Aim: Compressive forces on the medial femoral cortex and tensile forces at the lateral femoral cortex along with cortical comminution lead to a high risk of failure of surgical fixation of subtrochanteric fractures. The purpose of the study was to correlate the incidence of fracture healing complications to the surgical stabilisation method used.

Methods: A comprehensive search of various data sources extending from 1966 to October 2003 was conducted to identify appropriate studies using specific search terms. We also scanned the reference lists of eligible studies for potentially relevant reports. Articles of all languages were considered. Studies with a follow-up of less than six months, pathological fractures, fractures treated non-operatively and studies reporting on less than ten fractures were excluded. Abstracts were also excluded. Each eligible study was independently reviewed by authors for methodological quality. A methodological scoring system adapted from that of Detsky was used. Guidelines for reporting of meta-analysis, adapted from QUOROM statement were followed.

Results: 39 studies including 1835 fractures were analysed. For extramedullary devices, the incidence of non-union (35/673 – 5.2%), delayed union (11/221 – 4.7%), implant breakage ( 24/444 – 5.1%) and deep infection (14/459 – 3.0%) was statistically significantly higher than non-union (14/506 – 2.7%), delayed union (5/529 – 0.94%), implant breakage (12/628 –1.9%) and deep infection (9/764 – 1.2%) for intramedullary devices. Mortality and superficial infection were higher for extramedullary than intramedullary devices. However, this was not statistically significant. Malunion, shortening and implant cut out were higher for intramedullary than extramedullary devices. This was not statistically significant.

Conclusion: The incidence of fracture healing complications appear to be significantly less with intramedullary than extramedullary devices. Based on this study, we advocate the use of intramedullary surgical fixation devices for subtrochanteric fractures.