Aim, study design, participants, and setting
The present study primarily aimed to clarify the efficacy of preoperative LST program as prehabilitation for early QW among patients with TKA using propensity score matching analysis. We hypothesized that the LST group would have lower knee and thigh swelling and knee pain, resulting in lower QW and walking disability immediately after TKA, compared to the non-LST group (control group).
This single-center retrospective observational study evaluated the efficacy of preoperative LST program as prehabilitation for early QW after TKA at an orthopedic clinic in Japan. This study used data from participants who met the inclusion and exclusion criteria. The inclusion criteria were patients with knee osteoarthritis (Kellgren-Lawrence grade 3 or 4) who were scheduled to undergo unilateral TKA between August 2015 and January 2017. The exclusion criteria were those with missing outcome data (e.g., muscle strength and walking ability) due to their inability to perform tests. Participants were required to visit the orthopedic clinic three months and one month before surgery to undergo preoperative examination for TKA. Only those who were able to visit the orthopedic clinic regularly until surgery were prescribed prehabilitation once a week from three months before surgery. Outcome measurements among participants who underwent prehabilitation and those who did not undergo prehabilitation were evaluated approximately one week before surgery and approximately one month before surgery, respectively, and again on postoperative day 4. Data were collected by four physiotherapists with extensive training in performing measurements. Perioperative care was similar to that described in a previous study .
Individual prehabilitation was provided to participants who were divided into two treatment categories—by five physiotherapists at an orthopedic clinic. Participants in category 1 were those who could complete LST with light or moderate pain during exercise. Category 1 included LST and aerobic exercise with a cycle ergometer (heart rate < 120 beats per minute, 15–20 min). The LST program was based on some earlier programs [22, 23]. The types of resistance exercise in LST were squats, forward lunges, and bilateral knee extensions in a seated position. Squats and forward lunges were performed using bodyweight as the load. Bilateral knee extensions were performed at 30% of maximum isometric voluntary contraction with a bilateral isotonic knee extension machine (WT-L02; Minato Medical Science Co., Ltd., Osaka, Japan). Maximum isometric voluntary contraction was measured using the same bilateral knee extension machine at a knee angle of 90° (0° = full extension). The three types of LST resistance exercises were performed as 3 sets of 10 repetitions with slow movement and tonic force generation (3-s eccentric, 5-s isometric, and 3-s concentric actions without rest between repetitions). The rest period between resistance exercise items and between sets was 60 s.
Participants in category 2 were those unable to perform LST due to severe pain during exercise. Category 2 included general and light exercises, such as low-intensity knee extensions and knee range of motion exercises on a therapy couch bed (20 min). Participants in both categories received thermotherapy (hot pack, 10 min) and electrotherapy (Interferential Current Equipment, 10 min).
Pre- and postoperative evaluation comprised measurements of knee circumference at 1 cm and 10 cm proximal to the upper edge of the patella, knee pain during the quadriceps strength test (QST) as well as timed up and go test (TUG), quadriceps strength, and the TUG. Thigh volume was calculated using the truncated cone method through the following formula: V = 1/3πh (a2 + ab + b2), where V represents thigh volume, h represents the distance between two points in the knee circumference (i.e., 9 cm), a represents knee circumference (1 cm), and b represents knee circumference (10 cm). This method has shown excellent criterion-related validity and intra-rater reliability, as well as good inter-rater reliability . Stair climb test (SCT) and Japanese Knee Osteoarthritis Measure (JKOM) scores , an index of disease-specific and patient-derived quality of life, were measured only before surgery. Relative changes (%) in knee circumference (1 cm and 10 cm), thigh volume, quadriceps strength, and the TUG were calculated using the formula [(postoperative value − preoperative value) / preoperative value × 100] to determine knee swelling (1 cm and 10 cm), thigh swelling, QW, and ΔTUG, respectively. Given that knee pain involved 0, the absolute change was calculated using the formula (postoperative value − preoperative value) to determine Δknee pain. Preoperative characteristics of the participants and tourniquet time were obtained from their medical records. The primary outcome was QW. Secondary outcomes were knee and thigh swelling, Δknee pain, and ΔTUG.
Knee circumference among participants relaxed in the supine position with knees extended was measured at 1 and 10 cm proximal to the upper edge of the patella using a non-stretchable tape measure. Two measurements were performed with mean of the two recordings being used for analysis. Circumference measurements using a tape measure have been shown to have excellent intra-rater reliability and good inter-rater reliability [27, 28].
Quadriceps strength test
Quadriceps strength was measured using the maximum voluntary isometric contraction via a pull-type hand-held dynamometer (Mobie; Sakai Medical Co., Ltd., Tokyo, Japan) as previously described . Participants were tested in a seated position with a hip angle of approximately 90° and a knee angle of 75° (0° = full extension) while gripping both sides of the couch. Each participant performed two warm-up trials followed by three maximal contractions with a 1-min rest interval. The highest measurement of two valid trials was used for analysis. Subsequently, quadriceps strength was expressed as the maximum voluntary torque per kg body mass using the external lever arm length and body mass of each participant (Nm/kg). Similar quadriceps strength measurements have been reported to have excellent intra-rater reliability and good inter-rater reliability .
Performance test (timed up and go test and stair climb test)
The TUG required participants to rise from a chair (height: 45 cm), walk 3 m, turn around, walk back to the chair, and sit down. Meanwhile, the SCT required participants to ascend and descend a set of 12 steps (high: 18 cm). Participants were allowed to use a T-handle cane and/or a single handrail during both tests if necessary. Both tests were timed twice using a stopwatch, with their mean of the two recordings being used for analysis. Both tests have been found to have excellent reliability .
Visual analog scale test
Pain in and around the knee immediately after the QST and TUG were evaluated using the visual analog scale (VAS), which is presented as a 100-mm line anchored by verbal descriptors (usually ‘no pain’ and ‘worst imaginable pain’). Each participant was asked to make a mark on the 100-mm line that would indicate pain intensity. Participants’ scores were determined by measuring the distance from the zero anchor (“no pain”) to the participant’s mark. The strongest pain intensity during the QST and TUG was used for analysis. The VAS test has been reported to have excellent reliability .
Sample size calculation was performed with QW as the primary outcome. Because a minimal clinically important difference of early QW after TKA had not been clarified, a moderate effect size (0.6) was used. Using a significance level of 5% and a power level of 80%, we calculated that a minimum of 45 participants were required for each group.
Statistical analysis was conducted using the IBM SPSS version 26 statistical software package (IBM Corp., Armonk, N.Y., USA). Participants were divided into the LST group and control group. The LST group included participants who performed category 1 sessions (LST and aerobic exercise) more than seven times for three months prior to surgery. The control group included participants who performed less than eight category 1 sessions, category 2 sessions (a general and light exercise) or had no prehabilitation (no exercise) for three months prior to surgery. Propensity score matching was used to balance group characteristics that could affect the LST program’s instructions and formulae. Propensity scores were estimated using a logistic regression model where treatment status was regressed on age, gender, body mass index, and preoperative measurements, including quadriceps strength of the affected leg, knee pain during the QST and TUG, the TUG, the SCT, and JKOM scores. Propensity scores were subsequently used to match participants on a one-to-one basis using the nearest-neighbor method without replacement and a caliper width of 0.2 standard deviations of the logit of the propensity score. Between-group differences in preoperative characteristics and tourniquet time were analyzed before and after matching using standardized mean difference (SMD), with a value exceeding 0.1 indicating a meaningful imbalance.
Measurement changes in both groups and the use of a T-handle cane during the postoperative TUG were then compared using the independent samples t-test for normally distributed continuous data and Mann-Whitney U test for non-normally distributed data. Effect sizes (Cohen’s d) for measurement changes were calculated using the online software available at the following website: https://www.psychometrica.de/effect_size.html.
Furthermore, univariate and multivariable analyses were used to examine the effect of knee swelling, thigh swelling, and Δknee pain on QW and the effect of QW on ΔTUG, respectively. Firstly, correlations between measurement changes were quantified by using Spearman rank correlation coefficients given that all changes were non-normally distributed data except for QW. Thereafter, factors determined to be significantly related to QW and ΔTUG on univariate analysis were entered into multivariable regression models (Enter method) for QW and ΔTUG. To adjust for possible confounding factors, the regression model for QW and ΔTUG included age and tourniquet time (known risk factors for IR injury) , and age and gender (known covariates of interest) , respectively. Data were assessed for multicollinearity using correlation coefficients and variance inflation factor. A variance inflation factor > 10 indicated multicollinearity.