Importance of the Topic

More than 1.5 million patients suffer hip fractures worldwide each year [3], and annual direct and indirect costs are projected to exceed USD 130 billion globally by 2050 [6]. Thirty-day mortality approaches 5% in males and 9% in females, and many more patients experience significant functional loss and debilitation [4]. Approximately 50% of all hip fractures are extracapsular, which surgeons can treat using a wide variety of possible internal fixation strategies.

Extramedullary sliding hip screws were the standard of care from the 1950s to the 1990s, but many surgeons now prefer intramedullary nails that interlock proximally in the femoral head [2]. There has been a more than 20-fold relative increase in the utilization of intramedullary nails since 1999, and approximately two-thirds of new orthopaedic surgeons now select them routinely [1]. Several designs are on the market from different manufacturers, each varying in length, diameter, neck-shaft angle, number of locking screws or blades, ability to slide and/or compress, ability to control rotation, construction materials, start-point, and surgical technique. This review aimed to determine whether different nail designs are associated with unique benefits or harms in patients with extracapsular hip fractures. Twelve designs were evaluated among 2130 patients from 17 separate trials. The authors concluded that there were no important differences in function, mobility, pain, death, fracture fixation complications, or rates of revision surgery between most of the implants that were studied.

Upon Closer Inspection

Extracapsular hip fractures include pertrochanteric simple fractures (AO classification 31-A1), pertrochanteric multifragmentary fractures (31-A2), intertrochanteric fractures (31-A3), and subtrochanteric fractures (32-A/B/C [1–3].1). Reverse obliquity intertrochanteric fractures, intertrochanteric fractures with subtrochanteric extension, and subtrochanteric fractures each carry a worse prognosis than A1 or A2 fractures [7, 8], and unequal inclusion of patients with these fractures in the primary trials could bias the pooled results towards worse outcomes for the treatment group that has more of them. In order to avoid this problem, the authors attempted to separate the results for patients with subtrochanteric fractures from the results for patients with intertrochanteric and pertrochanteric fractures. These types of secondary analyses are commonly referred to as a subgroup analyses and they are frequently used to address specific questions about particular patient groups, types of intervention, or types of study designs.

There are two important ways that subgroup analyses can be problematic. First, subgroup analyses are rarely powered adequately, which leaves them at risk of failing to identify important differences between treatment groups (Type II error). Second, testing multiple subgroups raises the possibility of identifying spurious but statistically significant differences due to chance alone, simply because there have been more “rolls of the dice” (Type I error) [11]. Readers can distinguish credible from less-credible subgroup claims by looking for hypotheses that were formulated before collecting the data, are supported by background literature and/or biological rationale, and are consistent across studies. The direction of a potential subgroup effect should be also prespecified and only a small number of carefully considered hypotheses should be considered. The proposed subgroup for subtrochanteric fractures met many of these criteria, but there was insufficient data in the primary trials to test it. As such, this meta-analysis does not inform whether certain nails might be better suited to specific fractures [9].

Take-Home Messages

Overall confidence in the pooled results was addressed using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach [5]. Confidence ratings are important because they reflect variations in the design of the primary studies, the risk of bias within and across the primary studies, the degree of imprecision in the pooled results, inconsistency, and indirectness. The GRADE approach has been adopted by more than 70 major health research organizations because it is transparent, structured, and comprehensive enough to guide treatment recommendations. The quality of the evidence for most comparisons in this meta-analysis was low or very low, which indicates that our confidence in the pooled effect estimates is actually limited and that the pooled estimates might differ substantially from reality. In comparison to sliding hip screws, the potential advantages of nailing include a shorter and more rigid moment arm and the contention that insertion is less invasive. However, nails cost between USD 700 and USD 1700 more per implant than sliding hip screws [12] and patient-important functional outcomes across most fracture patterns are similar in the current literature [10]. A recent economic evaluation found that intramedullary nail fixation was routinely cost-effective only for A3 fractures while sliding hip screw fixation was favored for all A1 fractures and most A2 fractures [12], but these conclusions were highly sensitive to the fixation failure rates. Given that older nails were associated with more reoperations and intra and postoperative fractures, and that modern nails and increased clinical experience may have solved these issues [2], further research comparing modern nails against sliding hip screws is likely to be of substantial value. At least one adequately powered multicenter randomized trial is currently registered and underway (NCT01380444), but additional trials and other study designs will also be needed to evaluate varying fracture patterns across different patient populations.