The standard medial parapatellar approach has consistently yielded excellent and reproducible results in long-term studies of total knee arthroplasty1-6. Despite the success of this approach in providing ample visualization of the operative field to assist in component alignment and placement, concerns remain about patient functional outcomes. Some investigators have reported both short and long-term deficits in quadriceps muscle function following total knee arthroplasty using the medial parapatellar approach7,8. These deficits may be reflected in lower patient satisfaction with this procedure compared with that reported by surgeons9. Additionally, disruption of the blood supply as a result of patellar eversion with this approach has been implicated in worse outcomes10.
Several surgeons have developed alternative surgical approaches in an attempt to overcome the limitations of these standard techniques, with specific emphasis on limiting soft-tissue trauma, especially to the knee extensor mechanism. Collectively, these approaches are referred to as minimally invasive surgery for total knee arthroplasty. Specific benefits that have been ascribed to these techniques include better postoperative quadriceps muscle strength11,12, decreased postoperative pain11,13, shorter hospital stay14, and improved range of motion11,13,15-18. While recent reports of minimally invasive total knee arthroplasty are encouraging, they encompass a number of different surgical approaches and techniques, each with potential advantages and drawbacks, and possible differences in outcomes. Of these newer approaches, two of the most commonly reported are the mini-subvastus and mini-midvastus17,19. While both appear to provide distinct clinical advantages over the traditional medial parapatellar approach, it is not clear whether one provides superior clinical outcomes compared with the other.
The purpose of this study was to compare the clinical and radiographic outcomes, muscle strength, and patient preference between minimally invasive subvastus and midvastus approaches in bilateral total knee arthroplasty with use of a prospective randomized design. Additionally, we sought to determine whether there were any differences in intraoperative factors between the two approaches, specifically for total blood loss and operative time.
Fifty-one patients (one hundred and two knees) scheduled to undergo bilateral total knee arthroplasty during a single anesthetic session were enrolled in this study and were followed for two years postoperatively. The group included forty-two women and nine men, who had a mean age of seventy years (range, fifty-four to eighty-seven years) at the time of the index surgery and a mean body mass index of 31 kg/m2 (range, 19 to 44 kg/m2). All fifty-one patients were diagnosed with osteoarthritis in both knees. All operations were performed by the first author (P.M.B.), from January 2004 to July 2006, with the mini-subvastus approach used in one knee and the mini-midvastus approach used in the contralateral knee (Fig. 1). Baseline characteristics for the enrolled patients can be found in Table I. The operating surgeon had performed >1000 total knee arthroplasties using the mini-midvastus approach and >100 procedures using the mini-subvastus approach prior to the start of this study. The patients were informed that the operations would be done with use of a different muscle incision in each knee, and they consented to this procedure. Institutional review board approval was granted for the study of these patients. Additionally, this study was registered with the U.S. National Institutes of Health ClinicalTrials.gov database as clinical trial number NCT00848133.
Based on a power of 0.9 to detect a significant difference (p < 0.05, two-sided) of 3 points on the 100-point Knee Society pain and function scales, thirty-two knees (thirty-two patients) were required in each study group. The study was designed to show equivalence. To ensure a more than sufficient number of patients for evaluation at the time of final follow-up, we planned to enroll fifty patients. All included patients were approached for participation during regular preoperative clinical consultation at our institution for the treatment of late-stage osteoarthritis of both knees. Ninety-one patients with Kellgren-Lawrence grade-III or IV osteoarthritis bilaterally were assessed for eligibility20. Of those, ten patients were excluded because they were considered to be at higher risk for loss to follow-up as a result of residing outside our institution's home state. A further thirty patients were excluded because of previous knee surgery, concomitant unilateral or bilateral diagnoses that could affect the outcome of one or both of the total knee arthroplasties (such as foot or ankle abnormalities and unilateral radiculopathy), or because they were enrolled in one or more separate studies with protocols that were incompatible with the present study. No patient declined participation in the study. A graphical representation of patient recruitment and analysis can be found in Figure 2.
Patients were evaluated clinically, with determination of the Knee Society pain and function scores, from four to eight weeks prior to surgery (Table II). All patients were randomized in the operating suite as to which approach would be used in which knee, with the left knee always operated on first. The surgical approach for each knee was determined by the circulating nurse, who drew from an opaque container one of two pieces of paper that were identical except that one was labeled "SV" (representing the mini-subvastus approach) and the other was labeled "MV" (representing the mini-midvastus approach). To reduce the potential for manipulation of the choice of approach, both pieces of paper were folded in half with the approach written on the inside. The indicated approach was used in the left knee. The right knee would then be operated on following the left knee as part of the same anesthetic session, with use of the other surgical approach. The distribution of surgical approaches can be found in Table I. Operating time from incision to final closure for each knee was recorded as part of the standard protocol and was later extracted from the anesthesia record for analysis.
Patients were followed daily in the immediate postoperative period until the time of discharge and then were evaluated during outpatient follow-up visits at approximately four weeks, twelve weeks, six months, one year, and two years following the index arthroplasties. Inpatient evaluation consisted of daily evaluation of knee range of motion and the ability to perform a straight leg raise. Outpatient clinical evaluation included determination of knee range of motion and Knee Society pain and function scores. Anteroposterior and lateral radiographs of both knees made at the four-week, six-month, one-year, and subsequent annual follow-up visits were evaluated for progressive radiolucencies and component migration. Postoperative and final follow-up radiographs were evaluated by a single observer to determine the tibiofemoral angle on the anteroposterior radiograph and the tibial component slope on the lateral radiograph. Additionally, at the four-week follow-up visit, patients were asked which knee they were more satisfied with, if either, with the answer recorded in the patient chart as "left knee," "right knee," or "neither knee." At the twelve-week follow-up visit, patients underwent isokinetic quadriceps and hamstring muscle strength testing of both knees with use of a dynamometer (Lido Active II; Chattanooga Group, Hixson, Tennessee), with peak torque values used for quantitative comparison of muscle strength. Dynamometer strength testing was performed by an independent physical therapist, and all clinical evaluations were performed by an independent nursing staff.
Operative time and blood loss were compared between the two surgical approaches. Any clinical or radiographic complications noted during follow-up visits were recorded.
All knees in all of the patients underwent the same postoperative rehabilitation protocol, consisting of four days of postoperative inpatient rehabilitation (or fewer if all goals were met prior to postoperative day 4), followed by at least two to four weeks of outpatient physical therapy. Any patient who underwent a subsequent knee procedure (specifically, manipulation under anesthesia or arthroscopic evaluation for persistent stiffness or pain) underwent additional knee rehabilitation following the procedure.
Three patients died during the follow-up period from causes unrelated to the surgery. Data from any evaluations that were performed prior to their death were included in the interim results in this report but were excluded from our analysis at the time of the final follow-up.
Statistical Methods
All data were collected and recorded in either office or inpatient charts. Statistical analysis was performed, with use of the Student t, paired t, Mann-Whitney, and signed-rank tests, where appropriate. A p value of <0.05 was considered to be the threshold for significance in all comparisons.
Source of Funding
No external funding was provided in support of this study.
There was no significant difference between the groups with respect to either Knee Society pain (p = 0.939) or function (p = 0.313) scores at the time of the final follow-up. A comparison of postoperative clinical findings can be found in Table III. The mean Knee Society pain and function scores for the midvastus group improved from 49 and 49 points, respectively, at the preoperative evaluation to 95 and 94 points at the time of the final follow-up. The mean Knee Society pain and function scores for the subvastus group improved from 49 and 48 points, respectively, at the preoperative evaluation to 96 and 95 points postoperatively. There were no significant differences in Knee Society pain and function scores at the follow-up intervals, apart from a 1-point difference in the mean pain scores at the four-week follow-up evaluation. There was no significant difference in the ability to perform a straight leg raise on postoperative day 1 (p = 0.838), with nineteen subvastus and nineteen midvastus knees achieving an independent straight leg raise at that time. Similarly, an identical number of patients achieved straight leg raises in both groups on postoperative day 3.
There was no significant difference in range of motion at the final inpatient evaluation (p = 0.630), with mean flexion of 95° (range, 85° to 108°) in the subvastus group and mean flexion of 96° (range, 85° to 105°) in the midvastus group. Similarly, there was no difference in range of motion at the time of the final follow-up (p = 0.557), with mean flexion of 114° in both groups (range, 90° to 125° in the subvastus group and 90° to 123° in the midvastus group).
With the numbers studied, no significant differences were found at the time of the final follow-up between men and women with respect to the Knee Society pain (p = 0.443) and function (p = 0.964) scores or ranges of motion (p = 0.532) for the knees that had the subvastus approach. For the knees managed with the midvastus approach, men had significantly lower Knee Society pain scores at the time of final follow-up (p = 0.014) compared with women, but no significant differences were found in Knee Society function scores (p = 0.654) or range of motion (p = 0.469).
In the survey of patient preference at four weeks after the index surgery, no significant difference was found between the two approaches (p = 0.241), with nine patients preferring the knee that had the subvastus approach, fifteen preferring the knee that had the midvastus approach, and the remaining twenty-seven patients having no preference.
Isokinetic strength testing at twelve weeks postoperatively revealed no significant difference in the ratios of peak extensor torque (p = 0.749) or peak flexor torque (p = 0.919) to body weight between the two groups. The mean ratio of peak extensor torque to body weight was 0.15 Nm/kg (range, 0.07 to 0.27 Nm/kg) in the subvastus group and 0.15 Nm/kg (range, 0.07 to 0.24 Nm/kg) in the midvastus group. The mean ratio of peak flexor torque to body weight was 0.14 Nm/kg (range, 0.07 to 0.23 Nm/kg) in the subvastus group and 0.14 Nm/kg (range, 0.05 to 0.24 Nm/kg) in the midvastus group.
Analysis of intraoperative factors revealed no significant difference in total blood loss between the two approaches (p = 0.145). There was a significant difference in operating time between the two approaches (p = 0.028), with a mean of fifty-five minutes (range, forty to eighty-four minutes) for the subvastus group and a mean of fifty-one minutes (range, forty to sixty-eight minutes) for the midvastus group. However, this four-minute difference in mean operative time was not considered clinically important.
Radiographic analysis revealed no patient with progressive radiolucencies or component migration. Four knees in each group had tibial components with a single, incomplete, nonprogressive lucency in one zone. Four knees in the subvastus group had femoral components with a single incomplete, nonprogressive lucency in one zone, and one knee had a femoral component with incomplete, nonprogressive lucencies in two zones. Two knees in the midvastus group had femoral components with a single, incomplete, nonprogressive lucency in one zone. One knee in the midvastus group had a tibiofemoral angle of 1° of varus alignment measured on radiographs made at the time of the final follow-up. This patient later underwent revision for late onset collateral ligament instability. All of the remaining knees had a tibiofemoral angle ranging from 3° to 9° of valgus alignment on standing anteroposterior radiographs, and all tibial components were within 2° of neutral on the lateral radiographs.
Five patients underwent manipulation under anesthesia for a decreased range of motion, with four patients undergoing bilateral manipulations and one patient requiring manipulation of only the knee that had the subvastus approach. Two patients were less compliant with rehabilitation than expected, discontinuing outpatient therapy after two weeks, and a third had a metatarsal fracture soon after the index arthroplasty, limiting participation in rehabilitation. All three of these patients who had lower compliance with rehabilitation were doing well at the time of the final follow-up, with maximum knee flexion ranging from 102° to 122°.
One patient was hospitalized with a respiratory infection unrelated to the arthroplasty surgery and had a range of motion from 0° to 90° bilaterally at the time of the one-year follow-up. Prior to hospitalization, this patient had a range of motion of 0° to 115° bilaterally. This patient died prior to the two-year follow-up evaluation (fourteen months following the index arthroplasty).
Some authors have suggested that the subvastus approach may confer some benefit over other minimally invasive approaches for total arthroplasty of the knee21,22, including a quicker ability to perform a straight leg raise and faster rehabilitation owing to the avoidance of any incision of the extensor mechanism. However, the current study suggests that there are no significant differences in clinical outcomes or muscle strength between minimally invasive subvastus and midvastus approaches for total knee arthroplasty.
The limitations of this study include the relatively short follow-up period of twenty-four months, the modest number of patients enrolled, and the single-surgeon study design. Longer-term follow-up with the participation of a number of surgeons might reveal differences in clinical outcomes or failure rates not appreciated in our analysis. Nevertheless, we believe that this study effectively compares the clinical outcomes of these two approaches and demonstrates that they are equivalent in the short term. An additional limitation is the sex imbalance of the study group, with forty-two women (82% of the patients) and nine men. No sex-specific selection criteria were applied while recruiting for this study because this distribution is similar to the overall ratio of women to men who underwent simultaneous bilateral knee arthroplasty at our institution over the same time period (eighty-two [77%] of the patients were women and twenty-four were men). However, some authors have suggested that the subvastus approach might be less desirable in men23,24, particularly those with well-developed quadriceps muscles. In contrast, Schroer et al. recently reported good outcomes in an unselected patient group that included obese individuals17. In the present study, a comparison of men and women with respect to the Knee Society pain and function scores and ranges of motion for the knees that had the subvastus approach revealed no significant differences; however, we may not have had a sufficient number of study subjects to detect a difference. Additionally, several reports have suggested that women have a greater prevalence of knee osteoarthritis. For these reasons, we believe that the results of this study should apply to the overall population of patients undergoing total knee arthroplasty for late-stage osteoarthritis. Nevertheless, a modified mini-subvastus approach, utilizing a short, proximal extension of the arthrotomy into the insertion of the vastus medialis obliquus muscle on the patella might still be appropriate in a small number of very muscular patients with a distal insertion of the quadriceps muscles on the patella25, irrespective of sex. A final limitation is that no formal quadriceps strength testing was performed preoperatively to evaluate baseline function and confirm differences, if any, between the knees managed with the two surgical approaches. However, because surgical approaches were distributed evenly between both knees in each patient, and because the knees were randomized to the two surgical approaches evaluated in this study, it is unlikely that the results of the present study would be biased by a significant preoperative difference in quadriceps muscle function.
A number of reports have compared the outcomes of minimally invasive approaches for total knee arthroplasty, including several prospective, randomized studies18,21,26-29. Of those, only Kim et al. compared approaches on contralateral knees, specifically evaluating the outcomes of the standard medial parapatellar approach and those of the minimally invasive quadriceps-sparing approach29. Aglietti et al. compared a minimally invasive quadriceps-sparing approach with the mini-subvastus approach, randomizing sixty patients scheduled to undergo unilateral total knee arthroplasty to the two techniques21. They found an earlier achievement of straight leg raising in the subvastus group, but no difference in postoperative pain, range of motion, or component alignment. Chin et al. randomized ninety knees to one of three approaches: the standard parapatellar approach, the minimally invasive midvastus approach, and a capsular approach with use of side-cutting instruments28. However, those authors investigated only radiographic outcomes, finding poorer alignment in the direct lateral group and no difference between the standard and midvastus approach groups.
Several additional comparative studies have evaluated the outcomes of minimally invasive total knee arthroplasty11-17,19,30-35, but they focused on contrasting outcomes from these new techniques against those achieved with use of the standard parapatellar approach. To the best of our knowledge, this investigation is the first prospective, randomized study comparing clinical outcomes of two of the most common minimally invasive techniques: the mini-subvastus and mini-midvastus approaches.
Some authors have suggested advantages in using the subvastus approach, as it completely avoids incising the vastus medialis obliquus muscle21,22,36, as well as preserving the medial retinaculum and patellar blood supply37. This is in contrast to the midvastus approach that requires some splitting of the extensor mechanism. However, the present report demonstrated no clinical benefit to the mini-subvastus approach, despite the theoretical advantages. Notably, the results suggest that a small incision of the extensor mechanism made in-line with the muscle fibers to minimize tissue trauma may have no impact on postoperative extensor muscle function compared with complete avoidance of extensor mechanism disruption, as no significant difference was found in the ratio of extensor muscle peak torque-to-body weight at the time of the twelve-week follow-up.
The outcomes of a number of comparative studies of minimally invasive approaches for total knee arthroplasty may be influenced by the limited experience of the reporting surgeon with the minimally invasive approach under investigation. For example, Kim et al. had performed only twenty-five arthroplasties with a quadriceps-sparing approach prior to the start of their study29, while Aglietti et al. did not report their experience with either of the techniques they studied21. A recent report by several of the present authors suggests that the learning curve for minimally invasive approaches to total knee arthroplasty may be as high as 200 procedures30. If this is correct, some published comparisons may be biased by the relative inexperience of the reporting surgeons with minimally invasive techniques, especially compared with the standard parapatellar approach, with which knee surgeons almost universally have extensive experience. That was not the case in this study.
In conclusion, when the procedures were performed by a knee surgeon experienced with minimally invasive techniques for total knee arthroplasty, there were no significant differences in clinical outcomes between the mini-subvastus and mini-midvastus approaches, as measured by Knee Society pain and function scores, ability to perform straight leg-raising, isokinetic muscle-strength testing, range of motion, or a patient survey of the preferred approach. We await further results to confirm that these results are maintained over a longer follow-up period. Surgeons should choose the approach that they are most comfortable performing, without concern that one approach will lead to worse short-term outcomes for the patient.
Note: The authors thank Lindsay Rudert, Gina Zerrusen, and Joy Marlowe for their assistance with the preparation of this manuscript.