Physicians’ and patients’ decision-making process between autograft hamstring (HG) vs. patellar tendon (BTB) for ACL reconstruction (ACLR) may be influenced by patient gender, laxity level, sport played, and/or competition level in the young, active athlete. ACLR specific to high school and college-aged athletes with these aforementioned factors in mind has not been evaluated. Therefore, our objectives were twofold: first, to develop a simple web-based risk calculator as a decision-making aid to provide the best estimate of expected 2-year KOOS, IKDC, and Marx outcomes by gender, sport, level of competition, and knee laxity. Second, to identify whether autograft HG or BTB is the optimal graft choice given any combination of the aforementioned variables.
Our inclusion criteria in the MOON cohort were patients aged 11-22 who were injured in sport (football, soccer, basketball, other), who were due to have a unilateral primary ACLR with either an autograft HG or BTB, and who had a contralateral normal knee. Excluded were revisions, allografts, those with a contralateral ACLR and concomitant MCL/LCL/PCL surgery. Laxity was graded as increased (Lachman > 10 mm or a pivot lock) or normal based on the EUA. Our modeling controlled for BMI, ethnicity, and baseline measures of patient-reported outcomes. Our two year outcomes were the KOOS knee related quality of life subscale, KOOS sports and recreation subscale, IKDC, and Marx activity level. Our multivariable modeling for risk online calculator and nomograms was generated in two ways. The performances for our models were measured using R squared, calibration curves, and bootstrapping.
937 patients were eligible, 809 (86%) had 2 year follow-up data. The average age was 17, with 50% females, and the distribution of HG to BTB was 301/508 respectively. First, in evaluating our models for ACLR autograft choice, neither KOOS subscale models performed better than chance. The IKDC and Marx models predicted significantly better than chance. For the IKDC outcome the combined modeling strategy was preferred, and in the Marx model the separate model strategy predicted better. Second, the model results by autograft type for the two KOOS subscales, IKDC, and Marx are as follows. For KOOS quality of life the predictions for HG vs. BTB were significantly and highly correlated (0.77 P<0.001). In a scatter plot only 23 observations (9%) had a change in KOOS by 10 points. For KOOS sports/recreation the two autograft models were also significantly and highly correlated (0.62, p<0.001). In a scatter plot only 20 (7%) had a change in KOOS by 10 points. For the IKDC adjusted R2 in the combined model is 0.075 95% CI (0.026, 0.109). For the Marx (see figure scatter plot) the two autografts were significantly and highly correlated (0.78, p<0.001).
There were no significant or clinically relevant predicted differences between autograft hamstring vs. patellar tendon in the two KOOS subscales, IKDC and Marx activity level in 11-22 year old athletes .The choice between HG and BTB should be made on an outcome besides these endpoints, and there is little need for a calculator to predict these outcomes because patients will all be the same (HG vs BTB).