GLOBAL TILT AND LUMBAR LORDOSIS INDEX Two parameters to understand posi0ve balance analysis

Transcription

1 Two parameters to understand posi0ve balance analysis L. Boissiere (1, 2), M. Takemoto (1), J. M. Vital (1, 2), V Challier (1), A Bourghli (1), A. Alanay (2) F. Pellisé (2), F. J. S. Pérez- Grueso (2), E. R. Acaroglu (2), F. Kleinstuck (2), Obeid (1, 2), European Spine Study Group (2) 1. CHU Bordeaux Pellegrin Unité rachis 1 Bordeaux 2. ESSG European Spine Study Group EUROSPINE COPENHAGEN September 2 nd 2015

2 INTRODUCTION In Adult Spinal Deformity (ASD), symptoms severity increases in a linear fashion following posifve balance. (Glassman et al. Spine, 2005) Many radiologic parameters were reported. (PI, PT, LL, SS, SVA, LLI, GT ) In recent literature clinico- radiologic correlafon Mandatory to validate a parameter Tends to homogenize ASD whereas issues are varied

3 Literature Review Classifica0on SRS- Schwab des déforma0ons rachidiennes (Schwab et al, Spine 2012) SVA, PT, PI- LL ó Correlate with ODI

4 But SVA can t explain ASD ASD Pain Malalignment The rela0onship between Spinal Deformity malalignment and pain remains: Complex Difficult to generalize

5 Aims 1 / Valid the clinico- radiologic correlafon for Global Tilt and Lumbar Lordosis Index 2 / Understand how radiographic parameters impact ASD management and understanding

6 Material & Methods European Spine Study Group (ESSG) database } 6 European centers Ankara, Barcelone, Bordeaux, Istanbul, Madrid, Zurich } Since 2011 } 755 included pa0ents (with complete data on July 2014) } 613 women / 142 men } 431 non surgical candidates / 324 surgical candidates } Mean Age 52.1 (± 18.8)

7 Methods } Analyzed Variables } Dependent variables: HRQL scores } Demographic variables (DV): age, BMI, diagnosis, past spinal surgery, surgical vs non surgical candidates, coronal curve types } Radiographic parameters: SVA, GT, T1PA, T1ST, PT, PI- LL, LLI } Sta0s0c Analysis } ANOVAs differences in HRQLs among groups of DV } Spearman correla0on between radiographic parameters and HRQLs } MulFvariate linear regression models for each of the HRQLs with sagijal parameters and DV as independent variables } P<0.05 considered sta0s0cally significant

8 Univariate correla0on Sagittal parameters Mean ODI PCS MCS SRS-22 (SD) Rho P Rho P Rho P Rho P SVA (mm) 32(59) *** *** ** *** GT (degree) 23(17) *** *** * *** T1PA (degree) 18(13) *** *** * *** T1ST (degree) -3(5) *** *** * *** PT (degree) 21(11) *** *** ns *** PImLL (degree) 8(21) *** *** ns *** LLI (ratio) 0.88(0.42) *** *** ns *** Global Tilt and LLI correlate with HRQL scores

9 Dem ographic variables N (%) ODI PCS MCS SRS-22 Mean P Mean P Mean P Mean P (SD) (SD) (SD) (SD) Age (years) Gender BMI Past spine surgery Diagnosis Coronal curve type (SRS-Schwab classification) Surgical or non-surgical candidate Age <40 (young) 231 (31) 19(17) 46(9) 46(11) 3.6(0.7) 40 Age <60 (middle aged) 222 (29) 35(19) *** 38(9) *** 43(12) ** 3.0(0.7) *** Age 60 (old) 302 (40) 43(19) 35(9) 44(13) 2.9(0.7) Female, 613 (81) 34(21) 39(10) 44(12) 3.1(0.8) ** ns *** * Male 142 (19) 29(19) 40(10) 48(11) 3.3(0.7) BMI < (59) 27(19) 41(10) 46(11) 3.3(0.7) 25 BMI < (28) 38(20) *** 37(10) *** 43(13) ** 3.0(0.7) *** BMI (13) 49(20) 33(9) 42(14) 2.7(0.7) No 495 (66) 32(20) 40(10) 45(12) 3.2(0.7) * * ns ns Yes 260 (34) 35(22) 38(10) 44(12) 3.1(0.8) Idiopathic 400 (54) 26(18) 42(10) 46(11) 3.3(0.8) Degenerative 237 (31) 43(18) *** 35(9) *** 42(12) ** 2.9(0.7) *** Others 118 (15) 35(22) 38(10) 44(13) 3.0(0.7) Type N 244 (32) 39(21) 37(9) 42(12) 3.0(0.7) Type T 75 (10) 25(17) 44(10) 44(11) 3.4(0.8) *** *** ** *** Type L 171 (23) 37(20) 37(9) 44(13) 3.0(0.8) Type D 265 (35) 27(19) 41(11) 46(11) 3.3(0.8) Non-surgical candidate 431 (57) 27(19) 41(10) 46(11) 3.4(0.7) *** *** *** *** Surgical candidate 324 (43) 41(20) 36(9) 42(12) 2.8(0.7) Demographic Variables ó Correlate with HRQL scores

10 Poten0al bias with univariate analysis ODI BIAS DEMOGRAPHIC VARIABLES RADIOGRAPHICS PARAMATERS => Mul0variate Analysis Interest

11 Adjusted Likelihood ratio test Model Predictors N R 2 AIC R 2 Comparison Δχ 2, Δdf, P-value 1 DVs* a DVs+SVA Model *** 2b DVs+SVA+PT Model *** Model 2a ns 3a DVs+GT Model * 3b DVs+GT+PT Model *** Model 3a *** 4a DVs+T1PA Model ns 755 Model *** 4b DVs+T1PA+PT Model 4a *** 5a DVs+T1ST Model *** 5b DVs+T1ST+PT Model *** Model 5a ns 6 DVs+PT Model ns 7a DVs+PImLL Model * Model ** Model 7a * 8a DVs+LLI Model * b DVs+PImLL+PT b DVs+LLI+PT Demographic variables explain 40.1% of ODI variance All sagijal parameters correlate with the same part of ODI 60% of ODI variance remains unexplained Model * Model 8a ns Sagijal Parameters addi0onaly explain less than 1.5 % ODI variance

12 } More a parameter is related to severe malalignment and more its univariate correla0on will be important = Best Parameter?? } Useless adding- up compensatory mechanism to obtain an important correla0on factor with HRQL scores

13 No radiographic parameter explains more than 2% ODI variance independently Summing up more than one parameter doesn t impact more ODI variance Specific Radiographic parameters are mandatory to answer specific quesfons

14 Global Tilt Interest Understand balance when SVA and PT seem opposite Ø Nega0ve SVA Ø Important PT Ø GT increase => Anterior Malalignment Obeid et al. Eurospine, 2014

15 Lumbar Lordosis Index Interest Assess LL lack when malalignment is due to a LL issue (Boissière et al. Eur Spine Journal, 2013) Most correlated parameter with LL lack (Boissière et al. Eur Spine Journal, 2014)

16 Conflict of Interest } European Spine Study Group receives a grant from DePuy Synthes