Demographic and surgical characteristics in patients who do not achieve minimal important change in the KOOS Sport/Rec and QoL after ACL reconstruction: a comparative study from the Swedish National Knee Ligament Registry

Study design

This study was conducted as a comparative registry study, where data was analysed cross-sectionally and was reported according to the Reporting of studies Conducted using Observational Routinely collected Data statement,21 which is an extension of the Strengthening the Reporting of Observational Studies in Epidemiology guidelines22 with data from the Swedish National Knee Ligament Registry (SNKLR) during the period January 2005–May 2019.

The SNKLR was established in 2005, and covers over 90% of all ACLRs performed in Sweden every year.23 The SNKLR aims to promote superior treatment for individuals with ACL injuries by providing hospitals and clinics with accurate feedback with regard to surgical technique and outcomes after ACLR, and to determine prognostic factors for successful and poor outcomes.24 The SNKLR has two different sources of information. The primary source is registered by a medical professional, and contains information on demographics, injury-related factors and treatment-specific data for patients undergoing surgical and non-surgical treatment after an ACL injury. The secondary source contains information registered by the patient including PROs,24 including the KOOS5 and EuroQoL-5 Dimension.25 Only the KOOS was used in this study.

Data from both sections are registered in the SNKLR at baseline, which corresponds to the time of surgical treatment. Patients are asked to respond to questionnaires at 1-, 2-, 5- and 10-year follow-ups after ACLR.

Activity at the time of injury is registered by choosing one activity from predefined answer alternatives, which is then registered to the individual patient. The activities at the time of injury presented in this study are grouped based on type of activity, that is, pivoting sports (soccer, handball, basketball, American football, rugby, dancing, floorball, gymnastics, ice hockey, bandy, martial arts, racket sports, volleyball and wrestling), alpine skiing, non-pivoting sports (cross-country skiing, cycling, motocross/enduro, snowboarding, surfing, wakeboarding, horseback riding, skateboarding), other physical activity (other recreational sports), traffic-related injuries and other (outdoor and walking).

Study population

Patients eligible for inclusion were ≥15 years old, had undergone a primary ACLR and had KOOS data for the Sport/Rec and QoL subscales before surgery and at 1-year follow-up. Patients with previous knee surgery, associated fracture to the patella, femur, tibia or fibula, or concomitant nerve, vascular or additional injuries to any other knee ligaments, such as the posterior cruciate ligament, or had a surgically reconstructed medial or lateral collateral ligament (MCL, LCL) tear were excluded from this study. In addition, patients with high KOOS at baseline for Sport/Rec (>87.9) and for QoL (>81.7),26 and thereby not able to achieve the MIC, were excluded from this study. Data regarding patients (age, sex, body mass index (BMI) and activity at the time of injury), and surgical characteristics (graft type (hamstring-, patellar- or quadriceps tendon autograft, allograft) and associated injuries) were obtained from the registry. Included patients were divided into two groups, (1) the MIC group, consisting of patients who did achieve the MIC and (2) no MIC group, patients who did not achieve MIC for one or both of the KOOS subscales Sport/Rec and QoL, between preoperative values and 1-year follow-up.

Outcome measure

The main outcome of interest in this study was differences in demographical (eg, age, sex, BMI, activity at time of injury) and surgical factors (eg, type of graft, cartilage injuries, meniscal injuries, collateral ligament injuries) between patients who achieved the MIC and patients who did not achieve the MIC, for the KOOS subscales Sport/Rec and QoL, between preoperative ratings and 1 year after ACLR. The MIC published by Ingelsrud et al26 of ≥12.1 for Sport/Rec and ≥18.3 for QoL was used in this study. The MIC reflects the smallest measured change in score that patients perceive as being important, that is, the patient’s function has improved by treatment.27 The MIC refers to the change from baseline to a specific follow-up timepoint10 for a given PRO that is perceived as being important by a patient. The MIC used in this study was calculated using an anchor-based method and predictive modelling, as this method anchors the PRO change score to an external measure of important change.26 Furthermore, the anchor-based method has the best accuracy in predictive modelling and does not easily overestimate the MIC.

The KOOS was developed in 1998 for the assessment of an individual’s precepted knee function in patients with knee-related conditions.28 The KOOS aims to assess symptoms and functional limitations following a knee injury and consists of 42 items scored on five subscales: Pain, Symptoms, Function in Daily Living (ADL), Sport/Rec and QoL.29 All items are scored using a Likert scale, with five response options ranging from 0 meaning extremely positive to 4 meaning extremely negative response. Scores are then converted to a 0–100 scale, where 0 represents the worst possible knee function and 100 suggests the best possible knee function. The KOOS has interclass correlation coefficients that range from 0.85 to 0.90 for the five subscales,30 however, there is no evidence for the content validity of the KOOS.7

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Statistics

Statistical analyses were performed with SAS Statistics for Windows (V.9.4, SAS Institute, Cary, NC, USA). For comparison between groups, Fisher’s exact test and χ² test were used for categorical variables, while the Mann-Whitney U test was used for continuous variables. Comparative analyses were performed for demographic and surgical characteristics between patients who achieved MIC and patients who did not, from baseline to 1-year follow-up for both the KOOS Sport/Rec and QoL. Continuous variables were presented as mean and SD as well as median with minimum and maximum. Count (n) and proportion (%) were used for presenting categorical variables. The significance level was set at p<0.05. Logistic regression was used to adjust for differences in age at surgery and preoperative KOOS. Adjustment for age at surgery was made as patient outcomes differ depending on age, and surgical methods can differ depending on age. Adjustment for preoperative KOOS was made as scores can differ in terms of how easy it can be to achieve MIC, that is, the greater score preinjury, the more difficult to achieve MIC postinjury. Missing data were defined as ‘missing’ in actual tables.

Demographic differences and the KOOS Sport/Rec

Table 1 presents results for the KOOS Sport/Rec both for demographics and concomitant injuries. A higher number of male patients (51.5%) achieved MIC, while a higher number of female patients did not achieve MIC (49.2%) for KOOS Sport/Rec (p<0.0001). The no MIC group had a significantly greater BMI compared with the MIC group (24.8±3.5 vs 24.6±3.3, respectively, p<0.0001). A greater proportion of patients in the no MIC group were younger at the time of surgery (28.5±10.3 vs 29.1±10.8 years, respectively, p=0.0002) and had a longer time from injury to surgery (22.0±38.5 vs 19.3±36.6 months, respectively, p=0.0002) compared with patients in the MIC group. Furthermore, there was a smaller proportion of patients treated with a hamstring tendon autograft in the no MIC group (91.4%) compared with the MIC group (94.1%). On the other hand, there was a greater proportion of patients treated with patellar tendon autograft in the no MIC group (6.3%) (p=0.0023) compared with the MIC group (4.2%).

Concomitant injuries and the KOOS Sport/Rec

Table 1 presents results for the KOOS Sport/Rec subscale both for demographics and concomitant injuries. Patients in the no MIC group had a greater rate of concomitant cartilage injuries (27.7% vs 26.7%, respectively, p<0.0001) especially located on the lateral femoral condyle (22.7% vs 19.4%, respectively, p=0.001) or in the presence of a multicartilage injury (27.7% vs 26.5%, p<0.0001) compared with patients in the MIC group.

Preoperative scores for KOOS Sport/Rec

Table 2 presents preoperative and postoperative scores for the KOOS Sport/Rec subscale. Patients in the no MIC group had significantly greater scores preoperative (48.5±25.2) compared with the MIC group (31.6±22.6) p<0.0001. Patients in the no MIC group also had lowered their scores postoperatively and had significantly lower scores postoperative compared with the MIC group (48.2±26.5 vs 73.0±20.9, respectively, p<0.0001).

Demographic differences and the KOOS QoL

Table 3 presents results for the KOOS QoL subscale for demographics and concomitant injuries. Patients in the no MIC group for KOOS QoL had greater BMI (24.7±3.5 vs 24.6±3.2, respectively, p=0.0001) and were younger at the time of surgery compared with patients in the MIC group (27.4±9.8 vs 29.7±11.0, respectively, p<0.0001). There were no significant differences between patients in the no MIC or MIC group for the KOOS QoL subscale with regards to patient sex, time from injury to surgery, or type of autograft.

Concomitant injuries and the KOOS QoL

Table 3 presents patients in the no MIC group who had a greater rate of concomitant cartilage injuries to the lateral femoral condyle (23.3% vs 19.0%, p=0.01), however, lower rate of general cartilage injury (26.1% vs 27.5%, respectively, p=0.002), to the medial patella (29.9% vs 20.5%, p=0.03) and multicartilage injuries (26.0% vs 27.3%, p=0.002) compared with patients in the MIC group.

Preoperative scores for KOOS QoL

Table 4 presents the preoperative and postoperative scores for KOOS QoL. Patients in the no MIC group had significantly greater scores preoperative (35.4±17.4) compared with the MCI group (30.5±15.9), p<0.0001. Both patients in the no MIC and MIC group had increased their scores postoperatively, although patients in the no MIC group had significantly lower scores compared with the patients in the MIC group overall (43.1±21.4 vs 64.5±20.5, respectively, p<0.0001).

Activities at the time of injury

There were significant differences in activities at the time of injury between the no MIC and MIC groups for both the KOOS Sport/Rec and QoL subscales (p<0.0001). However, the differences were small, 0.1%–4.4% (online supplemental appendices tables 1 and 2).

Patient demographics

Although it has been previously stated that demographics were not able to predict which patients achieve MIC and not31 results from other studies show otherwise.32 Our study shows a greater percentage of females in the no MIC group for KOOS Sport/Rec or QoL, and had, increased BMI, were younger age at time of surgery, and had longer time from injury to surgery, compared with patients in the MIC group, although no predictions for the achievement of MIC was made in this study. Previous research has shown that female sex was predictive for lower scores on all KOOS subscales33 and that males have significantly higher baseline scores for KOOS QoL.34 These results are however somewhat contradictory to results from our study, as there was a greater percentage of females in the no MIC group for KOOS Sport/Rec or QoL and, higher baseline score, which might indicate that MIC is harder to achieve for males than females. Although, males have reported a more positive outlook on their knee related self-efficacy compared with females, which could be the reason a lower percentage of females achieve MIC at 1-year follow-up.35 Our study also shows that a lower percentage of patients who are younger at the time of surgery in the no MIC group, and younger age has been associated with higher baseline scores for KOOS QoL.34 Higher baseline scores may make it more difficult to achieve MIC after 1-year follow-up. Furthermore, studies show younger patients have higher expectations of their knee function compared with older patients,36 which might also be reflected in our results. These results might also imply that younger patients need longer time than 1 year to improve to achieve MIC, or that these younger patients need help adapting to lower expectations of knee function in rehabilitation up to 1-year follow-up. Results from our study also show a greater percentage of patients who had a longer time from injury to surgery did not achieve MIC at 1-year follow-up. Delaying ACLR has also been associated with the risk of development of secondary meniscal injuries,37 38 which may further affect perceived knee function.

In addition, a greater proportion of patients treated with the patellar tendon autograft or quadriceps tendon autograft were in the no MIC group compared with patients treated with hamstrings tendon autograft who were in the MIC group. Previous research has shown that patients treated with hamstrings tendon autograft display less knee pain after surgery compared with patients treated with patellar tendon autograft.39 40 Furthermore, the hamstrings tendon autograft has reported less comorbidity compared with the patellar-, and quadriceps tendon autografts,41 which might also explain clinical differences within these patient groups in the present study.

Concomitant injuries

Previous research has suggested that concomitant meniscal injuries may predict inferior PROs after ACLR.42 43 However, in our analysis, the results suggested a greater proportion of patients in the MIC group despite concomitant meniscal injuries. While the differences between patients in the no MIC and MIC group, with regard to meniscus injuries appear contradictory, this finding may allude to a high prevalence of preoperative knee-related disability. Consequently, patients might achieve MIC by moving from a low score on the KOOS to merely a slightly higher score, and thereby still not have achieved an acceptable knee function according to the patient.

Regarding cartilage injuries, Brophy et al42 and Cox et al43 previously highlighted the negative effect of concomitant cartilage injuries on postoperative outcomes after ACLR, which is also seen in the present study as a higher percentage of patients with cartilage injury, especially to the lateral femoral condyle, were in the no MIC group. Another reason for a lower percentage of patients with cartilage injuries achieving the MIC can be due to the advanced surgical procedures with associated postoperative restrictions, resulting in a more restrictive postoperative rehabilitation process.44 In addition, patients with cartilage injuries have been reporting greater levels of pain and knee-related symptoms compared with patients without cartilage injuries45 which might affect both sports and QoL. Despite results from previous studies,42 43 postoperative outcomes might reach MIC values, but do not necessarily relate to a clinical improvement in that is, function. The clinical implication from these results indicates that patients who have a concomitant cartilage injury might not be able to have the same expectations as patients without a cartilage injury, of knee function regarding Sport/Rec up to 1 year after ACLR, and might need longer time to adjust to these disabilities, or might not be as responsive to surgical treatment of the ruptured ACL compared with patients without cartilage injuries. Whether this depends on the complexity of the injury or the treatment, is unknown.

Limitations

Although this study is registry-based, using a large cohort size, registries have limitations. Due to data configuration, there can be missing data at different follow-ups or for different variables. Furthermore, data from the SNKLR may be limited in terms of generalisability, since data only provide information about the Swedish population with ACLR. Unfortunately, there are no anchor-based questions in the SNKLR and therefore we had no ability to calculate new MIC cut-offs based on the data from SNKLR, instead we used the MIC used by Ingelsrud et al.26 In addition, the MIC can be calculated in different ways to match different data sets, and MIC might not apply to every unique patient. This means that patients can achieve MIC but do not necessarily equal to acceptable knee function. One year might not be enough to achieve MIC for some patients, as cartilage injuries usually take longer to rehabilitate, and might have persisting knee symptoms compared with an isolated ACLR. However, the MIC used in this study determines changes over time, from baseline to a specific follow-up (6 months, 1 or 2 years), in this case 1 year, and the cut-offs have previously been used.26 Furthermore, the MIC values might not have identified all possible MICs after ACLR, as MIC can vary depending on the magnitude of baseline scores for Sports/Rec and QoL.

Using the adjustments, for this study, age at surgery and KOOS preoperatively, might affect the achievement of MIC, as seen for meniscal injuries, as these patients might have a lower preoperative score and therefore a greater chance of achieving MIC. Results from this study can also be difficult to apply in clinic settings since our study mainly studied demographic and surgical characteristics between the no MIC and MIC groups. However, our results may assist surgeons in understanding which patients might benefit from an ACLR and which patients who might not. In line with this, the results from this study might also assist physical therapists in conveying patient expectations for rehabilitation up to 1 year after ACLR.

Finally, the statistically significant differences found in this study were small, thus, significant results might not have been clinically relevant. In addition, there was a large number of missing data in several items when comparing the two groups (MIC vs no MIC), for example, there were 1500 cases of missing BMI data in the No MIC group in the KOOS Sport/Rec. Results from this study should therefore be interpreted with caution.