IMPACT OF THE SITTING POSITION ON LUMBAR LORDOSIS AND ITS CORRELATION WITH PELVIC PARAMETERS

Objective: To evaluateboth the correlation between lumbar accommodation and pelvic parametersin different types of lordosis and the participation of different lumbar segments in the accommodation of lordosis in the standing and sitting positions.Methods: A retrospective study analyzingpatient images in standing and sitting positions. Correlations were conducted among the measured data: Cobb angle of the lumbar lordosis (LL,type of lordosis, pelvic incidence (PI),sacral slope (SS),pelvic tilt (PT), and the angulation of the L1-L2/L2-L3/L3-L4/ L4-L5/L5-S1 segments. Results: Fortypatients were included, 20 men and 20 women. The mean age was 60.8 (±11.5). Of these patients, 10.3% were classified as Roussouly type 2, 35.9% as type 3, 25.6% as type 3A, and 28.2%as type 4.There was a weakcorrelation between LL and PT, however, an inverse correlation between the two (r=-0.183 and p=0.264) was observed. SS hadthe strongest correlation with LL (r> 0.75). Only the correlation between LL and PI was stronger when sitting than standing (p=0.014). The pelvic parameters and angulations of the segments and lumbar discs when standing and sitting were different (p<0.05). In both positions, there was a difference in the contribution of the segments to the LL (p<0.001). On average, the differences in LL between standing and sitting wereequal among theRoussouly classifications (p=0.332). Conclusions: There was a correlation between the LL and the pelvic parameters, being more evident with the SS than with the other parameters. There was no difference in the accommodation of the LL in the different Roussouly types either standing or sitting. Regardless of the position,the L4-S1 segments were predominant in the composition of LL. Level of evidence IV


INTRODUCTION
4][5] Pelvic and spinal parameters have been considered in the evaluation of the spine and determine the principles of treatment used to treat disease of the spine. 6he association between lumbar lordosis and pelvic parameters and its relationship with the hip was first described by During. 5Legaye et al. 7 described the angles with better clarity and proposed the mathematical relationship between pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS) as PI = PT + SS.They also presented the association between pelvic incidence and lumbar lordosis by correlating pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS) with lumbar lordosis (Figure 1), 8,9 thus expanding the original concept. 5,10,11umbar lordosis is subdivided into anatomical and functional lordosis.Anatomical lumbar lordosis is described as the angle formed by the upper surfaces of L1 and S1.Functional lordosis is described by several geometric methods (circular arc, elliptical quadrant).The term distal lordosis was proposed to express functional lordosis to avoid confusion around the description. 5,12,13erthonnaud 14 described a mathematical design for functional lumbar lordosis using the inflection point where the lordosis changes to kyphosis without using any specific anatomical reference.Lordosis is formed by two arcs: the upper and the lower.The lower arc is located between the horizontal line that passes through the apex of the lordosis and the endplate of S1.The upper arcis formed by the horizontal line that passes through the apex of the lordosis and the line perpendicular to the tangent of the inflection point (Figure 2). 5 The lower arc or distal lordosis, or "distal spinal lordosis" according to Roussouly, has a value equal to the sacral slope (SS) in normal individuals.Pelvic incidence varies according to the sacral slope (SS), so there is a correlation between pelvic incidence and the lower lumbar lordosis arc.Based on the definition of the inflection point, the location of the apex of the lordosis, and the description of the lower arc, a classification for lumbar lordosis was proposed, whichconsidersnot only the angle of the lordosis, but also its distribution.pes 1 and 2 have a low pelvic incidence (PI) value <50 degrees and types 3 and 4 have higher pelvic incidence (PI) values >50 degrees.There is a variant of type 3 that may present sacral slope (SS) between 35 and 45 degrees with pelvic tilt (PT) <5 degrees, called type 3 with anteversion (Figure 3). 18,19he pelvic parameters (pelvic incidence, pelvic tilt, and sacral slope) have a geometric correlation (PI=PT+SS).Pelvic incidence is a morphological parameter and does not change with the position of the pelvis.Pelvic tilt and sacral slope (SS) are parameters that are influenced by the position of the pelvis and reflect changes in their values according to the positioning of the pelvis.Pelvic anteversion decreases sacral slope and increases pelvic tilt. 20,21Pelvic retroversion decreases sacral slope and increases pelvic tilt to accommodate the loss of lordosis.][24][25] When standing or sitting there is a change in the functional pelvic parameters (pelvic tilt and sacral slope), while pelvic incidence remains constant as it is an anatomical parameter.][28][29] Sagittal balance is related to the spinal and spinopelvic complex parameters, creating a constellation of interactions throughout this complex.Lumbar lordosis is one of the elements of foremost importancein the spinopelvic complex and in the sagittal balance of the spine.][33][34] The objective of this study was to evaluate the correlation between lumbar accommodation and the pelvic parameters, the correlation between lumbar accommodation and the different types of lordosis, and how the different segments of the lumbar spine participate in the degree of lordotic accommodation between the standing and sitting positions.

METHODS
This is a retrospective, analytic study of radiographic images.We collected data from patients scheduled for total hip arthroplasty surgery and treated at a referral hospital in the state of São Paulo from 2010 to 2019.Convenience sampling was conducted, including all patients who met the inclusion and exclusion criteria, that is, patients of both sexes who had adequate image records were included and patients with previous hip or spine surgery were excluded.The study was approved by the Institutional Review Board (CAAE: 38124820.2.0000.5479).
The radiographs were taken preoperatively in the standing and sitting positions according to the protocol of the orthopedic service of the hospital for patients undergoing hip arthroplasty.The radiographic parameters evaluated werethe Cobb angle of the anatomical lumbar lordosis (L1-S1), lower lordosis (upper arc, lower arc, apex), type of lordosis (Roussouly classification), pelvic incidence, sacral slope, pelvic tilt, and the angulation of the L1-L2, L2-L3, L3-L4, L4-L5, and L5-S1 segments.Surgimap® software (Nemaris Inc.™, New York, US), validated as a tool for this type of measurement, was used to measure the parameters. 35atient demographic data were described as summary measurements (mean, standard deviation, median, minimum, and maximum) for age and as absolute and relative frequencies for the qualitative variables. 36ll the parameters evaluated were described according to the position assessed using summary measurements and compared between the positions using paired Student's t tests. 36The Pearson's correlation between the lumbar lordosis and the radiographic measurements of each position evaluated was calculated and the correlations between the positions were compared. 37The relative angular values of each level and lumbar disc were calculated, described, and compared between measurements using generalized estimating equations (GEE) with normal marginal distribution and identity link function, assuming an exchangeable correlation matrix between the measurements of the same individual. 38The results were followed by multiple Bonferroni comparisons to identify what the angle and the intradiscal distances contributed to the respective totals. 37The differences in lumbar lordosis between the positions (standing -sitting) were calculated, and the differences were described according to Roussouly's classification, and the differencesbetween the classifications were compared using analysis of variances (ANOVA). 37he analyses were performed using IBM-SPSS for Windows version 20.0 software and tabulated using Microsoft-Excel 2003 software.The tests were conducted with a significance level of 5%.
In Table 2, we can see that the correlation between lumbar lordosis and pelvic tilt was weak in both positions and statistically nonsignificant in the sitting position (r= -0.183 and p= 0.264).It was also the only parameter that had an inverse correlation with lumbar lordosis, that is, the higher the pelvic tilt, the lower the lumbar lordosis and vice-versa.Sacral slope was the parameter that presented the strongest direct correlation with lumbar lordosis in both positions (r > 0.75) in the comparison between standing and sitting.Only the correlation between lumbar lordosis and pelvic incidence was statistically stronger in the sitting than in the standing position (p = 0.014).
When we compared the parameters PT, PI, SS of segments L1-L2, L2-L3, L3-L4, L4-L5,and L5-S1 measured in the standing and sitting positions, we observed that most presented a statistically significant mean difference (p<0.05),all values being higher when standing than sitting, the only exception being PT in the sitting position, which on average was higher (Table 3).
Table 4 shows that in both positions there was a statistically significant mean difference between the contribution of the segments to lumbar lordosis (p < 0.001) and between the disc angulations in each position.The mean differences were also statistically significant(p < 0.001).Table 5 shows that the mean differences betweenstanding and sitting LL were statistically equal among the Roussouly classifications of the patients (p = 0.332).

DISCUSSION
In this study, we evaluated a homogeneous sample, evenly distributed between the sexes.Individuals with Roussouly types 3 and 3A lumbar lordosis made up a greater proportion, which is in agreement with the frequency encountered in previous studies. 39We observed a higher frequency of type 3A (25.6%) than that described in the literature, but justifiable as the population studied was composed of patients with osteoarthritis of the hip, who may present hip flexion contracture, described as hip spine syndrome. 18In Table 3, the comparison of the pelvic tilt and sacral slope parameters showed statistically significant differences between the standing and sitting positions.The same did not occur with pelvic incidence, because, as expected, it is an anatomical parameter that should not vary, regardless of the individual's position. 7e observed that most of the angulations, whether of the segments or the intervertebral discs, presented variations according to the position in which they were measured.All the parameters evaluated were greater with the patients standing than when sitting, except formean pelvic tilt, which was greater when seated and consistent with the compensatory mechanism which occurs: flexion of the hip with retroversion of the pelvis, an increase in pelvic tilt, a decrease in sacral slope, and a decrease in lumbar lordosis in the sitting position. 30nderstanding that lumbar lordosis accommodation is related to pelvic incidence, pelvic tilt, and sacral slope, we can demonstrate these findings in our sample (Table 2).The greater the sacral slope, the greater the slope of L5 and, consequently, of the upper levels, tends to be.Therefore, the slope of the L5-S1 disc is directly related to the pelvic incidence and the lordosis.
The higher the pelvic incidence, the greater the sacral slope and the lumbar lordosis. 8Table 2 shows that sacral slope had a stronger direct correlation with lumbar lordosis, both in the sitting and standing positions, than pelvic incidence, corroborating the findings of other authors. 18elvic tilt was weakly correlated with lumbar lordosis in both positions.However, it presented an inverse correlation with lumbar lordosis.This result can be explained by observing the pelvic incidence formula, expressed by the geometric equation PI=SS+PT.Therefore, the greater the pelvic tilt, the smaller the sacral slope.
][24][25] When comparing the correlations in the standing and sitting positions, only the correlation of lumbar lordosis with pelvic incidence was statistically stronger when seated than in the standing position, since patients with higher pelvic incidence have a compensatory mechanism.According to the description by Le Huec, 9 the pelvis can rotate around the femoral heads, following the hip axis.When the pelvis performs a retroversion, the pelvic tilt parameter increases.When the pelvis performs an anteversion, pelvic tilt decreases.Pelvic tilt and sacral slope are positional parameters.The possibility of the pelvis rotating around the axis of the femoral heads is a critical mechanism for the regulation ofsagittal balance. 9  When we evaluated the contribution that each vertebral segment and each intervertebral disc makes towards the composition of lordosis (Table 4), we realized that the lower levels, L4-L5 and L5-S1, as well as their discs, correspond to approximately 60% of lordosis, [40][41][42] demonstrating that, even with a change in position, the lower levels are fundamental for the maintenance of lumbar lordosis.
When we compared the changes in lumbar lordosis according to Roussouly's 17 morphological classification of types, there was no difference among the types, with the difference between standing and sitting positions being similar for the different types.

CONCLUSIONS
This study demonstrated that there is a correlation between lumbar lordosis and the pelvic parameters, being more evident inthe sacral slope than in the other parameters.There was no difference in the accommodation of lumbar lordosis among the Roussouly morphological types in the standing and sitting positions.We also showed that, regardlessof the position, the L4-S1 levels are predominant in the composition of lumbar lordosis.
All authors declare no potential conflict of interest related to this article.

Figure 2 . 5 Figure 3 .
Figure 2. Representation of the spinal curves according to Berthonnaud's inflection point. 5 5,[15][16][17] Type 1-Characterized by a low sacral slope (SS) value <35 degrees with the apex of the lordosis in the center of L5.The point of inflection is low and posterior and does not extend beyond the L2-L3 level.The lordosis is short.Type 2-Characterized by a low sacral slope (SS) value <35 degrees with the apex at the base of the body of L4.The inflection point is higher and anterior.The lordosis is more extensive and flatter.Type 3-Characterized by sacral slope values between 35 and 45 degrees with the apex at the upper partof L4 or the L3-L4 disc.The inflection point is in the thoracolumbar region and the curve is well-distributed.Type 4-Characterized by a high sacral slope (SS) value >45 degrees with the apex of the lordosis at the base of L3.The number of lordotic vertebraeis greater than or equal to five.β = SS; θ = 20 o

Table 1 .
Description of the demographic characteristics of the patients and the Roussouly classification.

Table 2 .
Results of the correlation of lumbar lordosis and the radiographic parameters of interest by position and the results of the comparisons of the correlations between positions.-Test of significance of the correlation different than zero * -Test of the comparison of correlations. p

Table 3 .
Description of the parameters evaluated by position and the results of the comparison between positions.Each author made significant individual contributions to this manuscript.RT and NA: study design, data collection, writing, creation of the research project; WZS: discussion of the results, review, and approval of the final version of the work; RGMM: discussion of the results, review, and approval of the final version of the work; AOG: discussion of the results, review, and approval of the final version of the work; MFSC: discussion of the results, review, and approval of the final version of the work; RM: study design,discussion of the results, review, and approval of the final version of the work.
CONTRIBUTIONS OF THE AUTHORS:

Table 4 .
Description of the relative contribution of the vertebral segments to lumbar lordosis and of the disc angulations in each position and the results of the comparison between the parts.

Table 5 .
Description of the differences between lumbar lordosis in the standing and sitting positions according to the Roussouly classification and the results of the comparison between the categories.SD 33.8 ± 21.631.1 ± 20 44.3 ± 12.2 35.9 ± 15.5 36.1 ± 17.3 median (min., max.) ANOVA