The most important finding of this retrospective study is that the human allogeneic cortical bone screw is totally transformed into the host bone after several months [12]. This allows us to conclude that after this time, the scaphoid looks like as never broken before (Fig. 4C) and as also as shown by recent publications [9–12]. Three of the four nonunions in our retrospective study were proximal pole fractures; excluding these, as done in the SWIFFT study [14], would have improved our union rate result to 98.7%. The results are in agreement with findings by Jaminet et al. [20], who described nonunion rates of 3% for the middle third and 18% for the proximal pole in pseudarthrosis patients. Others describe 14% [32] or 30% [22] nonunion when treating the proximal pole in pseudarthrosis patients. For pseudarthrosis patients in our patient cohort, the union rate was 96%, in agreement with results reported in the literature: 84–90% when using vascularized grafts [17, 36] and between 80% and 87% when using non-vascularized grafts. Hegazy et al. [37] described union rates of 90% and 94% for corticocancellous and cancellous-bone-only autografts, respectively, in pseudarthrosis patients.

Additionally, this retrospective study demonstrated a high union rate (94–96%), especially in pseudarthrosis patients (96%), when using the human allogeneic cortical bone screw for fixation. The two nonunions (4%) in the pseudarthrosis patients were due to non-compliance and a fracture gap bridging that was too short (Fig. 7). Median time to union for the pseudarthrosis group was 18 weeks, as described by Pinder et al. [17]. Mean time to union was prolonged in the presented study due to the coronavirus pandemic, as patients omitted follow-ups and union could only be recorded after the pandemic restrictions were canceled, at which point it was far beyond the normal time to union. Another factor for the long time to union could be the high number of patients with a classification of D2 or higher (17%), who show a longer time to union. This is in agreement with Higgins et al. [38], who described fracture location, duration since injury, previous nonunion, smoking and fracture characteristics (other than location) as the main preoperative factors which impact the success of scaphoid nonunion surgery.

Merrell et al. [28] described lower (47%) union rates with a non-vascularized wedge graft. Aibinder et al. [30] observed a 71% union rate for structural iliac crest [30]. Using autologous bone, taken from the tibia of the patient and shaped intraoperatively, to fix the scaphoid fracture with an autologous bone screw yielded a union rate of 80% with a high complication rate and long surgery time [6].

In the presented study, we recorded only one complication in 75 patients (1.3%, complex regional pain syndrome) and arthrosis at the last follow-up in 2 patients where union could not be achieved. This favorable outcome is related to the use of the human allogeneic cortical bone screw, which avoids donor site morbidity and problems related to metal hardware. Complication rates were reported to be between 21 and 26%, depending on the graft used [23, 37, 39]. The meta-analysis by Feeley et al. [2] regarding the use of bioabsorbable fixation methods revealed 23% complications, and the SWIFFT study reports 14% in the surgical group [14]. Reigstad et al. [24] described increased arthrosis around the implanted screw.

The union rate in the fresh fracture group (94%) was similar to values described by Dias et al. [14], even though they included only non-displaced or minimally displaced patients and excluded patients with proximal pole fractions. Median time to union for a fresh fracture group as described by Pinder et al. [17] was 16 weeks, with 18 weeks and 4 months described by Tada et al. [40], who mainly investigated patients who had their surgery between 6 weeks and 6 months after the injury. The percentage of unions in the fresh fracture group after 100 days (14 weeks) was 38% and was similar to results presented by the SWIFFT study after 12 weeks (47% achieved full union, 68% achieved almost full union [14]).

Brcic et al. [8] describes the human allogeneic cortical bone screw as a sterilized cortical bone allograft for osteosynthesis. Furthermore, they describe that the human allogeneic cortical bone screw has a fine thread with high rotational stability and accuracy of fit, and it offers the ideal conditions for the migration, proliferation and differentiation of all bone cells [8]. One aspect to consider with every application of the Shark Screw® is that bone grafting means bone reconstruction. The original bone is reconstructed. This cortical homoplasty stimulates the environment for bone renewal. The osteogenic response is clearly visible around the Shark Screw® in the CT. This is particularly important in the case of pseudarthrosis when bone defects with bone loss (cysts, necrosis) have occurred. In order to fill these bone defects with cortical bone in the best feasible way (cortical homoplasty), the largest possible diameter of the Shark Screw® should always be selected, even if the biomechanical conditions would allow a thinner screw. This allograft is remodeled without scars by the physiological bone metabolism and, on average, within a year, it is converted into the patient's own bone [8, 10, 12] (Fig. 4C).

Time between injury and surgery was reported as 7–50 (mean 14.8) weeks [40], similar to the median for our patient cohort (same day to 63 weeks).

Donor site morbidity, as described in the literature [6, 19, 41], can be avoided by using the human allogeneic cortical bone screw. Distal radius bone graft donor site complications are underreported or may be attributed to the recipient site because of their close proximity [17]. In our experience, and in agreement with the literature [6], it takes about an extra 30–45 min to prepare autologous cortical bone.

Hardware removal is another problem which is avoided by using the human allogeneic cortical bone screw. In 17% [41] to 67% [42] of the patients, the hardware had to be removed. Keller et al. [23] reported implant irritation which healed only after implant removal. Biomaterials are one alternative, but these materials and their degradation products may interfere with bone healing [2]. 21,000 hardware removals were performed in Austria in 2016 (personal communication, Gebietskrankenkasse Oberösterreich). Each hardware removal costs 2350.62 € (2016), the patient stays on average 3.5 days in the hospital, and 1 hospital day costs 923 € (2019). In Upper Austria, sick leave is a minimum of 15 days after hardware removal (and can be up to 35 days), with a daily income of 140 € (source of all data: personal communication, Gebietskrankenkasse Oberösterreich). In total, this comes to 7681.12 € per hardware removal. Estimating that 1% of the hardware removals in Austria were due to hardware removal after scaphoid fixation, this accumulates to 1,613,035.20 € per year, which could be saved by using the human allogeneic cortical bone screw. These are only the costs of hardware removal and sick leave; they do not include the time saved in the operation theater at the initial operation and other costs, such as those of additional visits and medication. In a very recent publication, Andersson et al. [15] described that the frequency of settled claims of scaphoid fractures was 20% higher than for general complaints [15]. Furthermore, they found that 56% of the claims were due to delayed diagnosis and 41% were due to misdiagnosis [15]. In 60% of the patients with a settled claim, it resulted in medical invalidity [15]. They further described that total costs of 15,700 € per patient (indirect and direct costs) were calculated for misdiagnosed and mistreated patients [15].

Langegger [43] describes a future perspectives on the use of the human allogeneic cortical bone screw, Pastl et al. [9] reported the first results obtained with the cortical bone screw in hand and foot surgery, and Amann et al. [10] and Hanslik-Schnabel et al. [12] described results of using the screw in foot surgery and Huber et al. [11] in shoulder surgery.

There are several limitations of this study. First, the retrospective nature of the study. Second, due to the explorative nature of the study and ethical concerns, no control group, using a metal screw, instead of the cortical bone screw was available for comparison. Third, no scores were available to interpret the results. Further prospective studies are needed to confirm these preliminary findings.

In conclusion, using the human allogeneic cortical bone screw (Shark Screw®) led to similar union rates to those presented in the literature for other scaphoid fracture fixation techniques and enabled a short and low-invasive procedure without any donor site morbidity and without the necessity to remove the hardware in a second surgery. The pseudarthrosis patient group received a particularly strong benefit from this new procedure. The physiological bone metabolism remodels the cortical bone screw without scars.