PURPOSE: Baba classification for periprosthetic femoral fracture focuses on the positional relationship between the implant design and fracture site. The objective of this study was to investigate whether the addition of CT images and implant information to plain radiograms increases the reliability of Baba classification to a level higher than that of Vancouver classification. METHODS: Twenty patients were randomly selected from 61 patients with periprosthetic femoral fracture between 2009 and 2014. After Vancouver and Baba classifications were fully explained to six orthopedic surgeons (three experts and three non-experts), plain radiograms, CT imaging and implant information at the time of injury were presented, and the inter- and intra-observer reliability based on the Vancouver and Baba classifications and accuracy rates of stem stability were investigated. RESULTS: When interobserver reliability was made based on only radiograms, the mean κ value of Baba classification-based judgments made by the experts was 0.76, and those of Vancouver classification-based judgments made by the experts was 0.41. When interobserver reliability was made based on the radiograms and CT images and implant information, the mean κ value of Baba classification-based judgments made by the experts was 0.94, and those of Vancouver classification-based judgments were 0.48. Intra-observer reliability of Baba classification was 0.81 in the experts. Validity analysis showed 95.0% agreement within all subgroups. CONCLUSIONS: Reliability and validity of Baba classification was improved when additional information was given. We believe that this new classification is useful to establish a therapeutic strategy for femoral fractures around the stem.
PURPOSE: Baba classification for periprosthetic femoral fracture focuses on the positional relationship between the implant design and fracture site. The objective of this study was to investigate whether the addition of CT images and implant information to plain radiograms increases the reliability of Baba classification to a level higher than that of Vancouver classification. METHODS: Twenty patients were randomly selected from 61 patients with periprosthetic femoral fracture between 2009 and 2014. After Vancouver and Baba classifications were fully explained to six orthopedic surgeons (three experts and three non-experts), plain radiograms, CT imaging and implant information at the time of injury were presented, and the inter- and intra-observer reliability based on the Vancouver and Baba classifications and accuracy rates of stem stability were investigated. RESULTS: When interobserver reliability was made based on only radiograms, the mean κ value of Baba classification-based judgments made by the experts was 0.76, and those of Vancouver classification-based judgments made by the experts was 0.41. When interobserver reliability was made based on the radiograms and CT images and implant information, the mean κ value of Baba classification-based judgments made by the experts was 0.94, and those of Vancouver classification-based judgments were 0.48. Intra-observer reliability of Baba classification was 0.81 in the experts. Validity analysis showed 95.0% agreement within all subgroups. CONCLUSIONS: Reliability and validity of Baba classification was improved when additional information was given. We believe that this new classification is useful to establish a therapeutic strategy for femoral fractures around the stem.