patients with severe/critical COVID-19 should be rescheduled at a minimum of 12 weeks after hospital discharge (with complete resolution of symptoms).Conclusions: Given the negative association between preoperative COVID-19 and postoperative complications, patients should have elective arthroplasty surgery rescheduled at differing timeframes based on their symptoms. In addition, a multidisciplinary and patient-centered approach to rescheduling patients is recommended. Further study is needed to examine the impact of novel COVID-19 variants and vaccination on timeframes for rescheduling surgery. [ABSTRACT FROM AUTHOR]

Background: The volume of total knee arthroplasties, including revisions, in young patients is expected to rise. The objective of this study was to compare the reasons for revision and re-revision total knee arthroplasties between younger and older patients, to determine the survivorship of revision total knee arthroplasties, and to identify risk factors associated with failure of revision in patients fifty years of age or younger. Methods: Perioperative data were collected for all total knee arthroplasty revisions performed from August 1999 to December 2009. A cohort of eighty-four patients who were fifty years of age or younger and a cohort of eighty-four patients who were sixty to seventy years of age were matched for the date of surgery, sex, and body mass index (BMI). The etiology of failure of the index total knee arthroplasty and all subsequent revision total knee arthroplasties was determined. Kaplan-Meier survival curves were used to evaluate the timing of the primary failure and the survivorship of revision knee procedures. Finally, multivariate Cox regression was used to calculate risk ratios for the influence of age, sex, BMI, and the reason for the initial revision on survival of the revision total knee arthroplasty. Results: The most common reason for the initial revision was aseptic loosening (27%; 95% confidence interval [CI] = 19% to 38%) in the younger cohort and infection (30%; 95% CI = 21% to 40%) in the older cohort. Of the twenty-five second revisions in younger patients, 32% (95% CI = 17% to 52%) were for infection, whereas 50% (95% CI = 32% to 68%) of the twenty-six second revisions in the older cohort were for infection. Cumulative six-year survival rates were 71.0% (95% CI = 60.7% to 83.0%) and 66.1% (95%CI = 54.5% to 80.2%) for revisions in the younger and older cohorts, respectively. Infection and a BMI of ≥40 kg/m2 posed the greatest risk of failure of revision procedures, with risk ratios of 2.731 (p = 0.006) and 2.934 (p = 0.009), respectively. Conclusions: The survivorship of knee revisions in younger patients is a cause of concern, and the higher rates of aseptic failure in these patients may be related to unique demands that they place on the reconstruction. Improvement in implant fixation and treatment of infection when these patients undergo revision total knee arthroplasty is needed. Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]

Background: Hepatitis C is present worldwide. Little is known about the outcome of joint arthroplasty in asymptomatic patients with hepatitis C. We evaluated the surgical complications following hip and knee arthroplasty in patients who were seropositive for hepatitis C in a matched study. Methods: Seventy-one patients with hepatitis C underwent forty total hip arthroplasties and thirty-two total knee arthroplasties from 1995 to 2006. The patients had normal preoperative liver function tests. Patients with human immunodeficiency virus infection, hepatitis B, and hemophilia were excluded. A control group was matched in a 2:1 ratio with the hepatitis-C group for age, body-mass index, sex, year of surgery, and medical comorbidities, including diabetes, rheumatoid arthritis, and immunosuppressive conditions. Results: In the group of patients with hepatitis C who were managed with total hip arthroplasty, six patients (15%) had wound complications requiring oral antibiotics or wound irrigation and debridement and four hips (10%) had mechanical complications, including implant loosening or dislocation. In the control group, three patients (3.8%) had wound complications requiring oral antibiotics or irrigation and debridement and three patients (3.8%) had mechanical complications, including dislocation, periprosthetic femoral fracture, and implant failure. In the group of patients with hepatitis C who underwent total knee arthroplasty, three patients (9.4%) had mechanical complications, including loosening and periprosthetic fracture requiring revision. In the control group, three patients (4.7%) had wound complications, one (1.6%) had a deep infection requiring two-stage revision, and one (1.6%) underwent revision because of a mechanical problem. The combined hepatitis-C group had significantly longer hospital stays and higher rates of surgical and mechanical complications, reoperation, and revision. Conclusions: Patients with hepatitis C undergoing joint arthroplasty had a higher rate of surgical complications and a longer hospital stay. The reason for the higher rate of complications in this group of patients is unknown, and further investigation is needed. Patients with hepatitis C should be counseled about the potential for a higher incidence of postoperative complications prior to undergoing joint anthroplasty. Level of Evidence: Prognostic Level III. See Instructions to Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]

Abstract: Thromboembolic disease is a relatively common and potentially devastating complication of joint arthroplasty. Mechanical and chemical prophylaxes are effective in reducing the incidence of this complication. Inferior vena cava (IVC) filters have been used to prevent the propagation and/or migration of venous emboli into the pulmonary circulation. This article reports on a cohort of joint arthroplasty patients either with confirmed pulmonary embolism or at high risk of developing pulmonary emboli in whom an IVC filter was implanted. Using the institutional database, 95 joint arthroplasty patients who received an IVC filter were identified. The IVC filter was effective in preventing fatal pulmonary emboli in all patients. The IVC filter is a valuable and effective modality in preventing a fatal outcome of thromboembolism in patients with established thromboembolism. [Copyright &y& Elsevier]

This article is a comprehensive review of the published literature on total hip replacement design. It seeks to provide a nonbiased view of the technology related to total hip implants and bearing surfaces. Implant designs and fixation methods are discussed with summaries of their respective long-term outcome studies. Fixation methods include cemented, cementless and hybrid techniques and are explained and presented with current outcome data and their respective failure modes that have promoted new technological development. The current data point toward cementless acetabular fixation as being superior to cemented fixation. As for stem fixation, there are good data to suggest that either method of fixation is acceptable in the proper setting. Also included in this article is a brief review of bearing surfaces and hip biomechanics. [ABSTRACT FROM AUTHOR]

Abstract: Anatomic aberrations of the femur and tibia secondary to trauma, congenital defects, and prior surgery present challenges for the reconstructive knee surgeon because of an altered mechanical axis and distorted anatomic landmarks. Five patients with arthritis of the knee and extra-articular femoral and/or tibial deformity, retained hardware, or intramedullary (IM) implants underwent total knee arthroplasty using a computer navigation system. The navigation system obviated the need for an IM guide, and the normal mechanical axis of the patients was restored. Extensive dissection for hardware removal or osteotomy was not necessary in these patients. In these 5 cases, a navigation system proved to be an effective tool for restoration of limb alignment in the presence of significant extra-articular deformities and/or IM hardware. Thus, it provides an alternative approach to the traditional IM instrumentation for treating these patients in an effective manner. [Copyright &y& Elsevier]

Abstract: Periprosthetic infections after total joint arthroplasty are infrequent but potentially devastating complications. Gram-positive bacteria is the typical causative organism. However, uncommon sources such as Mycobacterium fortuitum have been documented in the literature. We report a case, the first to our knowledge, of bilateral infections involving this organism after total knee arthroplasty. M. fortuitum is notoriously resistant to many standard antibacterial medications, and a delay in initial diagnosis due to inadequate incubation time has been reported. In poor surgical candidates, long-term antibiotic suppression may be a viable alternative. [Copyright &y& Elsevier]

Background: Traditional pain management after total knee arthroplasty (TKA) relies heavily on opioids. Although there is evidence that in-hospital multimodal pain management (MMPM) is more effective than opioid-only (OO) analgesia, there has been little focus on postdischarge pain management. The hypothesis of this study was that MMPM after TKA would reduce pain scores and opioid consumption in the 30-day period after hospital discharge.Methods: This is a prospective, 2-group, comparative study with a provider cross-over design comparing a 30-day OO prn regimen with a MMPM regimen and opioid medications prn. The primary outcome measure was visual analog scale pain score and opioid-related side effects. Secondary outcome measures included morphine milligram equivalents consumed, failure of the protocol, and opioid refills.Results: There were 216 patients included in the trial, with final data available for 143. There was no clinically meaningful difference in visual analog scale score between the 2 groups at any time. Average opioid consumption at 30 days was 582.5 and 386.4 morphine milligram equivalents for the OO and MMPM cohorts, respectively (P = .0006). Average number of opioid pills consumed at 30 days was 91.8 and 60.4 for OO and MMPM cohorts, respectively (P = .0004).Conclusion: A 30-day postdischarge multimodal pain regimen reduced opioid use after TKA while maintaining a similar level of pain control as the OO regimen. OO regimens are at an increased risk of needing additional medications to control pain.Level Of Evidence: Level II. REGISTRY NAME: www.clinicaltrials.gov.Trial Number: NCT04003350. [ABSTRACT FROM AUTHOR]

Total hip arthroplasty was originally indicated for older, sedentary patients. This was due to concern that catastrophic wear and failure would occur in younger and more active patients. Total hip arthroplasty has since become a viable option for these patients with advances in implant design, tribology, and surgical technique. However, further study is needed to delineate the long-term outcome in this patient population. [ABSTRACT FROM AUTHOR]

Abstract: Trephine reamers have been used in revision total hip arthroplasty to facilitate the removal of femoral prostheses without creating additional osteotomy sites. These reamers are available in graduated sizes so as to minimize bone loss. We present a complication not previously reported in the literature. While performing the extraction of a well-fixed femoral prosthesis, the trephine reamer broke at its distal end, expanded, and created a cortical defect distal to the prosthesis. [Copyright &y& Elsevier]