 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
The Journal of Bone and Joint Surgery (American). 2007;89:2828-2837. doi:10.2106/JBJS.G.01232 © 2007 The Journal of Bone and Joint Surgery, Inc. What's New in Adult Reconstructive Knee Surgery1 Booth, Bartolozzi, Balderston Orthopaedics, Pennsylvania Hospital, 800 Spruce Street, Philadelphia, PA 19107. E-mail address for J.H. Lonner: lonnerj{at}pahosp.com
Specialty Update has been developed in collaboration with the Council of Musculoskeletal Specialty Societies (COMSS) of the American Academy of Orthopaedic Surgeons.
The purpose of this review is to discuss the research presented on selected topics related to adult knee reconstruction during 2006. The articles referenced in this update were selected from both The Journal of Bone and Joint Surgery (American volume) and The Journal of Arthroplasty. Unreferenced studies in this update were reported as podium presentations at the meeting of the American Academy of Orthopaedic Surgeons (held on March 22 through 26, 2006, in Chicago, Illinois), the combined Specialty Day meeting (held on March 22 through 26, 2006, in Chicago, Illinois), the interim meeting of the Knee Society, and the meeting of the American Association of Hip and Knee Surgeons (held on November 3 through 5, 2006, in Dallas, Texas).
Vessely et al. studied trends in the utilization of total hip and knee arthroplasty between 1969 and 2003 in a single United States county to determine, by estimation, the national need for these procedures. The age and gender-adjusted incidence per 100,000 person-years increased over time, from twenty-nine in 1971 to 1975 to 157 in 2000 to 2003, representing a >400% increase in the incidence of total knee arthroplasty (as compared with a 55% increase in total hip arthroplasty during the same period of time). There was a significantly higher utilization rate for women over the course of the study. The incidence increased with increasing patient age for total knee arthroplasty, except in patients more than eighty years old. The largest percentage increase was in patients less than fifty years old. There was a significant increase in the proportion of total knee arthroplasties performed for the treatment of osteoarthritis, from 51% of total knee arthroplasties in 1971 to 1975 to 92% in 2000 to 2003. This reflects a reduction in the incidence of total knee arthroplasties performed for the treatment of rheumatoid arthritis over the period studied. The continued and rapid growth of utilization of total knee arthroplasty reflects a trend that will require additional resources in the future. Kurtz et al.1 used census data and a nationwide inpatient sample that provides an estimate of approximately 25% of inpatient hospitalizations for procedures performed in the United States, regardless of payment source, to predict the volume growth of primary and revision total knee and total hip arthroplasties until the year 2030. It is projected that the number of primary total knee arthroplasties will increase from 450,400 to 3.48 million by 2030, compared with a growth in the number of primary total hip arthroplasties from 208,600 to 572,100 during the same interval. The volume of revision total hip arthroplasties is projected to grow from 40,800 in 2005 to 96,700 in 2030 (a 137% increase), and the volume of revision total knee arthroplasties is projected to grow from 38,300 in 2005 to 268,200 in 2030 (a 600% increase). Given the relatively constant growth in the number of orthopaedic surgeons, the projected case load per surgeon is estimated to grow from 51.9 procedures in 2010 to 167 procedures in 2030. This dramatically increased demand for arthroplasty procedures will require additional discussion regarding the distribution of economic resources; the allocation of surgeons, facilities and resources; and improved operative efficiency. Additionally, given the growth in the number of procedures in younger, more active patients, implant longevity will require further enhancement.
The shifting demographics of patients with unicompartmental knee arthritis, including younger, more active patients, is a major impetus for the growing interest in conservative surgical alternatives such as unicompartmental and patellofemoral arthroplasties. Ongoing scrutiny and evaluation of these procedures, particularly as designs and techniques evolve, are critical. In a prospective study of 142 unicompartmental arthroplasties performed with the same minimally invasive approach, Fisher et al. found better knee scores and pain scores at one year in association with metal-backed tibial components as compared with all-polyethylene tibial components, although these differences were eliminated by three years. At both one year and three years, the prevalence of radiolucent zones beneath the tibial implant was somewhat higher in association with all-polyethylene implants, but this finding did not correlate with pain or function. Complications and reoperations were more frequent in association with all-polyethylene tibial components. Compared with seventy-five metal-backed unicompartmental arthroplasties that were performed through a traditional arthrotomy, those performed through a minimally invasive approach were associated with a better range of motion at three years, suggesting early and sustained benefits of the minimally invasive surgical approach. Carlsson et al.2 reported on forty-one patients undergoing a Miller-Galante unicompartmental arthroplasty with cement (Zimmer, Warsaw, Indiana) who were randomly assigned to be managed with either a minimally invasive approach or a standard approach. Patients who were managed with the minimally invasive approach had significantly faster rehabilitation and a reduction in hospitalization, although there were no other significant differences between the two groups in terms of clinical or radiographic outcomes. In the study by Hamilton et al.3, a series of 221 unicompartmental arthroplasties that were performed with use of a minimally invasive technique and standard instrumentation and without computer assistance were compared with their previous series of 514 medial unicompartmental arthroplasties that had been performed through a traditional extensile approach. In the minimally invasive group, nine knees (4.1%) were revised because of component loosening (eight knees) or a deep infection (one knee). Sixteen additional knees in the minimally invasive group required a total of eighteen additional nonrevision reoperations. While the authors noted accelerated recovery and reduced hospital stay in the minimally invasive group, the rate of revision due to aseptic loosening was 3.7% in the minimally invasive group as compared with 1.0% in the open group and the overall rate of reoperation was 11.3% in the former group as compared with 8.6% in the latter group. Whether unicompartmental knee arthroplasty is a cost-effective method for the treatment of unicompartmental knee arthritis requires a decision analysis. SooHoo et al. found that it is necessary for the survival of a unicompartmental knee replacement to be within three to four years of the assumed survival of a total knee implant in order to be cost-effective. However, Slover et al.4, using a different decision model, found that unicompartmental arthroplasty would be a cost-effective measure for elderly, low-demand patients as long as the annual probability of revision is <4%. The cost-effectiveness of unicompartmental arthroplasty can be impacted by survival rates, implant selection, hospital costs, perioperative mortality, and utility values achieved with each procedure. Patellofemoral arthroplasty is gaining popularity as a treatment method for isolated patellofemoral arthritis. Sisto and Sarin5 reported on twenty-five patellofemoral arthroplasties that had been performed with use of a custom implant and found that, after a mean duration of follow-up of seventy-three months, the mean Knee Society functional and clinical scores were 89 and 91 points, respectively. In that series, no patient required additional surgery and no knee had component loosening. While outcomes have improved in association with contemporary patellofemoral designs, failures resulting from patellofemoral maltracking or tibiofemoral arthrosis may be encountered. Lonner et al.6 reported on a series of twelve failed patellofemoral arthroplasties in ten patients who had a revision to total knee arthroplasty because of isolated tibiofemoral arthritis (six knees), isolated patellofemoral catching and maltracking (three knees) or a combination of both (three knees). Standard posterior stabilized implants were used in each knee, without a need for stems, augments, or structural bone grafts. After a mean duration of follow-up of 3.1 years, the mean Knee Society clinical and functional scores increased from 57 to 96 points and from 51 to 91 points, respectively. At the time of the most recent follow-up, there was no clinical or radiographic evidence of patellofemoral maltracking, loosening, or wear. It appears, therefore, that the results of total knee arthroplasty after a previous patellofemoral arthroplasty are not compromised.
The role of minimally invasive techniques for total knee arthroplasty continues to evolve, with proponents adopting the mini-midvastus, mini-subvastus, and mini-quad incision approaches more than the quadriceps-sparing approach given the greater ease of exposure and the equivalency of results. Despite the success of minimally invasive total knee arthroplasty as reported by many authors and the enthusiasm for the technique as reported in publications from 2005 until recently, not all studies have demonstrated clear advantages of minimally invasive techniques for total knee arthroplasty. In the study by Laskin et al., seventy-four total knee arthroplasties that had been performed without patellar eversion were compared with fifty-seven total knee arthroplasties that had been performed with patellar eversion. The authors found that an acquired patella baja occurred in 37% of patients with patellar eversion during surgery, compared with 12% of those without patella eversion during surgery. The presence of patella baja resulted in reduced flexion and increased pain. Kim et al.7 performed a nonrandomized study in which the short-term results of primary total knee arthroplasties that had been performed with a quadriceps-sparing approach were compared with those of procedures that had been performed with use of a standard incision. No significant differences were found between the groups with respect to knee scores, range of motion, closed suction drainage, blood loss, hospital stay, or radiographic parameters. Iatrogenic tears of the quadriceps tendon, measuring between 2 and 4 cm, often were observed in the quadriceps-sparing group. Aglietti et al.8 performed a prospective, randomized, double-blind study comparing postoperative recovery and early results in two groups of thirty patients undergoing total knee arthroplasty with use of either a mini-subvastus approach or a mini-medial parapatellar arthrotomy with a proximal incision extending 2 to 3 cm into the quadriceps tendon. The tourniquet time, estimated blood loss, range of motion, postoperative pain, and radiographic alignment were similar in the two groups. Kim et al. reported on 100 consecutive bilateral total knee arthroplasties in patients who were prospectively randomized to undergo either a standard parapatellar arthrotomy or a quadriceps-sparing arthrotomy. Pain scores, functional scores, range of motion, radiographic parameters, and blood loss were found to be similar between the two groups, whereas operating times, tourniquet times, and technical difficulties were greater in the quadriceps-sparing group. The investigators did not find any early benefit derived from the quadriceps-sparing approach. We believe that most current scoring systems, however, are poor determinants of early functional outcome and may not capture the early benefits of minimally invasive approaches, which have been reported by others.
Many centers are harnessing the power of computers in an attempt to increase the precision of total knee arthroplasty component implantation and soft-tissue tensioning. Most studies to date have focused on attempts to demonstrate a significant improvement in implant positioning, whereas few have aimed at identifying a clinical advantage. In a case-control study by Stulberg et al.9, seventy-eight patients underwent total knee arthroplasty with use of either a conventional or a computer-assisted approach. Postoperative radiographs were evaluated by a blinded observer to assess the anterior-posterior mechanical axis and sagittal component alignment. Additionally, patients were followed prospectively with regard to clinical and functional outcomes. There were no significant differences between the conventional and computer-assisted groups with regard to limb alignment or component positioning. At one and six months, there were no significant differences between the groups with regard to clinical, functional, or pain scores. The surgical time was increased by twenty-seven minutes in the computer-assisted group, and the quantity of blood transfused was slightly higher. The authors raised the possibility that the use of computed tomography for the assessment of limb alignment and component positioning could have identified small differences that radiographs did not discern. It is theoretically possible that these small differences could influence prosthetic wear and the long-term results of knee arthroplasty, an end-point that has yet to be evaluated. In the study by Seon and Song10, forty-nine navigation-assisted midvastus total knee arthroplasties were compared with fifty-three standard manual medial parapatellar total knee arthroplasties. Although some early differences in pain scores and time to functional milestones were appreciated in the navigation-assisted-surgery group, no differences were significant by two weeks. Assessment of postoperative radiographs identified no differences in the mean alignment of the prostheses but did reveal fewer "outliers" in the navigation-assisted-surgery group.
It is becoming increasingly clear that the early results after total knee arthroplasty are highly dependent on perioperative medical management. Hypotension, anemia, nausea, and pain have been targeted by recovery protocols in an effort to promote early walking after surgery and to limit the symptoms that delay the achievement of functional milestones. Vendittoli et al.11, in a prospective, randomized study, compared the postoperative morphine consumption by forty-two patients who had been randomized to receive either (1) multimodal analgesic injection or (2) standard parenteral morphine only. All forty-two patients received spinal anesthesia. The injection group received an intraoperative injection of a mixture containing high-dose ropivacaine, ketorolac, and epinephrine into the deep and subcutaneous tissues about the knee. One day after surgery, ropivacaine was again injected into the knee through a catheter that had been positioned intraoperatively. There were no complications in association with ropivacaine use, and plasma levels remained below the toxic threshold, despite the authors' use of doses greater than the manufacturer's recommended maximum. The patients who received the multimodal protocol consumed significantly less morphine during the first eight to forty hours after surgery, using only about half the amount used by the control group over the first twenty-four hours. Additionally, visual analog pain scores were significantly lower in the treatment group for as long as forty hours after surgery, and the mean duration of postoperative nausea was reduced by more than half. Discharge times and functional scores were not significantly different between the groups. Busch et al.12 also performed a randomized, prospective trial evaluating the efficacy of multimodal periarticular injections during surgery. A total of sixty-four patients were included, with the test group receiving an injection mixture of high-dose ropivacaine, ketorolac, epimorphine, and epinephrine into the deep and subcutaneous tissues during surgery. The authors observed significantly less mean patient-controlled anesthesia use at six and twelve hours after surgery as well as significantly increased patient satisfaction scores and decreased pain scores within four hours after surgery. There were no adverse events associated with the injections. At six weeks, there were no differences between the groups. A number of publications have also highlighted more specific strategies for perioperative pain management. Kullenberg et al.13, in a randomized study of eighty-six patients, evaluated the use of postoperative cryotherapy after total knee arthroplasty. The authors showed that cold compression therapy resulted in a significantly faster discharge from the hospital (4.8 days compared with 6.2 days) as well as significantly improved range of motion at the time of discharge and at three weeks. There were no differences between the groups with regard to pain scores. Hartrick et al.14, in a multicenter, randomized, double-blind study of 168 patients, evaluated the efficacy of extended-release epidural morphine injection (20 to 30 mg) before total knee arthroplasty. When compared with a sham epidural, the treatment group exhibited significantly reduced pain-recall scores at all intervals up to thirty hours after total knee arthroplasty. The treatment group also used threefold lower quantities of supplemental postoperative opioids, with some patients exhibiting no need for supplemental opioids. Four patients in the treatment group, all of whom were more than sixty-five years old, had serious respiratory depression. De Ruyter et al.15 studied the efficacy of continuous femoral perineural injection catheters. A total of fifty patients were studied retrospectively, with twenty-four patients receiving the continuous femoral injection and twenty-six receiving intravenous opioids only. Patients who were managed with the femoral nerve catheter used significantly less supplemental opioids and exhibited significantly greater range of motion for two days after surgery and had an improved ability to adjust from a supine to a sitting position. The length of stay was reduced from 4.2 to 3.6 days in this group as well. A number of authors also have studied perioperative blood management in an attempt to maintain acceptable hemoglobin levels through surgery. Preoperative administration of epoetin alfa (Amgen, Thousand Oaks, California) was compared with autologous blood donation before total knee arthroplasty. Deutsch et al.16, in a prospective, randomized study of fifty patients, found that patients who received epoetin alfa had higher hemoglobin levels through the second postoperative day. Twenty-two of twenty-five patients in the autologous donation group received an autologous transfusion after surgery, and two of twenty-five received allogeneic blood. None of the twenty-five patients in the epoetin alfa group received autologous blood (as dictated by the study protocol); however, seven of the twenty-five patients received allogeneic blood. This threefold increase in the number of allogeneic transfusions in the epoetin alfa group, however, did not reach significance. There were few adverse events, and the authors concluded epoetin alfa is a safe alternative to autologous donation in patients who are at risk for transfusion. However, in order to make a definitive conclusion, the risk derived from allogeneic blood and the potentially elevated risk of thrombogenesis associated with the use of erythropoietin should be considered. Clark et al.17 studied the efficacy of the Orthopedic Perioperative Autotransfusion System (Zimmer, Warsaw, Indiana) by enrolling five patient cohorts comprising a total of 398 patients. The cohorts included patients undergoing unilateral primary and revision hip and knee arthroplasty as well as bilateral total knee arthroplasty. The Orthopedic Autotransfusion System collects blood intraoperatively and immediately postoperatively and processes the red blood cells for reinfusion. The authors found that its use significantly reduced the need for allogeneic transfusion by twofold to threefold in patients without preoperative autologous blood donation undergoing primary or revision hip arthroplasty or knee arthroplasty. After analysis of patient groups, the authors concluded that the Orthopedic Perioperative Autotransfusion System is useful for defined subsets of patients, potentially obviating the need for preoperative blood donation. Work on the perioperative management of patients has led to decreasing discharge times, decreasing symptoms after surgery, and more rapid return to walking. It is quite possible that these improvements will lead to decreasing medical complications that result from immobility, narcotic use, and lengthy hospital stays.
Selection of the method of prophylaxis against venous thromboembolism remains a controversial topic among proponents of various methods and variable outcome measures. The debate regarding venous thromboembolism also centers considerably on the issue of the balance between the efficacy of the medication and its safety profile. Burnett et al. performed a prospective study to investigate the use of low-molecular-weight heparin for prophylaxis against venous thromboembolism in patients undergoing total hip or knee arthroplasty. After procedures in which a ten-day course of low-molecular-weight heparin was used, the prevalence of symptomatic deep-vein thrombosis was 3.8% and the prevalence of non-fatal pulmonary emboli was 1.3%. The rates of surgical site complications necessitating readmission, irrigation and débridement of a hematoma and the wound, or prolonged hospitalization for wound drainage were 4.7%, 3.4%, and 5.1%, respectively. Wound drainage occurred for four to seven days after 9.3% of the procedures and for more than seven days after 9.3% of the procedures, with more than seven days of drainage being highly predictive of readmission and wound reoperation. Furthermore, a body mass index of >35 was predictive of prolonged wound drainage when low-molecular-weight heparin was used. The risk of excessive anticoagulation is not without consequence. Parvizi et al. found that patients who have development of a postoperative hematoma, wound drainage, or both had a significantly higher risk of development of a deep periprosthetic infection than those who did not. Given equivalency of efficacy in the prevention of pulmonary emboli, surgeons may be willing to accept a higher prevalence of deep-vein thrombosis after total knee arthroplasty in association with the use of anticoagulants that are less likely to cause bleeding complications. Lonner et al.18 found no significant difference in the prevalence of postthrombotic syndrome in patients with asymptomatic deep-vein thrombosis after total knee arthroplasty as compared with those without venographically proven deep-vein thrombosis. Westrich et al.19 performed a randomized study of 275 patients undergoing unilateral total knee arthroplasty with spinal epidural anesthesia, pneumatic compression devices, and administration of either enoxaparin or aspirin postoperatively for four weeks. All patients received ultrasound screening on the third to fifth postoperative day and a second follow-up ultrasound examination four to six weeks after surgery. The overall prevalence of deep-vein thrombosis was not significantly different between the drug treatment groups. In the study by Gelfer et al.20, pneumatic compression combined with low-dose aspirin was compared with enoxaparin. The study included patients undergoing total hip arthroplasty as well as those undergoing total knee arthroplasty. Sixty patients were randomized to receive enoxaparin, and sixty-one were randomized to receive pneumatic compression and aspirin. Venographic evidence of deep-vein thrombosis was observed in 28.3% of the patients in the enoxaparin group, compared with 6.6% of those in the pneumatic compression/aspirin group. While this difference was significant, the prevalence of adverse events was similar in both groups. Pellegrini et al.21, in a series of 477 total knee arthroplasties, evaluated the rate of readmission for the treatment of venous thromboembolism after surgery. Patients who were discharged from the hospital without additional anticoagulant prophylaxis beyond the use of warfarin while in the hospital had a rate of readmission for the treatment of venous thromboembolism of 1.05%, compared with a rate of 0.21% for those who received six weeks of warfarin prophylaxis with a target International Normalized Ratio of 2.0. Extended prophylaxis was therefore advocated by the authors.
The Centers for Medicare and Medicaid Services (CMS) have initiated steps to standardize protocols in an attempt to enhance outcomes and reduce complications associated with hip and knee arthroplasty. Many of these recommendations are controversial and will not satisfy all clinicians. One controversial measure is the pay-for-performance initiatives linking reimbursement for hip and knee arthroplasty to the timing of dosing of perioperative antibiotics. The objective of this program is to reward hospitals that are compliant with the timing of administration of antibiotics before primary joint replacement surgery. The CMS-recommended protocol is that antibiotics should be given within one hour before the surgical incision is made. Individual surgeons and hospitals are instituting policies to improve compliance with this initiative. Bhattacharyya reported that 16% of patients undergoing total hip arthroplasty and 10% of those undergoing total knee arthroplasty failed to receive optimal antibiotic dosing (within one hour before the skin incision). Better compliance with optimal antibiotic dosing was seen among surgeons performing fifty or more joint replacements annually as compared with those performing fewer than fifty joint replacements annually. Additionally, longer induction times were associated with the administration of antibiotics outside the one-hour window. Another potential pay-for-performance initiative relates to prophylaxis against venous thromboembolism. Recommendations by the CMS are based on the controversial 2004 recommendations of the American College of Chest Physicians (ACCP), which emphasize a reduction in the rate of deep-vein thrombosis but accept a relatively high rate of bleeding and wound complications. For elective total knee arthroplasty, the ACCP-recommended thromboembolism prophylaxis includes the use of low-molecular-weight heparin, Factor Xa inhibitor, and warfarin (with a target International Normalized Ratio of >2). Patients who are considered to be at increased risk of bleeding and those who receive epidural or spinal anesthesia may pass performance measures if they receive mechanical prophylaxis alone, namely, intermittent pneumatic compression. The American Academy of Orthopaedic Surgeons (AAOS) has countered the CMS mandate, with a focus more on reducing the risks of symptomatic pulmonary emboli and less on the reduction of deep-vein thromboses, which are often of little clinical consequence. The rationale for these AAOS recommendations is that aggressive anticoagulation has been shown to be associated with an increased risk of hemorrhage and bleeding complications. Whether the recommendations of the AAOS will become an initiative that guides performance and is adopted by the CMS is unclear at this point. We believe that chemoprophylaxis for patients undergoing hip or knee replacement requires risk assessment for each patient regarding the possibility of pulmonary embolism and major bleeding. In the AAOS guidelines, aspirin, low-molecular-weight heparin, synthetic pentasaccharides, warfarin, and mechanical prophylaxis are each considered to have a role in the effective prevention of pulmonary emboli (www.aaos.org/Research/guidelines/PE_guideline.pdf).
Total knee arthroplasty remains the preferred surgical treatment for the majority of patients with end-stage joint disease. The overwhelming rates of long-term success and survival following total knee arthroplasty have set a high standard that the less-invasive unicompartmental approaches strive to match. Vessely et al.22 received the 2006 Ranawat Award from the Knee Society for their survivorship analysis of 1008 cemented cruciate-retaining modular condylar total knee replacements. The average fifteen-year survival rate was 97% with component removal for mechanical failure as the end point and 98.8% with component removal for aseptic loosening as the end point. The prosthetic survival rate was found to be significantly decreased for patients less than sixty years old. Rasquinha et al.23 reported on 150 consecutive cemented modular press-fit condylar cruciate-substituting total knee replacements. After a mean duration of follow-up of twelve years, the radiographic and clinical survival rate with mechanical failure as the end point was 98.3%. Sixty-one percent of the patients reported participation in recreational sports, and 95% were fully satisfied with the overall result. Lachiewicz and Soileau24 reported the ten-year follow-up results of primary knee arthroplasties that had been performed with constrained components. The patients had a mean age of sixty-seven years and a mean weight of 154 pounds (69.9 kg). The reasons for the use of constrained implants included severe valgus deformity with medial ligament insufficiency and severe flexion contracture with poor flexion/extension balance. Of the fifty-four knees that were studied, only two were revised, including one because of tibial component loosening and one because of femoral component loosening. The average range of motion was 1.7° to 97°, with mostly good to excellent results. The ten-year survival rate was 96%, even with the constrained implant. Sathappan et al.25 evaluated the intermediate-term results of 114 total knee arthroplasties involving posterior cruciate ligament recession or excision in conjunction with a conforming polyethylene tray (without a post). The goal of the implant was to provide articular conformity to limit polyethylene insert wear while dispensing with the polyethylene post to avoid cam-post wear. The patients had a mean age of sixty-eight years and a mean weight of 183 pounds (83.0 kg). After an average duration of follow-up of eight years, there was no significant difference in terms of the range of motion or the Knee Society or function scores when patients who had had posterior cruciate ligament excision were compared with those who had had posterior cruciate ligament recession. Survivorship analysis, with inclusion of two failures that resulted from a traumatic fracture, demonstrated a 95% rate of radiographic and clinical survival at ten years. Sheng et al.26 reported the results of revision total knee arthroplasty as recorded in the nationwide Finnish Arthroplasty Registry. Two thousand six hundred and thirty-seven revision total knee arthroplasties over a twelve-year period were reviewed. The mean age of the patients at the time of revision was sixty-nine years. The most common reasons for revision included loosening (32%) and patellar complications (32%). Survivorship analysis with repeat revision as the end point revealed a survival rate of 95% at two years, 89% at five years, and 79% at ten years. Age was inversely correlated with survival of the components at five years, with an 82% survival rate in patients younger than fifty-six years and a 92% rate for those older than seventy years. Additionally, the prosthetic survival rate for patients who had the revision surgery within five years after the primary arthroplasty was worse than that for patients who had the revision surgery more than five years after primary arthroplasty (85% compared with 92%). Patellar subluxation before revision was another significant predictor of failure after revision. In 2006, at both the annual meeting of the AAOS and the fall meeting of the American Association of Hip and Knee Surgeons (AAHKS), a number of speakers reported on outcomes after total knee arthroplasty. Rosenzweig reported the minimum ten-year follow-up results for thirty-two patients undergoing cruciate-retaining bilateral total knee arthroplasty with the left and right sides being randomized for patellar resurfacing or nonresurfacing. The authors noted equivalent results when considering range of motion, the Knee Society score, patellofemoral pain, and revision rates. There appeared to be a small preference among the patients for the nonresurfaced side. Mahoney presented the results of a randomized prospective study of 100 total knee arthroplasties in which mobile-bearing arthroplasties were compared with fixed-bearing arthroplasties at a minimum of two years. Although the results were generally equivalent with regard to complications and function, the mobile-bearing group demonstrated significantly better stair-climbing function. A number of reports were also presented on total knee arthroplasty in young patients. Duffy reported the twenty-year results of total knee arthroplasty in patients with rheumatoid arthritis who were less than fifty-five years of age. Analysis of the results of forty-seven total knee arthroplasties with cement revealed an implant survival rate of 93.7%. All six revisions occurred after seventeen years and were due to polyethylene wear and osteolysis. He also reported on the fourteen-year results of total knee arthroplasty in patients with degenerative arthritis who were younger than fifty-five years of age. The estimated implant survival rate was 96% at ten years and 85% at fifteen years. Polyethylene wear and osteolysis led to a decrease in the survival of the components after ten years. Novak presented the results of 1047 total knee arthroplasties, performed from 1991 to 2005, in patients who were fifty-five years of age or younger. When the cemented implants were analyzed in isolation, the survival rate was 85% at 14.3 years. Lower survival rates were observed in association with cementless total knee arthroplasty as well as for men. Additionally, there were many presentations on outcomes of total knee arthroplasty in specific patient populations. Kubiak reported on 111 cruciate-retaining total knee arthroplasties in patients with at least 10° of varus or 20° of valgus deformity. The mean age of the patients was seventy-one years, and a variety of implants were utilized. There was only one revision due to osteolysis, and there were no revisions due to instability. Lombardi reported the seven-year results of primary and revision total knee arthroplasty after patellectomy in forty patients with an average age of sixty-two years. A posterior stabilized arthroplasty design was used in twenty-five cases, constrained components were used in eleven cases, and a rotating hinge was used in four cases. Seven knees were revised because of infection, instability, wear, or extensor complications; seven patients had an extensor lag; and seven knees required manipulation. The author concluded that despite an increased rate of revision, most patients had substantial clinical improvement without a substantial extensor lag. He found that immobilization did not appear to minimize or prevent an extensor lag but that it was associated with stiffness resulting in the need for manipulation. Farfalli reported that the results of total knee arthroplasty after proximal tibial osteotomy were relatively inferior to those of primary total knee arthroplasty for the treatment of degenerative joint disease. Analysis of thirty-four cemented total knee replacements at a mean of 5.6 years revealed significantly less satisfactory results due to soft-tissue complications, osteoarticular defects, residual tibial deformity, and the presence of hardware and infection. The authors recommended great care in the decision to utilize proximal tibial osteotomy. Patel presented the results of sixty-eight total knee arthroplasties in patients with Child class-A liver cirrhosis with an average duration of follow-up of fifty-one months. Fifteen percent of the patients had a superficial or deep infection after surgery. Advanced age, platelet count, and hepatitis-B-related cirrhosis were independent risk factors for infection. The authors found that a history of hepatic decompensation or variceal bleeding was an independent risk factor for serious complications and suggested that total knee arthroplasty is contraindicated in these patients. When total knee arthroplasty is offered to patients who have cirrhosis and these specific risk factors, the authors empirically suggested the use of antibiotic-impregnated cement and the administration of intravenous antibiotics for three days after surgery to reduce the probability of infection.
SooHoo et al.27 analyzed California hospital admissions from 1991 through 2001 to identify factors associated with readmission for morbidity and mortality after total knee arthroplasty. Interestingly, age, gender, race, comorbidity index, insurance type, and hospital volume were independently associated with readmissions for mortality, pulmonary embolism, and infection. It is becoming clear that complication rates after total knee arthroplasty depend strongly on the demographic characteristics of the patient population that a surgeon is serving. Patellofemoral complications after total knee arthroplasty have been mitigated by a greater appreciation of component rotation and position as well as design improvements; however, they remain a potential source of problems. There has been a recent focus on intraoperative technique and the resultant patellofemoral effects. Ritter et al.28 evaluated the effect of lateral patellar contact with the femoral component on postoperative pain and outcome in a retrospective study of 980 consecutive total knee arthroplasties in which the mean duration of follow-up was 3.7 years. Paradoxically, lateral patellar contact was associated with a lower risk of postoperative pain. The overall rate of patellofemoral complications was extremely low, and the authors emphasized the benefits of medializing the patellar button without concern for lateral patellar contact. Clarke et al.29 reported on 238 total knee arthroplasties after a minimum duration of follow-up of two years, with a focus on the changing patterns of patellofemoral complications associated with improvements in femoral component design. Unlike the historical observation of a patellar clunk in association with a proportion (as high as 20%) of Insall-Burstein posterior stabilized prostheses (Zimmer, Warsaw, Indiana), the authors observed no patellofemoral complications in association with the use of the NexGen Legacy posterior stabilized prosthesis (Zimmer). The authors attributed this improvement in clinical results to changes in the sagittal geometry of the femoral component, the introduction of side-specific components with lengthening and deepening of the trochlea, and an increased variety of sizes. Newbern et al.30 evaluated the need for lateral retinacular release in a study of 2381 knees in which either the posterior condylar axis or the transepicondylar axis was used to determine the femoral component rotation. Although three-year outcomes (including stair-climbing, pain, and range of motion) and patellar radiographs were studied, the only difference between the groups was the rate of lateral retinacular release. Specifically, a lateral release was required for 57% of the knees in which the posterior condylar axis was used, compared with only 12% of those in which the transepicondylar axis was used. Thus, the authors recommended the routine use of the transepicondylar axis. Masri et al.31 studied the effect of retaining a patellar prosthesis on the outcome after revision total knee arthroplasty. At a minimum of two years of follow-up, 126 revision total knee arthroplasties were reviewed with regard to functional outcomes and patient satisfaction. There were no significant differences between the fifty-eight patients with a retained patellar component and the fifty-two patients who were left with just an osseous patellar shell. The authors suggested that it is reasonable to leave an osseous shell at the time of revision instead of taking great efforts to resurface a patella with suboptimal osseous architecture. A recent focus on range of motion after knee replacement has led to substantial research on the causes of stiffness and the intraoperative factors contributing to the amount of knee flexion. Gandhi et al.32 described the risk factors for the development of postoperative stiffness (defined as flexion of <90°) in an analysis of 1216 primary total knee arthroplasties. At one year, they found a 3.7% rate of stiffness. Matched case-control analysis revealed preoperative and intraoperative flexion as being predictive of eventual flexion of <90°. Additionally, decreased patellar height preoperatively and postoperatively was associated with stiffness. The authors found no correlations with medical comorbidities. Bengs and Scott33 studied the effect of patellar thickness on the passive flexion arc. Custom patellar button trials were utilized intraoperatively during thirty-one consecutive total knee arthroplasties to test the effect of patellar thickness on flexion. On the average, for every increase of 2 mm in patellar thickness, a decrease of 3° in maximum passive flexion resulted. The authors observed no effect of patellar thickness on tracking. They concluded that patellar thickness has minimal effect on postoperative flexion and should not cause the surgeon to remove excessive patellar bone stock in an effort to gain more flexion. Namba presented the results of early and late manipulation for the treatment of stiffness in flexion. One hundred and two patients underwent early manipulation (less than ninety days postoperatively) and ninety-three patients underwent late manipulation (more than ninety days postoperatively). In the early manipulation group, the average flexion improved from 68° to 101°. In the late manipulation group, the average flexion increased from 81° to 98°. All patients had a decrease in pain; however, only the early manipulation group had improvement in extension. The author concluded that although early and late manipulation can improve range of motion, the results of early manipulation are superior. Fehring described the results of revision total knee arthroplasty for the treatment of flexion contractures. He described nineteen patients with symptomatic flexion contractures of >15° who underwent a revision total knee arthroplasty. Postoperatively, the average extension was 5°, with eleven patients having a contracture of <5°. Infection continues to present a challenge in the management of patients who undergo total knee arthroplasty. Barrack reported the results for patients who had an unexpected positive bacterial culture at the time of revision total knee arthroplasty. Of 692 patients who had a revision for the treatment of aseptic complications, 5.3% had an unexpected positive intraoperative culture. Twenty-four percent of the patients who had an unexpected positive culture had other evidence that increased the suspicion of infection and were managed with six weeks of intravenous antibiotics postoperatively. The remainder of the patients who had an unexpected positive culture were not managed with antibiotics because the culture findings were regarded as false-positive results and because of the absence of any other information suggesting infection. These untreated patients had no occurrence of infection at a minimum of two years follow-up. Fulkerson et al.34 reviewed 194 positive cultures of specimens that had been obtained at the time of reoperation at the site of a total hip or knee arthroplasty in order to determine antibiotic sensitivity. Overall, they found that 88% of the bacteria were sensitive to gentamicin, 96% were sensitive to vancomycin, and 61% were sensitive to cefazolin. Acute postoperative infections and infections that had resisted previous antibiotic treatments exhibited the most antibiotic-resistant strains. Hematogenous infections were associated with a high sensitivity to both cefazolin and gentamicin. The authors recommended treating acute hematogenous gram-positive infections with cefazolin and gentamicin until the final culture results are available, and they recommended treating all chronic and acute postoperative infections with vancomycin until the final culture results are available. Della Valle presented data on the utility of a synovial fluid cell count for the diagnosis of infection before revision total knee arthroplasty. One hundred and five patients who were awaiting reoperation after total knee arthroplasty were tested for determination of the erythrocyte sedimentation rate, C-reactive protein level, synovial fluid cell count, and differential cell count. Intraoperative cultures, purulence, and histopathological findings were used as the criteria for infection. A synovial fluid cell count of >3000 was found to be the most useful preoperative test for infection, with a sensitivity of 100% and a specificity of 98%. Huang et al.35 reported on twenty-one patients who underwent two-stage treatment of an infection at the site of a total knee arthroplasty with an articulating spacer. A modified V-Y quadricepsplasty was needed to obtain adequate exposure in six cases. The average duration of follow-up was fifty-two months. The average range of motion was 85° with the articulating spacer in place and 98° after reimplantation. One knee became reinfected during the follow-up period. Trezies et al.36 reported on the fate of eleven articulating spacers (comprising a femoral component and polyethylene tray) that were left in situ without a planned reimplantation stage. One of the eleven spacers was revised because of infection, two spacers were revised because of pain, and two patients died with the spacer in situ. The spacers in the remaining six patients were intact after an average duration of follow-up of sixty-five months, at which time the average Knee Society Score was 167. These results suggest that articulating spacers do not necessarily require second-stage reimplantation in patients who are at a very high risk of complications from reimplantation.
The editorial staff of The Journal reviewed a large number of recently published research studies related to the musculoskeletal system that received a Level of Evidence grade of I. Over 100 medical journals were reviewed to identify these articles, which all have high-quality study design. In addition to the articles already cited in this Update, four level-I articles were identified that were relevant to reconstructive knee surgery. A list of those titles is appended to this review after the standard bibliography. We have provided a brief commentary about each of the articles to help guide your further reading, in an evidence-based fashion, in this subspecialty area.
Tsumara N, Yoshiya S, Chin T, Shiba R, Kohso K, Doita M. A prospective comparison of clamping the drain or postoperative salvage of blood in reducing blood loss after total knee arthroplasty. J Bone Joint Surg Br. 2006;88:49-53. In this prospective, randomized study, 212 patients with degenerative arthritis who were undergoing unilateral total knee arthroplasty were enrolled and assigned to one of two groups. In Group 1, following surgery, clamping of the closed suction drain was performed after the injection of saline solution with adrenaline (1:500,000) into the knee joint through the drain tube. In Group 2, postoperative blood salvage was utilized. The groups were matched with regard to the number of patients, the ratio of men to women, the age at the time of surgery, diagnosis, and comorbidities. Spinal anesthesia was used for 206 patients, and general anesthesia was used for six. All total knee arthroplasties were performed with use of a medial peripatellar approach and a tourniquet, which was not released before skin closure. In the blood salvage group, collected blood was reinfused within six hours after surgery, after which the drain remained in place without continuous suction. The mean postoperative drain blood volume was 352 mL after drain clamping, compared with 662 mL after blood salvage. There was no significant difference between the groups with regard to the reduction of the hemoglobin level. In the blood salvage group, drain blood was reinfused in 71.7% of the cases. There was no significant difference between the two groups with regard to the need for allogeneic blood transfusion. The authors concluded that drain clamping with intra-articular injection of saline solution with adrenaline was more effective than postoperative autologous blood transfusion for reducing blood loss during total knee arthroplasty, although the basis for the conclusion is not clearly established in the manuscript. The results of this study are dependent on hospital-specific triggers for an allogeneic blood transfusion as well as the specific type of blood salvage method utilized. Furthermore, the possible long-term risk of infection resulting from postoperative fluid injection through a drain tube is difficult to quantify. Keene G, Simpson D, Kalairajah Y. Limb alignment in computer-assisted minimally-invasive unicompartmental knee replacement. J Bone Joint Surg Br. 2006;88:44-8. Twenty patients undergoing simultaneous bilateral medial unicompartmental knee arthroplasty were randomly assigned to undergo unicompartmental arthroplasty with use of conventional instrumentation on one side and computer-assisted surgery through a comparable surgical approach on the contralateral side. One implant system was utilized in all cases, although there was a mixture of fixed-bearing all-polyethylene tibial implants, fixed-bearing metal-backed tibial implants, and mobile-bearing tibial implants. Preoperative hip-knee-ankle alignment and valgus stress radiographs were used to plan the desired postoperative alignment of the limb. In each patient, the planned postoperative alignment was the same in both knees, averaging 2.3° of varus (range, 0° to 5° of varus). The mean variation between the preoperative plan and the achieved correction was 0.9° for the limbs in the computer-assisted surgery group, compared with 2.8° for those in the conventional instrumentation group. The authors found that the difference in variation between the groups was significant. Lower limb alignment was within 2° of the preoperative plan in 60% of the limbs in the conventional instrumentation group, compared with 87% of those in the computer-assisted surgery group. The authors concluded that computer-assisted surgery significantly improves postoperative alignment of medial unicompartmental knee arthroplasty. The clinical importance of small changes in limb alignment remains unclear; however, there is evidence that larger deviations in limb alignment lead to more rapid polyethylene wear. Kalairajah Y, Cossey AJ, Verrall GM, Ludbrook G, Spriggins AJ. Are systemic emboli reduced in computer-assisted knee surgery? A prospective, randomised, clinical trial. J Bone Joint Surg Br. 2006;88:198-202. Erratum in: J Bone Joint Surg Br. 2006;88:1407. In this prospective, randomized study, fourteen total knee arthroplasties were performed with computer-assisted techniques and ten total knee arthroplasties were performed with a standard conventional surgical technique involving the use of tibial and femoral intramedullary alignment guides. In both groups, cranial embolization was evaluated with use of noninvasive transcranial Doppler devices. Other than the use of navigation or intramedullary alignment instruments, the surgical approach, surgeon, technique, and release sequences were identical between groups. All patients had spinal or general anesthesia and femoral nerve blocks. In the instrumentation group, techniques were used to reduce marrow embolization, including overdrilling of the holes and use of fluted rods to allow egress of marrow elements. Tourniquets were utilized in all limbs, with deflation of the tourniquets only after the dressings were applied. While transcranial Doppler signals were acceptable in all patients in the computer-assisted surgery group, they were considered to be acceptable in only 90% of those in the conventional surgery group. Additionally, high-intensity signals were detected in only seven (50%) of the fourteen patients in the computer-assisted group and in all nine of the patients in the conventional group who had acceptable signals. There was a significant difference in the prevalence of detectable emboli between groups (p = 0.0003). In the computer-assisted group, the mean number of emboli was 0.64, with no patient having more than two detectable emboli. In the conventional group, the mean number of emboli was 10.7 (range, one to forty-three), with six patients having more than two detectable emboli. Almost all emboli occurred at the time of, or soon after, insertion of the femoral rod in the conventional group and at the time of insertion of the trial prosthesis in the computer-assisted group. The authors concluded that the prevalence of systemic emboli detected with transcranial Doppler ultrasonography was significantly lower for computer-assisted total knee arthroplasty as compared with conventional total knee arthroplasty with use of intramedullary femoral and tibial jig-based systems. This conclusion supports the findings of previous studies suggesting that intramedullary systems are associated with a higher systemic embolization rate, a factor that must be carefully considered along with the other risks and benefits of computer-assisted arthroplasty. Cobb J, Henckel J, Gomes P, Harris S, Jakopec M, Rodriguez F, Barrett A, Davies B. Hands-on robotic unicompartmental knee replacement: a prospective, randomised controlled study of the acrobat system. J Bone Joint Surg Br. 2006;88:188-97. Twenty-seven patients undergoing twenty-eight unicompartmental knee arthroplasties were enrolled in a prospective, randomized, controlled trial in which conventional instrumentation was compared with a hands-on robotic system. In both groups, a minimally invasive approach was utilized. Five of thirteen patients in the robotic system group were women, compared with eight of the fifteen patients in the conventional instrumentation group. One surgeon performed each of the thirteen procedures with the robotic system, and four surgeons performed the fifteen conventional procedures. Preoperative and postoperative computed tomography demonstrated that all patients in the robotic system group had coronal plane tibiofemoral alignment within 2° of the planned position, whereas this level of accuracy was achieved in only 40% of the patients in the conventional instrumentation group. The mean tibiofemoral alignment in the coronal plane was 0.65° ± 0.59° (range, –1.6° to 0.3°) in the robotic system group, compared with –0.84° ± 2.75° (range, –4.2 to 4.2°) in the conventional instrumentation group. These results were significant (p < 0.001). The mean operating time was not significantly different between the groups. There was a significantly greater mean increase in the Knee Society score in the robotic system group as compared with the conventional instrumentation group at both six weeks and three months, although the Knee Society score is not typically an effective measurement of early functional recovery.
|
Top
Introduction
Epidemiology