With the aim to clarify the fracture patterns of proximal humeral fractures and to reevaluate the validity of using the revised Neer classification (2002), we analyzed 668 cases that had been treated by a total of 18 Active Members of the Japan Shoulder Society (JSS database). There were 174 males and 494 females, with a mean age of 63.4 years old. We found that all the fractures of the JSS database, with the exception of articular surface injuries, fell under one of the fracture patterns provided by Codman (1934). When the revised Neer classification was applied, we found that 654 of the 668 fractures (98 percent) had an appropriate category in this classification, including 19 cases (2.8 percent) of valgus impacted fracture. Those that could not be classified with the revised Neer classification (2 percent) were “3-part” with a complete anatomical neck fracture. We concluded as follows: 1) proximal humeral fractures, excluding articular surface fractures, occur along the original epiphyseal lines; 2) the revised Neer classification covers 98% of all proximal humeral fractures, and it would be appropriate to use this classification in a clinical practice; 3) when there are three displaced segments, close attention should be paid to the fracture line on the anatomical neck. If the fracture spans along the full length of the anatomical neck, it is not included in the revised Neer classification and is highly likely to be a fracture type that involves a high risk of humeral head necrosis.
