Verneuil is recognized as performing the first interpositional arthroplasty of the knee in 1863 when he inserted a flap of joint capsule between the distal femur and proximal tibia.⁶,⁷ The clinical outcomes of this initial interpositional arthroplasty are unknown. Others continued to develop and experiment with interpositional arthroplasty utilizing various materials, including fascia lata, muscle, chromicized pig bladder, silk, gutta percha, magnesium, gold, silver, rubber, cellophane, nylon, and other materials. Dr John D. Murphy, working at Rush Medical College and Northwestern Medical College in Chicago during the first decades of the 20th century, conducted experiments with interposition of local tissues, pedunculated fascia and fat, into ankylosed joints and reported on the histologic appearance of these tissues in animal models as well as performing interpositional arthroplasty in various joints of human patients.⁸,⁹ Dr Willis C. Campbell, practicing in Memphis, was also a proponent of interpositional arthroplasty and published widely on his technique, including indications, surgical technique, postoperative protocols, and outcomes (Fig. 1-2).¹⁰ Campbell reported on 12 cases and the majority of patients were purported to have reasonable results with most patients achieving near full extension and flexion from 50° to 90°.¹⁰ Duncan C. McKeever, working in Houston, Texas, experimented with cellophane interpositional arthroplasty in the 1940s, but results were unsatisfactory. John G. Kuhns and associates at the Robert B. Brigham Hospital in Boston reported on their experience with interpositional arthroplasty utilizing various synthetic materials.⁷ Their initial attempts to interpose “malleable, nonoxidized metallic sheets” proved unsuccessful, and they ultimately settled on the use of nylon interposition. The authors reported complications that included motion less than 60°, pain, re-ankylosis, infection, worsening arthritis, and others; however, they did report satisfactory results in 58 patients and unsatisfactory in 12 patients treated with their nylon interposition technique.⁷ Although interpositional arthroplasty, particularly with transfer of local autogenous tissues, provided encouraging outcomes over short-term follow-up, these
procedures were ultimately subject to degradation of the interpositional material, which later authors attributed to recurrence of pain and instability, and these procedures were unable to correct mechanical deformity.

Dissatisfied with results of soft-tissue interpositional arthroplasty and taking inspiration from Smith-Petersen’s technique of vitallium cup arthroplasty in the hip, surgeons began to design devices for hemiarthroplasty in the knee. Dr Campbell, in conjunction with his colleague Dr Harold B. Boyd, developed a vitallium prosthesis to resurface the distal femur in much the same way Campbell had resurfaced the distal femur with his soft-tissue interposition. The size of the implant required was estimated preoperatively using radiographs and was then fabricated for each patient. The prosthesis was affixed to the distal femur utilizing two flanges that hooked to the posterior condyles and then a vitallium screw inserted through the device into the anterior distal femur.¹¹ Boyd implanted the first vitallium distal femoral hemiarthroplasty in 1938 and Campbell published a preliminary report of the technique in 1940, which he had performed on two patients.¹² The results of the initial two vitallium distal femoral hemiarthroplasties were disappointing, with both patients achieving only limited motion postoperatively. The work of Boyd, Campbell, and Smith-Petersen would ultimately lead to the development of the femoral prosthesis used at the Massachusetts General Hospital, which included a stem for intramedullary fixation attached to the femoral resurfacing portion of the prosthesis that was intended to provide increased stability. The implant would become known as the MGH prosthesis (Fig. 1-3).

Charles C. Townley, who spent the majority of his career in private practice in Port Huron, Michigan, developed a tibial hemiarthroplasty in 1951.¹¹,¹³ Townley’s stainless steel tibial plate was asymmetric to match medial and lateral plateaus, necessitating a right and left prosthesis, was available in three sizes, and was fixed with screws to the anterior tibia. Among the initial 39 cases, 19 had follow-up of at least 2 years (range 2 to 9 years) and 14 (74%) were reported to have satisfactory results (“minimal or no pain and with useful motion”).¹¹

McKeever, not satisfied with the cellophane interpositional arthroplasty, developed a metallic, monoblock prosthesis to resurface either the lateral or medial tibial plateau (Fig. 1-4).¹⁴,¹⁵ The prosthesis was first implanted in April 1952.¹⁴ Several authors have reported mid- to
long-term follow-up of the McKeever unicompartmental hemiarthroplasty and results were good to excellent in approximately 70% of knees.¹⁵,¹⁶

McKeever also developed a patellar prosthesis and reported his results in 1955.¹⁷ His prosthesis was made of vitallium and came in three sizes, and fixation was achieved using two lips that captured the bony patella and augmented with a transfixion screw. McKeever reported using his prosthesis in 40 patients and reported 4 failures due to infection but no mechanical failures; however, the follow-up time was not specified.¹⁷ Harrington reported long-term results of the McKeever patellar prosthesis in a cohort of 28 patients with minimum 4-year follow-up (range 4 to 16 years) and reported excellent results in 17 of the 28 patients at 5-year follow-up.¹⁸

During the 1940s and 1950s, as surgeons in the United States were developing metallic hemiarthroplasty options, those working in Europe were reconsidering the concepts for hinged knee replacement first described by Themistocles Gluck some 70 years earlier. Judet and colleagues reported on their hinged knee replacement in 1947¹⁹ and Magnoni followed with his own report shortly after in 1949.²⁰ Throughout the 1950s the
concept of a hinged knee prosthesis were refined and further developed by Dr Börje Walldius and colleagues working at the Karolinska Institute in Stockholm, Sweden.²¹,²²,²³ Walldius developed his initial hinged knee prosthesis using an acrylic material called Bonoplex for the femoral and tibial components, which were then linked by a stainless steel axel.²¹ The articular portions of the femur and tibia required 38 mm bone resection, which was made perpendicular to the long axis of the bone. Fixation relied on a long intramedullary stem and two small pegs inserted in the cancellous medullary bone; cement was not used. Walldius reported reasonable success in his initial 32 knees (26 patients): 75% achieving pain relief and average motion of 84°, 4 required arthrodesis, and 2 required amputation.²² Walldius reported 8-year follow-up for his initial 64 knees (51 patients) and 74% were either “very good” or “good.”²³ He went on to describe modifications to his original prosthesis, which included changing the prosthesis from acrylic to vitallium due to problems with fracture, reducing the central articular height to 28 mm, and shortening the supplementary fixation pins to avoid conflict with cortical bone (Fig. 1-5).²³

The Walldius hinged prosthesis was appealing for several reasons. Its high degree of constraint enabled correction of large deformities about the knee and eliminated the need to retain soft-tissue structures about the knee, including the cruciate or collateral ligaments. This simplified the operative technique and it was relatively straightforward to achieve appropriate alignment. Significant drawbacks included its large size and high degree of constraint. The size of the endoprosthesis required significant bone resection that limited future salvage options, and infection was a prominent concern given the large amount of retained foreign material and resulting dead space between host tissues and the implant. The high degree of constraint resulted in significant force transmission to the implant-bone interface and led to bone destruction and ultimately loosening
of fixation. Shiers,²⁴ Young,²⁵ and others designed and implanted hinged knee prostheses; however, these did not achieve widespread usage.

The Guepar hinged prosthesis was developed by a group of surgeons in Paris in the late 1960s and early 1970s. The designing surgeons focused on six design principles: (1) minimal prosthetic bulk in width (avoid superficial soft tissues) and height (preserve bone), (2) avoid flexion limitation by contact between the two parts, (3) obtain “rolling” of the tibia under the femur in flexion, (4) preserve patellar motion, (5) dampen component-component contact in extension, and (6) reestablish normal knee axis by incorporating a valgus tilt to the femoral stem.²⁶ The prosthesis incorporated a posteriorized hinge toward achieving rolling of the tibia and silastic thrusters to absorb impact between components.²⁶ Clinically, the Guepar prosthesis performed quite well. Insall and others at the Hospital for Special Surgery (HSS) compared results from their experience with four different prosthetic designs, which they indicated for varying degrees of deformity and instability. The prostheses studied included a unicondylar prosthesis, two surface replacement prostheses (the Duocondylar and the Geometric), and the Guepar hinge. The Guepar was implanted in knees with the worst preoperative deformity and function and yielded the best postoperative function and had the lowest overall complication rate, but the authors noted that the Guepar prosthesis had the highest infection rate and was the most difficult prosthesis to effectively salvage.²⁷

In 1973, Larry Matthews and colleagues at the University of Michigan reported on the spherocentric knee, which they designed to address problems with previous hinged designs.²⁸ Previous hinged designs allowed motion only in the sagittal plane, with no motion in the coronal (varus/valgus) or axial plane (internal rotation/external rotation), which provided stability but placed increased force on the bone implant interface, leading to bone resorption and implant loosening. Another problem was metal-to-metal stops at terminal flexion and extension, which again provided stability but led to loosening and wear debris. Recognizing that polyethylene wear remained problematic, modularity was then incorporated into the design. The spherocentric knee included stemmed tibial and femoral components that were cemented into the bone; the tibial component included a central metal sphere in the intercondylar region and posterior to the midline of the femur and two tracks along the plateaus to accommodate polyethylene runners in which the femoral condyles would track. The femoral component incorporated a variable radius femoral condyle to provide gradual deceleration at extremes of motion, and the intercondylar notch contained a housing for the polyethylene component that captured the tibial sphere (Fig. 1-6). The authors reported on the longer-term outcome of their prosthesis and reported 5% infection rate, 11% loosening rate, 15% reoperation rate, and 32% incidence of radiolucent lines at an average 8-year follow-up.²⁹

The 1960s saw the recognition that the polymethylmethacrylate (PMMA) bone cement and high-density polyethylene utilized successfully by Sir John Charnley in hip arthroplasty might be incorporated into designs for knee arthroplasty. Dr Frank Gunston, a Canadian who had performed a fellowship with John Charnley, developed the first cemented surface knee replacement without a hinge in 1968 in Wrightington Hospital in Lancashire, England.³⁰ The polycentric knee arthroplasty consisted of two independent stainless steel runners to resurface the distal femoral condyles that articulated with high-density polyethylene tracks implanted into slots in the proximal tibia (Fig. 1-7).³⁰ All components were affixed to the bone utilizing PMMA cement. The articular surface was designed to permit both rocking and gliding movements in the sagittal plane as well as 20° of rotation in the axial plane. The cement fixation and allowance for axial rotation were both designed to disperse forces at the bone-implant interface. The initial experience with this prosthesis and technique in 22 knees (20 patients) was reportedly successful for pain relief and motion.³⁰ Gunston’s polycentric knee was implanted in larger numbers at the Mayo Clinic and there were 35 complications among the initial 450 cases, including 9 infections, 7 dislocations, 6 fractures, 6 VTEs, 4 skin necrosis events, and 4 cases of loose prosthetic components.³¹ Ten-year results revealed a 34% failure rate.³²

The 1970s saw two distinct design philosophies arise as surgeons and engineers sought to improve on designs for condylar total knee replacement. The anatomic camp generally sought to maintain soft-tissue structures about the knee and design the prosthesis around these soft tissues. The functional camp sought to design the prosthesis to substitute for the function of soft-tissue structures, which would permit a simplified surgical technique.

Professor Toshio Kodama and Sumiki Yamamoto, working at Okayama University Medical School, designed a prosthesis they called the Mark I, which was first implanted in 1970.³³ The femoral component was a modified Sbarbaro femoral condyle mold that incorporated an anterior flange for patellar tracking, and was made of COP alloy (Cr, Ni, Co, Mo, P).³³ The tibial component was minimally constrained with a cutout to preserve both cruciate ligaments.³³ The Mark I utilized no cement.³³ A subsequent iteration of the Mark I increased constraint between the femur and tibia, which prevented posterior sliding of the tibia and restricted flexion if the PCL was retained.³³ The Mark I was implanted in 43 knees. Yamamoto visited Freeman in 1974, and design alterations following this visit led to the Mark II prosthesis, which remained noncemented but reverted to less constraint between the femur and tibia to allow some rotatory motion in flexion but to restrict it in extension.³³

Theodore Waugh, MD, PhD, working with colleagues at the University of California at Irvine (UCI), developed the UCI total knee replacement and reported early results in 1973.³⁴ The design incorporated a single cobalt-chrome femoral component consisting of two “J” shaped runners, designed to mimic the instant center of rotation of the human knee, connected by a thin piece of metal anteriorly. The tibial component was horseshoe shaped to facilitate preservation of the cruciate ligaments and was made from ultra-high-density polyethylene. Intermediate results for 103 knees indicated most patients (78%) improved but mechanical complications occurred in 17.4%.³⁵

Dr Townley, not completely satisfied with results from his tibial hemiarthroplasty technique and recognizing failures with hinged total knee replacements, developed the Townley anatomic total knee and reported its use in
1974 (Fig. 1-8).³⁶ The femoral component for the first time incorporated an anterior flange designed not only to connect the condyles but to provide an articular surface for the patella. The all-polyethylene tibial component was again cut out to preserve both cruciates and had a dished articular surface conforming to the femoral condyles, and the undersurface was corrugated to facilitate PMMA interdigitation.³⁶

Seedhom and colleagues developed the Leeds knee following an anatomic approach.³⁷ The femoral prosthesis, attempting to replicate native anatomy, was asymmetric, with the medial condyle employing a J curve and the lateral condyle designed with a relatively uniform radius.³⁷ It was manufactured in three sizes. The polyethylene tibial component consisted of two disks connected anteriorly; the disks incorporated a dished design to replicate the menisci and the degree of dishing was modified with subsequent designs to facilitate increased motion. Results of the Leeds knee were never published by the surgeons.³⁸

The Duocondylar prosthesis was developed at HSS in New York City with major contributions from surgeon Chitranjan Ranawat and engineer Peter Walker.³⁹ It was first implanted in December 1971. The cobalt-chrome femoral component consisted of condylar surfaces closely replicating the normal knee connected with an anterior bar, each condyle had a peg affixed to the back for augmented fixation with PMMA.³⁹ The femoral component came in two sizes. The tibia was resurfaced with two independent polyethylene implants, which were dovetailed on the undersurface to facilitate cementation into the tibial bone.³⁹ The tibial articular geometry was designed to permit sagittal plane motion without limit but had an upward curvature toward the intercondylar notch to provide some stability in the coronal and axial planes.³⁹ The tibial components were affixed to a jig to maintain alignment with the femoral component during implantation. Ranawat described contraindications for the duocondylar prosthesis, which included sagittal plane instability, fixed coronal plane subluxation, hyperextension greater than 10°, and flexion contracture greater than 25°. Initial results were acceptable, but failure modes included instability and tibial loosening and led to a 5.5% revision rate at 3-year follow-up.⁴⁰ The Duocondylar prosthesis was subsequently modified to create the Duopatellar prosthesis, which included an anterior flange on the femoral component, a patellar resurfacing, and a single-piece tibial polyethylene component with increased conformity and a large fixation peg.⁴¹,⁴² Although initially satisfied with the Duopatellar prosthesis, Ranawat ultimately favored the Total Condylar Prosthesis, which was developed at the HSS contemporaneous with the Duopatellar. In 1974 Ranawat and Walker traveled to Boston to present the Duopatellar and Total Condylar Prosthesis to a group of surgeons at MGH in Boston who had experience with the Duocondylar prosthesis but were dissatisfied with nearly 20% of patients with residual patellofemoral symptoms; the MGH surgeons included Bill Harris, Bill Jones, Clement Sledge, Richard Scott, and others.⁴³ The Boston contingent, given their considerable experience and overall good results with the McKeever hemiarthroplasty, preferred retention of the PCL and favored the Duopatellar prosthesis and it was first implanted in Boston in 1974.⁴²,⁴³ The initial experience over short-term follow-up with the Duopatellar prosthesis in Boston was good, with a 2.8% revision rate among which half were for patella problems (pain, tracking, fracture) and another third were for aseptic tibial loosening.⁴² The Duopatellar prosthesis was subsequently modified in Boston over the next 7 years and would come to be called the Robert Brent Brigham Hospital (RBBH) knee; modifications included a deepened and valgus aligned trochlea, extension of the posterior condyles, incorporation of a one-piece tibial component with central stem to improve fixation (they would later incorporate a metal-backed monoblock tibial component), and flattening of the tibial polyethylene in the coronal plane to allow femoral rollback in flexion.⁴²

Clement Sledge and Frederick Ewald, among other surgeons working at the Robert Brent Brigham Hospital in Boston, worked with Peter Walker, who had left HSS to work as an engineer for Howmedica, to develop the Kinematic knee system, which was a modification of the RBBH knee and was first implanted by Ewald in 1978.⁴²,⁴⁴ Results at 10-year follow-up were generally good with 96% survivorship, but loosening of the patellar component was problematic.⁴⁵ The Kinematic knee would evolve into the Kinemax knee, which modified the articular geometry to increase contact area.⁴⁶ The Kinemax knee demonstrated acceptable midterm results with survivorship of 99% at 5 years and 97% after 9 years.⁴⁶ Other surgeons working in Boston, including Richard D. Scott and
Thomas S. Thornhill, who had previously worked on the Kinematic team, would go on to develop the cruciate-sparing Press Fit Condylar (PFC) knee system in conjunction with Johnson & Johnson.⁴⁷ Most current cruciate-sparing total knee prostheses have origins in these systems designed in Boston.