Relative pelvic version: an individualized pelvic incidence-based proportional parameter that quantifies pelvic...

Pelvic tilt (PT) is used as an indicator of pelvic version with increased values indicating retroversion and disability. The concept of using PT solely as an absolute numerical value can be misleading, especially for the patients with pelvic incidence (PI) values near the upper and lower normal limits. Relative pelvic version (RPV) is a PI-based individualized measure of the pelvic version. Relative pelvic version indicates the individualized spatial orientation of the pelvis relative to the ideal sacral slope as defined by the magnitude of PI.

The aim of this study was to compare RPV and PT for their ability to predict mechanical complications and their correlations with health-related quality of Life (HRQoL) scores.

A retrospective analysis of a prospectively collected data of adult spinal deformity patients was carried out. Mechanical complications (proximal junctional kyphosis or proximal junctional failure, distal junctional kyphosis or distal junctional failure, rod breakage, and implant-related complications) and HRQoL scores (Oswestry Disability Index [ODI], Core Outcome Measures Index [COMI], Short Form-36 Physical Component Summary [SF-36 PCS], and Scoliosis Research Society 22 Spinal Deformity Questionnaire [SRS-22]) were used as outcome measures.

Inclusion criteria were ≥4 levels fusion, and ≥2-year follow-up. Correlations between PT, RPV, PI, and HRQoL were analyzed using Pearson correlation coefficient. Pelvic incidence values and mechanical complication rates in RPV subgroups for each PT category were compared using one-way analysis of variance, Student t test, and chi-squared tests. Predictive models for mechanical complications with RPV and PT were analyzed using binomial logistic regressions.

A total of 222 patients (168 women, 54 men) met the inclusion criteria. Mean age was 52.2±19.3 (18-84) years. Mean follow-up was 28.8±8.2 (24-62) months. There was a significant correlation between PT and PI (r=0.613, p<.001), threatening the use of PT to quantify pelvic version for different PI values. Relative pelvic version was not correlated with PI (r=-0.108, p>.05), being able to quantify pelvic version for all PI values. Compared with PT, RPV had stronger partial correlations with ODI, COMI, SF-36 PCS, and SRS-22 scores (p<.05). Discrimination performance assessed by area under the curve, percentage accuracy in classification, true positive rate, true negative rate, and positive and negative predictive values was better for the model with RPV than for PT. For average PI sizes, the agreement between RPV and PT were moderate (0.609, p<.001), whereas the agreement in small and large PI sizes were poor (0.189, p>.05; -0.098, p>.496, respectively). When analyzed by RPV, each PT "0," "+," and "++" category was further divided into two or three distinct subgroups of patients having different PI values (p=.000, p=.000, and p=.029, respectively). Relative pelvic version subgroups within the same PT category displayed different mechanical complication rates (p=.000, p=.020, and p=.019, respectively).

Pelvic tilt may be insufficient or misleading in quantifying normoversion for the whole spectrum of PI values when used as an absolute numeric value in conjunction with previously reported population-based average thresholds of 20 and 30 degrees. Relative pelvic version offers an individualized quantification of ante-, normo-, and retroversion for all PI sizes. Schwab PT groups were found to constitute inhomogeneous subgroup of patients with different mean PI values and mechanical complication rates. Compared with PT, RPV showed a greater association with both mechanical complications and HRQoL.