Proximal humerus fractures (PHF) are the third most common fractures in the elderly. Treatment of complex PHF has remained challenging with mechanical failure rates ranging up to 35% even when state-of-the-art locked plates are used. Secondary (post-operative) screw perforation through the articular surface of the humeral head is the most frequent mechanical failure mode, with rates up to 23%. Besides other known risk factors, such as non-anatomical reduction and lack of medial cortical support, in-adverse intraoperative perforation of the articular surfaces during pilot hole drilling (overdrilling) may increase the risk of secondary screw perforation. Overdrilling often occurs during surgical treatment of osteoporotic PHF due to minimal tactile feedback; however, the awareness in the surgical community is low and the consequences on the fixation stability have remained unproved. Therefore, the aim of this study was to evaluate biomechanically whether overdrilling would increase the risk of cyclic screw perforation failure in unstable PHF.
A highly unstable malreduced 3-part fracture was simulated by osteotomizing 9 pairs of fresh-frozen human cadaveric proximal humeri from elderly donors (73.7 ± 13.0 ys, f/m: 3/6). The fragments were fixed with a locking plate (PHILOS, DePuy Synthes, Switzerland) using six proximal screws, with their lengths selected to ensure 6 mm tip-to-joint distance. The pairs were randomized into two treatment groups, one with all pilot holes accurately predrilled (APD) and another one with the boreholes of the two calcar screws overdrilled (COD). The constructs were tested under progressively increasing cyclic loading to failure at 4 Hz using a previously developed setup and protocol. Starting from 50 N, the peak load was increased by 0.05 N/cycle. The event of initial screw loosening was defined by the abrupt increase of the displacement at valley load, following its initial linear behavior. Perforation failure was defined by the first screw penetrating the joint surface, touching the artificial glenoid component and stopping the test via electrical contact.
Bone mineral density (range: 63.8 – 196.2 mgHA/cm3) was not significantly different between the groups. Initial screw loosening occurred at a significantly lower number of cycles in the COD group (10,310 ± 3,575) compared to the APD group (12,409 ± 4,569), p = 0.006. Number of cycles to screw perforation was significantly lower for the COD versus APD specimens (20,173 ± 5,851 and 24,311 ± 6,318, respectively), p = 0.019. Failure mode was varus collapse combined with lateral-inferior translation of the humeral head. The first screw perforating the articular surface was one of the calcar screws in all but one specimen.
Besides risk factors such as fracture complexity and osteoporosis, inadequate surgical technique is a crucial contributor to high failure rates in locked plating of complex PHF. This study shows for the first time that overdrilling of pilot holes can significantly increase the risk of secondary screw perforation. Study limitations include the fracture model and loading method. While the findings require clinical corroboration, raising the awareness of the surgical community towards this largely neglected risk source, together with development of devices to avoid overdrilling, are expected to help improve the treatment outcomes.