Does Compensatory Anterior Pelvic Tilt Decrease After Bilateral Periacetabular Osteotomy?

Document Type


Publication Date


Publication Title

Clinical Orthopaedics and Related Research


BACKGROUND: The kinetic link among the lumbar spine, pelvic tilt, and the hip has been hypothesized, but this relationship requires further study in acetabular dysplasia. Anecdotal reports suggest that patients may compensate for acetabular dysplasia with an involuntary increase in anterior pelvic tilt; it is not known if this relationship is affected by acetabular reorientation.

QUESTIONS/PURPOSES: (1) Does compensatory pelvic tilt decrease on preoperatively obtained standing AP pelvis radiographs compared with those obtained at a minimum of 6 months after bilateral periacetabular osteotomy (PAO)? (2) Does a modified surrogate measurement of pelvic tilt, the pubic symphysis to sacroiliac (PS-SI) index, correlate with a physical synthetic bones model in which pelvic tilt can be directly measured? (3) Can the PS-SI index demonstrate high interrater reliability?

METHODS: We assessed the surgical records of one surgeon, who participates in the longitudinally maintained Academic Network of Conservational Hip Outcomes Research (ANCHOR) registry, for patients who had undergone the second side of a staged bilateral PAO between 2007 and 2016; there were 113 such patients. Of those, 70 (62%) were lost to followup within 6 months of the second PAO or did not have adequate imaging studies, and another three (3%) were excluded for prespecified reasons, leaving 40 (35%) for evaluation in this retrospective study. Standing preoperative and most recent postoperative AP pelvis radiographs were used to measure the Tönnis angle, anterior wall index, posterior wall index, lateral center-edge angle, pubis symphysis-to-sacrococcygeal junction distance, and the PS-SI index. The most recent radiographs were obtained at a mean of 16 ± 6 months after the second PAO. We chose 6 months as the minimum because at this time point, the majority of patients have reached their maximum clinical improvement and are no longer limited by postoperative muscle dysfunction. Statistical analysis was performed using the intraclass correlation coefficient (ICC) for interrater reliability and paired t-tests for assessing change in measurements from pre- to postoperative. Additionally, a model was created using a physical synthetic bones model in which pelvic tilt could be directly measured. This model was secured through bilateral acetabuli on a mount and rotated through 5° increases in pelvic tilt. AP pelvis radiographs were obtained at each point, the PS-SI index was measured, and a regression analysis performed to evaluate for trend.

RESULTS: Overall, 37 of 40 patients (93%) had a decrease in pelvic tilt, as measured by the PS-SI index. The mean amount of pelvic tilt as measured by the PS-SI index decreased after surgery when comparing the preoperative with latest radiographs on this parameter (97 ± 14 mm versus 89 ± 13 mm, mean difference 8 ± 9 mm; 95% confidence interval, -11 to -5; range 17 increase to 24 decrease, p < 0.001). A linear relationship between pelvic tilt and PS-SI index (PS-SI index = 5.0° + 3.6° tilt, R = 0.99) was identified in the synthetic bones validation model. Finally, the interrater reliability was found to be excellent for the PS-SI index preoperatively (ICC = 0.986) and postoperatively (ICC = 0.988).

CONCLUSIONS: We found a modest reduction in anterior pelvic tilt after bilateral PAO. This finding suggests that acetabular reorientation affects pelvic position. In clinical practice, patients with acetabular dysplasia may compensate with dynamic and reversible changes in pelvic tilt. The PS-SI index is a reproducible tool to measure the height of the pelvic inlet as an assessment of pelvic tilt. In the future, clinical studies should evaluate the clinical implications of these radiographic findings, including the assessment of back pain, which although multifactorial may be influenced by pelvic tilt.

LEVEL OF EVIDENCE: Level III, therapeutic study.





First Page


Last Page






PubMed ID