Computer-assisted surgery (CAS) techniques have been shown to improve implant placement and reduce the radiation time during cephalomedullary nailing in cadaveric and pilot clinical studies of intertrochanteric hip fractures. However, clinical comparisons of CAS and conventional techniques are lacking. It is unclear whether CAS offers clear advantages in terms of radiation time, operative time, and accuracy of lag-screw placement in patients undergoing surgery for intertrochanteric hip fractures and whether any potential difference in accuracy is associated with a change in the risk of lag-screw cut-out postoperatively.
In patients undergoing cephalomedullary nailing for intertrochanteric hip fractures compared with the conventional technique, we asked: (1) Is the CAS technique associated with a decrease in tip-apex distance (TAD), with less variation and fewer outliers at the standard (25 mm) and lower (15 mm) TAD thresholds? (2) Is the CAS technique associated with a decrease in radiation and operative time? (3) If the CAS technique results in a decrease in TAD, is this decrease associated with a decrease in lag screw cut-out postoperatively?
Between Oct 2007 and June 2015, 964 stable and unstable intertrochanteric hip fractures were treated surgically at our institution. Of these, 23% (225 of 964) were isolated, acute intertrochanteric hip fractures managed by a single surgeon. Ninety-five percent (213 of 225) of hip fractures were surgically treated with the same cephalomedullary nail based on the general indications of displaced fractures, poor bone quality, and medical complexity. This same surgeon used a sliding hip screw device in the remaining 5% (12 of 225) of hip fractures for the treatment of nondisplaced and minimally displaced fractures in younger patients with fewer medical comorbidities and good bone quality. Between October 2007 and August 2011, all procedures were performed with conventional lag screw placement (n = 110), and between September 2011 and June 2015, all procedures were performed with CAS (n = 103) for lag screw placement. Postoperative radiographs were missing or unavailable for TAD analysis for 3% (3 of 110) of the conventional technique group and 6% (6 of 103) of the CAS group, so these patients were excluded. The remaining 97% (107 of 110) of conventional procedures and 94% (97 of 103) of CAS procedures were included in the TAD, radiation time, and operative time analysis. For the evaluation of cut-out postoperatively, 24% (26 of 107) of conventional patients and 25% (24 of 97) of CAS patients were excluded due to mortality and/or loss to follow-up at a minimum of 1 year. The remaining 76% (81 of 107) of conventional patients and 75% (73 of 97) of CAS patients were included in the cut-out analysis. A retrospective chart review was performed to obtain the data and then compare TAD, radiation time, operative time, and cut-out between the two cohorts.
The median TAD for the CAS procedures was lower than the median TAD for the conventional procedures (median 13 mm versus median 16 mm, median difference 3 mm; p < 0.001 power for difference = 85%). In addition, the TAD variation was also less for the CAS procedures compared with the conventional procedures (interquartile range [IQR] 4 mm versus IQR 9 mm, IQR difference 5 mm; p < 0.001, power for difference = 98%). A TAD greater than 25 mm was found in 1% (1 of 97) of the CAS procedures and 12% (13 of 107) of the conventional procedures. A difference between the proportions could be detected indicating a lesser chance of a TAD > 25 mm in the CAS cohort (odds ratio = 0.075 [95% confidence interval 0.010 to 0.587]; p = 0.002, power for difference 90%). A TAD > 15 mm was found in 23% (22 of 97) of the CAS procedures and 56% (60 of 107) of the conventional procedures, also indicating a lesser chance of a TAD > 15 mm in the CAS cohort (OR = 0.230 [95% CI 0.125 to 0.423], relative risk for TAD > 15 mm = 0.404 [95% CI 0.270 to 0.606]; p < 0.001, power for difference > 99%). The median radiation time for the CAS cohort was lower than the median radiation time for the conventional cohort (median 1.4 minutes versus median 1.7 minutes, median difference 0.3 minutes; p = 0.002, power for difference = 81%). No difference in median total operating time was found for the CAS procedures compared with the conventional procedures (median 36 minutes versus median 38 minutes, median difference 2 minutes; p = 0.227, power for difference = 18%, power for equivalency = 93%). There was no difference in cut-out noted with the use of the CAS compared with the conventional technique with the numbers available. Based on the current results, the upper 95% probability for a cutout complication ranges from 0% to 5% in the CAS cohort versus 0% to 9% in the conventional cohort (difference of upper 95% CI = 4%).
CAS use is associated with a decrease in median TAD with less variation and fewer outliers during cephalomedullary nailing. Compared with the conventional technique, fewer outliers were noted with the CAS at the standard TAD threshold of 25 mm and a lower TAD threshold of 15 mm. Additional research is needed to determine the association of TAD variation and outliers on cut-out and to determine if there is any clinical value to the decrease in TAD variation and outliers noted here. The patient and surgical team are exposed to less radiation with the CAS compared with the conventional technique, but this difference is small and it is unclear if this benefit justifies CAS use. Incorporating CAS into the cephalomedullary nailing procedure is not associated with a change in operative time, so there are no costs or risks associated with increased operative time. More procedures would be needed to provide adequate power to better analyze the risk of lag screw cut-out, allowing a more complete understanding of the value of this technology compared with its cost.
Level of Evidence
Level III, therapeutic study.