Proposal of a new indicator of hip compensation for spinopelvic–hip mismatch: a retrospective study in Japan

Article information

Asian Spine J. 2025;19(6):967-977

Publication date (electronic) : 2025 August 11

doi : https://doi.org/10.31616/asj.2024.0475

Ryo Fujita, Kohei Takahashi, Ko Hashimoto, Kazuyoshi Baba, Kenichiro Yahata, Takahiro Onoki, Takashi Aki, Keisuke Ishikawa, Toshimi Aizawa

Department of Orthopaedic Surgery, Tohoku University Graduate School of Medicine, Sendai, Japan

Corresponding author: Ryo Fujita, Department of Orthopaedic Surgery, Tohoku University Graduate School of Medicine, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan, Tel: +81-22-717-7245, Fax: +81-22-717-7248, E-mail: ryo.fujita.e7@tohoku.ac.jp

Received 2024 November 19; Revised 2025 April 13; Accepted 2025 April 17.

Abstract

Graphical Abstract

Introduction

Maintaining harmonious alignment of the lower limbs, pelvis, and spine is essential for upright posture in humans. The cone of economy concept by Dubousset [1] likens posture control to an inverted pendulum, where a smaller cone reduces muscle activity and improves energy efficiency [2]. Disruption of this alignment, such as that observed in anterior trunk inclination, is associated with increased muscle activity, balance impairment, and fall risk [3,4]. Spinopelvic parameters, such as pelvic incidence (PI) and lumbar lordosis (LL), are commonly used to assess sagittal alignment. In fact, PI–LL mismatch has been increasingly recognized as a key indicator of spinopelvic imbalance and is widely applied in surgical planning for adult spinal deformity [5–8].

Sagittal malalignment of the spine is compensated by the lower extremities, as shown by recent studies using EOS imaging [9–11]. Therefore, understanding the relationship between the lumbar spine, pelvis, and lower extremities is important. In particular, the hip joints, which are in close proximity to the spine, are involved in sagittal alignment, and their dysfunction can lead to a condition known as hip–spine syndrome [12]. However, only a few studies have investigated this relationship [13,14].

Pelvic tilt (PT) has traditionally been used to evaluate hip position relative to the pelvis [11,15–17]. However, the fact that it does not directly represent hip position and is influenced by knee flexion and extension limits its accuracy [18]. On the other hand, the pelvic femoral angle (PFA), which is defined as the angle between the femoral axis and a line connecting the femoral head to the sacral endplate, is a more direct representation of sagittal hip alignment. PFA increases with hip extension and decreases with hip flexion. Although PFA has been widely used in EOS imaging due to its high precision [9,19], it can also be measured using standard lateral radiographs [14,20]. Although EOS is a promising imaging modality, its high installation cost, need for specialized software and trained personnel, and limited availability, mainly in Europe and North America, hinder its widespread clinical use. Consequently, conventional radiography remains the standard for alignment assessment. Therefore, developing an indicator of lower limb compensation using whole-spine radiographs alone would be highly valuable for the diagnosis and treatment planning of spinal deformities.

In this study, we hypothesized the following: (1) PFA correlates with PI–LL until the compensatory capacity of the hip joints is exceeded; (2) this correlation diminishes when the spinopelvic mismatch exceeds the compensatory limit of the hip joints; and (3) the discrepancy between spinopelvic mismatch and sagittal hip alignment correlates with anterior trunk inclination. To assess hip joint compensatory ability without relying on EOS imaging, we propose PFA–(PI–LL) as a novel indicator and demonstrated its utility by examining its correlation with anterior trunk inclination and health-related quality of life (HRQOL).

Materials and Methods

Radiographic evaluation

For the lateral whole-spine plain radiographic view, patients were instructed to stand naturally with their arms raised horizontally forward at approximately 30° of shoulder flexion [21]. Various spinopelvic parameters, including SVA, thoracic kyphosis (TK) at T5–12, PI, PT, LL at L1–S1, and sacral slope (SS), were measured. PFA was used to represent sagittal alignment of the hip joints and was defined as the angle between the line from the midpoint of the bilateral femoral head centers to the midpoint between the sacral endplate and proximal femoral shaft axis. The proximal femoral shaft axis was drawn as a line connecting the midpoints of the most anterior and posterior cortices of the proximal and distal parts of the two femurs. Femoral inclination (FI) was defined as the angle between the plumb line and the proximal femoral shaft axis [17] (Fig. 1). In addition, we proposed and calculated a new parameter, PFA–(PI–LL), as an indicator of spinopelvic–hip mismatch to represent the discrepancy between spinopelvic mismatch and sagittal hip alignment.

Fig. 1

Measurement of spinopelvic parameters in the whole spine lateral radiograph. The following parameters were measured: sagittal vertical axis (SVA, the distance between plumb line passing through the center of C7 vertebra and posterior superior corner of sacral endplate); thoracic kyphosis (TK, the angle between the upper endplate of T1 and lower endplate of T12); lumbar lordosis (LL, the angle between the upper endplate of L1 and the endplate of sacrum); pelvic incidence (PI, the angle between a line perpendicular to the endplate of sacrum and a line drawn from the midpoint of the bilateral femoral head centers to the midpoint of the sacral endplate); pelvic tilt (PT, the angle between the plumb line and a line drawn from the midpoint of the bilateral femoral head centers to the midpoint of the sacral endplate); pelvic femoral angle (PFA, the angle between a line drawn from the midpoint of the bilateral femoral head centers to the midpoint of the sacral endplate, and the proximal femoral shaft axis); femoral inclination (FI, the angle between the plumb line and the proximal femoral shaft axis).

To verify the reliability of radiographic measurements, two spine surgeons with 9 years of clinical experience measured these parameters twice in a randomly selected group of 50 patients (16 with osteoporosis and 34 with spinal kyphosis). The intraclass correlation (ICC) coefficients for intraobserver agreement (ICC [1, 2]) and interobserver agreement (ICC [2, 1]) for each parameter were calculated using IBM SPSS Statistics software ver. 21.0 (IBM Corp., Armonk, NY, USA). The (ICC [2, 1]) was calculated in a round robin manner (i.e., 2 measurements×2 examiners=4 combinations).

Results

Radiographic measurement reliability

The (ICC [2, 1]) ranged from 0.83 to 0.99, and (ICC [1, 2]) ranged from 0.87 to 0.99. The average ICC values ranged from 0.87 to 0.99, indicating an almost perfect observer agreement (Table 1).

Table 1

The interclass correlation coefficients of inter-observer and intra-observer agreement

Comparison between the normal and malalignment groups

A comparison between the PI–LL <20° and PI–LL ≥20° groups revealed significant differences in all assessed parameters. Compared with the PI–LL ≥20° group, the PI–LL <20° group had significantly lower median (25th–75th percentile) SVA (8.5 mm [−8.4 to 27.1] vs. 127.3 mm [68.8 to 209.5], p<0.001); PI (47.6° [41.2 to 55.4] vs. 53.6° [45.1 to 59.9], p=0.001); PT (15.0° [11.0 to 18.9] vs. 36.2° [30.3 to 46.3], p<0.001); PFA (189.1° [185.1 to 193.4] vs. 202.7° [195.9 to 210.6], p<0.001); FI (6.3° [4.4 to 8.1] vs. 13.4° [9.2 to 19.3], p<0.001); and PI–LL (−0.5° [−6.4 to 4.7] vs. 44.6° [37.7 to 59.6], p<0.001) and significantly higher median (25th–75th percentile) TK (26.6° [19.2 to 34.4] vs. 17.7° [6.8 to 31.2], p<0.001); LL (48.6° [41.0 to 58.1] vs. 1.8° [−5.2 to 18.5], p<0.001); SS (32.6° [27.1 to 39.4], 16.7° [6.3 to 22.7], p<0.001); and PFA–(PI–LL) (189.5° [185.3 to 194.6] vs. 159.5° [142.0 to 168.4], p<0.001) (Table 2).

Table 2

The median values of each parameter for each group (25th, 75th percentile)

Correlation coefficients for each parameter in each group

The Pearson’s correlation coefficients for each parameter are listed in Table 3. PI–LL correlated with PFA (r=0.56, p<0.001) in the PI–LL <20° group but not in the PI–LL ≥20° group. SVA was correlated with TK (r=0.21, p=0.003); LL (r=–0.27, p<0.001); FI (r=0.38, p<0.001); PI–LL (r=0.35, p<0.001); and PFA–(PI–LL) (r=–0.47, p<0.001) in the PI–LL <20° group and with LL (r=–0.55, p<0.001); PFA (r=–0.36, p=0.004); FI (r=0.67, p<0.001); PI–LL (r=0.60, p<0.001); and PFA–(PI–LL) (r=–0.80, p<0.001) in the PI–LL ≥20° group. PFA–(PI–LL) and FI showed correlation in both the PI–LL <20° group (r=−0.56, p<0.001) and PI–LL ≥20° group (r=−0.73, p<0.001).

Table 3

Correlations of each parameter in each group

Coefficients of determination of SVA and each radiographic parameter

The R² values for the SVA of each parameter are listed in Table 4. The R² for PFA–(PI–LL) was the highest in both groups and across all patients. The respective R² values for PI–LL and PFA–(PI–LL) were 0.12 and 0.22 in the PI–LL<20° group and 0.36 and 0.64 in the PI–LL ≥20° group. In all patients, the R² for PI–LL and PFA–(PI–LL) were 0.68 and 0.77, respectively. Scatter diagrams between SVA and PI–LL and between SVA and PFA–(PI–LL) in all patients are shown in Fig. 2. For all patients, a linear correlation equation for SVA using PFA–(PI–LL) was calculated, as follows:

Table 4

Coefficients of determination for SVA of each parameter

Fig. 2

Scatter diagrams between SVA and PI–LL (A) and SVA and PFA–(PI–LL) (B). SVA, sagittal vertical axis; PI, pelvic incidence; LL, lumbar lordosis; PFA, pelvic femoral angle.

Predicted SVA=663-3.43×[PFA-(PI-LL)]

When the PFA–(PI–LL) values were 166° and 182°, the corresponding SVAs were 95 mm and 40 mm, respectively.

Correlation between radiographic parameters and JOABPEQ domain scores

Each parameter showed significant correlations with all JOABPEQ score domains (Table 5). Among all domains, the absolute correlation coefficient of PFA–(PI–LL) was the largest.

Table 5

Correlation coefficients for the PI–LL and PFA–(PI–LL) of each JOABPEQ domain scores

Case presentations (Fig. 3)

Fig. 3

Case presentations. (A) Case 1 (44-year-old female): The parameters were as follows: SVA, 11.3 mm; TK, 15.9°; LL, 59.0°; PI, 56.7°; PFA, 188.4°; PI–LL, −2.3°; PFA–(PI–LL), 190.7°; and predicted SVA, 8.9 mm. (B) Case 2 (80-year-old male): The parameters were as follows: SVA, 69.1 mm; TK, 43.7°; LL, 39.4°; PI, 33.6°; PFA, 176.7°; PI–LL, −5.9°; PFA–(PI–LL), 182.4°; and predicted SVA, 37.3 mm. (C) Case 3 (72-year-old male): The parameters were as follows: SVA, 27.3 mm; TK, 39.4°; LL, 30.5°; PI, 68.9°; PFA, 230.7°; PI–LL, 38.5°; PFA–(PI–LL), 192.3°; and predicted SVA, 3.4 mm. (D) Case 4 (78-year-old female): The parameters were as follows: SVA, 300.6 mm; TK, 38.9°; LL, 7.5°; PI, 51.2°; PFA, 171.1°; PI–LL, 43.8°; PFA–(PI–LL), 127.3°; and predicted SVA, 226.4 mm. SVA, sagittal vertical axis; TK, thoracic kyphosis; LL, lumbar lordosis; PI, pelvic incidence; PFA, pelvic femoral angle.

Case 1

A 44-year-old woman presented with an SVA of 11.3 mm, TK of 15.9°, LL of 59.0°, PI of 56.7°, and PFA of 188.4°. The PI–LL was −2.3°, and PFA–(PI–LL) was 190.7°. The predicted SVA was 8.9 mm.

Case 2

An 80-year-old man presented with an SVA of 69.1 mm, TK of 43.7°, LL of 39.4°, PI of 33.6°, and PFA of 176.7°. The PI–LL was −5.9°, and PFA–(PI–LL) was 182.4°. The predicted SVA was 37.3 mm.

Case 3

A 72-year-old man presented with an SVA of 27.3 mm, TK of 39.4°, LL of 30.5°, PI of 68.9°, and PFA of 230.7°. The PI–LL was 38.5°, and PFA–(PI–LL) was 192.3°. The predicted SVA was 3.4 mm.

Case 4

A 78-year-old woman presented with a markedly increased SVA of 300.6 mm, TK of 38.9°, LL of 7.5°, PI of 51.2°, and PFA of 171.1°. The PI–LL was 43.8°, and PFA–(PI–LL) was 127.3°. The predicted SVA was 226.4 mm.

Discussion

SVA is a commonly used parameter to quantify anterior trunk inclination [23] relative to the severity of clinical symptoms [23,24]. In this study, we proposed PFA–(PI–LL) as an indicator of the discrepancy between hip joint alignment and spinopelvic mismatch and established its utility for the assessment of hip compensatory function by demonstrating its correlation with SVA and HRQOL measures.

The relationships between SVA and various spinopelvic parameters, such as TK, LL, PT, and PI–LL have been investigated, with PI–LL having the strongest correlation with SVA (correlation coefficient: 0.418–0.596) [5–7]. However, none of these parameters can independently and accurately determine SVA, because anterior trunk inclination involves multiple segments [5–7]. Attempts have been made to predict SVA using approximate formulas with multiple parameters. Pan et al. [8] proposed the following predictive formula for SVA among patients aged >35 years who had various degenerative spinal diseases: −25.68+2.98×LL+2.37×PI+1.67×TK (R²=0.416). Lafage et al. [25] proposed a formula among 179 adult patients: SVA=−52.87+5.90×PI–5.13×[maximal LL]–4.45×PT–2.09×[maximal thoracic kyphosis]+0.513×[age] (R²=0.793) [26]. Although these formulas may appear complex, they offer high coefficients of determination and are useful in preoperative planning for spinal deformity correction surgery. Although the lower extremities significantly contribute to maintaining an upright trunk [11,20], there had been no general formula that represents both the spinal and lower extremity components of SVA.

In a radiological study, Diebo et al. [11] reported that pelvic retroversion increased with greater spinopelvic mismatch until the compensatory mechanisms of the hip joints reached their limit. When the limit is exceeded, compensation is mainly based on the significant contributions from the knee and posterior pelvic shift. Moreover, anterior trunk inclination is determined by the positional relationship between the cranium and pelvis. Posterior pelvic shift, which is associated with knee flexion, has been reported to cause upper body rotation around the ninth thoracic vertebra [27], resulting in anterior trunk inclination. These findings suggested that when hip compensation reaches its limit, further increase in PI–LL becomes strongly associated with anterior trunk inclination. In our study, we observed a correlation between PFA and PI–LL in the PI–LL <20° group, indicating that the hip joints compensated for the increased spinopelvic mismatch in this cohort. Conversely, the absence of correlation between PFA and PI–LL in the PI–LL ≥20° group was likely because of the mixture of patients with both sufficient and insufficient hip extension (Fig. 3C, D).

We hypothesized that PFA–(PI–LL) can serve as an indicator of the compensatory ability of the hip joints for spinopelvic mismatch. FI, which reflects knee flexion compensation on lateral radiographs, indicates such compensation when it exceeds 5°. In this study, the negative correlation between PFA–(PI–LL) and FI suggested that a small PFA–(PI–LL) indicated insufficient hip extension to adequately compensate for spinopelvic mismatch, leading to compensation through knee flexion. This finding underscored the usefulness of PFA–(PI–LL) in assessing the compensatory function of hip joints. Furthermore, PFA–(PI–LL) was found to be a more effective parameter than PI–LL in explaining SVA, as demonstrated in patients who had significant anterior trunk inclination despite minor spinopelvic mismatch (Fig. 3B) and those in whom the hip joint can sufficiently extend to maintain an upright posture despite a significant PI–LL (Fig. 3C). Our study introduced the first formula based on PFA–(PI–LL) that incorporates both spinal and hip joint parameters to predict SVA; it achieved an R² of 0.77, which was higher than that of the model by Pan et al. (R²=0.416) and comparable with the model by Lafage et al. (R²=0.793). These findings suggested that PFA–(PI–LL) is a reliable and highly precise predictor of SVA. In addition, the proposed formula is a practical tool for clinical application. The target PFA values obtained by rehabilitation to enhance hip extension functionality allow accurate predictions of improvements in anterior trunk inclination in patients with spinal kyphosis.

HRQOL measures reportedly differ among the three Scoliosis Research Society (SRS)-Schwab grades (0, +, ++) of the sagittal modifiers SVA, PT, and PI–LL [28]. In this study, HRQOL was shown to correlate with PFA–(PI–LL), and its strength of correlation was comparable with that of SVA, PT, and PI–LL. In addition, based on the proposed formulas, indices of 166° and 182° correspond to predicted SVAs of approximately 95 mm and 40 mm, respectively, which fall within grade + of the SRS-Schwab classification. Therefore, this range may serve as a useful reference for clinical application and can be used as a tool for determining the need for surgical intervention in patients with spinal kyphosis.

This study had several limitations. Although incorporating additional factors, such as TK, PT, SS, age, and body mass index, could have improved the accuracy the formula, our focus was to evaluate PFA–(PI–LL) as a tool for assessing hip joint compensation. Moreover, a more accurate formula could have been obtained by evaluating the leg and foot, which would require EOS imaging. Although we used FI, it only reflected femoral orientation and did not fully assess the knee joint. A more detailed evaluation of compensatory function might have been achieved by assessing the range of motion and muscle strength of the lower limb joints; however, such assessments were not conducted in this study. Inclusion of patients from an osteoporosis outpatient clinic limited the generalizability of this study, because the population cannot be considered a consecutive series of healthy individuals. Furthermore, a significant number of patients were excluded because of history of fragility fractures or use of gonadal shielding aprons, which obscured femoral imaging. Despite these limitations, we believe that PFA–(PI–LL), which represents spinopelvic–hip mismatch, can be a valuable indicator of the compensatory ability of hip joints.

Conclusions

PFA–(PI–LL), which represents spinopelvic–hip mismatch, may serve as a valuable indicator of the compensatory ability of hip joints in response to spinopelvic mismatch and was associated with anterior trunk inclination and HRQOL measures.

Article information Continued