Comment on the article “Proximal femoral nailing is superior to hemiarthroplasty in AO/OTA A2 and A3 intertrochanteric femur fractures in the elderly: a systematic literature review and meta-analysis”

Abstract

It was with great interest that we read the article entitled “Proximal femoral nailing is superior to hemiarthroplasty in AO/OTA A2 and A3 intertrochanteric femur fractures in the elderly: a systematic literature review and metaanalysis,” published in June 2020 in the International Orthopaedics. The authors performed a meta-analysis to compare the outcomes of proximal femoral nail (PFN) and bipolar hemiarthroplasties (BPH) in treatment of intertrochanteric femur fractures (IFF) in the elderly patients (> 70 years). The authors concluded that PFN were superior to BPH for unstable IFF in the elderly in terms of functional outcomes and overall mortality. We wish to commend the authors for bringing out this novel conclusion [1]. Although we read the article with pleasure, it must be stated that this meta-analysis raises some thought-provoking issues. Therefore, we wish to have the opportunity to comment on this article respectfully. Three electronic databases (PubMed, EMBASE, and SCOPUS databases) were systematically searched by the authors. It would make the outcomes more convincing by obtaining more literature if the authors searched other databases, like BIOSIS previews, clinicaltrials.gov, and NLM Gateway. The search methodology stated that the primary search was conducted on 28 January 2018, with no limitation on the year, language, or country of publication. However, three important studies which published online prior to 2017 was missed [2–4]. This was possibly related to the improper search strategy used in the study. Identifying and aggregating as many eligible studies as possible are the keys to improving the quality of meta-analysis. In the search methodology, the authors pointed out that “with no limitation on the year, language.” In the exclusion criteria, however, they emphasized that “We also excluded articles that were not in English.” This made us very confused. Could the authors give a reasonable explanation? Meanwhile, the results of this meta-analysis based only on seven eligible studies should probably be taken into account with some caution and there are notably unclear or high risk of bias in several included studies due to poorly described methodology. Any bias inherent in the studies may exaggerate the estimated intervention effects. Furthermore, the heterogeneity of this meta-analysis needs further discussion. In this article, the heterogeneity of the groups is high, but the source of heterogeneity has not been explored. Finally, the risk of bias of this study was confusing in both methodology and evaluation. Only three randomized clinical studies (RCTs) and 4 retrospective studies were included; why did the Cochrane risk of bias tool be used to evaluate all articles? This tool was only employed to assess the quality of RCTs, but not retrospective studies. The Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) assessment tool can be used to assess the quality of retrospective studies. Furthermore, the study of Tang et al. was a retrospective study [5]; why did you think it was a randomized study? We respectfully appreciate that Prasoon Kumar et al. provided us with an important meta-analysis which can provide a guide for clinical decision-making. However, more studies * Tao Sun [email protected]