The main finding of the present study is that MIPO using locking plate supplemented with intramedullary graft is associated with a significantly shorter operative time compared to ORIF. However the functional and radiological outcomes as well as complication rate were comparable for both surgical methods.
Open reduction with internal fixation via the conventional deltopectoral approach is the most commonly used fixation technique for unstable PHFs [7, 38, 41]. The advantages of ORIF include an excellent visibility of the anterior aspect of the proximal humerus, ease in applying the LP, valuable options for extension of the exposure both distally and proximally and reduced risk of neurovascular injuries [7, 41]. On the other hand the disadvantages of open plating are related to the inevitable soft tissue stripping, the significant risk of AVN (due to damage to the ascending branch of the anterior circumflex artery) and the limited exposure of the lateral aspect of the proximal humerus [41, 42]. The MIPO technique with the deltoid splitting approach requires less soft tissue stripping and has a lower risk of injury to the anterior circumflex humeral artery, resulting in a lower rate of AVN [41, 42, 43]. The MIPO technique provides direct access to the lateral aspect of the proximal humerus, but the procedure could be technically challenging as fragments need to be reduced indirectly, especially in the medial part of the proximal humerus [41]. Another disadvantage of the technique is the possible risk of axillary nerve damage [7, 41, 43]. Regardless of the method used, LP fixation in unstable PHFs is still associated with high complication rate [44].
The use of IMG as augmentation to the LP fixation was introduced by Garder et al. in order to reduce the number of complications [17]. Biomechanical studies conclude that IMG increases stiffness and failure load of the bone-implant constructs, and limits fracture displacement, [18–22] while simultaneously allowing interfragmentary micro-movements [18, 45]. This additional rather elastic stabilization meets Lill's criteria for setting of an ideal fixation device for PHFs [46]. Overall, the available clinical studies demonstrate good or excellent functional results with a small percentage of complication when using IMG with LP for unstable PHFs [23–36].
From the available literature, most authors use ORIF when using LP and IMG fixation for unstable PHFs [23–29, 31–36] and only one author uses MIPO technique with the deltoid splitting approach [30]. To our knowledge, the current study is the first comparing ORIF and MIPO techniques in the presence of LP fixation with IMG. In a prospective, randomized controlled trial, Sohn et al. compared the two fixation techniques in unstable PHFs and found no statistically significant differences in terms of functional and radiological outcomes, but the MIPO technique provided significantly shorter operation time than open plating [41]. Another prospective, randomized study by Bhayana et al. comparing the two surgical approaches came to a similar conclusion [43]. The results of the present study further support these findings, which are also relevant in the presence of IMG augmentation. Although functional and radiological outcomes do not appear to differ significantly, a shorter duration of surgery may be beneficial for patients who have serious comorbidities or multiple fractures due to trauma [41]. One disadvantage of the MIPO technique is believed to be the potential damage of the axillary nerve which invariably lies over the holes through which the calcar screws have to be inserted [43]. LP fixation with IMG through MIPO has an additional benefit of providing robust cortical support using the intramedullary graft which has been proven to eliminate the need for calcar screws [45, 47]. The suggestion that fracture reduction with the MIPO technique is technically demanding may be supported by our results, showing that anatomical reduction is significantly more prevalent in the ORIF group. However, the latter did not seem to affect the final functional outcomes between the two groups in the current study.
The functional outcomes for the both group of the current investigation are comparable with the reported values in the literature on LP fixation with IMG in terms of DASH score [23–26, 29, 33], CSindiv and CSrel [23–26, 29, 31, 32, 36]. In addition, the change in HHH for both groups is within the reported range from 0.3 mm to 2.14 mm [24, 26–28, 31, 32], being less than the critical level of 5 mm. In the literature, the critical change of NSA varies between 5° and 10° [28, 30, 48, 49] and some authors report values between 2.6° and 3.25° [28, 30–32, 34, 36] which are superior to our findings. However, although the change in NSA reflects the loss of reduction, it’s not the most accurate criterion for a varus deformation, as some patients have such a change of more than 10° but still have NSA within physiological borders (120°-150°). We support the conclusion of Shnetzke et al.[15] that NSA between 110° and 120° still represent an acceptable varus deformation, while values below 110° are considered a real varus deformation.
The rate of AVN in the previous investigations of PHFs fixed with IMG and LP varies from 0–10.6% [17, 23–29, 31, 32], which is less than the rates reported in the current study. Besides, secondary screw penetration through the humeral head was much higher versus the previous reports (0–4%) [17, 24, 28, 31]. Another complication identified in this study and not reported in previous work on PHFs treated with LP and IMG, was SAI and AVN of the GT. We were not able to establish the cause for AVN of GT. One reason might be the potential injury of the anterior circumflex vessels due to the deltopectoral approach used in 7 of 8 patients. [46] The other potential cause might be the allograft itself as Miyamura et al. [50] pointed out that fibular cortical strut may interfere with revascularization, resulting in tuberosity resorption. In all 8 patients with AVN of the GT a fresh frozen fibular allograft was used. The complication was diagnosed on the final follow-up x-rays, but the shoulder function was seriously altered in only 2 of the patients.
Although the current study showed higher complication rate compared with the previous reports, there were no significant differences between the ORIF and MIPO groups in terms of functional and radiological outcomes as well as complications rate.
Our study has some limitations. First, the study was limited by its retrospective design, which could introduce selection bias and the potential for confounding. Second, the sample size is relatively small and futher studies with a bigger sample size are therefore required to confirm these results. Third, the number of surgeons performing the procedure could be a source for a potential bias, influencing the external validity of the study. However, all surgeons are from the same department, they all had more than ten years of experience in the management of proximal humerus fractures before the beginning of this study and the surgical technique did not involve a long learning curve. Fourth, the duration of follow-up is also relatively short. The minimum follow-up period was 12 months, which may not be sufficient to identify all negative clinical outcomes after osteosynthesis of proximal humeral fractures. However, the mean duration of follow-up in our study was 28 months, which is similar to that reported in the literature.