To investigate the effects of proximal humeral fracture with rotator cuff injury after hemi-shoulder arthroplasty.

Thirty-three patients (age ≥50 years) with proximal humerus fractures treated by hemi-shoulder arthroplasty in the Department of Orthopedics, Second Affiliated Hospital of Xi'an Jiaotong University from December 2008 to November 2022 were retrospectively analyzed, and patients with a history of previous shoulder disease or surgery, old fractures, and pathologic fractures were excluded. 10 men and 23 women; age were 57 to 90 years; 18 cases on the left side and 15 cases on the right side; all were traumatic fractures. According to Neer's classification, six cases were two-part fractures, 19 cases were three-part fractures and eight cases were four-part fractures; 17 patients had combined rotator cuff injuries. The first postoperative X-ray indices (acromioclavicular humeral head spacing, shoulder joint gap and glenohumeral index) and complications were also recorded, and the final visual analogue scale (VAS), Neer shoulder score and Fudan University shoulder function score (FUSS) were followed up. Measures that conformed to normal distribution were described as (±s), and those that conformed to skewed distribution were described as M(P₂₅, P₇₅). Comparisons between groups obeying normal distribution were performed using two independent samples t test and ANOVA, and comparisons of non-normally distributed data were performed using rank sum test. Fisher exact probability test was used for comparison of count data. Dependent variables were analyzed by binary logistic regression. The differences in general condition, first postoperative radiographic index and postoperative shoulder function were evaluated by different fracture types and rotator cuff injuries, respectively.

The average age of enrolled patients was(73±8) years, and the follow-up time was (67 ± 54) months on average. The final VAS score was 1 (0, 2); FUSS score was 56 (50, 74). Neer score on the affected side was (64±13) and Neer score on the healthy side was (75±11), and the ratio of affected side to healthy side was 86.3% (44%-97%) for rotator cuff injury. The ratio of affected side to healthy side Neer score was not statistically significant in patients with and without rotator cuff injury (W=250.500, P>0.05). Seventeen of 33 patients (51.5%) had concomitant rotator cuff injury, and the rates of rotator cuff injury were 50.0%, 68.4%, and 12.5% in patients with two-, three-, and four-part fractures, respectively (P=0.034). The age of patients with combined rotator cuff injury (75±8) years was older than the patients without rotator cuff injury (69±6) years (t=-2.320, P=0.027), and age [odds ratio (OR)=1.124, 95% confidence interval (CI)(1.006, 1.256), P=0.039] was an independent risk factor for rotator cuff injury. Neer score on the affected side in patients with rotator cuff injury was (58±10), which was lower than (71±12) in the patients without rotator cuff injury (t=3.212, P=0.003); the daily function and muscle strength of patients with rotator cuff injury [9 (3, 9) and 6 (6, 9)] were lower than those without rotator cuff injury [15 (9, 25) and 9 (8, 15)], respectively (W=224.00, 220.50, both P<0.05). There was no statistically significant difference in the first postoperative X-ray indices between patients with or without rotator cuff injury (all P>0.05). Postoperatively, there were two cases of dislocation, one case of periprosthetic loosening, and one case of wound infection, and none of them had internal fixation failure.

The overall reconstructive effect of hemi-shoulder arthroplasty on proximal humeral fracture combined rotator cuff injury is good, and it shows almost the same effects in postoperative pain, shoulder range of motion, and doctor-patient satisfaction as proximal humeral fracture without rotator cuff injury. However, patients often have poor recovery of daily function and muscle strength, which is also related to age-induced rotator cuff degeneration and should be taken into account in clinical work.