Association of Quality of Life With Moderate-to-Vigorous Physical Activity After Anterior Cruciate Ligament Reconstruction

For this cross-sectional study, we recruited potential participants via orthopaedic surgeons in the University Health System of the University of North Carolina at Chapel Hill, flyers placed at outpatient physical therapy clinics and club soccer fields, and university classes. We assessed descriptive variables and the KOOS in participants during a single visit to the laboratory. We measured their daily steps and time in MVPA using an ActiGraph GT9X Link triaxial accelerometer (ActiGraph, LLC) worn for 7 consecutive days after the laboratory visit. The institutional review board at the University of North Carolina at Chapel Hill approved the study, and all participants provided written informed consent.

Individuals who had undergone unilateral ACLR at least 6 months before data collection, had physician approval to return to unrestricted participation in PA, were between the ages of 18 and 35 years, and had a body mass index ≤35 were eligible for the study. We excluded individuals who were pregnant, had been diagnosed with any form of inflammatory arthritis, needed a multiligament reconstruction, or had a history of ACLR of the involved or contralateral limb. Descriptive data and self-reported concomitant meniscal injury information are shown in Tables 1 and 2. Previous researchers¹⁷  have demonstrated an association (odds ratio = 0.41; 95% CI = 0.26, 0.63) between less self-reported sport-related activity using the Marx Activity Rating Scale and a lower KOOS in an ACLR cohort. Using G*Power (version 3.1.9.2),¹⁸  we determined, based on the effect size (Hedges g = −0.21) from an earlier investigation,¹⁷  that 66 participants would be needed to detect a moderate association (R² = 0.15) between the KOOS and daily steps and MVPA (α = .05, 80% power).

Participants completed the 5 subscales of the KOOS (Pain, Symptoms, Activities of Daily Living [ADL], Sport and Recreation Function [Sports], and Knee-Related Quality of Life [QOL]) via a health care informatics system (REDCap; Vanderbilt University). The KOOS is a valid and reliable assessment of functional status and QOL in individuals post–ACLR.^19,20  The KOOS subscale scores range from 0–100, with 100 representing the best possible score.

We used a definition developed by Englund et al¹⁵  in an ACLR cohort to dichotomize participants into those with (symptomatic) or those without (asymptomatic) clinically meaningful knee-related symptoms. Individuals reporting a KOOS-QOL ≤ 87.5 as well as ≥2 of the other 4 subscores below the cutoff values (KOOS Symptoms ≤ 85.7, KOOS Pain ≤ 86.1, KOOS-ADL ≤ 86.8, KOOS Sports ≤ 85.0) were considered symptomatic.¹⁵  We chose the KOOS-QOL subscore as our primary outcome because we were most interested in how overall knee-related QOL was influenced by PA, and KOOS-QOL was the most responsive subscore.^21–23 

Participants wore an ActiGraph GT9X Link triaxial accelerometer over the right anterior-superior iliac spine during waking hours for 7 consecutive days immediately after the study visit. The ActiGraph GT9X Link is a valid device for quantifying PA in free-living conditions.²⁴  We considered a data-collection period acceptable if it included ≥4 days (3 weekdays and 1 weekend day) of wear for ≥10 hours per day.²⁵  Participants removed their monitors for bathing and sleeping. They returned their monitors after the wear period, at which time we assessed the data for fidelity and consistency with wear guidelines.²⁵  The accelerometer did not specifically account for non–weight-bearing activities such as swimming and bicycling.

We processed average daily steps and time in MVPA (minutes) using ActiLife software (version 6.13.3; ActiGraph, LLC). We evaluated total wear time using recommendations from Choi et al²⁵  (zero-count threshold during a nonwear time interval, 90-minute time window for consecutive zero or nonzero counts, and allowance of a 2-minute interval of nonzero counts with the upstream and downstream 30-minute consecutive zero-counts window for detection of artifactual movements) to ensure that each participant wore the ActiGraph for ≥4 days for ≥10 hours per day. We used Freedson Adult triaxial vector magnitudes cut points to calculate time in MVPA (2690–9642 counts/min).²⁶  Daily steps were assessed using vertical acceleration data measured using the ActiGraph monitor. We averaged variables over valid wear days to create a daily average.¹¹ 

Means and SDs were calculated for all continuous descriptive and outcome variables and frequencies for all noncontinuous variables. Before analysis, we identified outliers via box plots as any data points >3 times the interquartile range and removed them before our primary analyses. We conducted independent t tests for all descriptive and outcome variables between asymptomatic and symptomatic groups.

For our primary analyses, we performed separate multiple regressions with each PA variable (daily steps and MVPA) as a separate predictor variable and KOOS-QOL as the criterion variable. Age was associated with PA,²⁷  and time post–ACLR was associated with the KOOS.¹⁷  Therefore, we included age and time post–ACLR as covariates in each regression before the KOOS. A history of concomitant meniscal injury may also be associated with KOOS outcomes post–ACLR. Concomitant meniscal injury was self-reported, and 6 participants selected unknown when asked about meniscal injury. Thus, we removed these 6 participants and conducted a sensitivity analysis evaluating the associations between the KOOS-QOL and PA variables with 3 covariates: age, time post–ACLR, and history of concomitant meniscal injury.

For our secondary analyses, we calculated separate multiple regressions between each PA variable and the KOOS-QOL as described for the primary aim in each group (ie, asymptomatic or symptomatic) with the inclusion of age and time post–ACLR as covariates. Similarly, we performed a sensitivity analysis to evaluate the associations between the KOOS-QOL and PA variables in each group (asymptomatic and symptomatic) including 3 covariates: age, time post–ACLR, and history of concomitant meniscal injury. For the supplementary analyses, we conducted similar regressions for the other KOOS subscores (KOOS Symptoms, KOOS Pain, KOOS-ADL, and KOOS Sports).

Recent evidence²⁸  suggested that female, but not male, participants post–ACLR were less physically active than uninjured control participants. Therefore, we stratified our cohort by sex and symptoms into 4 subgroups: symptomatic women, asymptomatic women, symptomatic men, and asymptomatic men. We computed 1-way analyses of variance with Bonferroni post hoc comparisons for all descriptive and outcome variables among subgroups before the post hoc analyses. We then conducted separate univariate linear regressions with each PA variable as a separate predictor variable and the KOOS-QOL as the criterion variable. For supplementary analyses, we calculated similar post hoc regressions for the other KOOS subscores (KOOS Symptoms, KOOS Pain, KOOS-ADL, and KOOS Sports). The α level for all analyses was set a priori at ≤.05, and SPSS (version 21.0; IBM Corp) was used for all analyses.

We screened 150 individuals via telephone or email to determine study eligibility (Figure 1) and scheduled 76 participants for a laboratory visit. A total of 66 individuals completed all aspects of the study. For our secondary aim, 36 individuals were classified as asymptomatic and 30 as symptomatic. The symptomatic group demonstrated lower KOOS subscores (t₆₅ range = 5.42–8.80; P ≤ .05) and less time post–ACLR (t₆₅ = 2.22, P = .03) than the asymptomatic group did (Table 1). No differences existed between groups for the other outcome or descriptive variables.

In our post hoc analyses, 19 men were asymptomatic, 11 men were symptomatic, 17 women were asymptomatic, and 19 women were symptomatic. Differences between groups were noted for age (F_3,62 = 3.31, P = .03), height (F_3,62 = 25.11, P < .001), mass (F_3,62 = 14.30, P < .001), time post–ACLR (F_3,62 = 5.71, P < .01), and KOOS subscore (F_3,62 range, 8.11–29.44; P < .001; Table 2).

We did not identify any associations between the KOOS-QOL and daily steps (ΔR² = 0.01, β < 0.01, P = .50) or MVPA (ΔR² = 0.01, β = 0.07, P = .36). No associations were present between the KOOS-QOL and daily steps (ΔR² = 0.01, β < 0.01, P = .32) or MVPA (ΔR² = 0.01, β = 0.09, P = .29) after our sensitivity analysis that accounted for meniscal injury history. Additional R and β values for each regression can be found in Supplemental Table 1 (http://dx.doi.org/10.4085/1062-6050.0670-20.S1).

More time in MVPA was associated with a higher KOOS-QOL in the symptomatic (ΔR² = 0.12, β = 0.21, P = .05) but not the asymptomatic (ΔR² = 0.01, β = −0.05, P = .47; Figure 2) group. We did not identify associations for daily steps and the KOOS-QOL in either the asymptomatic (ΔR² = 0.02, β < −0.01, P = .40) or the symptomatic (ΔR² = 0.08, β < 0.01, P = .12) group. More time in MVPA was associated with a higher KOOS-QOL in the symptomatic (ΔR² = 0.13, β = 0.23, P = .05) but not the asymptomatic (ΔR² = 0.01, β = –0.04, P = .58) group after our sensitivity analysis that accounted for meniscal injury history. No associations were observed for daily steps and the KOOS-QOL in either the asymptomatic (ΔR² < 0.01, β < 0.01, P = .71) or the symptomatic (ΔR² = 0.07, β < 0.01, P = .15) group after our sensitivity analysis. Additional R and β values for each regression can be found in Supplemental Table 1. Results of the supplementary regression analyses conducted for the other KOOS subscales are included in Supplemental Table 2 (http://dx.doi.org/10.4085/1062-6050.0670-20.S2).

More time in MVPA was associated with a higher KOOS-QOL (R² = 0.230, β = 0.251, P = .04) in the symptomatic female group (Table 3). We did not identify other associations between KOOS-QOL and MVPA (R² range = 0.008–0.196; P ≥ .17) or daily steps (R² range = 0.012–0.133; P ≥ .12) in the asymptomatic male, symptomatic male, or asymptomatic female group (Table 3). Results of the supplementary post hoc regression analyses conducted for the other KOOS subscales are provided in Supplemental Table 3 (http://dx.doi.org/10.4085/1062-6050.0670-20.S3).

Contrary to our primary hypothesis, no associations existed between the KOOS-QOL and daily steps and MVPA in the entire cohort. Consistent with our secondary hypothesis, more time in MVPA was associated with better KOOS-QOL in the symptomatic group. When we stratified by sex, an association between more time in MVPA and higher KOOS-QOL persisted in the symptomatic female group. The KOOS-QOL is an important outcome to monitor post–ACLR, because it is associated with better outcomes on clinical tests, such as greater distance on the single-legged hop,²⁹  and in psychological aspects, such as greater self-efficacy for rehabilitation²⁹  and readiness to return to sport.³⁰  However, we are the first to demonstrate an association between more time in MVPA and a better KOOS-QOL in individuals with symptoms post–ACLR. To our knowledge, no PA promotion interventions have been conducted in individuals post–ACLR, yet it is possible that increasing time in MVPA among individuals with symptoms post–ACLR may improve knee function. In a previous study,¹²  individuals with KOA demonstrated increases in the KOOS after an 8-week aerobic exercise intervention. We hypothesize that increasing PA will result in improvement in the KOOS among individuals post–ACLR, similar to those with KOA. Future authors should determine whether increasing daily steps and time in MVPA via PA promotion influences knee function.

More time in MVPA was associated with the KOOS-QOL in the symptomatic group. Daily steps, which were not associated with the KOOS-QOL in our study, reflect total PA volume.³¹  Only 30% of participants (8 in the symptomatic and 12 in the asymptomatic group) met the common threshold of “10 000 daily steps,” suggesting that many individuals with ACLR may need to engage in more daily steps. Although daily steps may still be an important outcome post–ACLR, the KOOS-QOL may be associated with MVPA because MVPA indicates higher-intensity movement.³¹  Individuals with symptoms may not make decisions about light PA on the basis of their perceived QOL; however, they may modify time spent in MVPA to avoid the pain and difficulty with knee function that occurs during higher-intensity activities. Moderate-to-vigorous physical activity is defined as purposeful movement with a metabolic equivalent of task ≥ 3.0, such as brisk walking, weight training, or an activity that causes an elevated heart rate.³¹  In our cohort, 97% of individuals (29 in the symptomatic and 35 in the asymptomatic group) met PA recommendations of 150 min/wk of MVPA for adults aged 18–65 years.³²  Our findings are consistent with those of Bell et al,¹¹  who found that 97% of the ACLR cohort engaged in at least 150 min/wk of MVPA. Return to sport is one of the most commonly discussed clinical goals for patients post–ACLR;⁹  yet little is known about engagement in MVPA post–ACLR. Despite most participants engaging in 150 min/wk of MVPA,¹¹  approximately 56% of patients still did not return to the same level of competitive sport post–ACLR.³  Thus, individualized MVPA targets rather than a uniform target of 150 min/wk may be an important clinical outcome to consider during recovery post–ACLR.

In our study, the asymptomatic group post–ACLR did not demonstrate an association between the KOOS and daily steps and MVPA. The asymptomatic group displayed high scores on all KOOS subscores, which decreased the variability of the KOOS in the asymptomatic group. Bell et al¹¹  noted that knee function measured using the International Knee Documentation Committee score was not associated with daily steps or MVPA in individuals post–ACLR. They did not stratify participants based on symptoms, and most of the cohort exhibited knee function >85%, suggesting that the cohort was mostly asymptomatic, which may explain why no associations were identified.¹¹  Our investigation was specifically powered to stratify the ACLR cohort on the basis of knee symptoms. We hypothesize that individuals with symptoms are more likely than individuals without symptoms to consciously modify activities to avoid knee pain and discomfort and to adapt to limited function. An alternative hypothesis is that individuals without symptoms modified engagement in PA before enrollment in the study. An individual's decision to engage in MVPA is multifactorial, and how individuals with ACLR make decisions regarding MVPA based on perceived knee function remains unknown. Our results reflect the need for future longitudinal research to evaluate engagement in PA and KOOS-QOL throughout different stages of ACLR rehabilitation.

Our study had some limitations that could better inform future work. Our design was cross-sectional, and we could not determine whether MVPA influenced the KOOS-QOL or the KOOS-QOL influenced MVPA. It is possible that the KOOS-QOL and MPVA can be targeted simultaneously in a clinical setting to improve both outcomes during recovery post–ACLR. Our sample was 6 to 146 months post–ACLR. We included time post–ACLR as a covariate in all regression models in the primary and secondary analyses because it is associated with the KOOS-QOL, daily steps, and MVPA. The asymptomatic group demonstrated more time since ACLR (39 ± 40 months) than the symptomatic group (16 ± 15 months). Knee-related symptoms may differ in origin based on time post–ACLR; in particular, individuals who are 6 months post–ACLR may experience persistent joint effusion or anterior knee pain specifically related to the ACL injury or reconstruction surgery during some activities. Future authors should assess patients earlier in the recovery process or stratify participants based on time post–ACLR. In our post hoc analyses, the asymptomatic and symptomatic male groups were older and had greater height and mass than the asymptomatic female group. We did not include these descriptive variables as covariates in our post hoc analyses because the preliminary analyses should be viewed as hypothesis generating rather than hypothesis confirming. Although our results indicated that the associations between time in MVPA and knee function may be influenced by sex and symptoms, future studies involving larger sample sizes should be conducted to also determine how multiple covariates affect these relationships. In addition, our findings suggested the need to longitudinally evaluate the KOOS-QOL and MVPA throughout recovery post–ACLR.

No association existed between the KOOS-QOL and daily steps and MVPA in the entire cohort. When we stratified by symptoms, individuals with symptoms who spent more time in MVPA demonstrated a better KOOS-QOL. When we further stratified by sex, women with symptoms displayed the strongest association between more time in MVPA and a better KOOS-QOL. Patient-reported QOL and MVPA may be outcomes to target via therapeutic interventions in individuals with symptoms post–ACLR, either separately or simultaneously.

This study was funded by a grant from the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health (R21 AR074094; Drs Franz and Pietrosimone).