What Factors and Patient-reported Outcome Measures Are Associated With Stress Fracture After Periacetabular Osteotomy?
Academic Article
Overview
abstract
BACKGROUND: Although periacetabular osteotomy (PAO) is a commonly used procedure with generally good intermediate and long-term outcomes, complications such as stress fractures of the ischium or pubis have been reported. A limited number of studies have investigated stress fracture after PAO, but the results lack consensus and do not thoroughly explore lifestyle factors or patient-reported outcome measures (PROMs). QUESTIONS/PURPOSES: Among patients treated with PAO: (1) What percentage of patients and hips developed a stress fracture after PAO? (2) What preoperative and intraoperative factors were associated with stress fractures? (3) Did PROMs or the minimum clinically important difference (MCID) and the patient acceptable symptom state (PASS) achievement differ between patients with stress fractures and patients without at most recent follow-up? METHODS: Six hundred seventy-five hips (546 patients) were treated with PAO with or without hip arthroscopy for symptomatic acetabular dysplasia between February 2016 and October 2024 by one surgeon in a mature hip preservation practice. Patients were excluded if the index procedure for those who had bilateral PAOs occurred before the study period. Patients who underwent concomitant femoral osteotomy, surgical hip dislocation, or relative neck lengthening were also excluded, yielding 90% (608 of 675) of hips from 487 patients as potentially eligible for analysis. Ninety-four percent (574 of 608) of PAOs were performed in women, and 65% (396 of 608) were performed without concomitant hip arthroscopy. The mean ± SD age was 26 ± 8 years, and the mean BMI was 23.2 ± 3.9 kg/m2. All hips had 6-week, 3-month, and 6-month postoperative radiographs available for review. Two patients with stress fractures and one without converted to THA, leaving 81% (393 of 484) of patients with available minimum 1-year PROMs after their most recent PAO surgery. Stress fracture diagnoses were tallied by hip and by patient to establish the percentages. Exploratory analyses included age, BMI, preoperative vitamin D levels, magnitude of lateral center-edge angle (LCEA) correction, magnitude of anterior center-edge angle correction, Tönnis grade, sex, marijuana use, nicotine use, screw type, prior ipsilateral surgery, PAO with or without concomitant hip arthroscopy, initial PAO versus subsequent contralateral PAO, and diagnoses of Ehlers-Danlos syndrome (EDS) or hypermobility. Factors with p < 0.1 were considered in the multivariate analysis. To determine the association between stress fractures and postoperative outcomes, univariate regression was performed with the presence of stress fractures as the independent variable. Multivariate regressions were performed to determine whether stress fractures were associated with modified Harris hip score (mHHS) and International Hip Outcome Tool-12 (iHOT-12) improvement after controlling for factors identified in the previous analysis. Similarly, regression models were used to determine whether stress fractures were associated with achievement of the MCID and PASS. RESULTS: Stress fractures occurred in 8% (37 of 487) of patients and in 7% (40 of 608) of hips. For patients who underwent bilateral PAOs, 11% (13 of 121) experienced a stress fracture after the second surgery. Bilateral stress fractures occurred in 2% (3 of 121) of patients. After controlling for potential confounding variables such as age, BMI, LCEA correction, substance use, and EDS or hypermobility, we found that several factors were associated with developing stress fractures. Increasing age was associated with higher odds of stress fracture (OR 1.05 [95% confidence interval (CI) 1.01 to 1.09] for each year of increasing age; p = 0.03). Higher BMI was associated with greater odds of stress fracture (OR 1.09 [95% CI 1.002 to 1.19] for each kg/m2 increase in BMI; p = 0.046). Greater magnitude of LCEA correction was associated with higher odds of stress fracture (OR 1.05 [95% CI 1.01 to 1.10] for each degree of correction; p = 0.02). Current marijuana users had greater odds of stress fracture compared with nonusers (OR 3.06 [95% CI 1.2 to 8.0]; p = 0.02), as did current nicotine users (OR 6.41 [95% CI 1.2 to 34]; p = 0.03). Patients with EDS or hypermobility diagnoses also had higher odds of stress fracture (OR 2.88 [95% CI 1.3 to 6.0]; p = 0.01). Although the proportion was higher, no difference was found between stress fracture occurrence after the first PAO and second PAO in patients who underwent bilateral procedures (OR 2 [95% CI 1 to 4]; p = 0.07). After controlling for factors found to be associated with stress fractures, preoperative PROM scores, and time since most recent PAO, stress fractures were found to be associated with lower preoperative to postoperative improvements in mHHS and iHOT-12 scores, along with decreased odds of achieving the PASS for iHOT-12 and the MCID for mHHS. Patients with stress fractures had mean 6 points less improvement in mHHS than patients without (95% CI -11.6 to -0.84; p = 0.02). Patients with stress fractures had mean 12 points less improvement in iHOT-12 scores than patients without (95% CI -20.6 to -2.45; p = 0.01). Patients with stress fractures had lower odds of achieving the PASS for iHOT-12 (OR 0.36 [95% CI 0.15 to 0.86]; p = 0.02) and lower odds of achieving the MCID for mHHS (OR 0.33 [95% CI 0.13 to 0.83]; p = 0.02). Stress fractures were not associated with achieving the PASS for mHHS (OR 0.52 [95% CI 0.18 to 1.53]; p = 0.24) or the MCID for iHOT-12 (OR 0.86 [95% CI 0.33 to 2.24]; p = 0.76). CONCLUSION: Increasing age, higher BMI, greater LCEA correction, marijuana use, nicotine use, and EDS or hypermobility were associated with increased risk of stress fracture development after PAO for symptomatic acetabular dysplasia. At minimum 1-year follow-up, stress fractures were associated with smaller improvements in mHHS and iHOT-12, as well as lower odds of achieving the PASS for iHOT-12 and the MCID for mHHS. Rather than serving as barriers to surgery, these factors can guide surgeon-patient discussions to provide personalized counseling and rehabilitation, including guidance on potential substance cessation, the use of extended nonweightbearing periods, and realistic expectations for early functional gains. With larger samples of stress fractures, potentially through multicenter registries, future studies should aim to establish clinically meaningful thresholds for associated factors and evaluate the long-term relationship between stress fractures and PROMs, including the influence of fracture healing and fracture location. LEVEL OF EVIDENCE: Level III, therapeutic study.
