Critical appraisal of bone graft meta-analysis in tibial plateau fractures

Abstract

We read with great interest the meta-analysis by Alshahrani et al, which compared autologous bone grafts and bone substitutes in the management of tibial plateau fractures. This comprehensive review offers valuable data; however, several methodological aspects should be carefully considered to strengthen the interpretation of its findings. Tibial plateau fractures exhibit wide variability in terms of injury mechanism and prognosis. In the included studies, fracture classification was not consistently reported, and studies on simple, mixed, and complex patterns were pooled together, which may have introduced bias. Furthermore, the bone substitutes assessed, namely, calcium phosphate cement, calcium sulfate, bioactive glass, and various composites, differ greatly in composition, mechanical properties, and degradation behavior. The primary outcome of joint depression was analyzed in very small subgroups, limiting statistical power, and was treated as continuous data despite the small sample sizes. The follow-up durations were short, although bone graft incorporation may take months to years, and shorter follow-up may miss later complications. Finally, the search strategy did not specify language criteria, which may have contributed to the inability to locate one of the included studies during our search, potentially indicating the inclusion of non-English articles. Addressing these methodological concerns would increase the robustness, validity, and clinical applicability of this important meta-analysis.

Key Words: Tibial fractures; Bone substitutes; Bone transplantation; Treatment outcome; Meta-analysis

Core Tip: Meta-analyses of tibial plateau fractures may benefit from careful consideration of fracture complexity, graft material heterogeneity, and follow-up duration when interpreting clinical outcomes. Attention to these factors could enhance the reliability and generalizability of pooled results.

TO THE EDITOR

We read with great interest and appreciation the meta-analysis by Alshahrani et al[1], which compared autologous bone grafts and bone substitutes in the management of tibial plateau fractures. The authors systematically analyzed seven randomized controlled trials involving 424 patients and reported that bone substitutes provide similar structural outcomes to autologous grafts while offering significant procedural advantages, such as reduced blood loss and shorter operative times, particularly in Asian populations.

Tibial plateau fractures are represent 1%-2% of all fractures. These fractures occurred primarily in men, and genearlly the affected individuals were younger than 40 years of age[2]. Different fracture types have been shown to yield different prognoses regardless of the treatment applied: Simple, low-energy patterns such as Schatzker type II-III or AO type B fractures are generally associated with better functional outcomes (higher KOOS and HSS scores), shorter hospital stays, and lower complication rates, whereas complex, high-energy patterns such as Schatzker type V-VI or AO type C fractures tend to result in worse functional recovery, longer operation times, greater muscle atrophy, and a greater risk of posttraumatic osteoarthritis on follow-up radiographs[3]. This emphasizes the importance of considering fracture classification and energy level when interpreting pooled outcome data in meta-analyses. Therefore, future meta-analyses should stratify or reclassify cases by fracture type and injury energy, as these factors may interact with graft material performance and influence pooled outcomes.

The bone substitutes used in the included studies included various materials such as calcium phosphate cement, calcium sulfate, bioactive glass, and different composite formulations. Several studies have demonstrated that bone substitute materials developed in different eras exhibit markedly different mechanical properties. For example, injectable calcium phosphate cement composites with bioactive glass additives have shown substantially improved compressive strength compared with traditional cpcs, and newer bioactive glasses or sol-gel-derived materials often provide better control of porosity, different degrees of crystallinity, and faster or slower degradation depending on their formulation[4-6].

Although the authors stated that their literature search covered studies from inception to April 20, 2025, no lower time limit was specified. This is important because the composition, manufacturing techniques, and mechanical properties of bone substitute materials have evolved substantially over the years, and newer-generation products may perform differently from older products. For example, in the study by Russell et al[7], although published in 2008, patients were operated on between 1999 and 2002 using endothermic calcium phosphate cement, whereas Hofmann et al[8] used biphasic hydroxyapatite combined with calcium sulfate cement. These are fundamentally different materials developed in different eras, which may confound the pooled analysis by introducing heterogeneity related to technological progress over time. For example, Pizzoli et al[9] demonstrated that in more complex Schatzker type V-VI fractures, the use of β-TCP/PMMA (Cal-Cemex) allowed early weight-bearing and yielded satisfactory functional outcomes. These findings suggest that treatment results may vary according to the class of bone substitute (e.g., CPC, β-TCP, HA) and fracture type, thereby underscoring the need for subgroup analyses in future meta-analyses.

One of the key concerns is the inclusion of studies with tibial plateau fractures for which the fracture classification was not reported. Furthermore, studies involving simple fractures such as Schatzker types II-III fractures were analyzed together with studies including mixed or complex fracture patterns. Type II-III fractures are typically low-energy injuries, whereas complex fractures often result from high-energy trauma and differ substantially in their structural, mechanical, and supportive characteristics. Combining studies with unknown classification together with simple, complex, and mixed fracture patterns may introduce bias and limit the validity of the pooled estimates.

In the present meta-analysis, the primary outcome of joint depression was evaluated in only a very limited number of cases with known fracture types-20 simple fractures (9 treated with autologous grafts and 11 with bone substitutes) and 25 complex fractures (11 autologous, 14 bone substitutes). Such small subgroups would be considered insufficient to provide reliable statistical power even in a single-center study, let alone in a meta-analysis. Moreover, the authors appear to have analyzed this outcome as continuous data; however, given the very small sample sizes, a categorical comparison (e.g., displacement present vs absent) with a χ² or Fisher’s exact test would have been more appropriate. The conclusion that there is “no difference” between autologous grafts and bone substitutes on the basis of such limited numbers seems overly assertive and should be interpreted with caution.

The follow-up duration varied across studies, ranging from six months to 24 months. However, considering that the primary outcome of this meta-analysis was joint depression, such follow-up periods are relatively short. Additionally, the time required for bone grafts to become organized and fully integrated into the host bone varies depending on the type of material used, its biological properties, and patient-related factors, and may range from several months to several years[10]. For studies of this nature (other than case reports), at least one year of follow-up is generally needed. Both the patient’s return to preinjury functional status and the biological incorporation and remodeling of the implanted graft typically take considerable time to be reliably assessed. Shorter follow-up periods may miss later complications or structural changes.

Previous methodological studies have shown that over 90% of orthopedic trauma randomized control trial are underpowered, with small sample sizes leading to high type II error rates and false equivalence between treatments, whereas nearly one-third of trials with nonsignificant findings are inadequately powered and prone to bias, particularly when follow-up is too short to capture late complications or functional decline[11,12]. Together, these findings underscore that underpowered design and limited follow-up can distort meta-analytic conclusions and should be explicitly recognized in any appraisal of fracture-healing evidence.

Another important point is that the search strategy and study selection section does not specify the languages of the publications considered eligible for inclusion. This should have been explicitly stated, as language restrictions can substantially influence study selection and reproducibility. This lack of clarity may also explain why one of the included studies Chen et al[13] could not be located through English-language searches. Clearly, defining the language criteria would enhance the transparency and verifiability of the review process.

We appreciate the authors’ valuable contribution to the literature, and we believe that addressing these methodological points may further strengthen the interpretation of their findings and guide future research on this important topic.