Problem: The vascularity of the articular fragment is of key importance for the final outcome in three- and four-part fractures of the humeral head. Displacement of the articular segment may compromise the arterial blood supply to the humeral head and result in avascular necrosis. There is still controversy as to whether three-and four-part fracture dislocations (articular fragment outside the glenoid) have an even worse prognosis than displaced three- and four-part fractures.
Patients and methods: Between January 1985 and May 1993, 102 patients with three- and four-part fractures of the humeral head were treated by ORIF (mostly tension band wiring) at our institution. In a retrospective study we analysed the functional (Constant 100 point score) and radiological outcome of 67 (66%) of these patients. There were 21 patients with fracture dislocations (FD), n = 5 type B2X, n = 5 type B3X, n = 3 type C2X, n = 8 type C3X, according to the classification of Habermeyer . The "X" represents the dislocation of the articular fragment, whereas the classification to each type is done after reduction of the head. The remaining 46 patients presented with displaced, but not dislocated, three- and four-part fractures (DF), n = 24 type B2, n = 7 type B3, n = 3 type C2, n = 12 type C3. Average follow-up was 25 months (7-72 months). Patients with FD were significantly younger (average age 50 years) than patients with DF (average age 63 years, P < 0.05) and showed a significantly higher incidence of traumatic nerve or plexus lesions (FD 19%, DF 2%, P < 0.05).
Results: Concerning the functional results, there was no statistically significant difference between the two groups. The FD patients even showed a slight tendency to better results than patients with DF. This was true for the three-part fractures (average Constant score 78 versus 67 points), as well as for the four-part fractures (average Constant score 62 versus 55 points). The significantly younger age of the FD patients may explain their better results. The entire group of patients with three-part fractures showed a significantly better functional outcome (average Constant score 68 points) than patients with four-part fractures (average Constant score 55 points, P < 0.05). The rate of partial and total avascular necrosis of the humeral head was strongly correlated to the fracture type (number of fragments, fracture of the anatomical or surgical neck, according to the classification of Habermeyer),but again there was no difference between the FD and DF group (B2X: 20%, B3X: 20%, C2X: 33%, C3X: 63%; B2: 25%, B3: 29%, C2: 33%, C3: 67%). Astonishingly, the FD were not associated with an increased rate of avascular necrosis of the humeral head. Three (axillary nerve) out of the five observed primary nerve and plexus lesions had a full neurological recovery after 6-12 months; the two patients with alterations of the brachial plexus showed a slow tendency of improvement at follow-up (12 and 18 months), but still had gross muscular atrophy and impaired sensory function.
Conclusion: In displaced three-and four-part fractures of the humeral head the dislocation of the articular segment does not seem to increase the risk of avascular necrosis, if treated by timely and careful ORIF with respect to the vascularity. Even with the increased risk of primary nerve and plexus lesions in fracture dislocations, good functional results can be achieved by early operative nerve decompression and fracture stabilization in this middle-aged patient group. However, older patients with displaced or dislocated four-fragment fractures through the anatomical neck (type C3) have a poor chance of a favourable outcome, and therefore primary prosthetic replacement should be considered.