IntroductionThe arthroscopic bone block has shown reliability and reproducibility regarding functional scores and shoulder mobility compared to the open bone block technique. The recovery of muscle strength, especially the strength ratio external rotator/internal rotator (ER/IR), is crucial to recovering satisfactory function. This ratio should be as near to 1 as possible, meaning a good strength balance. Little is known about the difference in strength recovery between the open and arthroscopic techniques. HypothesisArthroscopic Latarjet reduces surgical stress and improves the strength recovery and strength ratio. ObjectivesTo compare arthroscopic and open Latarjet procedures for shoulder muscle strength and functional outcomes. Material and methodsIt was an observational longitudinal and prospective cohort follow-up. Two groups of patients were accessible for comparison: 35 in an arthroscopy group and 38 in an open group. The main outcome was the muscle strength of shoulder muscles measured with a dynamometer and expressed in Newton (N) at day 21 (D21), D45, D90, D180, and D365 in the operated and contralateral shoulders. The measurements were made for the pectoralis major, the three deltoid fascicles, and the subscapularis. The shoulder ER/IR strength ratio was calculated. Other variables were the range of motion (ROM), the Walch-Duplay (WD), the Western Ontario Shoulder Instability score (WOSI), and the Visual Analogic Scale for pain assessment. The strength, ROM, and functional scores were compared between open and arthroscopy with linear mixed models. ResultsThe median strengths at 52 weeks were significantly higher than at 3 weeks (P<10−4): anterior deltoid (AD) 8N (Q1:7, Q3: 9) versus 4N (Q1:2, Q3: 5), lateral deltoid (LD) 9N (Q1:9, Q3: 11) versus 6N (Q1:4, Q3: 7), posterior deltoid (PD) 14N (Q1:12, Q3: 15) versus 9N (Q1:8, Q3: 10), subscapularis 10N (Q1:9, Q3: 12) versus 7N (Q1:5, Q3: 8), and pectoralis major (PM) 11N (Q1:9, Q3: 12) versus 7N (Q1:5, Q3: 10). The overall strengths were lower in the open group compared to the arthroscopy group: AD −2.1N (CI95%[−3.1−−1.2], p=0.0005), LD −1.3N (CI95% [−2.4−−0.15], p=0.03), PD −0.35N (CI95% [−1–0.9], p=0.52), subscapularis −2.1N (CI95% [−3.3−−0.7], p=0.006), and PM −1.4N (CI95% [−2.2−−0.02], p=0.03). The ER/IR ratio was stable throughout the follow-up for both the operated and contralateral shoulders (p>0.5). The overall mean ratio was 1.3 (median 1.2, Q1: 1, Q3:1.45) for the operated shoulder and 1.1 (median 1, Q1: 0.9, Q3:1.3) for the contralateral shoulder (p=0.0004). The average ER/IR ratio was 0.27 points higher in the open group (CI95% [0.1–0.46], p=0.003). The ROM was similar between the two groups, and there was no correlation between the ER/IR ratio and the ROM (p>0.5). The VAS < 3 weeks and WD > 12 weeks were significantly poorer in the open group: +0.61 (CI95% [0.03–1.16] p=0.02), and −7.3 points (CI95% [−13−−0.01], p=0.05), on average, respectively. ConclusionThe patients in the arthroscopy group had a better ER/IR strength ratio (closer to 1) and better WOSI after 12 weeks. The strength and the ROM were not correlated with each other. Level of evidenceIII.