AFTER TOTAL KNEE REPLACEMENT

Transcription

1 CORONAL ALIGNMENT AFTER TOTAL KNEE REPLACEMENT ROBERT S. JEFFERY, RICHARD W. MORRIS, ROBIN A. DENHAM From Queen Alexandra Hospital, Portsmouth Maquet s line passes from the centre of the femoral head to the cenfre of the body of the talus. The distance of this line from the centre of the knee on a long-leg radiograph provides the most accurate measure of coronal alignment. Malalignment causes abnormal forces which may lead to loosening after knee replacement. We report a series of 115 Denham knee replacements performed between 1976 and 1981 using the earliest design of components, inserted with intramedullary guide rods. Patients were assessed clinically and long-leg standing radiographs were taken before operation, soon after surgery and up to 12 years later. In two-thirds of the knees (68%) Maquet s line passed through the middle third of the prosthesis on postoperative films and the incidence of subsequent loosening was 3%. When Maquet s line was medial or lateral to this, an error of approximately ± 3#{176}, the incidence of loosening at a median period of eight years was 24%. This difference is highly significant (j =.1). Accurate coronal alignment appears to be an important factor in prevention of loosening. Means of improving the accuracy of alignment and of measuring it on long-leg radiographs are discussed. On an anteroposterior long-leg radiograph, a line from the centre of the femoral head to the centre of the body of the talus normally passes through the middle third of the knee (Maquet 1972). If this is not the case after total knee replacement, compression forces on the concave side and tensile forces on the convex side of the joint may lead to loosening (Denham and Bishop 1978). Bargren, Blaha and Freeman (1983) have demonstrated in cadaver studies that eccentric loading of the tibial component caused failure at between a third and a half of the force required to produce failure by central loading, depending on the type of prosthesis and the position of the load. In the clinical situation, malaligned prostheses are subjected to cyclical eccentric loading. At total knee replacement, it is difficult to obtain consistently acceptable tibiofemoral alignment by eye because of the tourniquet, drapes and subcutaneous fat. Externaijigs or intramedullary guide rods are required. R. S. Jeffery, FRCS, Orthopaedic Registrar R. A. Denham, FRCS, Emeritus Consultant Department of Orthopaedics, Queen Alexandra Hospital, Cosham, Portsmouth, Hampshire P6 3LY, England. R. W. Morris, MSc, Lecturer in Medical Statistics Department of Public Health Medicine, United Medical and Dental Schools of Guy s and St Thomas Hospitals, Guy s Hospital, London Bridge, London SE1 9RT, England. Correspondence should be sent to Mr R. Jeffery British Editorial Society ofbone and Joint Surgery 31-62X/91/5l85 $2. JBoneJoint Surg[Br] 1991 ; 73-8: We have studied 1 15 Denham knee replacements aligned with intramedullary guide rods, performed between 1976 and Full-length standing radiographs were taken before operation, postoperatively and on alternate years thereafter. Maquet s line was used to assess coronal alignment on all these films and the results have been related to the incidence of loosening after a median follow-up period of eight years. PATIENTS From 1976 to April 1981, a total of 139 Denham total knee replacements were performed at Queen Alexandra Hospital, Portsmouth under the care of the senior author (RAD) using the original pattern of components and a standard technique. Of these, one patient with bilateral prostheses emigrated and the radiographs of one patient (2 knees) have been lost. In 2 knees long-leg alignment radiographs were not taken before or after operation. For 13 knees pre-operative radiographs were not taken because the patients were unable to stand or had fixed knee flexion of more than 2#{176}. No postoperative standing radiographs were made of eight knees because of infection or poor general condition. There was one postoperative death due to pulmonary embolus. Excluding these 24 knees, the study included full prospective records for 12 patients with 115 prostheses. Details are given in Table I. All patients were followed VOL. 73-B, No. 5, SEPTEMBER

2 71 R. S. JEFFERY, R. W. MORRIS, R. A. DENHAM up until revision, death or a minimum of eight years postoperatively. Forty-four patients died from causes unrelated to their knee replacement, the surviving 58 patients were contacted (68 knees). METHODS Operative technique. The Denham knee replacement uses intramedullary guide rods to achieve coronal alignment (Thacker and Fulford 1986). The femoral template has an 18 cm intramedullary stem and the femoral component has a similar intramedullary stem, tapering from a 3 mm x 1 mm oval to a 5 mm x 5 mm square section. The instruments used for preparation of the tibial plateau include an intramedullary guide rod : the tibial component has a central hole which admits a full-length 6 mm diameter intramedullary guide rod, which is removed when the cement has set (Fig. 1). Table I. Details of the series of 12 patients with 115 knee replacements Female : male 85:17 OA:RA 5:52 Age at operation (years) Median (range) 66 (21 to 85) Interquartile range 6 to 71 Previous reconstructive surgery Knee replacement 5 Tibial osteotomy 3 Benjamin osteotomy 4 Hip replacement 3 Hiparthrodesis I Fig. I The tibial component with a removable intramedullary guide rod, and two views ofthe femoral component. Abnormal femoral or tibial diaphyses Abnormal bowing 2 Previous fracture 2 Length of follow-up (years) Median (range) 8 ( to 12) Interquartile range 4 to 1 one patient also had a femoral osteotomy Follow-up. The patients had been assessed clinically and radiographically pre-operatively, postoperatively and on alternate years until they died or until the prosthesis was removed. All patients who were too old or frail to attend were contacted at the time of the review to establish whether the knee was pain free, and whether function and movement had remained the same or had deteriorated. ning of the tibial component was defined as occurring when there was pain associated with a fracture in the cement or an increasing radiolucent line. Alignment radiographs. The patient stands, trying to take weight equally on both feet, with the knees in maximal extension, and positioned to face the X-ray tube at a distance of 18 cm. A 1 ma,.5 second exposure is used at 1 to 115 kv, depending on leg thickness. The 15 cm X 35.5 cm Cronex film (DuPont) varies from Fig. 2 Measurements on an alignment radiograph. THE JOURNAL OF BONE AND JOINT SURGERY

3 CORONAL ALIGNMENT AFTER TOTAL KNEE REPLACEMENT 711 High Plus, through Standard, to High Definition so that a single exposure will produce optimal images of the hip, knee and ankle. Each long film is displayed on a large viewing screen Table II. Alignment data Variation of tibiofemoral angle from 7#{176} valgus (degrees) U Maquet s line passes within the middle third of the knee. Maquet s line passes through the knee. E Maquet s line passes outside the knee. Distance of Maquet s line from the centre of the knee (mm) Mean (SD) Mean (SD) Pre-operative (n = 1 1 5) 12.4 (8.9) 42 (32) Postoperative (n = 1 1 5) 2.6 (2. 1) 9 (9) 8-to 12yearfollowup(n52) 3.3(3.4) 11 (12) and the mid-medullary lines of the femur and of the tibia are drawn, passing through points which bisect the diameter of the bone at a quarter of its length from each end. Maquet s line is drawn from the centre of the femoral head to the centre of the body of the talus. Measurements are then made of the tibiofemoral angle and of the distance (D) from the centre of the tibia! plateau to Maquet s line (Fig. 2). The radiographic image of the tibia! prosthesis is 7 mm wide, so Maquet s line passes through the middle third of the knee when the distance D is 1 1 mm or less ; D is recorded as negative for varus angulation and positive for va!gus. An error of rotation of the limb of up to 2#{176} during radiography introduces a projectiona! error of less than 2#{176} in the tibiofemoral angle, provided that the knee is fully extended (RWM). The combination ofrotation and knee flexion may introduce much larger errors. All leg alignment films should therefore be taken in maxima! extension ; a lateral view was taken in this position in each case, to confirm that the alignment film was acceptable PREOPERATIVE (n115) of knees Table III. Pre-operative deformity related to postoperative angulation Postoperative tiblofemoral e (degrees) Pre-operative tibiofemoral angle(degrees) Mean (SD) > 3 varus valgus (3.7) 3 valgus to 3 varus valgus (3.2) 4to lvalgus(normal) valgus (3.1) 1 1 to 17 valgus valgus (3.4) > 17 valgus valgus (2.9) one way analysis of variance p = YEARS ( n 52) knees to to to to to to to to to to to to to to to to Varus TIBIAL COMPONENT Valgus h_ Fig. 3 D(mm) Alignment before and after operation and at a minimum of eight years postoperatively, related to the width of the tibial component. In order to calculate the error due to measurement of the same angle on different films, we recorded the angle between the femoral mid-medullary line and a line from the centre of the femoral head to the centre of the knee. This should remain constant if the shape and rotation of the femur is unchanged. RESULTS Alignment. Before operation, Maquet s line passed through the middle third of the knee in only 13% of the 1 15 radiographs. After operation, this was true of 68% of the early radiographs and 65% of those taken at more than eight years after knee replacement (Table II and Figure 3). In these cases the corresponding tibiofemoral angle was from about 4#{176} to about 1#{176} valgus. The standard deviation for the measurement of the same angle on different films ofthe same patient was.8#{176}. Table III shows that failure to achieve accurate VOL. 73-B, No. 5, SEPTEMBER 1991

4 712 R. S. JEFFERY, R. W. MORRIS, R. A. DENHAM alignment in this early series occurred mostly in patients with pre-operative varus deformity. Twenty patients had bowing or malunited fractures of the tibia! or femoral shaft or had had previous reconstructive operations (Table I). In ten of these (5%), Maquet s line passed outside the middle third of the knee postoperatively and in four it was over 23 mm from the centre of the tibia! plateau. Subsequent loosening. Eleven of the 1 15 knees (1%) had either been revised for loosening or showed definite clinical and radiological signs of loosening at follow-up. Only two of the 1 1 knees which loosened had been correctly aligned postoperatively (Fig. 4). We found no overallassociation between subsequent loosening and pre-operative malalignment, but there was a strong association with poor postoperative alignment (p =.1, Tables IV and V). When divided according to diagnosis and varus or valgus malalignment (Tables VI and VII), there was a difference between osteoarthritis and rheumatoid arthritis but this was not significant (p =.9). Knees with varus deformity pre-operatively were more likely to loosen than those with valgus Table IV. Pre-operative alignment related to subsequent loosening Maquet s line ning the middle third 15 1 (7%) Valgus or varus 1 1(1%) Total p = 1. (Fisher s exact test) Table V. Postoperative alignment related to subsequent loosening Maquet s line ning middle third 78 2(3%) Valgusorvarus 37 9(24%) Total p =.1 (Fisher s exact test) of knees U Prostheses which loosened Prostheses which did not loosen Table VI. Pre-operative alignment related to diagnosis and loosening Maquet s line Osteoarthritls Rheumatoid Nwnber Varus middlethird Valgus arthritis ning : varus versus valgus, p =.5 ; osteoarthritis versus rheumatoid arthritis, p =.9 (Fisher s exact test). < >35 to to to to to to Varus Valgus I TIBIAL Postoperative D (mm) Fig. 4 _ COMPONENT Postoperative alignment related to subsequent loosening. (p =.5), but postoperatively the difference was not significant. The association of loosening with postoperative malalignment was independent of pre-operative deformity (stratified chi-squared = 9.43, 2 df; p <.1, Table VIII). Conversely, when corrected for postoperative alignment, the effect of pre-operative deformity was not significant (stratified chi-squared = , 2 df; p >.5). Clinical review. Forty-one patients with 49 prostheses were reviewed clinically and radiographically eight to 12 years postoperatively (RSJ). Scores for pain, function and the findings on examination were added to give the BASK score (Aichroth et a! 1978). We found no independent relationship between either the early postoperative or the present alignment and the BASK score, although the two loose, unrevised prostheses were painful and had the lowest marks (Fig. 5). A low score for function often reflected general disability rather than an unsuccessful operation. Complications. The only major complications other than THE JOURNAL OF BONE AND JOINT SURGERY

5 CORONAL ALIGNMENT AFTER TOTAL KNEE REPLACEMENT 713 Table VII. Postoperative alignment related to diagnosis and loosening Osteoarthrith Rheumatoid arthritis (3) and poor postoperative alignment : Maquet s line was outside the middle third of the prosthesis in five of the 11 available postoperative ifims. Maquet s line Nwnber Varus middle third Valgus ning : varus versus valgus, p =.28 ; varus versus middle third, p=o.4; valgusversusmiddlethird, p =.26 (Fisher s exact test) Table VIII. Pre-operative and postoperative deformity related to loosening Postoperative Pre-operative Vanis middle third Nwnber Varus middlethird Valgus in >- Cl) Cl) #{149}#{149} #{149}:#{149},s Postoperative : Varus Fig. 5 D (mm) BASK knee score at eight- to 12-year follow-up related to postoperative alignment. loosening were one case of infection and one of severe translocation. One knee showed loosening ofthe femoral as well as the tibial component. In the 24 knee replacements performed during the same period and excluded from the main study for the reasons given previously, there were three cases of loosening, three of infection and one of translocation. Since 13 of these cases were excluded because of severe pre-operative deformities, it is of interest to note that they had a disproportionately high rate of both loosening Valgus DISCUSSION Measurement of alignment. Short-arm goniometers (Tew and Waugh 1985) and short radiographs (Bonnici and Allen 1991) are accurate only to 5#{176}. When the patient is carefully positioned and the knee is in full extension, fulllength standing radiographs can be used to measure tibiofemoral angles to within 2#{176}. Measuring this angle to 5#{176} accuracy would not appear to be sufficiently precise to detect the moderate degree of malalignment which can affect the result. ning and postoperative alignment. Some previous studies have included moderately malaligned knees within the range regarded as acceptable. This obscures the effect of minor degrees of malalignment, though clearly showing major errors. Bargren et a! (1983) found an 1 1% failure rate in 18 knees between 1#{176} and 15#{176} valgus and an 86% rate in 14 knees outside this range. Our study has shown a clear association between minor degrees of postoperative malalignment and subsequent loosening. Intramedullary guide rods. Our study has shown that intramedullary guide rods can restore the tibiofemoral angle with reasonable accuracy in otherwise normal limbs. Whiteside (1989) reported achieving a tibiofemoral angle of between 4#{176} and 1#{176} valgus in 87% of 226 prostheses using an intramedullary system. Manning,.,. Elloy and Johnson (1988) reported 73% within 3#{176} of the mechanical axis, also using intramedullary guide rods... In our series, before the modified tibia! components were introduced, Maquet s line was within the middle third of the prosthesis in 68% of postoperative films. For patients who have had previous surgery or have bony deformity, special precautions are needed, including careful study oflong-leg radiographs. Improvementsto the prosthesis and technique. At operation it is important to locate accurately the surface marking Valgus of the mid-medullary canal on the articular surface. Better alignment would be possible with a longer intramedullary guide rod and a wider tip for the femoral stem ; but greater length might compromise a future hip replacement. The post-1981 design of the femoral stem has a wider tip and might be further improved by a centring device. The largest size of prosthesis which can be accommodated should be used ; this increases the width ofthe middle third. We feel that the added complexity of using femoral stems with varied angles is not justified ; the angle between the line ofthe femoral shaft and the line between the centres of the hip and knee was between 6#{176} and 8#{176} valgus in 87% of our pre-operative radiographs. Because of the poor correction of severe preoperative deformity in some early cases, modified tibia! components were introduced. These had a 1 5#{176} or 3#{176} VOL. 73-B, No. 5, SEPTEMBER 1991

6 714 R. S. JEFFERY, R. W. MORRIS, R. A. DENHAM polyethylene wedge on the undersurface to reduce the need for a wedge of cement on an eroded tibia! condyle. This wedge ofcement may fracture and lead to loosening. A 2#{176} or 4#{176} inclination of the surface of the plateau was later added to these special prostheses to overcompensate for severe pre-operative deformity. A permanent tibia! intramedullary stem is now used for cases with major instability. Who needs alignment radiographs? Alignment radiographs are expensive, bulky to store and time-consuming to measure, but they are essential in the development and assessment of prosthetic design. In clinical practice, these radiographs are necessary for patients with bony deformity or a previous operation. In all cases, a single postoperative alignment radiograph is valuable for audit and prognosis. CONCLUSIONS The theoretical need for Maquet s line to pass through the middle third of the knee is recognised. This corresponds to an error of up to 3#{176} from the normal tibiofemoral angle of 7#{176} valgus. In practice, a much wider range of angles has been regarded as acceptable. Our results for the Denham prosthesis confirm that there is a highly significant increase in loosening of the tibia! component when Maquet s line does not pass through the middle third of the knee. Intramedullary guide rods are one of the best available methods of alignment, and enabled us to achieve this higher standard in over two-thirds of postoperative radiographs. Since the reported series, modifications to the prosthesis and technique have improved accuracy in difficult cases. Careful study of long-leg radiographs and the selective use of special components are necessary for patients with previous injuries, operations or deformities of the lower limb. We thank Mr Wai Ng for his assistance with statistical presentation, Mrs L Blackwell and Mrs C McBride for clerical help, the photography and radiology departments at Queen Alexandra Hospital, Portsmouth and the medical illustration department at Southampton General Hospital. One or more of the authors have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. In addition benefits have also been or will be directed to a research fund, foundation, educational institution, or other non-profit institution with which one or more of the authors is associated. REFERENCES Aichroth P, Freeman MAR, Smillie IS, Souter WA. A knee function assessment chart. J Bone Joint Surg [Br] 1978 ; 6-B :38-9. Bargren JH, Blaha JD, Freeman MAR. Alignment in total knee arthroplasty : correlated biomechanical and clinical observations. ClinOrthop 1983; 173: Bonnici AV, Allen PR. Comparison of long leg and simple knee radiographs in assessment of knees prior to surgery. J Bone Joint Surg[Br] 1991 ; 73-B Supp 1:65. Denham RA, Bishop RED. Mechanics of the knee and problems in reconstructive surgery. J Bone Joint Surg [Br] 1978 ; 6-B : Manning M, Elloy M, Johnson R. The accuracy of intramedullary alignment in total knee replacement. J Bone Joint Surg [Br] 1988; 7-B :852. Maquet P. Biom#{233}canique de la gonarthrose. Ada Orthop Beig 1972; 38 :Suppl.I :S33-S54. Tew M, Waugh W. Tibiofemoral alignment and results of knee replacement. J Bone Joint Surg [Br] 1985 ; 67-B : Thacker C, Fulford P. Assessment of the Denham knee replacement. J Bone Joint Surg [Br] 1986; 68-B :6-7. Whiteside LA. Intramedullary alignment of total knee replacement : a clinical and laboratory study. Orthop Rev Suppi 1989; 18:9-12. THE JOURNAL OF BONE AND JOINT SURGERY