**3. Results**

The average head and trunion size in the deltopectoral group was 42, with a range from 39 to 47; and the mean size in the PI group was 41, with a range from 39 to 43. In the deltopectoral group, six small, two medium, and one large screw were implanted, while in the PI group, three small, four medium, and two large screws were used. In both groups, five small, three medium, and one large glenoid component were trialed.

While performing the PI approach, the identification of the internervous plane was poor in four cases, acceptable in three cases, and excellent in two cases. The average distance of the internervous plane to the axillary nerve was 33 mm, with a minimum distance of 25 mm and a maximum distance of 45 mm. The posteroinferior dislocation of the humeral head through the internervous interval and inferior to the posterior deltoid was possible in all cases. The exposure of the humeral head was excellent in five cases, acceptable in three cases, and poor in one case. The possibility of resection of the humeral head was excellent in three cases, acceptable in three cases, and poor in three cases. In all cases, with a medial T-shaped incision of the capsule instead of a lateral T-shaped incision, an acceptable or excellent exposure and resection opportunity were identified. The exposure of the glenoid and simulated reaming was excellent in four cases, acceptable in three cases, and poor in two cases. The surgeon satisfaction rating with the procedure when performing the PI approach displayed a learning curve with a positive impact of the switch from a lateral T-shaped to a medial T-shaped incision of the capsule between cases 3 and 4 (Figure 8).

The intraclass correlation coefficients for the radiographic measurements of the two independent observers was almost perfect for six parameters, substantial for two parameters, moderate for two parameters, and fair for two parameters (Table 2).

While the COR was positioned significantly (*p* = 0.031) more medial in the PI approach group (3.7 ± 3.4%, range: −2.3% to 8.7%) than in the deltopectoral approach group (−0.2 ± 3.6%, range: −6.9% to 4.1%) on average, none of the remaining measured radiographic parameters were significantly different between both groups (Table 3).

**Figure 8.** Change in surgeon's satisfaction from the first case to the ninth case when performing total shoulder arthroplasty via a posteroinferior approach. Between cases 3 and 4, a switch from a lateral T-shaped to a medial T-shaped incision of the capsule was made.



**Table 3.** Comparison of the postoperative radiographic parameters between the anatomical arthroplasties performed via a deltopectoral approach and a postero-inferior approach.



**Table 3.** *Cont.*
