Knee Manipulation Following Total Knee Arthroplasty Analysis of Prognostic Variables
Daniel Daluga, MD,* Adolph V. Lombardi, Jr., MD, FACS, *t Thomas H. Mallory, MD , FACS,** and Bradley K. Vaughn, MD, FACS*t
Abstract: In an effort to identify prognostic indicators for knee man ipulation . the authors retrospectively reviewed the records of 60 osteoarthritic patients with posterior stabilized knee impl ant s who required man ipulation (94 knees) between January 1984 and December 1986. They also stud ied the records of 28 consecutive osteoarthritic patients who were implanted with 41 posterior stabilized knees between January 1985 and September 1985 who se knees did not require man ipulation (control group ). In both patient groups the following parameters were assessed and compared : overall knee alignment. joint line elevation. anterior to posterior (AP) dimension of the knee. AP placement of th e tibial component, patellar height. obesity, age, preoperative flexion, time of manipulation, single vs bilateral knee implants . final flexion. final Hospital for Special Surgery (HSS) score, and the development of heterotopic ossification . The findings of this study showed that an increase in the AP knee dimen sion by 12% or greater was a critically independent variable that significantly predisposed patients to manipulation. They also show that quadriceps adhe sions were another major factor leading to manipulation, and that rupturing of the se adhesions led to an increase in heterotopic ossification. This review also indicated that 3 months after knee arthroplasty was a significant time for evaluation because knee flexion and HSS score at this point in th e patient's recovery positively correlated with the final HSS score. Key words: knee manipulation. total knee arthroplasty, posterior stabilized knee implants.
Adequate range of motion is critical for a successful total knee arthroplasty (TKA). It has been shown that 67° of flexion are required for the swing phase of gait, 83° to climb stairs. 90° to descend stairs. and
at least 93° to rise from a chair (I I). While aggressive physical therapy is an important component of knee arthroplasty, manipulation may be necessary to supplement physical therapy when the knee stops gaining comfortable. active flexion in the early postoperative period. Manipulation is considered by some to be a surgical complication. It is also associated with risks, which can include anesthetic complications, supracondylar fractures (9). wound dehiscence (7). patellar ligament avulsions (9), and hemarthrosis (7).
From * Joint Implant Surgeons. Inc.. fDivision of Orthopaedic Surgery, The Ohio State University: Attending Staff. tSection of Joint lmplant Surgery. Grant Medical Center. Columbus. C!hio
All surgical procedures were performed at SaintAnthony Medical Center. Columbus. Ohio. Reprint requests: Adolph V. Lombardi. Jr., MD . FACS, Joint Implant Surgeons, Inc ., 720 East Broad Street, Columbus, OH 43215.
119
120
The Journal of Arthroplasty Vol. 6 No.2 June 1991
It has also been hypothesized that manipulation resulted in an increased incidence of heterotopic bone formation and that this adversely affected the final range of motion and the final HSS score (10). If factors predisposing a patient to the need for manipulation could be identified, it is likely that the incidence of manipulation and the associated complications could be reduced. Two studies have investigated factors relating to manipulation. Figgie et al. reported that joint line elevation positively correlated with an increased incidence of manipulation (6). Fox and Poss found age to be a factor, with patients over 70 years of age at greater risk (7). No other factors have been reported to correlate with the need for manipulation. In an effort to identify prognostic indicators for manipulation, we retrospectively reviewed the records of osteoarthritic patients who underwent primary total knee arthroplasty using the Insall-Burstein Posterior Stabilized Total Knee System (Zimmer, Warsaw, IN) that required manipulation. We also reviewed the records of a comparable group of osteoarthritic patients who underwent the same surgery but whose knees did not require manipulation (control group). We limited our study to osteoarthritic patients in order to create a more homogeneous group and avoid potentially confounding variables. We assessed overall knee alignment. joint line elevation, anterior to posterior (AP) dimension of the knee, AP placement of the tibial component, patellar height, obesity, age, and preoperative flexion, as well as the time of manipulation and single vs bilateral knee implants. The relationship between these variables and final flexion and HSS scores was also assessed. In both groups of patients. the heterotopic ossification was also assessed and analyzed.
Materials and Methods Between January 1984 and December 1986, 94 posterior stabilized knee arthroplasties were manipulated in 60 patients. There were 45 women and 15 men; the mean age was 66 ± 8 years (range, 4782 years). We also evaluated a control group of 41 consecutive posterior stabilized knee arthroplasties in 28 patients performed between January 1985 and September 1985. which did not require manipulation. This time period was chosen so that follow-up evaluation would be similar for both groups. There were 22 women and 6 men in the control group; the mean age was 70 ± 9 years (range. 38-85 years).
The follow-up period averaged 34.8 months (range, 24-50 months) for the manipulation group and 37 months (range, 24-40 months) for the control group. An obesity index was determined for each patient, according to the method described by the U.S. Department of Health and Human Services (21). Overweight was defined in terms of body mass index (BMI), which was determined by dividing weight in kilograms by height in meters squared. Overweight was then defined as BMI equal to or greater than that of the 85th percentile of men and women aged 20-29 years. Severe overweight was defined as a BMI equal to or greater than that at the 95th percentile. For patients in the manipulation group, the time of total knee arthroplasty, the time of manipulation, and any complications were recorded. Range of motion was determined before operation, prior to manipulation, and at manipulation; at 1 month, 3 months, 6 months, 12 months, and 24 months after manipulation; and at the most recent (final) office visit. Patients in the control group were measured at I, 3, 6, 12, and 24 months after operation, as well as during their most recent (final) office visit. In all patients, knee function was evaluated using the HSS knee scoring system. The HSS score was determined at 3, 6, 12, and 24 months after operation and during the most recent follow-up (final) office visit. Knees with less than 60 points were considered failures.
Radiographic Assessment
A standing AP radiograph (14" x 17") and an 8" x 10"lateral radiograph were made before operation and at each postoperative office visit. Before operation, the varus or valgus alignment was noted on the AP radiograph; the AP dimension of the femur and patella, as well as the joint line, were measured on the lateral radiograph (Fig. 1). The joint line was measured from the fibular head to the subchondral plate of the tibial plateau. Variations in radiographic penetration made it impossible to consistently use the tibial tubercle in this study. Similar measurements were obtained after operation. Radiographs that were rotated on the lateral precluded accurate measurement and were excluded from the study. The change in overall AP dimension of the knee was determined by noting the difference between the preoperative and postoperative patella plus femoral width divided by the addition of a preoperative AP patellar and femoral width. The postoperative anterior to posterior dimension of the patella did not
Knee Manipulation Following TKA
\
~
•
Daluga at al.
121
The height of the patella was measured from the inferior aspect of the patellar component to a line parallel to the weight-bearing surface of the tibial dish. This number was always positive in this study (Fig. 4). Heterotopic bone was evaluated radiographically in the postoperative study on the lateral radiograph in both patient groups. The area (ern") of heterotopic bone was noted for each knee. The time and location of first radiographic appearance of the heterotopic bone was also noted (Fig. 5) .
Physical Therapy
Postoperative physical therapy was based upon the status of the wound, with primary focus on quadriceps strengthening follow ed by knee flexion. From
(9+A' -A) + (B' - B) A+B
Fig. 1. Preoperative lateral radiographicmeasurements. A, patellar width; B , femoral width; JL, joint line, distance from fibular head to subchondral plate of tibial plateau. include the patellar prosthesis, since this could not be consistently measured in each radiograph. The patellar implants in this study measured 9 rom in each case and, therefore, this was entered into the numerator of the equation (Fig. 2). Each radiograph was taken from a set position, and magnification was noted to be insignificant, as judged by comparing preoperative and postoperative dimensions. This has been found to be the case in other studies, as well (6). A measure of the AP position ofthe tibial component was tabulated. A method similar to that described by Figgie et al. (6) was used by noting the distance from the anterior aspect of the tibial cortex to the posterior aspect of the cortex, and then calculating a center position. A center position was calculated for the tibial prosthesis, and the distance from the center of the prosthesis to the center of the tibial plateau represented the displacement. A positive number represented posterior placement of the component (Fig. 3).
Fig. 2. Postoperative lateral radiographic measurements. 9, width of patellar implant; A', postresection patellar width; 9 + A', width of patellar implant composite; B I ,
postarthroplasty femoral width; = % change in knee Width.
(9+A'-A)+(B'-B) A +B
122
The Journal of Arthroplasty Vol. 6 No.2 June 1991
JL'
Fig. 4. Postoperative lateral radiograph. P, height of patella
measured from inferior aspect of the patellar component to a line parallel to the weight-bearing surface of the tibial component; JL, preoperative joint line; JL', postoperative joint line. Fig. 3. Postoperative lateral radiograph. X, distance be-
tween prosthesis center and tibia center. 1984 to June of 1985, knees were placed in a Jones dressing until day 3. If the wound was dry and not swollen, low arc continuous passive motion (CPM) was started on day 3 along with knee slings for gentle flexion. Active flexion was initiated on day 7, and passive flexion on day 8 or 9. This regimen was changed in July 1985, to begin low arc CPM on day 1 and active flexion on day 5. Patients underwent manipulation if they had not achieved at least 70° of flexion by the time of discharge (early manipulation group) or were not progressing at a satisfactory rate. Patients in the intermediate group were discharged with 65°-75° of flexion, but subsequently stopped progressing or regressed. Patients in the late manipulation group were those who were unable to achieve 80°-85° by 3 months after discharge.
Statistical Analysis Methods of statistical analysis were applied to the following eight independent variables: preoperative flexion, change in the position of joint line, change
Fig. 5. Typical location of heterotopic bone.
Knee ManipUlation Following TKA •
in AP dimension of the knee, anterior to posterior placement of the tibial plateau. patellar height. body mass index (BMI), knee alignment. and patient age. The two dependent continuous variables, knee flexion and HSS knee scores, were recorded for each patient at the intervals described under Radiographic Assessment. The Pearson correlation coefficient was used to determine associations between each of the two dependent variables and the eight independent variables for each stud y group. Regression analysis was done separately for each of the two dependent variables with each of the eight independent variables . Knees with heterotopic ossification were analyzed with their respective group and as a separate group. The use of both knees in bilateral cases was not felt to be an independent observation. Analysis revealed a strong correlation between right and left knees of bilateral cases (r = .79, P = .000 1). Therefore. in an effort to prevent a biased evaluation of this population. only values for one knee were ana1yzed. The knee analyzed was chosen on a random basis. Method of Manipulation
Manipulation was performed with the patients under general anesthesia. Maximum muscle relax ation was obtained by using succinylcholine. Once muscle relaxation was reached. the hip was flexed and the knee was firmly, but not forcibly, flexed. Breaking of the adhesions could often be heard or palpated. This was done until a firm end point was reached or at least 100° of flexion was achieved. If adequate flexion was not achieved. the thigh was grasped and the lower leg allowed to fall. flexing the knee by gravitational forces from full extension to a flexed position. This was repeated until at least 100° of flexion were reached or an unyielding end point was felt. Physical therapy was begun immediately in the recovery room with passive flexion exercises. The motion was maintained with a continuous passive motion machine. and ice was applied liberally, as necessary. Active and passive flexion exercises were again carried out the morning after the manipulation, and patients were discharged from the hospital under the supervision of the physical therapist.
Results The incidence of manipulation in our osteoarthritic patients undergoing total knee arthroplasty using the tnsall-Burstein Posterior Stabilized Total
Daluga et al.
123
Knee System was 12% during the period reponed in this study . This incidence remained constant despite the change in our physical therapy protocol to begin low arc CPM and active flexion earlier. The overall anterior-posterior alignment on a 14" x 17" standing AP radiograph was not significantly different in the two groups. However. when data from the two groups were combined, pat ients with an alignment of 3°_9° of valgus had a mean HSS score of 89.5 ± 7. as compared to the final mean HSS score of 82.9 ± 9 for those patients with an alignment of less than 3° of anatomical valgus. While this was significant (P = .008). there was no correlation with final flexion. The overall lateral alignment of the femoral and tibial components was not significantly different in the two groups. The joint line in the manipulation group had a mean elevat ion of3.9 ± 4 mm (range. - 3-14 mrn). with 10 of the 80 knees measured having an elevation of 10 mm or greater. The control group's mean joint line elevation was 4.2 ± - 4 mm (range. - 5-12 mm) , with 9 of 36 knees having joint line elevations of 10 mm or greater. The differences between the groups were not statistically significant. and there was no correlation between joint line elevation and final flexion or final HSS scores. The AP diameter in the manipulation group had a mean percentile change of 6.6 ± 6% (range, - 3.8-17.2 %) . This change was greater than for the control group (4.6% ± 4.7 %; range, - 5-11 %). but was not statistically significant (P = .15) . No correlation could be found between increasing AP diameter and decreasing range of motion or HSS scores. When evaluating knees with a 9% or greater increase in an AP dimension, a trend toward manipulation was seen: 18 of 22 patients with a greater than or equal to 9% increase in AP dimension required manipulation (P = . 18). This trend became significant at a 12% increase in AP dimension, as all 10 patients in this subgroup required manipulation (P = .026). The final mean flexion and HSS score of these 10 knees were comparable to the remainder of the manipulation group (102° and 84°, respectively). Two of the four failures in the manipulation group were in this subgroup; one patient experienced persistent pain and stiffness and the other patient developed tibial loosening. In the manipulation group. the tibial tray had a mean position of 1.5 mm posterior to the midline (range, - 4 - 5 mm) and in the control group, 1.8 mm (range, - 3-7 mm) . Again, no statistical differences were found between the two groups and no correlation was evident between the tibial tray placement and the final HSS score or final flexion. The mean patellar height for the manipulation group was 24 mm ± 7 mm (range, 5-36 mm) and
124
The Journal of Arthroplasty Vol. 6 No.2 June 1991
10°, which is increased to a mean amount of flexion of 96° ± 9° at manipulation. At the most recent (final) evaluation, the mean amount of flexion finally achieved by the manipulation group was 103° ± 10° (range, 75°-120°), as compared to the control group, which achieved a mean flexion of 109° ± 7° (range, 95°-120°). This difference was statistically significant (P = .02). The greatest increase in range of motion occurred during the first postoperative month for both patient groups. The amount of flexion gradually increased for both groups until 1 year after surgery (Fig. 6). The manipulation group was divided into early (0-21 days), intermediate (22-90 days), and late (greater than 90 days) groups, according to when postoperative manipulation was performed. In the early and intermediate groups there was no significant difference between final flexion and final HSS scores. There was, however, a significant difference between the early and late groups in final mean flexion achieved (104° vs 97°) and the final mean HSS score (89 vs 81) (P == .08). Patients who underwent bilateral procedures had no greater incidence of manipulation. Thirty-four of the 60 patients (57%) in the manipulated group vs 13 of the 28 patients (46%) in the control groups
for the control group, 23 mm ± 5 mm (range, 1132 mm). The difference between the two groups was not significant, but there was a positive correlation between patellar height and final flexion (r = .3; P = .004). This correlation was strongest between an elevation of 10-24 mm of the joint line. Obesity was similar in each group and the correlation between an increase in obesity and a decrease in range of motion or decrease in HSS could not be made. Massive obesity (greater than the 95th percentile) was also evaluated but without a positive correlation for either group. The mean age of patients in the manipulation group was 66 years ± 8 years (range, 47-82 years) vs the mean age of 70 years ± 9 years for the control group. This was significant (P == .05) and contradicts an earlier study that implicated age greater than 70 years to be a factor in the need for later manipulation (7). No correlation was found between age and final flexion and final HSS score. The degree of preoperative flexion had no relationship to postoperative manipulation and was actually slightly greater in the manipulation group (average 107° ± I S°) than in the control group (average 104° ± 24°) (P = .04). In the manipulation group, the mean flexion prior to manipulation was 61° ±
110
..........0
.0· ..·········
---- ---
.......
......, . .
•
• '0' . ' •
..: •
.:...:~:..·_·o __ -:.&-~----·A
•
100
.-.... Manipulated
90
0····0
Control
A---·AHO
80
70
~--=---+----~'::-----------:::':-.",/---:=:""-:-7" I
3
6
12 Months Post-operative
24
Fig. 6. Average postoperative flexion as a function of time.
Final
Knee Manipulation Following TKA •
Daluga et al.
125
90
p
.
85
.!
f . .: t::f*"" :
:
.:
:
80
:
.: :
:
:'
,,
/,' 0'
z
I
.r"
... .,,"
...
;,.;.----
~__~ -6f. ... rus._-.,..,.,........------- .
o
------6
, ,, , ,
I
•
• Manipulated 0·······..0 Control 6.----6. HO
75
70
,
L---!-----:!::------:-"=-----------::L.:--fiI'~:-'-":'""
12 Months Post-operative
24
'
Final
Fig. 7. Average postoperative HSS scores as a function of time.
had bilateral knee arthroplasties performed. The difference, however, was not significant. The mean final HSS score was 87 ± 7 (range, 5999) for the manipulation group and 88 ± 8 (range, 66-98) for the control group. This difference was not statistically significant. There was a positive correlation with the amount of flexion at 3 postoperative months and the final HSS score (r = .37; P = .0009). The 3-month HSS score also correlated positively with the final HSS score (r = .38; P = .002). HSS and flexion made significant improvements until I year after surgery for both patient groups (Figs. 6 and 7). Heterotopic ossification was noted in 17 of the 60 patients (28 .3%) whose knees were manipulated, and in I of the 28 patients in the control group (3.7%). This difference is significant (P < .001). Thirteen of the 17 patients with heterotopic ossification had bilateral procedures performed. but only three had heterotopic ossification in both knees. When comparing the bilateral knee scores of the 10 patients with heterotopic ossification, the knee with heterotopic bone did as well or better as the knee without heterotopic bone in 7 of the 10 bilateral cases. There was no significant difference in final flexion achieved
or in final HSS score between the two knees (Figs. 6 and 7) . The mean final HSS scores in the patients with heterotopic ossification was 86 ± 7, and the mean final flexion achieved was 106° ± 6°. Flexion in the heterotopic ossification group lagged initially but eventually surpassed the final flexion of the manipulated group; however, this was not statistically significant. The area of heterotopic bone (range, 1.2 cm-16 crrr') could not be correlated with final flexion . Two locations of heterotopic ossification were seen. All but one was juxtaposed to the distal femur 0-2 em proximal to the femoral component (Fig. 8) . The remaining patient had bone deposition in the quadriceps tendon with bone adjoining the patella. This patient had decreased flexion of the knee and noted a generalized stiffness. This was the only patient in the group who noted a significa nt difference (Fig. 9). There were four failure s in the manipulation group. Two were in the subgroup of knees with a 12% or greater increase in the AP dimension and the other two were revised to a more constrained pros thesis secondary to instability. The three failures in the control group were all secon dary to patellofemoral problems.
126
The Journal of Arthroplasty Vol. 6 No.2 June 1991
study , the largest improvement was seen in the first month in both patient groups. Range of motion continued to improve until I postoperative year had passed. The mean final flexion achieved by patients in the manipulation group was less than that achieved by patients in the control group (103° vs 108°, respectively). This is contrary to the report by Fox and Pass (7). Nonetheless, the mean final flexion at followup examination, 103 °, is still quite adequate for all activities of daily living. The mean final HSS score was essentially the same for both the manipulation and control groups (87 vs 88, respectively) . It has been theorized that reestablishing the preoperative dimension of the knee joint is crucial in achieving good results .in total knee arthroplasty (18) . Increasing the overall AP dimension of the knee will anteriorly displace the extensor mechanism and tighten the quadriceps. The net effect is a functionally shortened quadriceps, resulting in a reduction of flexion (16). Our study failed to find a correlation between a change in the AP measurement and range of motion, but when a 9% increase was calculated,
Fig. 8. Two-year postoperative radiograph showingtypical location of heterotopic ossification.
Discussion Postoperative manipulation has been reported to be necessary in as many as 54-60% of total knee arthroplasties (19, 20) . While the incidence is much lower today, manipulation is still the most common sequela of total knee arthroplasty. It has been shown that manipulation, while being an important therapeutic adjuvant, does not increase the ultimate flexion that can be achieved (7). Achievement of flexion greater than 105°, though, is seldom indicated and may actually increase patellofemoral problems. Fox and Poss (7) have reported that manipulated knees eventually achieve a flexion similar to nonmanipulated knees and found only age as a predisposing factor. Figgie et al. (6) found that joint line elevation, especially when equal to or greater than 10 mrn, significantly increased the chances for the need of manipulation. Schurman, Parker, and Ornstein (17) reported that the greatest increase in range of motion occurred during the first 3 months following surgery. In our
Fig. 9. Four-year postoperative radiograph of a patient with atypical heterotopic ossification limiting range of motion.
Knee ManipUlation Following TKA •
a trend toward manipulation was seen (P = . 18). At a 12% increase in AP dimension, all 10 knees required manipulation (P = .026) . Two of these knees went on to failure-one because of stiffness and the other due to tibial component loosening. The remaining eight knees showed no long-term effects; their mean final flexion of 106° and HSS score of 88 were better than the mean flexion (103°) and HSS score (87) of the manipulation group as a whole. This indicates that excessively increasing the depth of the knee ("stuffing the joint") compromises the initial gains in flexion but does not prevent an adequate flexion or exceIlent HSS score from being achieved. Motion is delayed in these knees until the quadriceps can be adequately stretched . If the depth of the knee is excessively increased, quadriceps stretching may be beyond the capabilities of the patient. and manipulation may be necessary. The increase in AP depth may also predi spose to a higher failure rate but there were too few patients in this subgroup to make a conclusive statement. Patients 70 years of age and older were previously reported to be more likely to require manipulation (7). This was not substantiated in our study. as the manipulation group was actually younger than the control group (P = .05). Further results of our study indicate that the manipulation group and the control group were indistinguishable on the basis of joint line elevation. AP position of the tibial prosthesis, patellar height, postoperative alignment. weight of the patient, and degree of preoperative flexion. When patients with joint line elevation of 10 mm or more were specifically assessed, a relationship still could not be found in regard to influencing manipulation. In fact. joint line elevation of 10 mm or more was actually more prevalent in the control group. A relationship was found between patellar height and degree of flexion. A previous study has recommended that optimal patellar height should be between 10 mm and 30 mm (6). Our study supports that recommendation, but shows that the greater the elevation, the better the ultimate flexion. The correlation is strongest between 10 mm and 24 mm . In other words, as patellar height increases up to 24 mm, flexion improves. No further correlation was seen for elevations from 24 mm to 30 mm. Therefore. it seems that for the best final flexion. a range of 2430 rnm for patellar height is preferable. No correlation was found with patellar height and final HSS score. Manipulations are not always performed within the first month after surgery and. from the findings of our study . no difference could be found between the early (21 days) and intermediate (22-90 days)
Daluga et al.
127
groups . However, overall flexion and HSS scores are lower in knees manipulated beyond 3 postoperative months. This is not surprising, since 3 postoperative months seems to be an important period for knee evaluations. At 3 months, knee flexion and HSS scores both closely correlate with the final HSS score . If a knee is still doing poorly 3 months after surgery, the overall prognosis must be guarded. Heterotopic bone is commonly found after total hip arthroplasty, especially in hypertrophic osteoarthritic patients (L 2, 4, 13). It is also increased in patients with elbow trauma when combined with passive range of motion (3,8). Reports of heterotopic ossification have ranged from an incidence of 1042% following total knee arthroplasty (5, 12). It was not found to be progressive nor did it affect the final outcome in these studies. Predisposing factors included patients with posttraumatic arthritis, a preoperative combined deformity of greater than 15°, and manipulation in the postoperative period (5). We found manipulated knees to be significantly more predisposed to heterotopic ossification , but no relationship to preoperative alignment was found. Even though a wide range of heterotopic bone was found , no classification was attempted since no correlation between the amount of heterotopic ossification and final flexion or final HSS score could be found. In fact, patients with heterotopic ossification actually had an improved overall flexion as compared to the manipulated group as a whole. The development of heterotopic ossification after manipulation lends credence to the likelihood that scarring of the quadriceps is a significant factor in restricted postoperative motion. Michelsson et al. (14) have shown that immobilization and forceful manipulation of the knee lead to heterotopic bone formation secondary to microscopic tearing of the quadriceps and subsequent hematoma formation. In another study using a rabbit model, Michclsson and Rauschning found that bleeding and subsequent heterotopic ossification were localized mainly under the fascia in the vastus intermedius (15) . The heterotopic ossification in our study was also found under the distal portion of the vastus intermedius. In knees with a slow recovery rate, adhesions are formed. Once formed . no further significant gains in range of motion can be realistically hoped for, and manipulation is inevnable. Many physical therapists can easily identify this patient as one with a "hard end feel." These patients will require manipulation and, although there are no significant differences found between early and intermediate manipulations. once a hard end point is realized, the knee should be manipulated as soon as possible so that the patient can quickly achieve a functional range of motion and not
128
The Journal of Arthroplasty Vol. 6 No.2 June 1991
become discouraged with fruitless efforts during physical therapy. It is possible that the true value of continuous passive motion will be in preventing these adhesions from forming and ultimately decreasing the manipulation rate. Vince et al. (22) , in a prospective study, found manipulations unnecessary in knees treated w ith aggressive continuous passive motion. Continuous passive motion was used as a physical thcrapy modality in our study but initially only at a low arc. Increase of the AP dimension of the knee by 12% or more is a critical independent variable that significantly predisposes the patient to manipulation and may increase the likelihood of failure of the implant. Quadriceps adhesions and scarring are other major factors that predispose to knee stiffness. Manipulation is necessary to break these adhesions if an adequate range of motion is to be achieved . The microtrauma from the breakage of these adhesions results in a significant percentage of knees with heterotopic ossification. A period of 3 months was found to be a prognostically significant time for evaluation since, at 3 postoperative months, knee flexion and HSS score correlated positively with the final HSS score, and knees manipulated after 3 months did not do as well . Heterotopic os sification affected neither the final range of motion nor the final HSS scores and was, therefore, regarded as an incidental finding in thi s stu dy.
References 1. Cella JP, Salvati EA, Sulco TP: Indomethacin for the
2.
3.
4.
5.
6.
prevention of heterotopic ossification following total hip arthroplasty: effectiveness. contra indications and adverse effects. J Arthropla sty 3:229, 1988 Delee J, Ferrari A, Charnley J : Ectopic bone formation following low friction arthroplasty of the hip. Clin Orthop 121:53, 1976 Delee JC, Green DP, Wilkins KE: Fractures and dislocations of the elbow . p. 613. In Rockwood CA Jr. Green DP (eds): Fractures in adult s. 2nd ed. JB Lippincott Co, Philadelphi a, 1984 Eftekhar NS, Kiernan HA. Stinchfield FE: Systemic and local complications following low -friction arthro plasty of the hip. Arch Surg I I I : ISO, 1976 Figgie III HE, Goldberg VM, Heiple KG et al: The incidence and significance of heterotopic ossification following total knee anhroplasty. Adv Orthop Surg 10: 12, 1986 Figgie III HE, Goldberg VM, Heiple KG et al: The influence of tibial-patellofemorallocation on function
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18,
19.
20.
21.
22.
of the knee in patients with the posterior stabilized condylar knee prosthesis. J Bone Joint Surg 68A:1035, 1986 Fox .n, Poss R: The role of man ipulation following total knee replacement. J Bone Joint Surg 63A:357, 1981 Hotchkiss RN: Elbow and forearm : trauma in adult and pediatric patients . p. 223. In Fitzgerald RHJr (ed] : Orthopaedic knowledge update 2. American Academy of Orthopaedic Surgery, Park Ridge, II., 1987 Insall J, Scott WN, Ranawat CS: The total condylar knee prosthesis: a report of two hundred and twenty cases. J Bone Joint Surg 61A:173, 1979 Ivey M: Myositis ossificans of the thigh following manipulation of the knee : a case report . Clin Orthop 198: 102, 1985 Kettlekarnp DB: Gait characteristics of the knee: normal. abnormal and postreconstruction. p. 47. In American Academy of Orthopaedic Surgeons: Symposium on Reconstructive Surgery of the Knee. CV Mosby, St. Louis, 1976 Lovelock JE, Griffiths HJ, Silverstein AM et al: Complications of total knee replacement. Am J Roentgenol 142:985, 1984 Matos M, Amstutz HC, Finerman G: Myositis ossificans following total hip replacement. J Bone Joint Surg 57A:137, 1975 Michelsson JE, Granroth G, Andersson LC: Myositis ossificans following forcible manipulation of the leg: a rabbit model for the study of heterotopic bone formation. J Bone Joint Surg 62A:81 I. 1980 Michelsson JE, Rauschning W: Pathogenesis of experimental heterotopic bone formation following temporary forcible exercising of immobilized limbs. Clin Orthop 176:265, 1983 Rand JA, Gustilo RB: Technique of patellar resurfacing in total knee arthroplasty. Techniques Orthop 3:57, 1988 Schurman OJ, Parker IN, Ornstein 0: Total condylar knee replacement: a study offactors influencing range of motion as late as two years after arthroplasty. J Bone Joint Surg 67A:lO06, 1985 Scott RD: Treatment of patellar instability associated with total knee replacement. Techniques Orthop 3:9, 1988 Shoji H, Yoshino S, Komagamine M: Improved range of motion with the Y/S total knee arthroplasty system. Clin Orthop 2 I8: 150, 1987 Soudry M, Binazzi R, Insall IN et al: Successive bilateral total knee replacement. J Bone Joint Surg 67A:573, 1985 U.S. Department of Health and Human Services: Anthropometric reference data and prevalence of overweight. Vital and Health Statistics. DHHS Publication No. (PHS) 87·1688 Vince KG, Kel1y MA, Beck J et al: Continuous passive motion after total knee arthroplasty. J Arthroplasty 2:281, 1987
Daniel Daluga, MD,* Adolph V. Lombardi, Jr., MD, FACS, *t Thomas H. Mallory, MD , FACS,** and Bradley K. Vaughn, MD, FACS*t
Abstract: In an effort to identify prognostic indicators for knee man ipulation . the authors retrospectively reviewed the records of 60 osteoarthritic patients with posterior stabilized knee impl ant s who required man ipulation (94 knees) between January 1984 and December 1986. They also stud ied the records of 28 consecutive osteoarthritic patients who were implanted with 41 posterior stabilized knees between January 1985 and September 1985 who se knees did not require man ipulation (control group ). In both patient groups the following parameters were assessed and compared : overall knee alignment. joint line elevation. anterior to posterior (AP) dimension of the knee. AP placement of th e tibial component, patellar height. obesity, age, preoperative flexion, time of manipulation, single vs bilateral knee implants . final flexion. final Hospital for Special Surgery (HSS) score, and the development of heterotopic ossification . The findings of this study showed that an increase in the AP knee dimen sion by 12% or greater was a critically independent variable that significantly predisposed patients to manipulation. They also show that quadriceps adhe sions were another major factor leading to manipulation, and that rupturing of the se adhesions led to an increase in heterotopic ossification. This review also indicated that 3 months after knee arthroplasty was a significant time for evaluation because knee flexion and HSS score at this point in th e patient's recovery positively correlated with the final HSS score. Key words: knee manipulation. total knee arthroplasty, posterior stabilized knee implants.
Adequate range of motion is critical for a successful total knee arthroplasty (TKA). It has been shown that 67° of flexion are required for the swing phase of gait, 83° to climb stairs. 90° to descend stairs. and
at least 93° to rise from a chair (I I). While aggressive physical therapy is an important component of knee arthroplasty, manipulation may be necessary to supplement physical therapy when the knee stops gaining comfortable. active flexion in the early postoperative period. Manipulation is considered by some to be a surgical complication. It is also associated with risks, which can include anesthetic complications, supracondylar fractures (9). wound dehiscence (7). patellar ligament avulsions (9), and hemarthrosis (7).
From * Joint Implant Surgeons. Inc.. fDivision of Orthopaedic Surgery, The Ohio State University: Attending Staff. tSection of Joint lmplant Surgery. Grant Medical Center. Columbus. C!hio
All surgical procedures were performed at SaintAnthony Medical Center. Columbus. Ohio. Reprint requests: Adolph V. Lombardi. Jr., MD . FACS, Joint Implant Surgeons, Inc ., 720 East Broad Street, Columbus, OH 43215.
119
120
The Journal of Arthroplasty Vol. 6 No.2 June 1991
It has also been hypothesized that manipulation resulted in an increased incidence of heterotopic bone formation and that this adversely affected the final range of motion and the final HSS score (10). If factors predisposing a patient to the need for manipulation could be identified, it is likely that the incidence of manipulation and the associated complications could be reduced. Two studies have investigated factors relating to manipulation. Figgie et al. reported that joint line elevation positively correlated with an increased incidence of manipulation (6). Fox and Poss found age to be a factor, with patients over 70 years of age at greater risk (7). No other factors have been reported to correlate with the need for manipulation. In an effort to identify prognostic indicators for manipulation, we retrospectively reviewed the records of osteoarthritic patients who underwent primary total knee arthroplasty using the Insall-Burstein Posterior Stabilized Total Knee System (Zimmer, Warsaw, IN) that required manipulation. We also reviewed the records of a comparable group of osteoarthritic patients who underwent the same surgery but whose knees did not require manipulation (control group). We limited our study to osteoarthritic patients in order to create a more homogeneous group and avoid potentially confounding variables. We assessed overall knee alignment. joint line elevation, anterior to posterior (AP) dimension of the knee, AP placement of the tibial component, patellar height, obesity, age, and preoperative flexion, as well as the time of manipulation and single vs bilateral knee implants. The relationship between these variables and final flexion and HSS scores was also assessed. In both groups of patients. the heterotopic ossification was also assessed and analyzed.
Materials and Methods Between January 1984 and December 1986, 94 posterior stabilized knee arthroplasties were manipulated in 60 patients. There were 45 women and 15 men; the mean age was 66 ± 8 years (range, 4782 years). We also evaluated a control group of 41 consecutive posterior stabilized knee arthroplasties in 28 patients performed between January 1985 and September 1985. which did not require manipulation. This time period was chosen so that follow-up evaluation would be similar for both groups. There were 22 women and 6 men in the control group; the mean age was 70 ± 9 years (range. 38-85 years).
The follow-up period averaged 34.8 months (range, 24-50 months) for the manipulation group and 37 months (range, 24-40 months) for the control group. An obesity index was determined for each patient, according to the method described by the U.S. Department of Health and Human Services (21). Overweight was defined in terms of body mass index (BMI), which was determined by dividing weight in kilograms by height in meters squared. Overweight was then defined as BMI equal to or greater than that of the 85th percentile of men and women aged 20-29 years. Severe overweight was defined as a BMI equal to or greater than that at the 95th percentile. For patients in the manipulation group, the time of total knee arthroplasty, the time of manipulation, and any complications were recorded. Range of motion was determined before operation, prior to manipulation, and at manipulation; at 1 month, 3 months, 6 months, 12 months, and 24 months after manipulation; and at the most recent (final) office visit. Patients in the control group were measured at I, 3, 6, 12, and 24 months after operation, as well as during their most recent (final) office visit. In all patients, knee function was evaluated using the HSS knee scoring system. The HSS score was determined at 3, 6, 12, and 24 months after operation and during the most recent follow-up (final) office visit. Knees with less than 60 points were considered failures.
Radiographic Assessment
A standing AP radiograph (14" x 17") and an 8" x 10"lateral radiograph were made before operation and at each postoperative office visit. Before operation, the varus or valgus alignment was noted on the AP radiograph; the AP dimension of the femur and patella, as well as the joint line, were measured on the lateral radiograph (Fig. 1). The joint line was measured from the fibular head to the subchondral plate of the tibial plateau. Variations in radiographic penetration made it impossible to consistently use the tibial tubercle in this study. Similar measurements were obtained after operation. Radiographs that were rotated on the lateral precluded accurate measurement and were excluded from the study. The change in overall AP dimension of the knee was determined by noting the difference between the preoperative and postoperative patella plus femoral width divided by the addition of a preoperative AP patellar and femoral width. The postoperative anterior to posterior dimension of the patella did not
Knee Manipulation Following TKA
\
~
•
Daluga at al.
121
The height of the patella was measured from the inferior aspect of the patellar component to a line parallel to the weight-bearing surface of the tibial dish. This number was always positive in this study (Fig. 4). Heterotopic bone was evaluated radiographically in the postoperative study on the lateral radiograph in both patient groups. The area (ern") of heterotopic bone was noted for each knee. The time and location of first radiographic appearance of the heterotopic bone was also noted (Fig. 5) .
Physical Therapy
Postoperative physical therapy was based upon the status of the wound, with primary focus on quadriceps strengthening follow ed by knee flexion. From
(9+A' -A) + (B' - B) A+B
Fig. 1. Preoperative lateral radiographicmeasurements. A, patellar width; B , femoral width; JL, joint line, distance from fibular head to subchondral plate of tibial plateau. include the patellar prosthesis, since this could not be consistently measured in each radiograph. The patellar implants in this study measured 9 rom in each case and, therefore, this was entered into the numerator of the equation (Fig. 2). Each radiograph was taken from a set position, and magnification was noted to be insignificant, as judged by comparing preoperative and postoperative dimensions. This has been found to be the case in other studies, as well (6). A measure of the AP position ofthe tibial component was tabulated. A method similar to that described by Figgie et al. (6) was used by noting the distance from the anterior aspect of the tibial cortex to the posterior aspect of the cortex, and then calculating a center position. A center position was calculated for the tibial prosthesis, and the distance from the center of the prosthesis to the center of the tibial plateau represented the displacement. A positive number represented posterior placement of the component (Fig. 3).
Fig. 2. Postoperative lateral radiographic measurements. 9, width of patellar implant; A', postresection patellar width; 9 + A', width of patellar implant composite; B I ,
postarthroplasty femoral width; = % change in knee Width.
(9+A'-A)+(B'-B) A +B
122
The Journal of Arthroplasty Vol. 6 No.2 June 1991
JL'
Fig. 4. Postoperative lateral radiograph. P, height of patella
measured from inferior aspect of the patellar component to a line parallel to the weight-bearing surface of the tibial component; JL, preoperative joint line; JL', postoperative joint line. Fig. 3. Postoperative lateral radiograph. X, distance be-
tween prosthesis center and tibia center. 1984 to June of 1985, knees were placed in a Jones dressing until day 3. If the wound was dry and not swollen, low arc continuous passive motion (CPM) was started on day 3 along with knee slings for gentle flexion. Active flexion was initiated on day 7, and passive flexion on day 8 or 9. This regimen was changed in July 1985, to begin low arc CPM on day 1 and active flexion on day 5. Patients underwent manipulation if they had not achieved at least 70° of flexion by the time of discharge (early manipulation group) or were not progressing at a satisfactory rate. Patients in the intermediate group were discharged with 65°-75° of flexion, but subsequently stopped progressing or regressed. Patients in the late manipulation group were those who were unable to achieve 80°-85° by 3 months after discharge.
Statistical Analysis Methods of statistical analysis were applied to the following eight independent variables: preoperative flexion, change in the position of joint line, change
Fig. 5. Typical location of heterotopic bone.
Knee ManipUlation Following TKA •
in AP dimension of the knee, anterior to posterior placement of the tibial plateau. patellar height. body mass index (BMI), knee alignment. and patient age. The two dependent continuous variables, knee flexion and HSS knee scores, were recorded for each patient at the intervals described under Radiographic Assessment. The Pearson correlation coefficient was used to determine associations between each of the two dependent variables and the eight independent variables for each stud y group. Regression analysis was done separately for each of the two dependent variables with each of the eight independent variables . Knees with heterotopic ossification were analyzed with their respective group and as a separate group. The use of both knees in bilateral cases was not felt to be an independent observation. Analysis revealed a strong correlation between right and left knees of bilateral cases (r = .79, P = .000 1). Therefore. in an effort to prevent a biased evaluation of this population. only values for one knee were ana1yzed. The knee analyzed was chosen on a random basis. Method of Manipulation
Manipulation was performed with the patients under general anesthesia. Maximum muscle relax ation was obtained by using succinylcholine. Once muscle relaxation was reached. the hip was flexed and the knee was firmly, but not forcibly, flexed. Breaking of the adhesions could often be heard or palpated. This was done until a firm end point was reached or at least 100° of flexion was achieved. If adequate flexion was not achieved. the thigh was grasped and the lower leg allowed to fall. flexing the knee by gravitational forces from full extension to a flexed position. This was repeated until at least 100° of flexion were reached or an unyielding end point was felt. Physical therapy was begun immediately in the recovery room with passive flexion exercises. The motion was maintained with a continuous passive motion machine. and ice was applied liberally, as necessary. Active and passive flexion exercises were again carried out the morning after the manipulation, and patients were discharged from the hospital under the supervision of the physical therapist.
Results The incidence of manipulation in our osteoarthritic patients undergoing total knee arthroplasty using the tnsall-Burstein Posterior Stabilized Total
Daluga et al.
123
Knee System was 12% during the period reponed in this study . This incidence remained constant despite the change in our physical therapy protocol to begin low arc CPM and active flexion earlier. The overall anterior-posterior alignment on a 14" x 17" standing AP radiograph was not significantly different in the two groups. However. when data from the two groups were combined, pat ients with an alignment of 3°_9° of valgus had a mean HSS score of 89.5 ± 7. as compared to the final mean HSS score of 82.9 ± 9 for those patients with an alignment of less than 3° of anatomical valgus. While this was significant (P = .008). there was no correlation with final flexion. The overall lateral alignment of the femoral and tibial components was not significantly different in the two groups. The joint line in the manipulation group had a mean elevat ion of3.9 ± 4 mm (range. - 3-14 mrn). with 10 of the 80 knees measured having an elevation of 10 mm or greater. The control group's mean joint line elevation was 4.2 ± - 4 mm (range. - 5-12 mm) , with 9 of 36 knees having joint line elevations of 10 mm or greater. The differences between the groups were not statistically significant. and there was no correlation between joint line elevation and final flexion or final HSS scores. The AP diameter in the manipulation group had a mean percentile change of 6.6 ± 6% (range, - 3.8-17.2 %) . This change was greater than for the control group (4.6% ± 4.7 %; range, - 5-11 %). but was not statistically significant (P = .15) . No correlation could be found between increasing AP diameter and decreasing range of motion or HSS scores. When evaluating knees with a 9% or greater increase in an AP dimension, a trend toward manipulation was seen: 18 of 22 patients with a greater than or equal to 9% increase in AP dimension required manipulation (P = . 18). This trend became significant at a 12% increase in AP dimension, as all 10 patients in this subgroup required manipulation (P = .026). The final mean flexion and HSS score of these 10 knees were comparable to the remainder of the manipulation group (102° and 84°, respectively). Two of the four failures in the manipulation group were in this subgroup; one patient experienced persistent pain and stiffness and the other patient developed tibial loosening. In the manipulation group. the tibial tray had a mean position of 1.5 mm posterior to the midline (range, - 4 - 5 mm) and in the control group, 1.8 mm (range, - 3-7 mm) . Again, no statistical differences were found between the two groups and no correlation was evident between the tibial tray placement and the final HSS score or final flexion. The mean patellar height for the manipulation group was 24 mm ± 7 mm (range, 5-36 mm) and
124
The Journal of Arthroplasty Vol. 6 No.2 June 1991
10°, which is increased to a mean amount of flexion of 96° ± 9° at manipulation. At the most recent (final) evaluation, the mean amount of flexion finally achieved by the manipulation group was 103° ± 10° (range, 75°-120°), as compared to the control group, which achieved a mean flexion of 109° ± 7° (range, 95°-120°). This difference was statistically significant (P = .02). The greatest increase in range of motion occurred during the first postoperative month for both patient groups. The amount of flexion gradually increased for both groups until 1 year after surgery (Fig. 6). The manipulation group was divided into early (0-21 days), intermediate (22-90 days), and late (greater than 90 days) groups, according to when postoperative manipulation was performed. In the early and intermediate groups there was no significant difference between final flexion and final HSS scores. There was, however, a significant difference between the early and late groups in final mean flexion achieved (104° vs 97°) and the final mean HSS score (89 vs 81) (P == .08). Patients who underwent bilateral procedures had no greater incidence of manipulation. Thirty-four of the 60 patients (57%) in the manipulated group vs 13 of the 28 patients (46%) in the control groups
for the control group, 23 mm ± 5 mm (range, 1132 mm). The difference between the two groups was not significant, but there was a positive correlation between patellar height and final flexion (r = .3; P = .004). This correlation was strongest between an elevation of 10-24 mm of the joint line. Obesity was similar in each group and the correlation between an increase in obesity and a decrease in range of motion or decrease in HSS could not be made. Massive obesity (greater than the 95th percentile) was also evaluated but without a positive correlation for either group. The mean age of patients in the manipulation group was 66 years ± 8 years (range, 47-82 years) vs the mean age of 70 years ± 9 years for the control group. This was significant (P == .05) and contradicts an earlier study that implicated age greater than 70 years to be a factor in the need for later manipulation (7). No correlation was found between age and final flexion and final HSS score. The degree of preoperative flexion had no relationship to postoperative manipulation and was actually slightly greater in the manipulation group (average 107° ± I S°) than in the control group (average 104° ± 24°) (P = .04). In the manipulation group, the mean flexion prior to manipulation was 61° ±
110
..........0
.0· ..·········
---- ---
.......
......, . .
•
• '0' . ' •
..: •
.:...:~:..·_·o __ -:.&-~----·A
•
100
.-.... Manipulated
90
0····0
Control
A---·AHO
80
70
~--=---+----~'::-----------:::':-.",/---:=:""-:-7" I
3
6
12 Months Post-operative
24
Fig. 6. Average postoperative flexion as a function of time.
Final
Knee Manipulation Following TKA •
Daluga et al.
125
90
p
.
85
.!
f . .: t::f*"" :
:
.:
:
80
:
.: :
:
:'
,,
/,' 0'
z
I
.r"
... .,,"
...
;,.;.----
~__~ -6f. ... rus._-.,..,.,........------- .
o
------6
, ,, , ,
I
•
• Manipulated 0·······..0 Control 6.----6. HO
75
70
,
L---!-----:!::------:-"=-----------::L.:--fiI'~:-'-":'""
12 Months Post-operative
24
'
Final
Fig. 7. Average postoperative HSS scores as a function of time.
had bilateral knee arthroplasties performed. The difference, however, was not significant. The mean final HSS score was 87 ± 7 (range, 5999) for the manipulation group and 88 ± 8 (range, 66-98) for the control group. This difference was not statistically significant. There was a positive correlation with the amount of flexion at 3 postoperative months and the final HSS score (r = .37; P = .0009). The 3-month HSS score also correlated positively with the final HSS score (r = .38; P = .002). HSS and flexion made significant improvements until I year after surgery for both patient groups (Figs. 6 and 7). Heterotopic ossification was noted in 17 of the 60 patients (28 .3%) whose knees were manipulated, and in I of the 28 patients in the control group (3.7%). This difference is significant (P < .001). Thirteen of the 17 patients with heterotopic ossification had bilateral procedures performed. but only three had heterotopic ossification in both knees. When comparing the bilateral knee scores of the 10 patients with heterotopic ossification, the knee with heterotopic bone did as well or better as the knee without heterotopic bone in 7 of the 10 bilateral cases. There was no significant difference in final flexion achieved
or in final HSS score between the two knees (Figs. 6 and 7) . The mean final HSS scores in the patients with heterotopic ossification was 86 ± 7, and the mean final flexion achieved was 106° ± 6°. Flexion in the heterotopic ossification group lagged initially but eventually surpassed the final flexion of the manipulated group; however, this was not statistically significant. The area of heterotopic bone (range, 1.2 cm-16 crrr') could not be correlated with final flexion . Two locations of heterotopic ossification were seen. All but one was juxtaposed to the distal femur 0-2 em proximal to the femoral component (Fig. 8) . The remaining patient had bone deposition in the quadriceps tendon with bone adjoining the patella. This patient had decreased flexion of the knee and noted a generalized stiffness. This was the only patient in the group who noted a significa nt difference (Fig. 9). There were four failure s in the manipulation group. Two were in the subgroup of knees with a 12% or greater increase in the AP dimension and the other two were revised to a more constrained pros thesis secondary to instability. The three failures in the control group were all secon dary to patellofemoral problems.
126
The Journal of Arthroplasty Vol. 6 No.2 June 1991
study , the largest improvement was seen in the first month in both patient groups. Range of motion continued to improve until I postoperative year had passed. The mean final flexion achieved by patients in the manipulation group was less than that achieved by patients in the control group (103° vs 108°, respectively). This is contrary to the report by Fox and Pass (7). Nonetheless, the mean final flexion at followup examination, 103 °, is still quite adequate for all activities of daily living. The mean final HSS score was essentially the same for both the manipulation and control groups (87 vs 88, respectively) . It has been theorized that reestablishing the preoperative dimension of the knee joint is crucial in achieving good results .in total knee arthroplasty (18) . Increasing the overall AP dimension of the knee will anteriorly displace the extensor mechanism and tighten the quadriceps. The net effect is a functionally shortened quadriceps, resulting in a reduction of flexion (16). Our study failed to find a correlation between a change in the AP measurement and range of motion, but when a 9% increase was calculated,
Fig. 8. Two-year postoperative radiograph showingtypical location of heterotopic ossification.
Discussion Postoperative manipulation has been reported to be necessary in as many as 54-60% of total knee arthroplasties (19, 20) . While the incidence is much lower today, manipulation is still the most common sequela of total knee arthroplasty. It has been shown that manipulation, while being an important therapeutic adjuvant, does not increase the ultimate flexion that can be achieved (7). Achievement of flexion greater than 105°, though, is seldom indicated and may actually increase patellofemoral problems. Fox and Poss (7) have reported that manipulated knees eventually achieve a flexion similar to nonmanipulated knees and found only age as a predisposing factor. Figgie et al. (6) found that joint line elevation, especially when equal to or greater than 10 mrn, significantly increased the chances for the need of manipulation. Schurman, Parker, and Ornstein (17) reported that the greatest increase in range of motion occurred during the first 3 months following surgery. In our
Fig. 9. Four-year postoperative radiograph of a patient with atypical heterotopic ossification limiting range of motion.
Knee ManipUlation Following TKA •
a trend toward manipulation was seen (P = . 18). At a 12% increase in AP dimension, all 10 knees required manipulation (P = .026) . Two of these knees went on to failure-one because of stiffness and the other due to tibial component loosening. The remaining eight knees showed no long-term effects; their mean final flexion of 106° and HSS score of 88 were better than the mean flexion (103°) and HSS score (87) of the manipulation group as a whole. This indicates that excessively increasing the depth of the knee ("stuffing the joint") compromises the initial gains in flexion but does not prevent an adequate flexion or exceIlent HSS score from being achieved. Motion is delayed in these knees until the quadriceps can be adequately stretched . If the depth of the knee is excessively increased, quadriceps stretching may be beyond the capabilities of the patient. and manipulation may be necessary. The increase in AP depth may also predi spose to a higher failure rate but there were too few patients in this subgroup to make a conclusive statement. Patients 70 years of age and older were previously reported to be more likely to require manipulation (7). This was not substantiated in our study. as the manipulation group was actually younger than the control group (P = .05). Further results of our study indicate that the manipulation group and the control group were indistinguishable on the basis of joint line elevation. AP position of the tibial prosthesis, patellar height, postoperative alignment. weight of the patient, and degree of preoperative flexion. When patients with joint line elevation of 10 mm or more were specifically assessed, a relationship still could not be found in regard to influencing manipulation. In fact. joint line elevation of 10 mm or more was actually more prevalent in the control group. A relationship was found between patellar height and degree of flexion. A previous study has recommended that optimal patellar height should be between 10 mm and 30 mm (6). Our study supports that recommendation, but shows that the greater the elevation, the better the ultimate flexion. The correlation is strongest between 10 mm and 24 mm . In other words, as patellar height increases up to 24 mm, flexion improves. No further correlation was seen for elevations from 24 mm to 30 mm. Therefore. it seems that for the best final flexion. a range of 2430 rnm for patellar height is preferable. No correlation was found with patellar height and final HSS score. Manipulations are not always performed within the first month after surgery and. from the findings of our study . no difference could be found between the early (21 days) and intermediate (22-90 days)
Daluga et al.
127
groups . However, overall flexion and HSS scores are lower in knees manipulated beyond 3 postoperative months. This is not surprising, since 3 postoperative months seems to be an important period for knee evaluations. At 3 months, knee flexion and HSS scores both closely correlate with the final HSS score . If a knee is still doing poorly 3 months after surgery, the overall prognosis must be guarded. Heterotopic bone is commonly found after total hip arthroplasty, especially in hypertrophic osteoarthritic patients (L 2, 4, 13). It is also increased in patients with elbow trauma when combined with passive range of motion (3,8). Reports of heterotopic ossification have ranged from an incidence of 1042% following total knee arthroplasty (5, 12). It was not found to be progressive nor did it affect the final outcome in these studies. Predisposing factors included patients with posttraumatic arthritis, a preoperative combined deformity of greater than 15°, and manipulation in the postoperative period (5). We found manipulated knees to be significantly more predisposed to heterotopic ossification , but no relationship to preoperative alignment was found. Even though a wide range of heterotopic bone was found , no classification was attempted since no correlation between the amount of heterotopic ossification and final flexion or final HSS score could be found. In fact, patients with heterotopic ossification actually had an improved overall flexion as compared to the manipulated group as a whole. The development of heterotopic ossification after manipulation lends credence to the likelihood that scarring of the quadriceps is a significant factor in restricted postoperative motion. Michelsson et al. (14) have shown that immobilization and forceful manipulation of the knee lead to heterotopic bone formation secondary to microscopic tearing of the quadriceps and subsequent hematoma formation. In another study using a rabbit model, Michclsson and Rauschning found that bleeding and subsequent heterotopic ossification were localized mainly under the fascia in the vastus intermedius (15) . The heterotopic ossification in our study was also found under the distal portion of the vastus intermedius. In knees with a slow recovery rate, adhesions are formed. Once formed . no further significant gains in range of motion can be realistically hoped for, and manipulation is inevnable. Many physical therapists can easily identify this patient as one with a "hard end feel." These patients will require manipulation and, although there are no significant differences found between early and intermediate manipulations. once a hard end point is realized, the knee should be manipulated as soon as possible so that the patient can quickly achieve a functional range of motion and not
128
The Journal of Arthroplasty Vol. 6 No.2 June 1991
become discouraged with fruitless efforts during physical therapy. It is possible that the true value of continuous passive motion will be in preventing these adhesions from forming and ultimately decreasing the manipulation rate. Vince et al. (22) , in a prospective study, found manipulations unnecessary in knees treated w ith aggressive continuous passive motion. Continuous passive motion was used as a physical thcrapy modality in our study but initially only at a low arc. Increase of the AP dimension of the knee by 12% or more is a critical independent variable that significantly predisposes the patient to manipulation and may increase the likelihood of failure of the implant. Quadriceps adhesions and scarring are other major factors that predispose to knee stiffness. Manipulation is necessary to break these adhesions if an adequate range of motion is to be achieved . The microtrauma from the breakage of these adhesions results in a significant percentage of knees with heterotopic ossification. A period of 3 months was found to be a prognostically significant time for evaluation since, at 3 postoperative months, knee flexion and HSS score correlated positively with the final HSS score, and knees manipulated after 3 months did not do as well . Heterotopic os sification affected neither the final range of motion nor the final HSS scores and was, therefore, regarded as an incidental finding in thi s stu dy.
References 1. Cella JP, Salvati EA, Sulco TP: Indomethacin for the
2.
3.
4.
5.
6.
prevention of heterotopic ossification following total hip arthroplasty: effectiveness. contra indications and adverse effects. J Arthropla sty 3:229, 1988 Delee J, Ferrari A, Charnley J : Ectopic bone formation following low friction arthroplasty of the hip. Clin Orthop 121:53, 1976 Delee JC, Green DP, Wilkins KE: Fractures and dislocations of the elbow . p. 613. In Rockwood CA Jr. Green DP (eds): Fractures in adult s. 2nd ed. JB Lippincott Co, Philadelphi a, 1984 Eftekhar NS, Kiernan HA. Stinchfield FE: Systemic and local complications following low -friction arthro plasty of the hip. Arch Surg I I I : ISO, 1976 Figgie III HE, Goldberg VM, Heiple KG et al: The incidence and significance of heterotopic ossification following total knee anhroplasty. Adv Orthop Surg 10: 12, 1986 Figgie III HE, Goldberg VM, Heiple KG et al: The influence of tibial-patellofemorallocation on function
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18,
19.
20.
21.
22.
of the knee in patients with the posterior stabilized condylar knee prosthesis. J Bone Joint Surg 68A:1035, 1986 Fox .n, Poss R: The role of man ipulation following total knee replacement. J Bone Joint Surg 63A:357, 1981 Hotchkiss RN: Elbow and forearm : trauma in adult and pediatric patients . p. 223. In Fitzgerald RHJr (ed] : Orthopaedic knowledge update 2. American Academy of Orthopaedic Surgery, Park Ridge, II., 1987 Insall J, Scott WN, Ranawat CS: The total condylar knee prosthesis: a report of two hundred and twenty cases. J Bone Joint Surg 61A:173, 1979 Ivey M: Myositis ossificans of the thigh following manipulation of the knee : a case report . Clin Orthop 198: 102, 1985 Kettlekarnp DB: Gait characteristics of the knee: normal. abnormal and postreconstruction. p. 47. In American Academy of Orthopaedic Surgeons: Symposium on Reconstructive Surgery of the Knee. CV Mosby, St. Louis, 1976 Lovelock JE, Griffiths HJ, Silverstein AM et al: Complications of total knee replacement. Am J Roentgenol 142:985, 1984 Matos M, Amstutz HC, Finerman G: Myositis ossificans following total hip replacement. J Bone Joint Surg 57A:137, 1975 Michelsson JE, Granroth G, Andersson LC: Myositis ossificans following forcible manipulation of the leg: a rabbit model for the study of heterotopic bone formation. J Bone Joint Surg 62A:81 I. 1980 Michelsson JE, Rauschning W: Pathogenesis of experimental heterotopic bone formation following temporary forcible exercising of immobilized limbs. Clin Orthop 176:265, 1983 Rand JA, Gustilo RB: Technique of patellar resurfacing in total knee arthroplasty. Techniques Orthop 3:57, 1988 Schurman OJ, Parker IN, Ornstein 0: Total condylar knee replacement: a study offactors influencing range of motion as late as two years after arthroplasty. J Bone Joint Surg 67A:lO06, 1985 Scott RD: Treatment of patellar instability associated with total knee replacement. Techniques Orthop 3:9, 1988 Shoji H, Yoshino S, Komagamine M: Improved range of motion with the Y/S total knee arthroplasty system. Clin Orthop 2 I8: 150, 1987 Soudry M, Binazzi R, Insall IN et al: Successive bilateral total knee replacement. J Bone Joint Surg 67A:573, 1985 U.S. Department of Health and Human Services: Anthropometric reference data and prevalence of overweight. Vital and Health Statistics. DHHS Publication No. (PHS) 87·1688 Vince KG, Kel1y MA, Beck J et al: Continuous passive motion after total knee arthroplasty. J Arthroplasty 2:281, 1987
