The ideal site of cement application in cement augmented sacroiliac screw fixation: the biomechanical perspective

In older adults fractures of the pelvis are mainly caused by physiological loads during activities of daily living, induced by the patient’s own body weight or by low-energy trauma, in case of an extreme reduction in bone mass (e.g. severe osteoporosis) [1‐4]. Moreover, older people are at risk for simple falls from standing positions: Thirty percent of the  people over 65 years of age and 50% of those over 80 years of age fall at least once each year, and older adults who fall once are likely to fall again within one year [5]. A simple fall can result in a fragility fracture of the pelvis (FFP) [3].

Once osteoporotic pelvic fractures have occurred, the destabilized pelvis causes a sudden onset of immobilizing pain. The loss of mobility determines the length of hospital stay and the discharge type; patients with pelvic fractures are representative for long hospitalization periods [6]. The amount of independently living patients decreased from 89% (before hospital admission) to 64% (at the point of discharge) leading to a loss of autonomy [7]. Finally, pelvic fractures in older adults with increased mortality rates are reported to range from 13 to 27% one year after the injury [8‐11]. In summary, these fractures are indubitably severe injuries in older adults.

Decision making for treatment is guided by the inability of the pelvic ring to withstand physiological loads without displacement as determined by the degree of instability of various fracture types [3, 12]. The majority of FFPs present moderate instability caused by a non-displaced fracture of the posterior pelvic ring with or without anterior ring injuries (FFP Type II; e.g. FFP IIB, Fig. 1). In these injuries, the use of percutaneously placed sacroiliac screws with low access morbidity has been recommended for fixation [3]. However, a diminished screw purchase of sacroiliac screws as a monocortical screw device in cancellous, osteoporotic bone is a concern [13, 14]. Therefore, cement augmentation of sacroiliac screws at the level of the body of S1 vertebra was proposed to increase the anchorage between bone and implant and was a matter of preclinical research [15‐20]. However, neither an advantage with regard to cyclic loading, nor a distinct clinical benefit of sacroiliac screw augmentation was evident in recent systematic reviews [21, 22] whereas shortened length of hospital stay and a reduced risk for general complications was noted in sacroiliac screws augmented with bone cement [23]. In contrast, the lateral sacral mass was identified clearly as the weakest part of the sacrum previously in cadavers of older adults [4, 24]. Accordingly, the sacral lateral mass serves as an optimal site for cement augmentation of sacroiliac screws. However, to our knowledge no study assessed this site for cement augmentation of sacroiliac screws. Thus, we hypothesize that cement augmentation of sacroiliac screws facilitated at the weakest part (lateral mass of S1) enhances construct stability compared to current practice (body of S1) and previous research. The aim of this study was to compare two different sites for cement augmentation of sacroiliac screws in their fixation strength in a biomechanical pelvic fracture model exposed to cyclic loading.

Morphometrically, BMD in the massa lateralis was homogeneously distributed between the sites having the screw tip (90.7 ± 12.5 mgHA/cm³) or the midshaft (81.1 ± 15.6 mgHA/cm³) augmented, p = 0.267. Screw augmentation was associated with cement leakage into the first neuroforamen following screw tip augmentation (group A) in one case and in all cases in the fracture gap but not into the true pelvis in group B. The initial axial construct stiffness was comparable for screw tip (37.1 ± 7.7 N/mm) or midshaft (39.5 ± 7.1 N/mm) augmentation, p = 0.729.

All outcome measures analyzed over the five time points after 10,000, 20,000, 30,000, 40,000, and 50,000 cycles are summarized in Table 1. Complementary continuous values over the first 50,000 cycles are shown for (i) “gap angle” (Fig. 5), (ii) “screw tilt ilium” (Fig. 6), and (iii) “screw tip cutout” (Fig. 7). For the pooled data sets over the five time points, relative interfragmentary movements in terms of “gap angle” were associated with significantly higher values for group A (screw tip augmentation; median [CI] 2.4° [2.3, 4.9]) versus group B (midshaft augmentation; median [CI] 1.4° [1.0, 1.6]); p < 0.001). Similarly, the pooled outcome measure “screw tilt” was associated with significantly higher values following instrumentation of group A (median [CI] 3.3° [3.0, 8.9]) versus group B (median [CI] 1.4° [1.2; 2.9]; p = 0.003). However, for “screw tip cutout” no significant difference was obtained between group A (median [CI] 0.6 mm [0.6, 1.3]) and group B (median [CI] 0.8 mm [0.8, 1.8]; p = 0.376). Finally, comparison of the reproducibility revealed significantly higher pooled SD's for group A (median [CI] 1.2 [0.9, 4.6]) versus group B (median [CI] 0.5 [0.4, 1.4]; p = 0.017).

Pelvic (fragility) fractures resulting from low-energy trauma are increasingly frequently observed in patients older than 60 years of age [33]. Moreover, the incidence of pelvic fractures in elderly people (> 80 years) rose from 73 to 364/100.000 individuals from 1970 to 2013 [34]. Consequently, a 127% increase of the total number of hospitalizations following pelvic fractures in older patients (≥ 65 years) between 1986 and 2011 has been reported [35]. A 2.4 fold increase for the occurrence of fractures is predicted for the year 2030 [34] making fragility fractures of the pelvis an increasingly important public health care issue and economic burden [36].

Fragility fractures of the pelvis with moderate instability (FFP IIB) account for the majority of these injuries and surgical treatment with sacroiliac screws is the clinical practice. However, failure rates of sacroiliac screws occurred in 14% to 17% of patients [37, 38]. Accordingly, cement augmentation of sacroiliac screws at the level of the body of S1 vertebra was proposed to increase the anchorage between the bone and the implant and was a matter of preclinical research [15‐20]. In addition, a 25% reduced time from surgery to discharge was noted following cement augmentation [23]. However, there are still some controversies and open questions in the existing literature with regard to the efficiency of cement augmented sacroiliac screws [21, 22].

To our knowledge, the site for cement augmentation of sacroiliac screws in FFP IIB fractures providing the highest construct stability has not been defined clearly yet. We hypothesized that cement augmentation of sacroiliac screws at the weakest part (lateral mass of S1) enhances construct stability compared to current practice (body of S1). The rationale for that hypothesis and the presented study is given by previous consistent findings as follows: The lowest bone quality was noted in the lateral portions of S1 which contain yellow marrow and an alar void, associated with minimal regional bone density [24, 39‐41]. To that effect, sacral stress fractures occur at this site if the applied shear stresses exceed the mechanical strength of the sacral ala, e.g. due to an increased strain caused by osteoporotic defects in this region [42, 43]. Therefore, medial cement augmentation (vertebral body of S1) at the tip of the screw will not stabilize the lateral alar void defect.

The presented biomechanical analysis with cyclic loading demonstrates that augmentation of the lateral mass compared to the one in the center of the vertebral body of S1 results in less angular displacement between the sacral fragments medial and lateral to the fracture, as well as between the screw and the pelvic bone itself. In summary, the screws with a cement-augmented lateral mass provided a construct being more stable under cyclic loading.

The results confirm the hypothesis of the study, as the cementation at the weakest part of the sacrum enhanced stability. This finding might parallel the noted so-called “windshield-wiper effect” in reconstruction surgery of the anterior cruciate ligament (ACL) as follows: Transverse motion of the ACL graft in the proximal tibial bone tunnel results in widening of the osseous tunnel at risk for failure of ACL reconstruction surgery [44, 45]. Accordingly, the cement augmentation at the tip of the sacroiliac screw might not reduce the long lever arm of the screw with its poor anchorage in the sacral alar. Thus, transverse motion might result in loosening of the screw with backing out. This mechanism might be a potential explanation for the failure of the sacroiliac screw, however, needs to be assessed in other biomechanical experiments. Finally, the technique of cement augmentation at the region “from-screw-head-to-midshaft” rather than at the level of the screw tip has been introduced successfully for management of odontoid fractures in osteoporotic bone earlier [46].

Another explanation might be that the osteosynthesis in these pathologic fractures acts comparably with a compound osteosynthesis [47]. Using this technique, the healing of the pathological fracture is rather not the main purpose (although possible) of the treatment than the durability of the osteosynthesis providing long term stability. The management of FFP in osteoporotic fractures might be comparable as fracture healing is uncertain and the main treatment goals are prevention from fracture progression and the so-called “fracture disease” [12, 48].

A limitation of the study might be concerns about the clinical application of cement injection at the sacral lateral mass using fully threaded screws with the potential risk for cement leakage in the fracture gap and/or no compression at the fracture site preventing fracture healing. However, the following arguments might abolish these concerns: (1) Cement leakages are observed (e.g. as “retrograde” leakage) regardless of the site of cement application [21], (2) surgeons rather should be aware of the physical properties of the cement and the “bone permeability”[49, 50] and using a balloon guided cement augmentation was successful in leakage prevention [51], (3) using sacroplasty for sacral insufficiency fractures appears to be safe and efficient in current available literature, however, controlled randomized clinical trials are lacking [52], (4) the need and the surgical feasibility to obtain compression at the fracture site is questionable as (i) the FFP IIB fracture is a result of (already occurred) lateral compression, (ii) from a surgeon`s perspective there is the risk for stripping of the screw or cut-through of the washer through the weak innominate bone, (iii) a recent biomechanical study did not reveal any differences in four out of nine cadavers in a compression test analyzing fully versus partially threaded screws even in rather young (mean age 68 years), male donors (atypical patients in terms of FFP) whereas an overall significant difference of 1 mm in fracture site compression (questionable benefit versus risk for screw stripping) was reported in favor of partially threaded screws [53], (5) cement augmentation at the weakest site with filling the void is a technique known as “compound osteosynthesis” in pathological fractures of extremities in patients with limited life expectancies without claiming for fracture healing [47]. In summary, there are many arguments to repeal these concerns above. Another limitation might be the low number of specimens and the biomechanical setup per se. However, there is a limited availability of fresh-frozen human cadavers; the experimental setup was designed to detect differences between two study groups to provide evidence for a design modification and the rationale to reconsider current clinical practice. Potential further limitations of the present study arise from the nature of the biomechanical protocol, e. g. the resection of the symphysis does not reflect clinical reality. However, the rationale to resect the symphysis was to obtain a reproducible setup for both sides allowing for internal rotation of both hemipelvises in the horizontal plane. For placement of the sacroiliac screw no insertion guide was used. However, CT scans were performed after screw placement and ruled out any relevant side-to-side differences in screw positioning and further confirmed correct intraosseous screw placement. In addition, we developed a preoperative planning approach for sacroiliac screw placement [54], which supports reproducible screw positioning and procedures were performed by a senior consultant surgeon experienced with this technique.

The strengths of the study are: (1) we used fresh-frozen human cadavers to reflect the clinical situation rather than synthetic bone models without any ligaments attached to the bone, (2) to ensure homogeneity and comparability of these specimens clear inclusion criteria were defined and analyzed using CT scans; including a phantom to detect differences in bone mineral density, (3) to allow for direct comparison between the two groups, both techniques were tested within the same bone quality within one cadaver, (4) a FFP IIB fracture was tested by generating “a posterior undisplaced sacral crush injury with an anterior ring disruption” to reflect the most frequent fracture pattern, (5) for fixation, we used fully threaded screws as these provided more stability in osteoporotic bone by improved anchorage and by fixation within the cortical bone of the sacroiliac joint; accordingly, we obtained a high fixation strenght at baseline before application of cement, (6) all osteosynthetic constructs were tested under cyclic loading to reflect the physiological loading as much as possible with shear stresses concentrated at the sacral lateral mass.

The presented study demonstrates less fragment and screw displacement in a FFP IIB fracture model under physiologic, cyclic loading by cement augmentation of sacroiliac screws at the level of the lateral mass compared to the center of vertebral body of S1 for the first time. The study results provide the rationale to assess the potential benefit of this technique in a clinical setting.