After a decade of the Delaware-Oslo ACL Cohort treatment algorithm: Measurement properties and outcomes after reconstruction or rehabilitation alone for anterior cruciate ligament injury

Abstract

Introduction: Anterior cruciate ligament (ACL) injuries are either treated surgically with ACL reconstruction (ACLR) or nonsurgically with rehabilitation alone. High-quality literature is sparse but shows similar long-term outcomes after both treatment strategies. We lack long-term studies that evaluate long-term outcomes (including knee osteoarthritis) following a clinically relevant treatment algorithm, including rehabilitation before shared decision-making about treatment. Patient-reported outcome measurements (PROMs) are recommended as primary endpoints of ACL injury, whereas knee muscle strength tests provide complementary, objective information on knee function. Clinicians and other decision-makers are challenged to interpret which PROM scores patients perceive as satisfactory. In addition, there is no consensus on which knee muscle strength tests should be used based on measurement properties. The overall aims of this dissertation were (1) to optimise the interpretability of PROM scores and evaluate the measurement properties of knee muscle strength test outcomes, and (2) to describe the 10-year outcomes of participants who followed the Delaware-Oslo ACL Cohort treatment algorithm.

Methods: This dissertation comprises four papers based on two research projects: Paper II is a systematic review whereas Papers I, III, and IV originate from a prospective cohort, the Delaware-Oslo ACL Cohort. This cohort study includes 276 young athletes with a first-time unilateral ACL injury from Delaware and Oslo (Norwegian arm). All participants followed a treatment algorithm including a five-week rehabilitation programme, patient education, and functional testing followed by a shared decision-making process for treatment (ACLR or rehabilitation alone). In Paper I, we calculated thresholds representing patient acceptable symptom state (PASS) for PROM scores for all participants at the 10-year follow-up. We used adjusted predictive modeling to calculate the thresholds and then validated their performance in identifying PASS. Paper II summarised and critically appraised the measurement properties of knee extensor and flexor muscle strength tests after ACL and/or meniscus injury. In Paper III, we described clinical (PROMs), functional (muscle strength, hop performance), sports participation, and osteoarthritis (symptomatic and radiographic) outcomes of all participants at the 10-year follow-up. Additionally, we compared outcomes between participants who either chose (1) early ACLR (within six months) with preoperative and postoperative rehabilitation, (2) delayed ACLR (after six months) with preoperative and postoperative rehabilitation, or (3) progressive rehabilitation alone. In Paper IV, we compared patient-reported outcomes between participants from the Norwegian arm who followed our treatment algorithm and those from the Norwegian Knee Ligament Registry who followed usual care at the 10-year follow-up after ACLR. PROM scores and percentages of participants who scored above the predefined PASS thresholds were compared.

Main results: Predictive modeling thresholds resulted in accurate percentages of PASS. Isokinetic concentric strength tests on computerised dynamometry had the best measurement properties, showing sufficient intrarater reliability (very low evidence). Isotonic knee strength tests using conventional weight machines showed sufficient criterion validity, while isometric knee extensor strength tests with handheld dynamometry had insufficient criterion (high evidence). The follow-up rate at the 10-year follow up was 69%. Of the entire cohort, 78% (n=126) reported PASS, 72% n=109) had symmetrical knee extensor strength, and 92% (n=162) had returned to some kind of sports. Only 1% (n=1) of participants had symptomatic osteoarthritis and 12% (n=17) had radiographic osteoarthritis. At follow-up, 60% (n=114) of the participants underwent early ACLR, 12% (n=24) delayed ACLR, and 28% (n=53) progressive rehabilitation alone. Participants who chose progressive rehabilitation alone had outcomes similar to those who chose early ACLR. Participants who underwent delayed ACLR had statistically significantly lower scores in knee function in sport and recreation, knee-related quality of life, and hop performance compared to the other two treatment groups. There were no statistically significant group differences for any other outcomes. Participants who followed our treatment algorithm had higher PROM scores and PASS percentages than those who followed usual care, except for knee-related quality of life.

Conclusion: Predictive modeling provided valid PASS thresholds for determining satisfactory outcomes at the 10-year follow-up after ACL injury. Isokinetic concentric strength tests using computerised dynamometry are most recommended, with isotonic strength tests being a good alternative. Participants after ACL injury who followed our treatment algorithm had high percentages of satisfaction, symmetrical muscle strength and knee function, high sports participation rates, and low prevalence of osteoarthritis. Participants who underwent progressive rehabilitation alone or early ACLR did equally well on all outcomes. Participants who underwent delayed ACLR reported lower knee-related function, quality of life, and hop performance compared to the other two treatment groups. Those who followed our treatment algorithm reported lower symptoms, higher knee function, and higher percentages of satisfaction compared to those who followed usual care. The Delaware-Oslo ACL Cohort treatment algorithm likely holds potential to optimise long-term outcomes after ACLR compared to usual care.