Anterior Cruciate Ligament Graft Removal Versus Retention in the Setting of Septic Arthritis After Reconstruction: A Systematic Review and Expected Value Decision Analysis

Abstract

Purpose: To provide further guidance on the optimal decision between anterior cruciate ligament (ACL) graft removal versus retention in the setting of septic arthritis following reconstruction using an expected value decision analysis.

Methods: A systematic review and expected value decision analysis and sensitivity analyses were performed to quantify the clinical decision. A decision tree was created with 5 outcomes of interest: nonoperative complications, revision surgery, early reoperation, late reoperation, and "well." Pooled probabilities of each outcome were generated through a systematic literature review. We included only peer-reviewed studies, published in English, with at least 6 months of follow-up. One hundred randomly selected volunteers were given descriptions of the clinical scenario, the 2 treatment options, and outcomes of interest. Patients younger than 18 and older than 50 years and those previously treated for either ACL injury or septic arthritis, or both, were excluded from the analysis to minimize bias. These hypothetical patients indicated preferences for each outcome on a visual analog scale and responses were averaged to generate overall "utility values." Fold-back analysis summed products of pooled outcomes probabilities with respective averaged utility values. The resulting overall expected values for graft removal and debridement were compared, with the highest expected value considered to be superior. We then performed 1-way sensitivity analyses to mitigate sample bias.

Results: Fold-back analysis revealed graft removal to be strongly favored over retention, with overall expected values of 17.2 and 8.64, respectively. The most important contributor to the difference in overall expected values was late reoperation (8.59 vs 2.50 for removal and retention, respectively). Despite adjustments made to the rates of revision and early reoperation during the 1-way sensitivity analyses, graft removal remained the optimal strategy.

Conclusions: This expected value decision analysis revealed that ACL graft removal was strongly favored by patients over graft retention in the setting of postoperative septic arthritis when consideration was given to the probabilities of wellness, nonoperative complications, revision surgery, early reoperation, and late reoperation. Sensitivity analysis revealed that although variation in rates of other outcomes did not impact this preference, the rate of late reoperation had a substantial impact. Only a sizable increase in the probability of late reoperation (from 0% to 60%) after graft removal would cause potential patients to favor graft retention.

Level of evidence: Level IV, systematic review and decision analysis.