Purpose:

This study sought to analyze factors that predict postoperative shoulder balance based on clinical photography.

Methods:

Based on inclusion criteria, 132 AIS patients were selected. Age, sex, and BMI of each patient were recorded. The following parameters were recorded from radiographs: clavicle angle, T1 tilt, the upper instrumented vertebra (UIV), lowest instrumented vertebra (LIV) thoracic kyphosis, lumbar modifier, preoperative and postoperative proximal thoracic Cobb angle, preoperative and postoperative main thoracic Cobb angle, and preoperative and postoperative thoracolumbar Cobb angle, if applicable. Two spine surgeons independently assigned the photographs shoulder balance grades based on the WRVAS (1-2 = Acceptable, 3-5 = Unacceptable). Surgeons were blinded as to whether the photographs were taken preoperatively or postoperatively. The shoulders were also graded as right high, left high, or balanced.

Results:

Of all variables analyzed, only main thoracic Cobb angle correction (MTCAC) showed a statistically significant relationship with postoperative shoulder balance (p = 0.01). Odds of having unacceptable shoulder balance increase by 21% for every 5° increase in MTCAC (Adjusted OR = 1.21, 95% CI 1.015-1.452). The odds of unbalanced shoulders are 4.7 times higher for patients whose MTCAC is 40° or more (p = 0.001). Inter-rater reliability was excellent (k =0 .7). Intra rater reliability was perfect for Surgeon 1 (kappa = 1.0) and showed substantial agreement for Surgeon 2 (kappa = 0.8) CONCLUSIONS: Greater correction of main thoracic Cobb angle predicts unacceptable postoperative shoulder balance with 40° of correction signifying a major dichotomy between acceptable and unacceptable.