Journal of Orthopaedic Research 86471 Raven Press, Ltd., New York 0 1990 Orthopaedic Research Society

Stability of Initial Fixation of the Tibia1 Component in Cementless Total Knee Arthroplasty Hitoshi Shimagaki, Joan E. Bechtold, Robert E. Sherman, and Ramon B. Gustilo Orthopedic Biomechanics Laboratory, Hennepin County and Metropolitan-Mount Sinai Medical Centers, and Department of Orthopaedic Surgery, University of Minnesota, Minneapolis, Minnesota, U.S.A.

Summary: This study measured the vertical displacement of three kinds of cementless tibial components [Porous Coated Anatomical (PCA), Tricon, and Whiteside], under eccentric loading up to 2,225 N. Displacement between the tibial tray and the proximal tibia was measured with linear variable differential transformers at the anterior and posterior side when anteriorly or posteriorly loaded, and at the medial and lateral side when medially or laterally loaded. The general pattern of motion was sinking at the loaded side and lift-off at the opposite side. Lift-off opposite the loaded side was fairly small for all components at all measurement sites. Among the three components, the Whiteside showed the smallest displacements. The Tricon (when anteriorly or posteriorly loaded), and the PCA (when medially or laterally loaded) showed sinking at the loaded side. Anterior screw fixation of the PCA was not effective in preventing anterior lift-off. The tilting motion of the tibial components observed in this study implies instability of the initial fixation, which could possibly compromise bony ingrowth. Furthermore, this tilting could cause uneven distribution of load, and potentially result in fracture of the underlying bone. Key Words: Cementless tibial component-Initial stability-Micromovement.

Total knee arthroplasty has been a reliable treatment for severe arthritis of the knee joint. Recently cementless fixation has been developed for prosthetic replacement. Biological fixation with bony ingrowth was introduced with the aim of increasing the functional longevity of implants, and has been in clinical use in total knee arthroplasty . Encouraging results in cementless total knee arthroplasty have been reported (7,9,12,15,17). Hungerford et al. (9) reported satis-

factory results in 94.5% of 93 knees treated with cementless porous coated anatomical (PCA) prosthesis with 2-5 years’ follow-up. No tibial component loosening was found in his series. Rosenqvist et al. (18) found late loosening of beads in eight of nine cementless PCA tibial components, suggesting that there was relative motion of the implant and that bone ingrowth had not occurred. Ryd (19) demonstrated inducible displacement of the PCA tibial component at 2 years’ follow-up, implying failure to achieve sufficient bony ingrowth. Histologic analyses of retrieved implants have found little bony ingrowth occurring in tibial components (14,22). Thomas et al. (22), in an analysis of retrieved components from humans, reported that in no case was more than 10% of the total surface area of the tibial component ingrown with bone. Thus, a reliable method of achieving

Address correspondence and reprint requests to Dr. J. E. Bechtold at Orthopedic Biomechanics Laboratory, D 651, Metropolitan-Mount Sinai/Hennepin County Medical Center, 900 South Eighth Street, Minneapolis, MN 55404, U.S.A. Dr. Shimagaki’spresent address is Department of Orthopaedic Surgery, Niigata University, School of Medicine, Niigata, Japan. This study was presented in part at the 34th annual meeting of the Orthopedic Research Society, Atlanta, Georgia 1988.

64

H . SHIMAGAKI ET AL.

rigid fixation of a cementless tibial component to enhance bony ingrowth is still unresolved. Prerequisites for induction of bony ingrowth into a porous surface include an initial stable fixation of a component to the bone (3,4,21). If rigid initial fixation cannot be achieved, fibrous tissue will grow into the porous surface rather than bone. Initial fixation of a cementless tibial component depends largely on the design of the component, and especially its method of fixation to the cancellous bone of the proximal tibia. Although several kinds of posts or pegs have been developed for cementless fixation (1,68,12,16,20,26,27), few studies have reported the effect of pegs and central posts on the initial stability in cementless knee-joint replacement (2,5,23,24,28). The objective of this project is to study the stability of three kinds of cementless tibial components (PCA, Tricon, and Whiteside) in vitro. In addition, the contribution of screw fixation to stability of the PCA component is studied. MATERIALS AND METHODS Cadaver tibiae without evidence of musculoskeletal disease were kept frozen at - 20°C except when being tested. The distal shaft was potted in dental Diestone (Columbus Dental, St. Louis, MO 63188) for fixation to the testing machine. The proximal tibia was resected 5 mm below the lowest articular surface with the osteotomy line perpendicular to the long axis of the tibia, using an electrical band saw. An effort was made to cut the same amount of bone from right and left tibiae of a pair. The cut surface of the tibiae appeared to have no gross irregularities. Indentation tests were carried out to compare the strength of cancellous bone in the proximal tibia among the specimens. Three points each in medial, lateral, and central parts of the resected surface were loaded with a 2-mm-diameter indentor. Relative strength of cancellous bone was expressed as the magnitude of the compressive load required to indent to a depth of 0.5 mm. Then the tibial components were implanted according to standard instrumentation and technique for each device. Tibia1 components studied were: PCA (Howmedica, Rutherford, NJ 07070), Tricon (Richards Medical, Memphis, TN 38116), and Whiteside (Dow Corning-Wright, Arlington, TN 38002). The PCA, Tricon, and Whiteside are metal-backed prostheses with porous coating, nonconstrained articular surface, and with posterior cruciate ligament retention.

65

The PCA component has two small porous coated pegs on each of the medial and lateral sides, which are sloped posteriorly by 30". The pegs are 15 mm in length and 9 mm in diameter. The PCA has a central hole anteriorly for supplemental screw fixation. The Tricon component has two straight 20mm-long, 9-mm-diameter posts, each with eight polyethylene flanges (Freeman peg). The Whiteside component has a smooth central stem 48 mm in length with a polyethylene sleeve 44 mm in length and 18 mm in diameter. This component has six small peripheral pegs. Three pairs of fresh cadaver tibiae were used to compare the PCA and the Tricon. The PCA was implanted on one side of a pair; the Tricon was implanted on the contralateral side. The PCA was tested both with and without 50 mm A-0 cancellous screw fixation. After the tests with the PCA component, the Whiteside tibial component was implanted into the same tibia. Because the stem of the Whiteside component is central, the medial and lateral holes left in the cancellous bone by the pegs of the PCA did not interfere with its fixation. (It is conceivable, however, that the support of the tray and its stability may be influenced by these holes.) The stability of a cementless tibial component was assessed as the displacement between the tray of the tibial component and the proximal tibia, at the level of the cut (25). Displacement was measured with linear variable differential transformers (LVDTs) (Model MHR 250, Schaevitz, Pennsauken, NJ 08110; linearity 0.12-0.25%; full scale = 0.625 mm). The LVDTs were mounted to a metal ring attached with screws immediately distal to the resected surface. The LVDT cores were attached to the metal tray at the anterior, posterior, medial, and lateral sides (Fig. 1). The body and the core of the LVDT were positioned perpendicular to the horizontal plane of the cut surface of the tibia. Four different loading configurations were applied using an electrohydraulic material test system (Model 810, MTS Systems Corporation, Minneapolis, MN 55424). Eccentric anterior, posterior, medial, and lateral loads were applied to the upper surface of the tibial component. Anterior load was applied along the midline at the anterior one-third of the tibial plateau, and posterior load was applied along the midline at the posterior one-third. In the PCA and Tricon study, medial load was applied along the anterior-posterior midline at the medial

J Orthop Res, Vol. 8, N o . 1, 1990

CEMENTLESS TIBIAL COMPONENT INITIAL STABILITY

66

POLYETHYLEE TBIAL COMPONENT,

LVDT

FIG. 1. Schematic diagram of the linear variable differential transformer (LVDT) attachment to the prosthesis and to the tibia.

one-fourth, and lateral load at the lateral onefourth. In the Whiteside study, the deepest point along the midline in the anterior-posterior direction was chosen. Maximum load of 2,225 N (500 lb, or approximately three times body weight) was applied in increments of 445 N. RESULTS Indentation tests showed that the strength of the underlying bone was comparable within a pair, but varied among the pairs. This variation was particularly evident at the medial aspect (Table 1). All components tested appeared to be snugly seated without visible space between the tray and the underlying bone. All components were correctly sized, with no overhang of the cortical rim of the tibia. The outline of both the resected surface of the tibia and the metal tray was traced for all tibiae and each type of the component. Percent coverage of the resected surface was similar for each type of component: 78.4 1.6% for the PCA; 77.9 6.1% for the Tricon; and 71.7 k 6.0% for the Whiteside. Each loading configuration was repeated three times and the displacement readings were averaged. All tibiae tolerated eccentric load without visible fracture or subsidence in either the cortical or can-

*

J Orthop Res, Vol. 8, No. 1, 1990

*

cellous bone of the proximal tibia, with the exception of the last sequence of one test in which the PCA loaded laterally subsided slightly into the cancellous bone. This run was excluded from the statistical analysis. Displacement readings of LVDTs were repeatable for all components and for all loading configurations. Because the secondary horizontal displacements were not measured, displacements represent only the primary vertical displacement. Vertical displacements for all loadings were measured at an offset from the periphery of the tibial components. The offset was dictated by proximal tibial anatomy and was 12 mm for the anterior and posterior LVDTs and 5 mm for the medial and lateral LVDTs . Displacement of the tray with respect to the tibia at the anterior and posterior sides was analyzed when anteriorly or posteriorly loaded, and displacement at the medial and lateral sides was analyzed when medially or laterally loaded. Because offset was the same at each side for the three tibial components, displacement with the offset under maximum load of 2,225 N was compared among the PCA with screw fixation, the Tricon, and the Whiteside components. Displacement was compared using analysis of variance, assuming mechanical strength to be comparable within a pair of bones. Mean displacement at the loaded side and opposite the loaded side were graphed for the PCA, the Tricon, and the Whiteside components (Fig. 2). To facilitate direct comparison, displacements were adjusted for the offset (Table 2). Comparison of Displacement Among PCA, Tricon, and Whiteside Anterior Load (2,225 N )

When anteriorly loaded, posterior displacements were quite small for all components. Anterior sinkTABLE 1. Maximum load ( N ) to compress to a depth of 0.5 rnm (indentor-2mm diameter) Location of indentation Pair

Medial

Lateral

Central

1

104.1 81.0 40.9 31.2 49.4 39.2

43.6 46.7 30.7 24.9 46.7 25.4

16.0 12.0 8.9 8.9 8.5 14.7

2 3

H . SHIMAGAKI ET AL.

M M A L LOAD

LATERAL LQlrD

IF

mm

5-

0.1

5-

67

c

0ff.01

LATERAL SINKING

LATERAL LIFT-OFF

mm

MEDIAL LIFT-OFF

0.1

0ff.01

0.2

0.2

0.3

0.3

0.4

1

1

+”

A N T E R M LOAD

I

ANTERIOR S l Y l

0.4

POSTERIOR LOAD

o-2

mm

mm 0.1

COlTERIOR LIFT-OFF

0.2 0.3

AYTERfOR LIFT-OFF

FIG. 2. Mean displacement of porous coated anatomical prosthesis (PCA) with screw fixation, Tricon, and Whiteside tibial component index eccentric loading (2,225 N).

for both sinking and lift-off in these sets of bones. It should be noted that the analysis of variance only shows the difference among the three components, not specific differences between two components.

ing was greatest for the Tricon, and smallest for the PCA. Statistical analysis using analysis of variance showed a significant difference among components

TABLE 2. Mean displacement of tibial trays adjusted f o r offset (2,225 N load) Posterior load

Anterior load

PCA wlscrew PCA wlo screw

Anterior displ.

Posterior displ.

Anterior displ.

Posterior displ.

J 0.029 (10.0 15) J0.074

J 0.028 (20.073)

J0.024 (10.043) t 0.066

(50.055)

(20.035) J0.086 (10.058) j’ 0.006 (20.020)

(1_0.010)

J 0.053 (20.054) T 0.001

J 0.187 (10.074) J 0.176 (-t0.071) J0.264 (20.017) J 0.044

(20.004)

(k0.020)

Medial displ.

Lateral displ.

Medial displ.

J0.278

f 0.142 (20.016) j’0.096 ( 20.015) J 0.007 (10.005)

J. 0.299

Tricon

(20.050) T 0.048 (LO.140)

Whiteside ~~~

t 0.052

~

Lateral load

Medial load

PCA wlscrew

( 20.088)

Tricon Whiteside

LO. 100 (t0.060) L O . 157 (20.064)

t 0.140 t 0.091 (20.050) t 0.009

Lateral displ.

J 0.264

(20.029)

(20.065)

(20.009)

(10.016) J 0.167 (20.036)

10.052

t : Lift off, J : sinking, (mm).

J Orthop Res, Vol. 8, No. 1 , 1990

68

CEMENTLESS TIBIAL COMPONENT INITIAL STABILITY Posterior Load (2,225 N )

When posteriorly loaded, anterior lift-off was greatest for the PCA with screw fixation. The Whiteside showed a minimal anterior displacement. Posterior sinking was greatest for the Tricon and smallest for the Whiteside. Statistical analysis using analysis of variance showed a significant difference among these three tibial components for both anterior lift-off and posterior sinking. Medial Load (2,225 N ) When medially loaded, lateral lift-off was greatest for the PCA and smallest for the Whiteside. Medial sinking was greatest for the PCA. Medial sinking with medial load did not show a statistically significant difference among the PCA, the Tricon, and the Whiteside, whereas lateral liftoff did. Lateral Load (2,225 N ) When laterally loaded, medial lift-off was greatest for the PCA and smallest for the Whiteside. Lateral sinking was greatest for the PCA and smallest for the Tricon. Statistical analysis showed that both medial liftoff and lateral sinking were significantly different among the three components for the subject bones tested.

Tricon When anteriorly or posteriorly loaded, sinking at the loaded side was >0.2 mm (up to 0.36 mm), whereas displacement opposite the loaded side showed either lift-off