In the series of Brislin et al. [3], published in 2007, the most common complication was recognized as persistent stiffness; failure of healing, infection, reflex sympathetic dystrophy, deep venous thrombosis and death were considered as less frequent complications.

The failure rate, reported in the literature, after rotator cuff repair varies from 13 % [4] to 94 % [5]. These figures are important because they highlight the problem, but raise questions about their actual value because they are strongly affected by: (a) number of examined patients, (b) original size of the lesion, (c) patient age, (d) fatty degeneration of their muscles, (e) acromio-humeral distance, (f) patient habits, (g) gleno-humeral arthritis [6–10].

In 2011, Chung et al. [11] stated that bone mineral density, as well as fatty infiltration of the infraspinatus and amount of retraction, was an independent determining factor affecting post-operative rotator cuff healing.

In the past, many patients underwent surgical cuff repair although they had important predisposing factors to tear and consequently to re-tear (advanced age, diabetes, hypercholesterolemia, metabolic syndrome, smoking and alcohol habit). For these patients, failure is almost the rule. Therefore, error does not consist in the used surgical technique for repair, but in the indication for surgery.

In 2013, Iannotti et al. [12] observed that re-tears primarily occur between 6 and 26 weeks after arthroscopic rotator cuff repair, and few additional tears occur thereafter. A substantial number of re-tears occur between 12 and 26 weeks after repair. Two years before, Miller et al. [13] had reached the same conclusions, although they had examined a lesser number of patients.

At the beginning of the learning curve, the re-tear can be the consequence of a wrong repair technique: (a) incorrect repair construct/procedure, (b) not tensioned knot, (c) incorrect number of used anchors, (d) reduced distance between two anchors resulting in the possibility to create a single large hole in the greater tuberosity that reduces the tightness of both anchors.

An incorrect positioning of the anchor may be the cause of subsequent failure. The angle of incidence to the bone is crucial. Anchors are ideally placed with a deadman angle, as described by Burkhart [14, 15], of less than 45°. If the insertion angle is too vertical, it will enter the softer bone of the greater tuberosity rather than the dense subchondral bone of the humeral head, increasing the risk of anchor pull-out (Fig. 2a–c). Instead, if the insertion angle is too horizontal (Fig. 3a–d) there is the possibility that the anchor protrudes from the articular surface and the suture is significantly weak. Benson et al. [16] believe that there is a minimal risk of suture anchor pull-out in small- to medium-sized tears; however, this risk increases with larger tear sizes. According to Kirchhoff et al. [17], placement of suture anchors in a medialized way at the border to the articular surface might guarantee a better structural bone stock.

Many patients over 60 years of age have dystrophic changes or pseudo-tumours in subcortical bone of the greater tuberosity (Fig. 4a–c). Obviously, in this area the holding capacity of the infixed anchor is decreased and the possibility of a pull-out in the intra-operative or in the immediate post-operative period is high.