Objectives
The purpose of this investigation was to identify clinical and demographic variables that may predict compliance, defined as patient follow-up at 1-year and 2-year postoperative appointments, after anterior cruciate ligament reconstruction (ACLR).
Methods
A total of 107 patients undergoing primary ACLR across four centres were prospectively followed for 2 years. Demographic and clinical data were recorded preoperatively and postoperatively, including patient-reported outcomes such as the Marx Activity Rating Scale, Cincinnati Occupational Rating Scale (CORS), International Knee Documentation Committee Subjective Knee Evaluation Form and Activities of Daily Living Scale (ADLS). Compliance was retrospectively defined using the presence or lack of outcome measures at 1-year and 2-year postoperative visits. Univariate analysis was done to compare demographic and clinical variables between compliant and non-compliant patients at 1-year and 2-year follow-up. Significant variables were entered into a logistic regression model. Significance was set at p<0.05.
Results
The overall rate of compliance at 1-year and 2-year postoperative appointments was 83.2% (89/107) and 57.0% (61/107), respectively. Regression analysis showed that residence in Kobe, Japan compared with Pittsburgh, USA (OR 10.28; 95% CI 1.0003 to 105.28), and ‘very strenuous’ (OR 16.74; 95% CI 3.21 to 87.43) and ‘strenuous’ (OR 18.78; 95% CI 2.01 to 175.78) preinjury activity level were independent factors associated with compliance at 1-year follow-up. At 2 years follow-up, younger age (OR 0.95; 95% CI 0.91 to 0.997), and greater preoperative score on CORS (OR 1.03; 95% CI 1.01 to 1.06) and on ADLS (OR 1.04; 95% CI 1.01 to 1.07) were independently associated with compliance.
Conclusion
While patients with greater level of strenuous activity were more likely to follow-up after ACL surgery at 1 year, patients who were younger and had higher preoperative function measured by CORS and ADLS were more likely to follow-up at 2 years. Differences in follow-up rates among the USA, Italy, Sweden and Japan may highlight important cultural, socioeconomic and infrastructural differences across international healthcare systems. Orthopaedic surgeons may consider the results of this study during preoperative discussion with their patients and design of future ACL clinical research trials.
Level of evidence
Level III, prospective cohort.
What are the new findings
- •The rate of 2-year compliance with postoperative follow-up, after anterior cruciate ligament reconstruction was greatest in Göteborg, Sweden (72.2%), followed by Kobe, Japan (66.7%), Bologna, Italy (50.0%), then Pittsburgh, USA (45.5%).
- •Younger age (OR 0.95; 95% CI 0.91 to 0.997), higher preoperative scores on Cincinnati Occupational Rating Scale (OR 1.03; 95% CI 1.01 to 1.06) and Activities of Daily Living Scale (OR 1.04; 95% CI 1.01 to 1.07) were predictive of 2-year compliance.
- •Very strenuous and strenuous preinjury level of activity (OR 16.74; 95% CI 3.21 to 87.43 and OR 18.78; 95% CI 2.01 to 175.78, respectively) were predictive of 1-year compliance.
- •Factors that did not influence compliance rates included: (1) degree of rotational knee instability, as measured by clinical and quantitative pivot shift exam, (2) concomitant meniscus or cartilage injury, (3) sex, (4) smoking history and (5) employment status.
Introduction
A pervasive criticism of orthopaedic surgery research is the current lack of high-quality (ie, randomised controlled trials) clinical trials with long-term follow-up.
1
, 2
, 3
, 4
, 5
An increasingly recognised reason for this is patient non-compliance with follow-up appointments or patient loss to follow-up.2
, 5
, 6
Compared with other surgical specialties, orthopaedic research trials have unfortunately been hampered by inadequate follow-up rates,5
and by a lack of reporting of non-compliance rates.7
, 8
Recently, loss to follow-up has been investigated in the setting of orthopaedic trauma,2
, 5
revealing risk factors for non-compliance such as male sex, young age, high-risk alcohol and drug use and treatment in a high-income country with predominantly privately funded healthcare system. Compliance in the setting of anterior cruciate ligament (ACL) surgery and orthopaedic sports medicine, however, is comparatively less well understood.Clinically, patients who are lost to follow-up after ACL surgery may be at risk for worse outcomes, delayed recognition of complications or may otherwise miss critical adjustments to physical therapy protocols with regard to range of motion, strength, proprioception and/or running/jumping mechanics. These are all pertinent and vital components of postoperative care following ACL surgery, particularly with regard to return to sport/work and prevention of ACL graft failure. An understanding of who may or may not be more likely to follow-up postoperatively can inform the preoperative conversation between orthopaedic surgeon and patient. Second, orthopaedic sport surgeons may develop better strategies for improving compliance rates if these risk factors are understood. From a research perspective, improving compliance rates minimises bias in evaluating outcomes and may allow a more transparent evaluation of the current standard of care.
Therefore, the goal of this study was to identify demographic and clinical variables associated with compliance with follow-up after ACL reconstruction. Based on previous reported risk factors for compliance after orthopaedic surgery,
2
, 5
it was hypothesised that patients undergoing ACL reconstruction may be (1) more compliant with 1-year and 2-year follow-up visits if they were female, older, non-cigarette smokers and participants outside of the USA and (2) equally compliant irrespective of their employment status.Methods
A retrospective analysis of data collected during the Prospective International Validation of Outcome Trial (PIVOT) international, multicentre study was performed. As previously described,
9
107 patients, aged 14–50 years, undergoing unilateral, primary ACL reconstruction using hamstring autograft, were enrolled in a 2-year prospective study across four international sites: (1) University of Pittsburgh, Pittsburgh, Pennsylvania, USA; (2) Instituto Ortopedico Rizzoli, Bologna, Italy; (3) Sahlgrenska University, Göteborg, Sweden and (4) Kobe University, Kobe, Japan. Institutional review board approval was obtained in all sites. Inclusion and exclusion criteria for this investigation were unchanged from previous.9
Various subjective and objective clinical parameters were measured preoperatively, intraoperatively and 1 year and 2 years postoperatively as part of the multicentre PIVOT trial. Preoperatively, patients' age, sex, body mass index (BMI), employment status, work level of activity, date of injury, injury mechanism, preinjury level of strenuous activity (on a five-choice qualitative scale based on the International Knee Documentation Committee (IKDC) Subjective Knee Form, ranging from ‘very strenuous activities like jumping or pivoting as in basketball or soccer’, ‘strenuous activities like heavy physical work, skiing or tennis’, ‘moderate activities like moderate physical work, running or jogging’, ‘light activities like walking, housework or yard work’ or 'unable to perform any of the above activities due to knee'), frequency of activity ('4–7 times per week', ‘1–3 times per week’, ‘1–3 times per month’, ‘less than one time per month‘), cigarette use, patient- reported outcomes (PROs) (Marx Activity Rating Scale, International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, Cincinnati Occupational Rating Scale (CORS) and Activities of Daily Living Scale (ADLS) scores), passive and active knee flexion and extension and grade on Lachman test were recorded. Intraoperatively, patients were examined under general anaesthesia and the clinician's grade of pivot shift exam was recorded. To provide an objective measurement of rotatory knee laxity, a quantitative pivot shift exam was performed as previously described, which provided information on the translation and acceleration of the lateral compartment during pivot shift manoeuvre.
9
During diagnostic arthroscopy, medial and/or lateral meniscus injury was documented, and articular cartilage was assessed using the Outerbridge Classification. For the purposes of this investigation, any unstable tear of the meniscus was considered a meniscus injury, and cartilage of grade≥2 on Outerbridge Classification was deemed injured. Outcome measurements recorded at 1-year and 2-year postoperative appointments were: active and passive flexion and extension, Lachman, pivot shift and anterior drawer tests, quantitative pivot shift exam (as measured using a previously validated iPad application and tibia accelerometer9
) and PROs (Marx Activity Rating Scale, IKDC, CORS and ADLS scores).Definition of compliance in the PIVOT trial
Compliance was defined retrospectively, using the presence or lack of outcome measures acquired during the PIVOT multicentre study. One-year and 2-year compliance was defined when a patient met either criteria at their 1-year and 2-year postoperative appointments: (1) any physical examination parameter was recorded (ie, a patient was considered compliant if there was a single measurement recorded for the Lachman test but none for pivot shift or anterior drawer tests, active or passive flexion, active or passive extension) and (2) any one of the PROs surveys was completed (ie, a patient completed the Marx Activity Rating Scale, but not the IKDC, CORS or ADLS scores).
Strategies to maximise compliance in the PIVOT trial
Numerous strategies were employed to maximise patient follow-up. Experienced research coordinators were largely responsible for communicating with patients. If an appointment was missed, research coordinators contacted patients both via phone and email up to three times to encourage follow-up. If a patient was unable to attend a follow-up appointment, PRO surveys were mailed to their home address. When patients attended their follow-up meetings, an experienced research coordinator was present in order to record all outcome measurements.
Statistical methods
Statistical analyses were performed using SAS V.9.4 (SAS, Cary, North Carolina, USA). A univariate analysis was performed to identify differences in demographic and clinical variables between patients who were compliant with 1-year and 2-year follow-up appointments, and those who were non-compliant. Continuous variables were analysed using a Wilcoxon rank sum test, whereas categorical variables were compared using Χ 2 or the Fisher's exact test. Significance was set at p<0.05. Clinical and demographic variables that were found to be significant were subsequently entered into a logistic regression model. Multivariable regression models were established using a backward elimination approach.
Results
All 107 patients (mean age 24.8±9.0 years; 43% female) enrolled in the PIVOT multicentre trial were eligible for this retrospective analysis. The majority of patients (90%) suffered an ACL injury from a sport-related event. Most patients reported either a very strenuous (69%) or strenuous (18%) level of activity preinjury. The number of patients enrolled in (1) Bologna, Italy, (2) Göteborg, Sweden, (3) Kobe, Japan and (4) Pittsburgh, USA were: 26 (24%), 18 (17%), 30 (28%) and 33 (31%), respectively. All demographic data are presented in table 1 and baseline PROs and physical examination are presented in table 2.
Table 1Patient demographic information (n=107)
| Demographic variable | Mean±SD Or N (%) |
|---|---|
| Age (years) | 24.8±9.0 |
| Sex | |
| Female | 46 (43%) |
| Male | 61 (57%) |
| Body mass index | 24.1±3.8 |
| City, country of residence | |
| Bologna, Italy | 26 (24%) |
| Göteborg, Sweden | 18 (17%) |
| Kobe, Japan | 30 (28%) |
| Pittsburgh, USA | 33 (31%) |
| Employment status | |
| Regular duty—full time | 45 (42%) |
| Regular duty—part time | 6 (6%) |
| Temporarily off work due to health | 2 (2%) |
| Unemployed | 4 (4%) |
| Student | 50 (47%) |
| Work level of activity | |
| Mostly sedentary | 23 (22%) |
| Sedentary | 21 (20%) |
| Moderately active | 32 (30%) |
| Demanding | 30 (28%) |
| Time from injury to surgery (day) | 125.3±90.1 |
| Injury mechanism | |
| Sports | 96 (90%) |
| Basketball | 20 (21%) |
| Soccer | 25 (26%) |
| Skiing | 14 (15%) |
| American football | 9 (9%) |
| Softball | 4 (4%) |
| Volleyball | 4 (4%) |
| Other | 20 (21%) |
| Work | 1 (1%) |
| Activities of daily living | 4 (4%) |
| Other | 6 (6%) |
| Preinjury activity level | |
| Very strenuous | 74 (69%) |
| Strenuous | 19 (18%) |
| Moderate | 10 (9%) |
| Light | 2 (2%) |
| No activity | 1 (1%) |
| Frequency of preinjury activity | |
| 4–7 times/week | 56 (52%) |
| 1–3 times/week | 35 (33%) |
| 1–3 times/month | 11 (10%) |
| <1/month | 2 (2%) |
| Cigarette smoking | |
| Current smoker | 6 (6%) |
| Ever smoked | 5 (5%) |
| Medial meniscus injury | 40 (37%) |
| Lateral meniscus injury | 42 (39%) |
| Articular cartilage injury | |
| Medial knee compartment | 8 (7%) |
| Lateral knee compartment | 5 (5%) |
| Patellofemoral compartment | 2 (2%) |
Table 2Baseline patient-reported outcomes and physical examination (n=107)
| Variable | Mean±SD Or N (%) |
|---|---|
| Preoperative Marx Activity Rating Score | 11.2±5.3 |
| Preoperative IKDC | 56.8±16.2 |
| Preoperative CORS | 29.4±14.9 |
| Preoperative ADLS | 76.7±19.0 |
| Passive knee extension | 2±4 |
| Active knee flexion | 133±20 |
| Passive knee flexion | 135±20 |
| Lachman test (mm) | |
| >10 | 5 (5%) |
| 6 to 10 | 56 (56%) |
| 3 to 5 | 38 (38%) |
| −1 to 2 | 1 (1%) |
| Pivot shift—clinical examination | |
| Gross | 2 (2%) |
| Clunk | 34 (34%) |
| Glide | 58 (58%) |
| Equal | 6 (6%) |
| Quantitative pivot shift | |
| Acceleration ipsilateral knee (m/s2) | 5.9±4.4 |
| Acceleration side-to-side difference (m/s2) | 2.9±4.4 |
| Translation ipsilateral knee (mm) | 3.0±2.1 |
| Translation side-to-side difference (mm) | 2.0±2.1 |
ADLS, Activities of Daily Living Scale; CORS, Cincinnati Occupational Rating Scale; IKDC, International Knee Documentation Committee.
One-year compliance
The overall rate of compliance at the 1-year postoperative appointment was 83.2% (89/107) (Bologna, Italy: 76.9%; Göteborg, Sweden: 94.4%; Kobe, Japan: 96.7%; Pittsburgh, USA: 69.7%). Using a univariate analysis, the following factors were shown to significantly influence compliance: (1) city, country where treatment was performed, (2) injury mechanism, (3) preinjury level of activity, (4) frequency of preinjury activity (table 3) and (5) preoperative score on the Marx Activity Rating Scale (p<0.05) (table 4). Age, sex, BMI, employment status, level of activity at work, time from injury to surgery, smoking, concomitant meniscus or articular cartilage injury (table 3), preoperative IKDC, CORS, ADLS, passive knee extension, active and passive knee flexion, grade on Lachman and pivot shift exam and lateral compartment tibia acceleration and translation as measured by quantitative pivot shift, were not shown to influence compliance at 1 year (table 4).
Table 3Univariate comparison of demographic factors associated with compliance at 1-year follow-up
| Variable | Compliant at 1 year | Non-compliant at 1 year | P values |
|---|---|---|---|
| (n=89) | (n=18) | ||
| Age (years) | 24.5±9.2 | 26.5±7.9 | 0.14 |
| Sex | |||
| Female | 37 (80.4%) | 9 (19.6%) | 0.51 |
| Male | 52 (85.2%) | 9 (14.8%) | |
| Body mass index | 24.1±3.8 | 24.5±4.2 | 0.85 |
| City, country of residence | |||
| Bologna, Italy | 20 (76.9%) | 6 (23.1%) | 0.01* |
| Göteborg, Sweden | 17 (94.4%) | 1 (5.6%) | |
| Kobe, Japan | 29 (96.7%) | 1 (3.3%) | |
| Pittsburgh, USA | 23 (69.7%) | 10 (30.3%) | |
| Employment status | |||
| Regular duty—full time | 36 (80%) | 9 (20%) | 0.35 |
| Regular duty—part time | 5 (83.3%) | 1 (16.7%) | |
| Temporarily off work due to health | 1 (50%) | 1 (50%) | |
| Unemployed | 3 (75%) | 1 (25%) | |
| Student | 44 (88%) | 6 (12%) | |
| Work level of activity | |||
| Mostly sedentary | 21 (91.3%) | 2 (8.7%) | 0.53 |
| Sedentary | 16 (76.2%) | 5 (23.8%) | |
| Moderately active | 26 (81.3%) | 6 (18.8%) | |
| Demanding | 26 (86.7%) | 4 (13.3%) | |
| Time from injury to surgery (days) | 125.53±85.01 | 119.17±114.62 | 0.42 |
| Injury mechanism | 0.003* | ||
| Sports | 84 (87.5%) | 12 (12.5%) | |
| Work | 1 (100%) | 0 (0%) | |
| Activities of daily living | 2 (50%) | 2 (50%) | |
| Other | 2 (33.3%) | 4 (66.7%) | |
| Preinjury activity level | 0.001* | ||
| Very strenuous | 66 (89.2%) | 8 (10.8%) | |
| Strenuous | 17 (89.5%) | 2 (10.5%) | |
| Moderate | 5 (50%) | 5 (50%) | |
| Light | 0 (0%) | 2 (100%) | |
| No activity | 1 (100%) | 0 (0%) | |
| Frequency of preinjury activity | 0.04 | ||
| 4–7 times/week | 49 (87.5%) | 7 (12.5%) | |
| 1–3 times/week | 29 (82.9%) | 6 (17.1%) | |
| 1–3 times/month | 9 (81.8%) | 2 (18.2%) | |
| <1/month | 0 (0%) | 2 (100%) | |
| Smoking | |||
| Current smoker | 4 (66.7%) | 2 (33.3%) | 0.27 |
| Ever smoked | 3 (60%) | 2 (40%) | 0.18 |
| Medial meniscus injury | 34 (85%) | 6 (15%) | 0.69 |
| Lateral meniscus injury | 35 (83.3%) | 7 (16.7%) | 0.97 |
| Articular cartilage injury | |||
| Medial knee compartment | 7 (87.5%) | 1 (12.5%) | 1 |
| Lateral knee compartment | 5 (100%) | 0 (0%) | 0.59 |
| Patellofemoral compartment | 1 (50%) | 1 (50%) | 0.31 |
Data are shown either as mean±SD or N (%).
Continuous variables were compared using Wilcoxon rank sum test.
Categorical variables were compared using either Χ2 or Fisher's exact test.
* P<0.05.
Table 4Univariate comparison of scores/physical examination associated with compliance at 1-year follow-up
| Variable | Compliant at 1 year | Non-compliant at 1 year | P values |
|---|---|---|---|
| (n=89) | (n=18) | ||
| Preoperative Marx Activity Rating Scale | 11.6±5.2 | 8.9±5.0 | 0.03 |
| Preoperative IKDC | 58.3±15.0 | 49.1±20.2 | 0.08 |
| Preoperative CORS | 28.8±14.2 | 32.4±18.6 | 0.6 |
| Preoperative ADLS | 78.5±17.3 | 67.4±24.6 | 0.07 |
| Passive knee extension | 2±4 | 2±2 | 0.66 |
| Active knee flexion | 133±21 | 134±14 | 0.97 |
| Passive knee flexion | 135±21 | 137±14 | 0.84 |
| Lachman test (mm) | |||
| >10 | 4 (80%) | 1 (20%) | 0.84 |
| 6 to 10 | 47 (83.9%) | 9 (16.1%) | |
| 3 to 5 | 33 (86.8%) | 5 (13.2%) | |
| −1 to 2 | 1 (100%) | 0 (0%) | |
| Pivot shift—clinical examination | |||
| Gross | 2 (100%) | 0 (0%) | 0.93 |
| Clunk | 28 (82.4%) | 6 (17.6%) | |
| Glide | 50 (86.2%) | 8 (13.8%) | |
| Equal | 5 (83.3%) | 1 (16.7%) | |
| Quantitative pivot shift | |||
| Acceleration ipsilateral knee (m/s2) | 6.0±4.8 | 5.4±1.4 | 0.21 |
| Acceleration side-to-side difference (m/s2) | 3.0±4.9 | 2.4±1.2 | 0.19 |
| Translation ipsilateral knee (mm) | 2.9±2.1 | 3.5±2.3 | 0.15 |
| Translation side-to-side difference (mm) | 1.9±2.2 | 2.5±1.9 | 0.17 |
Data are shown either as mean±SD or N (%).
Continuous variables were compared using Wilcoxon rank sum test.
Categorical variables were compared using either Χ2 or Fisher's exact test.
ADLS, Activities of Daily Living Scale; CORS, Cincinnati Occupational Rating Scale; IKDC, International Knee Documentation Committee.
* P<0.05.
When entered into a logistic regression model, the country where treatment was performed and level of strenuous activity prior to injury were identified as independent variables associated with compliance (table 5). Specifically, a significantly higher rate of compliance was found only when comparing patients treated in Kobe, Japan with those treated in Pittsburgh, USA when controlling for covariates (OR 10.28; 95% CI 1.0003 to 105.28); no significant differences were found if treated in Bologna, Italy or Göteborg, Sweden compared with Pittsburgh, USA (table 5). Patients who self-reported a preinjury level of strenuous activity of ‘very strenuous’ and ‘strenuous’ were 16.74 (95% CI 3.21 to 87.43) and 18.78 (95% CI 2.01 to 175.78) more likely to comply with 1-year follow-up appointments than those who self-reported a preinjury activity level of ‘moderate’ (table 5). Contrarily, injury mechanism, frequency of preinjury activity and preoperative scores on the Marx Activity Rating Scale were no longer found to be independently significant.
Table 5Logistic regression model for compliance at 1-year follow-up
| Covariate | OR | 95% CI |
|---|---|---|
| Bologna, Italy vs Pittsburgh, USA | 0.51 | 0.12 to 2.17 |
| Göteborg, Sweden vs Pittsburgh, USA | 6.11 | 0.56 to 66.41 |
| Kobe, Japan vs Pittsburgh, USA | 10.28 | 1.0003 to 105.28 |
| Very strenuous vs moderate/light/none preinjury activity level | 16.74 | 3.21 to 87.43 |
| Strenuous vs moderate prenjury activity level | 18.78 | 2.01 to 175.78 |
Two-year compliance
Overall rate of compliance at the 2-year postoperative appointment was 57.0% (61/107) (Bologna, Italy: 50.0%; Göteborg, Sweden: 72.2%; Kobe, Japan: 66.7%; Pittsburgh, USA: 45.5%). From univariate analysis, the following factors were shown to influence compliance: (1) age (table 6), (2) preoperative CORS, (3) preoperative ADLS, (4) baseline active knee flexion and (5) baseline passive knee flexion (table 7). Sex, BMI, city, country of residence, employment status, level of activity at work, time from injury to surgery, injury mechanism, preinjury level of activity, frequency of preinjury activity, smoking, concomitant meniscus or articular cartilage injury (table 6), preoperative Marx Activity Rating Scale, passive knee extension, grade on Lachman or pivot shift test and lateral compartment tibia acceleration and translation as measured by quantitative pivot shift, were not shown to influence compliance with 2-year follow-up appointments (table 7).
Table 6Univariate comparison of demographic factors associated with compliance at 2-year follow-up
| Variable | Compliant at 2 years | Non-compliant at 2 years | P values |
|---|---|---|---|
| (n=61) | (n=46) | ||
| Age (years) | 23.3±8.9 | 26.8±8.9 | 0.01 |
| Sex | |||
| Female | 27 (58.7%) | 19 (41.3%) | 0.76 |
| Male | 34 (55.7%) | 27 (44.3%) | |
| BMI | 23.5±3.3 | 24.9±4.4 | 0.15 |
| City, country of residence | |||
| Bologna, Italy | 13 (50%) | 13 (50%) | 0.16 |
| Göteborg, Sweden | 13 (72.2%) | 5 (27.8%) | |
| Kobe, Japan | 20 (66.7%) | 10 (33.3%) | |
| Pittsburgh, USA | 15 (45.5%) | 18 (54.5%) | |
| Employment status | |||
| Regular duty—full time | 22 (48.9%) | 23 (51.1%) | 0.61 |
| Regular duty—part time | 4 (66.7%) | 2 (33.3%) | |
| Temporarily off work due to health | 1 (50%) | 1 (50%) | |
| Unemployed | 2 (50%) | 2 (50%) | |
| Student | 32 (64%) | 18 (36%) | |
| Work level of activity | |||
| Mostly sedentary | 11 (47.8%) | 12 (52.2%) | 0.39 |
| Sedentary | 10 (47.6%) | 11 (52.4%) | |
| Moderately active | 21 (65.6%) | 11 (34.4%) | |
| Demanding | 19 (63.3%) | 11 (36.7%) | |
| Time from injury to surgery (days) | 124.7±84.0 | 126.1±98.6 | 0.87 |
| Injury mechanism | |||
| Sports | 57 (59.4%) | 39 (40.6%) | 0.37 |
| Work | 0 (0%) | 1 (100%) | |
| Activities of daily living | 2 (50%) | 2 (50%) | |
| Other | 2 (33.3%) | 4 (66.7%) | |
| Preinjury activity level | |||
| Very strenuous | 46 (62.2%) | 28 (37.8%) | 0.18 |
| Strenuous | 11 (57.9%) | 8 (42.1%) | |
| Moderate | 4 (40%) | 6 (60%) | |
| Light | 0 (0%) | 2 (100%) | |
| No activity | 0 (0%) | 1 (100%) | |
| Frequency of preinjury activity | |||
| 4–7 times/week | 36 (64.3%) | 20 (35.7%) | 0.11 |
| 1–3 times/week | 17 (48.6%) | 18 (51.4%) | |
| 1–3 times/month | 8 (72.7%) | 3 (27.3%) | |
| <1/month | 0 (0%) | 2 (100%) | |
| Smoking | |||
| Current smoker | 4 (66.7%) | 2 (33.3%) | 1 |
| Ever smoked | 3 (60%) | 2 (40%) | 1 |
| Medial meniscus injury | 26 (65%) | 14 (35%) | 0.2 |
| Lateral meniscus injury | 24 (57.1%) | 18 (42.9%) | 0.98 |
| Articular cartilage injury | |||
| Medial knee compartment | 6 (75%) | 2 (25%) | 0.46 |
| Lateral knee compartment | 1 (20%) | 4 (80%) | 0.16 |
| Patellofemoral compartment | 1 (50%) | 1 (50%) | 1 |
Data are shown either as mean±SD or N (%).
Continuous variables were compared using Wilcoxon rank sum test.
Categorical variables were compared using either Χ2 or Fisher's exact test.
* P<0.05.
Table 7Univariate comparison of scores/physical examination associated with compliance at 2-year follow-up
| Variable | Compliant at 2 years | Non-compliant at 2 years | P values |
|---|---|---|---|
| (n=61) | (n=46) | ||
| Preoperative Marx Activity Rating Score | 11.9±5.3 | 10.2±5.1 | 0.05 |
| Preoperative IKDC | 59.5±14.7 | 53.2±17.6 | 0.06 |
| Preoperative CORS | 31.5±13.8 | 26.6±16.1 | 0.03* |
| Preoperative ADLS | 81.0±14.4 | 71.0±22.7 | 0.03* |
| Passive knee extension | 2±3 | 2±5 | 0.19 |
| Active knee flexion | 137±10 | 127±27 | 0.01* |
| Passive knee flexion | 140±11 | 130±27 | 0.02 |
| Lachman test (mm) | |||
| >10 | 3 (60%) | 2 (40%) | 0.5 |
| 6 to 10 | 36 (64.3%) | 20 (35.7%) | |
| 3 to 5 | 16 (42.1%) | 22 (57.9%) | |
| −1 to 2 | 1 (100%) | 0 (0%) | |
| Pivot shift—clinical examination | |||
| Gross | 1 (50%) | 1 (50%) | 0.1 |
| Clunk | 20 (58.8%) | 14 (41.2%) | |
| Glide | 30 (51.7%) | 28 (48.3%) | |
| Equal | 5 (83.3%) | 1 (16.7%) | |
| Quantitative pivot shift | |||
| Acceleration ipsilateral knee | 6.5±5.5 | 5.1±2.3 | 0.65 |
| Acceleration side-to-side | 3.5±5.4 | 2.3±2.9 | 0.74 |
| Difference | |||
| Translation ipsilateral knee | 2.6±1.6 | 3.4±2.7 | 0.27 |
| Translation side-to-side | 1.7±1.8 | 2.4±2.5 | 0.33 |
Data are shown either as mean±SD or N (%).
Continuous variables were compared using Wilcoxon rank sum test.
Categorical variables were compared using either Χ2 or Fisher's exact test.
ADLS, Activities of Daily Living Scale; CORS, Cincinnati Occupational Rating Scale; IKDC, International Knee Documentation Committee.
* P<0.05.
When entered into a logistic regression model, age, preoperative CORS and preoperative ADLS were identified as independent variables associated with compliance; contrarily, baseline active and passive knee flexion were no longer found to be significant predictors for compliance (table 8). Age exhibited an inverse correlation with 2-year compliance (OR 0.95; 95% (CI 0.91 to 0.997) (table 8). Patients with higher preoperative CORS and ADLS scores were more likely to follow-up 2 years postoperatively (OR 1.03; 95% CI 1.01 to 1.06) and OR 1.04; 95% CI 1.01 to 1.07, respectively) (table 8).
Table 8Logistic regression model for compliance at 2-year follow-up
| Covariate | OR | 95% CI |
|---|---|---|
| Age | 0.95 | 0.91 to 0.997 |
| Preoperative CORS | 1.03 | 1.01 to 1.06 |
| Preoperative ADLS | 1.04 | 1.01 to 1.07 |
ADLS, Activities of Daily Living Scale; CORS, Cincinnati Occupational Rating Scale.
Discussion
The main findings from this retrospective analysis of compliance in the prospective cohort, PIVOT trial were as follows: cumulative 1-year and 2-year compliance rates following ACL reconstruction were 83.2% and 57.0%, respectively, patients with strenuous preinjury level of activity were more likely to follow-up at 1 year, and patients who were younger and had greater preoperative function as measured by CORS and ADLS were more likely to follow-up at 2 years. Additionally, clear discrepancies exist among the USA, Italy, Sweden and Japan in compliance rates, whether or not the logistic regression model was able to capture statistical differences.
Regression analysis demonstrated that patients who were younger, and had greater CORS and ADLS preoperative scores were more likely to follow-up 2 years postoperatively. A greater score on ADLS and CORS, which place emphasis on an individual's ability to perform functional knee movements, was consistent with the greater degree of preoperative passive and active knee flexion seen in these same patients. We hypothesise that perhaps these patients had easier and/or more affordable access to care and/or were more motivated to manage their knees preoperatively and thus postoperatively. Unfortunately, no assessment on insurance, annual income, socioeconomic status or psychological factors could be made from the available demographic data in this study. Furthermore, as the ORs were increased on the scale of one-hundredths, further studies are needed to corroborate our results and investigate how preoperative functional capacity of ACL-injured patients may affect long-term compliance.
Regression analysis demonstrated that a ‘very strenuous’ or ‘strenuous’ preoperative level of activity was independently associated with 1-year compliance. This was consistent with univariate analysis, which showed that compliant patients were more likely to suffer ACL injury from sports and score higher on Marx Activity Rating Scale. This trend was lost at 2 years, perhaps explained by a drop in follow-up of patients who injured their ACL during sports—from 87.5% at 1-year follow-up to 59.4% at 2-year follow-up. This lack of follow-up is particularly concerning considering high rates of second ACL injury in younger, more active patients,
10
and because only a minority of patients successfully fulfil return to sport battery testing when attempting to return to sport earlier than 1 year.11
Furthermore, continued follow-up may be necessary to fulfil objective strength tests associated with a return to same level of sports participation.12
Considering low rates of return to the same competitive level sport after ACLR (55%),13
it may be important for orthopaedic surgeons and/or physical therapists to continue follow-up past sport clearance, in order to assess deficiencies and optimise postsurgical rehabilitation.Clear discrepancies existed among countries at 1 year and 2 years following ACL surgery, which may be attributed to differences in healthcare infrastructure and cultural norms. A recent systematic review revealed significantly higher rates of loss to follow-up in the USA (13.8%) versus other countries (9.4%) in patients undergoing orthopaedic surgery.
8
Privately funded healthcare systems, used in Pittsburgh and in many US hospitals, may make it difficult for patients to follow-up if they are uninsured.14
This may be less of an issue in Japan, Sweden or Italy since each country has largely publicly funded healthcare systems (>70% of funding is public) and cost to the patient is not as much of a factor.15
, 16
, 17
Apart from healthcare policy, cultural differences in the physician-patient relationship among countries should be acknowledged. In the primary care setting, physicians in the USA spent more conversation time on treatment and follow-up than physicians in Japan, who focused more on physical examination and diagnosis.18
Given that physicians in the USA focus more on follow-up and still have lower follow-up rates suggests that patients may have different expectations for follow-up visits than in other countries. In the USA, patients may be less inclined to follow-up once they have fulfilled their own goals (ie, a return to sport), as evidenced by the fact that patients in the USA were more likely to be lost to follow-up even if pay per participant was higher than those of other countries.2
, 19
Although 1-year compliance rate (83.2%) in the PIVOT multicentre trial was above the historically accepted threshold of 80%,
20
and comparable to rates reported in other ACL trials, the 2-year compliance rate (57.0%) was significantly lower than numerous other published studies.2
, 21
, 22
, 23
, 24
, 25
, 26
, 27
While patients were contacted via email and telephone for follow-up, more rigorous participant-retention strategies may be required. A Cochrane review regarding the effect of numerous strategies to improve retention of patients in randomised trials observed that the only strategies that significantly improved trial retention were the ones with monetary incentives.19
Thus, a potential factor that might have prevented the significant decline between 1-year and 2-year compliance rates may have been compensation of patients for enrolment. Another study evaluating loss to follow-up demonstrated that loss-to-follow-up rates could be reduced by using several strategies of patient retention such as tying payments of sites to completion of follow-up, collecting more than one piece of contact information and trying to contact patients on the weekend.2
Still, an important result of the present study was the identification of patient characteristics related to higher risk of loss-to-follow-up in the ACL-injured population, which may serve as a first step to developing targeted strategies to these patients.The data from this study may be used to consider ways to improve compliance rates after ACL injury. Orthopaedic surgeons might consider emphasising the importance of follow-up in older patients, patients with lower preoperative PROs and patients with less strenuous activity. On a health policy level, follow-up may be improved in the USA if the cost of postoperative appointments can be reduced. Clinicians should be wary of rates of loss to follow-up in published studies, if they are reported, and how outcome measurements may be affected by them. For example, a recent systematic review of all ACL randomised controlled trials up to 2016 showed that only 8% of studies used an intent-to-treat analysis to address patient loss to follow-up.
28
In the design of ACL-research studies, rigorous strategies should be employed to achieve adequate follow-up rates, perhaps even >80%.6
This study is not without limitations. Compliance was defined retrospectively, which may not capture the true rate of follow-up. However, an exhaustive search of documented, postoperative physical examination and PROs was used to define compliance in this investigation. Second, no analysis was performed on the time from injury to day of presentation to clinic (when most preoperative variables were recorded), which is an important confounding variable that could affect baseline measurements. Lastly, although an achievement of this study was the identification of factors that predict postoperative compliance with follow-up, no analysis was performed on the effect compliance may have had on outcomes (ie, ACL reinjury or revision surgery rates).
Conclusion
While ACL-injured patients with a greater level of preoperative strenuous activity were more likely to follow-up with their orthopaedic surgeons at 1 year, ACL-injured patients who were younger, and with higher preoperative function as measured by CORS and ADLS were more likely to follow-up at 2 years. Differences in follow-up rates among the USA, Italy, Sweden and Japan may highlight important cultural, socioeconomic and infrastructural differences across international healthcare systems, and motivate changes to improve compliance rates. Orthopaedic surgeons may consider the results of this study during preoperative discussion with their patients and design of future ACL clinical research trials.
References
- A quality assessment of randomized clinical trials in pediatric orthopaedics.J Pediatr Orthop. 2007; 27: 573-581
- Predicting and preventing loss to follow-up of adult trauma patients in randomized controlled trials: an example from the FLOW trial.J Bone Joint Surg Am. 2017; 99: 1086-1092
- Loss to follow-up matters.J Bone Joint Surg Br. 1997; 79: 254-257
- Does a “Level I Evidence” rating imply high quality of reporting in orthopaedic randomised controlled trials?.BMC Med Res Methodol. 2006; 6: 44
- Loss of follow-up in orthopaedic trauma: who is getting lost to follow-up?.J Orthop Trauma. 2015; 29: 510-515
- Loss of follow-up in orthopaedic trauma: is 80% follow-up still acceptable?.J Orthop Trauma. 2013; 27: 177-181
- The quality of reporting of randomized trials in the Journal of Bone and Joint Surgery from 1988 through 2000.J Bone Joint Surg Am. 2002; 84-A: 388-396
- Loss to follow-up in orthopaedic clinical trials: a systematic review.Int Orthop. 2016; 40: 2213-2219
- Validation of quantitative measures of rotatory knee laxity.Am J Sports Med. 2016; 44: 2393-2398
- Should return to sport be delayed until 2 years after anterior cruciate ligament reconstruction? biological and functional considerations.Sports Med. 2017; 47: 221-232
- Functional assessments for decision-making regarding return to sports following ACL reconstruction. Part II: clinical application of a new test battery.Knee Surg Sports Traumatol Arthrosc. 2015; 23: 1283-1291
- Young athletes cleared for sports participation after anterior cruciate ligament reconstruction: how many actually meet recommended return-to-sport criterion cutoffs?.J Orthop Sports Phys Ther. 2017; 47: 825-833
- Fifty-five per cent return to competitive sport following anterior cruciate ligament reconstruction surgery: an updated systematic review and meta-analysis including aspects of physical functioning and contextual factors.Br J Sports Med. 2014; 48: 1543-1552
- What factors influence follow-up in orthopedic trauma surgery?.Arch Orthop Trauma Surg. 2015; 135: 321-327
- Sweden health system review.Health Syst Transit. 2012; 14: 1-159
- OECD reviews of health care quality: Italy 2014: raising standards.OECD Publishing, 2014
- OECD reviews of health care quality: Japan 2015: raising standards.OECD Publishing, Paris2015
- Doctor-patient communication: a comparison of the USA and Japan.Fam Pract. 2003; 20: 276-282
- Strategies to improve retention in randomised trials.Cochrane Database Syst Rev. 2013; 12MR000032
- Evidence-based medicine: how to practice and teach EBM. 1997;
- Predictors of activity level 2 years after anterior cruciate ligament reconstruction (ACLR): a Multicenter Orthopaedic Outcomes Network (MOON) ACLR cohort study.Am J Sports Med. 2010; 38: 2040-2050
- Patellar tendon or semitendinosus tendon autografts for anterior cruciate ligament reconstruction? A prospective randomized study with a two-year follow-up.Am J Sports Med. 2003; 31: 19-25
- A randomized trial of treatment for acute anterior cruciate ligament tears.N Engl J Med. 2010; 363: 331-342
- No difference in knee function or prevalence of osteoarthritis after reconstruction of the anterior cruciate ligament with 4-strand hamstring autograft versus patellar tendon-bone autograft: a randomized study with 10-year follow-up.Am J Sports Med. 2010; 38: 448-454
- The prevalence of patellofemoral osteoarthritis 12 years after anterior cruciate ligament reconstruction.Knee Surg Sports Traumatol Arthrosc. 2013; 21: 942-949
- The prognosis and predictors of sports function and activity at minimum 6 years after anterior cruciate ligament reconstruction: a population cohort study.Am J Sports Med. 2011; 39: 348-359
- Risk of tearing the intact anterior cruciate ligament in the contralateral knee and rupturing the anterior cruciate ligament graft during the first 2 years after anterior cruciate ligament reconstruction: a prospective MOON cohort study.Am J Sports Med. 2007; 35: 1131-1134
- A historical analysis of randomized controlled trials in anterior cruciate ligament surgery.J Bone Joint Surg Am. 2017; 99: 2062-2068
Article info
Publication history
Accepted:
June 16,
2018
Footnotes
Contributors JL, JVN, NKP and ACP contributed to the acquisition, analysis and interpretation of data, and are responsible for drafting the work and revising it critically. RK, SZ, KS and VM have substantially contributed to the revision of the manuscript and organization of information. All authors have given their final approval of the manuscript to be published. In addition, all authors are in agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding This study was funded by an ISAKOS/OREF research grant (research grant no. 708661) and by Wallace Coulter Foundation contributions to the University of Pittsburgh Medical Center.
Competing interests RK reports personal fees from Medacta International and Arthrex. VM reports other (consulting) from Smith & Nephew, outside the submitted work.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Collaborators PIVOT study group authors: Andrew Sheean, Jeremy M Burnham, Jayson Lian, Clair Smith, Adam Popchak, Elmar Herbst, Thomas Pfeiffer, Paulo Araujo, Alicia Oostdyk, Daniel Guenther, Bruno Ohashi, James J Irrgang, Volker Musahl, Freddie H. Fu (Department of Orthopaedic Surgery, University of Pittsburgh, Pittsburgh, PA, USA), Kouki Nagamune, Masahiro Kurosaka, Yuichi Hoshino, Ryosuke Kuroda (Department of Orthopaedic Surgery, Kobe University, Kobe, Japan), Alberto Grassi, Giulio Maria Marcheggiani Muccioli, Nicola Lopomo, Cecilia Signorelli, Federico Raggi, Stefano Zaffagnini (Instituto Ortopedico Rizzoli, Laboratorio di Biomeccanica eInnovazione Tecnologica, Bologna, Italy), Eleonor Svantesson, Eric Hamrin Senorski, David Sundemo, Haukur Bjoernsson, Mattias Ahlden, Neel Desai, Kristian Samuelsson, Jon Karlsson (Department of Orthopaedics, Sahlgrenska University Hospital, Molndal, Sweden).
Identification
Copyright
© 2018 THE AUTHORS. Published by Elsevier Inc. on behalf of International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
