Background:
Several studies have compared different methods for fixation of
the midpart of the humeral shaft, but there are only scattered data
regarding which type of plate construct provides the best fixation
for humeral nonunion. The objectives of this study were (1) to obtain
objective data on the performance of four different plate constructs
used for fixation of humeral nonunion, and (2) to report our clinical
experience with plate fixation of thirty-seven nonunions of the
midpart of the humeral shaft.
Methods:
In the first part of the study, four plate constructs were compared
in a Sawbones model. The groups consisted of (1) a posterior limited-contact
dynamic compression plate alone; (2) a posterior limited-contact
dynamic compression plate and an interfragmentary screw; (3) a posterior
limited-contact dynamic compression plate, a lateral 3.5-mm reconstruction plate,
and an interfragmentary screw; and (4) a posterior limited-contact
dynamic compression plate and a lateral 3.5-mm reconstruction plate.
Tests were performed with use of an MTS Bionix machine in anterior-posterior
four-point bending, medial-lateral four-point bending, and external
rotation torque.
In the second part of the study, the charts of thirty-seven consecutive
patients in whom a nonunion of the midpart of the humeral shaft
had been treated with plate fixation were reviewed retrospectively.
The average age of the patients was forty-eight years (range, thirteen
to seventy-eight years). Nineteen patients were treated with a single
posterior plate, and eighteen were treated with a two-plate construct
with the plates parallel and lying at 90° to each other.
All of the nonunions were treated with bone-grafting, and an interfragmentary
screw was used in thirty-six of the thirty-seven patients. Radiographs
and the clinical status were evaluated at an average of thirteen
months postoperatively.
Results:
The biomechanical testing showed that the two-plate constructs
were significantly stiffer than the single-plate constructs in all
test modes (p < 0.05). In the clinical part of the study, thirty-four
(92%) of the nonunions healed without complications at an average
of 4.8 months. Two nonunions treated with the two-plate construct
and one treated with one plate failed to heal.
Conclusions:
No significant difference in the healing rate was found between
the two clinical groups (p = 0.4, ß = 0.9), and the overall
healing rate was 92%. However, a two-plate construct with the plates
at right angles is mechanically stiffer than a single-plate construct,
which might be helpful if rigid stabilization of the humerus at
the midshaft level is needed.