Adherence to Rehabilitation After Anterior Cruciate Ligament Reconstructive Surgery : Implications for Outcome

Objective: To investigate the relationship between adherence to rehabilitation and outcome after reconstructive surgery of the anterior cruciate ligament (ACL). Design: A prospective cohort study with adherence to rehabilitation evaluated over 8 weeks correlated with outcomes at 9 and 12 months postsurgery. Participants: 68 patients who had undergone ACL-reconstructive surgery. Main Outcome Measures: Adherence was measured to and during appointments and by a self-report diary of home exercise. Outcomes were measured by 6 knee-function scales and 2 hop tests. Results: There was a signifi cant relationship between home-exercise adherence and many outcomes for participants under 30 years of age (r s = .33– .44). For participants age 30 and over there was a negative relationship between home-exercise adherence and outcome. There were no signifi cant relationships between adherence to and during physical therapy appointments and outcome after ACL-reconstructive surgery. Conclusion: Participants under 30 years of age who adhered to their home-exercise regimen had better functional outcome, whereas adherent participants age 30 and over experienced worse outcome with better home-exercise adherence.

Rupture of the anterior cruciate ligament (ACL) is a common and debilitating injury that is considered the most costly injury in sport. 1,2Surgical reconstruction of the ACL is generally indicated in patients involved in sports that require cutting and pivoting.Patients with very active lifestyles or physical occupations also tend to opt for surgical intervention over conservative management. 3ehabilitation after ACL reconstruction has in recent years changed from a traditional, conservative approach with greater restrictions on activity to a more accelerated approach promoting early mobilization and return to activity. 4The change from a traditional to an accelerated approach appears to be based on anecdotal evidence that nonadherent patients were returning to function sooner than patients adhering to the traditional regime.Shelbourne and Nitz 5 observed that patients who started weight bearing earlier or increased their range of knee motion

Participants
Ethical approval was sought and granted by the Faculty Human Ethics Committee, La Trobe University.Participants were recruited and provided informed consent before or just after ACL-reconstructive surgery.Surgeons provided the primary investigator with information regarding potential participants on a weekly basis.Patients meeting selection criteria were invited to participate.Participants included in the study were 68 patients (42 men, 26 women) of 4 orthopedic surgeons.All patients had undergone ACL-reconstructive surgery using either an autogenous hamstring tendon or an autogenous bone-patellar tendon-bone graft within 12 months of their ACL injury.Participants were included if they were attending one of the 7 physical therapy clinics involved in the study for follow-up rehabilitation.The age of participants ranged from 16 to 52 years, with a mean of 28.8 ± 8.3 years.The mean time between the injury and surgery was 4.1 ± 4.5 months.Most participants were not married (68%), were involved in competitive sport before injury (63%), and had injured their ACL while participating in sport (97%).Participants were about equally likely to have a manual (53%) or sedentary occupation (47%) and to have a patellar-tendon (46%) or hamstring-tendon graft (54%).
To reduce the likelihood of a poor outcome for reasons other than those related to rehabilitation adherence, patients were excluded if they (1) displayed collateral ligamentous laxity greater than grade II, as judged by the orthopedic surgeon at the time of surgery; (2) required surgery for repair of the posterior cruciate ligament; (3) displayed chondral lesions with exposed subchondral bone or radiographic or arthroscopic evidence of osteoarthritis; or (4) had a prior ACL injury to either knee.
Of 76 potential participants, 8 people were excluded, 5 because they did not attend a participating physical therapy clinic for rehabilitation, 1 because chondral lesions were discovered during surgery, 1 who was unable to attend assessment sessions, and 1 who declined the invitation to participate, citing lack of desire to be in the study, leaving 68 participants in the study.

Adherence
Three facets of adherence were measured: adherence to appointments, adherence during appointments, and adherence to home-exercise prescription.Adherence to appointments was measured by calculating the percentage of appointments attended of those scheduled.This method of adherence measurement is simple and objective. 10articipant adherence during each appointment was rated by the treating physiotherapists using the Sport Injury Rehabilitation Adherence Scale (SIRAS). 11he SIRAS is a 5-point scale comprising 3 indicators of in-clinic adherence: the intensity with which the partipicant completed exercises during treatment (1 = minimum effort, 5 = maximum effort), the frequency with which the participant followed instructions (1 = never, 5 = always), and the degree to which the participant was receptive to progressions or changes in the rehabilitation program (1 = very unreceptive, 5 = very receptive).The 3 items were summed to give a score out of 15.The SIRAS has been reported to possess adequate levels of test-retest reliability, interrater reliability, internal consistency, 12 construct validity, and interrater agreement. 13dherence to home-based exercise was assessed using a self-report (diary) method.Participants were required to record the number of exercise sessions and the number of different exercises prescribed by their physiotherapist and the number of sessions and exercises performed each day.The percentage adherence to the prescribed exercise regimen was calculated by dividing the number of exercises performed by the number of exercises prescribed and multiplying by 100.Prescribed exercises as recorded by the participants were cross-checked with the physiotherapist's record of exercise prescription.
To reduce possible inaccuracy resulting from reliance on memory, participants were instructed to complete the diary daily and return the relevant section of it on a weekly basis.5][16][17] In an attempt to reduce any bias in reporting in this investigation, participants were advised that accurate completion of the log would add to the worth of the investigation and were assured that their physiotherapist and surgeon would not be privy to their home-exercise-completion information.

Outcome Measures
Clinical Evaluation.The International Knee Documentation Committee (IKDC) Clinical Examination Form was used to evaluate knee outcome in terms of impairment.The IKDC system has recently been divided into a subjective section and a clinical-examination section. 18The clinical-examination section evaluates knee-ligament injuries by measuring effusion, range of motion, and ligament laxity and categorizing each impairment into 1 of 4 categories: normal, nearly normal, abnormal, and severely abnormal.The worst rating of any impairment serves as the fi nal overall evaluation of knee outcome.To date, all psychometric evaluations of the IKDC have been performed on the original IKDC system. 19,20estionnaires.Although there are a variety of knee-rating scales reported throughout the literature to examine the results of ACL-reconstructive surgery, there appears to be no single optimal method of reporting ACL-surgery outcomes. 19,21ifferent scales are proposed to be weighted toward different outcome features and produce quite different ratings for the same person. 22The absence of a gold standard led to 3 questionnaires being used in this research project to examine outcome.First, the new subjective section of the IKDC was used.This scale examines symptoms of pain, swelling, "giving way" during activity, and the impact of the knee's condition on the functional capacity of the respondent. 18Second, 4 scales from the Cincinnati Knee Rating System (CKRS) were used to examine various aspects of knee outcome: the symptom-rating scale, the change-in-sports-activity scale, the activities-of-daily-living (ADL) scale, and the function-sports scale.These scales allow assessment and scoring of symptoms at different levels of activity, changes in sporting activity since injury, and diffi culties with ADL and sport participation. 19hird, the Knee-Injury and Osteoarthritis Outcome Score (KOOS) was also used to assess subjective knee outcome. 23This 42-item self-administered questionnaire is made up of 5 sections labeled pain, symptoms, ADL, sport and recreation function, and knee-related quality of life. 23tivity.The 6-m timed hop test 24 was one of the 2 hop tests used to examine functional outcome.Participants were timed hopping over a distance of 6 m.A limb-symmetry index was calculated by dividing the mean time (2 trials) in seconds of the uninvolved limb by the mean time of the involved limb and multiplying by 100.
The triple-cross-over hop test 25 was also used to examine knee function.Participants were instructed to hop 3 consecutive times on 1 foot, crossing over a 15-cmwide strip on each hop.The total distance hopped was measured.A limb-symmetry index was calculated by dividing the mean distance (2 trials) in centimeters of the involved limb by the mean distance of the uninvolved limb and multiplying by 100.Both the 6-m hop test and triple-cross-over hop test have been shown to be reliable and have demonstrated evidence of construct validity. 24

Procedure
At the time of recruitment participants were provided with a folder containing details of the study, 12 postage-paid reply envelopes, and a 12-week exercise logbook.Participants were instructed to keep a daily log of home-based exercise completion and return the appropriate home-exercise sheet on a weekly basis for 12 weeks after surgery.
All participants undertook a standardized rehabilitation program based on the similar protocols of the 4 participating orthopedic surgeons.The protocols emphasized some of the key principles of an accelerated program, such as immediate weight bearing as tolerated, early restoration of full knee extension, and early closed kinetic chain exercises, with the overall aim of returning to activities such as sport after 6-9 months.Although the key principles of the accelerated program were followed, specifi c details of each program were left to the clinical judgment of the treating physical therapist.Table 1 presents a summary of the goals of the rehabilitation protocols used.
Adherence to appointments and therapist ratings of adherence during appointments were measured at every appointment for 12 weeks after surgery.At 9 and 12 months after surgery participants were reviewed and all outcome measurements taken.

Statistical Analysis
Adherence Calculations.
Home-exercise adherence was the average homeexercise completion over the fi rst 8 weeks of rehabilitation.The decision to use the fi rst 8 weeks of diary entries, rather than the fi rst 12 weeks as planned, was based on the higher return rate of diaries at the 8-week mark (96%) than at the 12-week mark (59%).It appeared that many participants only attended regular physical therapy and followed a prescribed program for the fi rst 8 weeks after surgery and began sport-specifi c training after that time, making the exercise diaries diffi cult to complete after 8 weeks.

Relationships Between Adherence and Outcome.
To address the primary aim of exploring the relationship between adherence and outcome, Spearman correlation coeffi cients were applied to explore trends in the data.After checking for assumptions of ratio of cases to variables, multicollinearity, singularity, normality, linearity, homoscedasticity, and independence of residuals, 26 we performed multivariate analyses to examine the contribution of the 3 adherence variables (attendance, SIRAS, home exercise) to the outcome measures (KOOS, 4 Cincinnati scales, IKDC subjective, IKDC objective, 6-m hop, triple-cross-over hop).The standardized regression coeffi cients calculated using multiple regression were used to interpret how much weight each variable contributed to the dependent variable, 27 and the signifi cance of each coeffi cient was analyzed using a t test.The proportion of the total variance in each dependent variable that could be explained by the combined independent variable was also analyzed in terms of the squared multiple correlation coeffi cient (R 2 ).The individual importance of each of the adherence variables was examined by calculating the squared semipartial correlations (Sr 2 ).This represents the decrement in R 2 that would result from the elimination of each adherence variable from the regression model.

Relationships Between Other Factors and Outcome.
In the event of a nonsignifi cant relationship between adherence and outcome, the data were explored to identify potentially important variables infl uencing outcome.The following factors were investigated for their infl uence on outcome: age, gender, number of physiotherapy appointments, surgeon, physical therapy clinic, graft type, job type (manual or sedentary), marital status (married/de facto or single), preinjury sporting level, and meniscal damage.Univariate analysis was used for dichotomous variables, and bivariate analysis was used for continuous variables.Any variables found to infl uence outcome were used to partition the data, and the relationship between adherence and outcome was reexamined using correlation and multiple regression.

Results
Attendance, SIRAS scores, and home-exercise data can be viewed in Table 2.No signifi cant relationships with Spearman correlation coeffi cient were found between the outcome measures and adherence at 9 months, and only 1 weak and inverse relationship emerged between attendance scores and IKDC clinical examination at 12 months (r s = -.31).In addition, none of the multiple-regression equations predicting outcome were signifi cant at 9 and 12 months for the combined cohort of 68 participants, with all R 2 values equal to or less than .10(Table 3).This suggests that prediction of outcome based on the values of the 3 measures of adherence was no better than prediction based on chance.The Spearman correlation coeffi cient and multiple-regression examination of the infl uence of adherence on outcome did not demonstrate a relationship until the participants were separated into 2 groups  based on age.Even so, this relationship was present only between home-exercise adherence and outcome.There were no signifi cant associations between adherence to and during appointments and outcome when the sample was separated by age.The age partition was performed because there was a signifi cant difference in outcomes in participants age 30 years and over and those under age 30.Participants under 30 scored better on all outcome measures at 9 and 12 months.This was signifi cant for KOOS scores, Cincinnati symptoms, IKDC subjective scores, 6-m hop test, and the triple-cross-over hop test at 9 months and Cincinnati symptoms scores and the 6-m hop-test scores at 12 months (Table 4).For other factors explored there was either no relationship with outcome (gender, graft type, surgeon, physical therapy clinic) or there were very few relationships and they were inconsistent in direction at the 12-month follow-up (job type, marital status, preinjury sporting level, meniscal damage).
Adherence to home exercise was negatively correlated with all outcome measures except 1 in participants age 30 years and over at 9 and 12 months.For participants under age 30 home-exercise adherence was positively correlated with most outcomes at both 9 and 12 months (Table 5).When participants under and over age 30 were combined, there were no apparent or signifi cant relationships between adherence and outcome (Table 5).
Multiple regression further supported home-exercise adherence as a positive predictor of outcome in participants under 30 years of age and a negative predictor of outcome in those over 30.Separate multiple-regression equations were calculated using the 3 adherence measures as independent variables and each continuous outcome measure at 9 and 12 months as the dependent variable.Three equations were signifi cant in the under-30 group at 12 months.The equations predicting Cincinnati symptoms scores (F 3,32 = 4.67, P < .01),Cincinnati ADL scores (F 3,32 = 2.92, P < .05),and Cincinnati sport scores (F 3,32 = 4.08, P < .05)were signifi cant, and home exercise was the sole signifi cant predictor of outcome scores in each equation (Table 6).Altogether, 30% (R 2 = .30) of the variance in Cincinnati symptom scores, 22% of the variance in Cincinnati ADL scores, and 28% of the variance in Cincinnati sport scores could be predicted by knowing scores on the 3 adherence measures.The squared semipartial correlation (Sr 2 ) shows that if home-exercise adherence were removed from the equation the predicted variance would be decreased by 15% for symptom scores, 19% for ADL scores, and 21% for sport scores.In the 30-years-and-over group there was only 1 signifi cant equation when adherence was regressed onto outcome.The equation predicting Cincinnati sport scores at 9 months (F 3,19 = 4.06, P < .05)was signifi cant.Home-exercise adherence was the only signifi cant predictor of Cincinnati sport scores (β = -0.60)and was a negative predictor, suggesting that greater home-exercise adherence was predictive of poorer scores on the scale.

Comments
The fi ndings of the present investigation suggest that for participants under 30 years of age, adhering to a home-exercise program in the fi rst 8 weeks after ACL reconstruction was correlated with improved outcome at 9 and 12 months postsurgery.Conversely, for participants age 30 years and over, adhering to a home-exercise program after ACL-reconstructive surgery was somewhat detrimental to outcome.Adherence in terms of attendance at rehabilitation and adherence during appointments (as measured by the SIRAS) did not predict outcome scores.
9][30][31] These 4 investigations noted no signifi cant difference in outcomes between participants over age 40 and those under age 40.The contrasting fi ndings between these studies and the present investigation can potentially be explained by the differing methods of measuring outcome.The outcome measures of choice in the 4 projects referenced included the Lysholm knee score (a subjective evaluation of impairment and activity), the IKDC clinical examination (effusion, range of motion, laxity, single hop for distance), and radiographic outcome.Such measures might not adequately discriminate outcomes to identify differences between age groups.
The Lysholm knee score has been identifi ed in the literature as being more specifi c to activities of daily living than sporting activities 22 and therefore might not identify problems during more strenuous activities, resulting in most participants scoring high on the measure despite signifi cant problems.Measures of impairment (IKDC clinical) did not demonstrate differences between age groups in the present investigation, so this fi nding is consistent with the lack of any signifi cant difference found between age groups using the IKDC clinical examination in the other investigations.The signifi cance of radiographic outcome after ACL reconstruction is not known.It is plausible that previous investigations have not identifi ed a difference in outcomes between age groups because of the use of outcome measures that cannot discriminate at this stage of rehabilitation.
Although previous investigations [28][29][30][31] have not identifi ed a difference in outcome after ACL-reconstructive surgery in older and younger participants, the fi ndings of the present investigation are consistent with age-related changes in tissue and lifestyle.These age-related changes could negatively infl uence the recovery of older participants and could also have infl uenced the initial baseline level of such participants.Although it is feasible that older participants might not fare quite as well as their younger counterparts after ACL-reconstructive surgery, it does not explain why older patients who adhere to rehabilitation experience a worse outcome than older patients who do not adhere.
The fi ndings of the current investigation agree with some of the fi ndings of a similar investigation reported by Brewer et al, 7 who examined the adherence and outcome of 95 patients (mean age = 27 ± 8.2) attending rehabilitation after ACL reconstruction.Similar indices of adherence were used, although the method of measuring home-exercise completion was different, and adherence to home cryotherapy (ice) was also measured.Outcome measures differed from the present investigation, with 3 outcome measures being used: the KT1000 (laxity measure), the 1-leg hop test, and the Lysholm scale.Results showed no association between home-exercise adherence and any of the outcome measures.This fi nding is comparable to the fi ndings of the present investigation with the sample combined (ie, all ages included in the calculations).
Rehabilitation protocols place a high demand on patients, with particular focus on the knee joint.The high exercise demands coupled with the limited recovery time evident in current ACL-rehabilitation protocols could be detrimental to patients over 30 years old.For example, the American College of Sports Medicine (ACSM) 32 guidelines for exercise prescription suggest a recovery time of 48 hours between strengthening-exercise sessions.For strengthening the quadriceps muscle, exercising 3 days per week is recommended. 33Current ACL protocols exceed these recommendations for optimal muscle fi tness and do not take into account the reduced ability of tissues such as cartilage, tendon, and bone to respond to normal loading or sustained exercise with age. 34Based on home-exercise diaries used in this study, adherent participants typically completed knee exercises for more than 10 hours per week over 6 days of every week.
Older athletes might require even more recovery time based on evidence of soft-tissue changes with age. 35A number of age-related changes in soft tissue have been reported in the literature and have been hypothesized to infl uence the recovery capacity of older athletes. 34Different healing rates have been reported in the literature for patients under and over 30 when looking at meniscus healing after ACL reconstruction. 36Tenuta and Arciero 36 found that on arthroscopy, patients over 30 who had previously undergone ACL reconstruction demonstrated decreased meniscal healing rates compared with patients under 30 years of age.Furthermore, the investigators found that patients undertaking a "conservative" rehabilitation program consisting of limited weight bearing and range of motion for the fi rst 6 weeks after surgery demonstrated a higher rate of completely healed meniscal repairs than did participants undertaking an "accelerated" program.The authors concluded that accelerated programs might not be appropriate after ACL reconstruction with simultaneous meniscal repair.Although only 2 participants (3%) in the present study underwent meniscal repair, the fi ndings of the Tenuta and Arciero 36 investigation lend weight to 2 explanations for the fi ndings of the present investigation.First, that there is a difference in the tissue of participants age under and over 30 years that infl uences healing and recovery rates and, second, that the current method of ACL rehabilitation might not be appropriate for all participants.
Although it is important to acknowledge that correlation does not equal causation, the clinical implications of these fi ndings are that rehabilitation after ACL reconstruction for older athletes should be decreased in intensity and spread out over a longer period.In addition, the promotion of home-exercise adherence among younger patients by physical therapists could help improve outcomes in this population.The fi ndings of this study highlight the need for research on the effi cacy of the accelerated rehabilitation approach after ACL reconstruction, particularly for older patients.A randomized study comparing the accelerated protocol with a less intense protocol based on ACSM guidelines in a sample of ACL-reconstructed patients over 30 years would be ideal.This would allow hypothesized changes in outcome to be attributed to the protocol.
A possible limitation of the current investigation was the method of data exploration.Unstructured searching through the data in an attempt to identify confounding variables could be considered a poor analysis technique.Considering the exploratory nature of the study, however, this method allowed the identifi cation of an important confounding variable that could be further examined in future research.
The use of self-report diaries to monitor home-exercise adherence could also be considered a potential limitation of this study.The notion that self-reported adherence is subject to bias in a socially desirable direction 37 was addressed by informing participants that only the researchers would be privy to diary information, and even then information would be coded.To reduce the likelihood of inaccuracy because of poor recall, 37 participants were asked to complete the diary daily and send in the completed section at weekly intervals.This method ensured that at the very worst, diaries would have been completed weekly, a time frame that still depends on short-term recall.Because a wide range of home-exercise adherence was reported (31-107%) and the distribution was within normal limits, this was not regarded as a confounding problem.

Conclusion
The introduction of an accelerated approach to ACL rehabilitation promoting earlier weight bearing and movement and faster progression through rehabilitation milestones has been based on anecdotal evidence, with limited investigation of its effects.One explanation for the results of this study is that current rehabilitation might be too vigorous for older athletes.Current regimens might need to be reviewed and revised using accepted recommendations for exercise prescription.

Table 3 Standard Multiple Regression of Adherence and Outcome at 9 and 12 Months for all Participants*
*R 2 indicates multiple correlation squared; ß, standardized beta coeffi cient; SIRAS, Sport Injury Rehabilitation Adherence Scale; KOOS, Knee-Injury and Osteoarthritis Outcome Score; ADL, activities of daily living; and IKDC, International Knee Documentation Committee.

Table 4 Signifi cant Group Differences on Outcome Using Independent Groups t Test*
*KOOS indicates Knee-Injury and Osteoarthritis Outcome Score, and IKDC, International Knee Documentation Committee.

Table 5 Correlation (r s ) Between Home-Exercise Adherence and Outcome*
*KOOS indicates Knee-Injury and Osteoarthritis Outcome Score; ADL, activities of daily living; and IKDC, International Knee Documentation Committee.†P< .05.