Comparison of autograft and allograft outcomes in adolescent ACL reconstruction: a propensity score analysis

To assess 3-year clinical and functional outcomes, complication rates and graft survival in adolescents with closed growth plates subjected to primary unilateral anterior cruciate ligament reconstruction (ACLR) with a hamstring autograft or an allograft. The failure rate reported in the literature reaches up to 20%.

We reviewed the records of 170 patients subjected to ACLR with a single-bundle autograft or an allograft from 2017 to 2020 with 3-year follow-up. Application of 1:2 Propensity Score Matching (PSM) considering age, sex, body mass index, injury to surgery interval, graft diameter and pre-injury Tegner Activity Scale (TAS) scores and Lysholm Knee Score (LKS) yielded 38 allografts and 62 autografts. We compared functional outcomes using the TAS, LKS and International Knee Documentation Committee (IKDC) at 18 and 36 months and the failure rates, defined as need for revision surgery, before and after PSM.

Mean age ranged from 15.8 (standard deviation [SD], 1.2) to 15.9 (SD, 0.9) years and the pre-injury TAS ranged from 7.3 (SD, 1.0) to 7.3 (SD, 1.2) in allograft and autograft patients, respectively. Analysis of the LKS, TAS and IKDC values demonstrated comparable levels both before and after PSM, except for the pre-matching IKDC score and the post-matching TAS score at 36 months, which were significantly higher in autograft than allograft patients (p = 0.024 and p = 0.039, respectively). As regards graft failure, before PSM significantly more allografts (12/54) than autografts (5/116) required revision surgery (p < 0.001). After matching, the difference (6/38 vs 3/62, p = 0.079) was no longer significant.

Our cohort of active adolescent who underwent primary unilateral ACLR had a higher absolute number of failures in the allograft group compared to the autograft group, emphasizing the need for careful clinical consideration. However, after applying PSM, this difference, while numerically still present, lost statistical significance. This finding may suggest that when an allograft is deemed the most appropriate choice based on patient characteristics, its use may not necessarily expose the patient to a definitively higher risk of failure.