T1 Slope Minus Cervical Lordosis (TS-CL), the Cervical Answer to PI-LL, Defines Cervical Sagittal Deformity in Patients Undergoing Thoracolumbar Osteotomy

INTRODUCTION

There is an interplay of regional spinal alignment in the economy of standing posture that has been referred to as the “chain of correlation.”^1,2 In order to achieve harmonious spinopelvic alignment, large pelvic incidence requires larger lumbar lordosis, larger thoracic kyphosis, and greater cervical lordosis, which is also important in the maintenance of horizontal gaze.³ The use of modern spinal deformity surgical techniques such as the pedicle subtraction osteotomy has enabled large global deformity corrections across fused segments. However, there has been increasing recognition that reciprocal changes in alignment occur in adjacent regions of the spine. Reciprocal thoracic hyperkyphosis occurring across unfused segments following lumbar osteotomy has also been demonstrated.^4,5 Conversely, thoracic osteotomies and limited fusions for the correction of thoracic kyphosis have been shown to decrease lumbar lordosis.⁴ Smith et al⁶ exhibited that patients with thoracolumbar sagittal imbalance with sagittal vertical axis (SVA) > 5 cm had reciprocal cervical hyperlordosis that resolved following correction of the underlying thoracolumbar deformity. A recent study revealed that among patients with thoracolumbar deformities, there was a prevalence of cervical deformities such as increased C2-C7 plumb line (CPL) and cervical kyphosis.⁶ A CPL or cervical offset of greater than 4 cm has been shown to be associated with greater disability by validated outcome measures.⁷ However, cervical offset may not be the best descriptor of primary cervical deformity in the setting of an underlying thoracolumbar deformity. Smith et al⁸ demonstrated that the correction of the thoracolumbar sagittal malalignment also led to improved cervical offset with a decrease in the CPL. Thus, deformity in the other portions of the spine can insidiously affect the cervical alignment and the respective measurements of cervical deformity.

With greater recognition of the influence of global spinal alignment on cervical deformity, several authors have shown that cervical kyphosis can be physiologic, depending on the underlying global spinopelvic alignment.^9,10 Pelvic incidence and lumbar lordosis are two measures that have been well established in playing a role in balancing global spinal alignment.^11–13 The pelvic incidence minus lumbar lordosis (PI-LL) parameter correlates with health-related quality of life measures in patients with thoracolumbar deformity.^11–13 In fact, lumbar lordosis is more highly correlated to sacral slope (SS) than pelvic incidence in asymptomatic patients.^2,14 An analogous cervical correlate is the balance between cervical lordosis (CL) and T1 slope (TS). A direct relationship between these two parameters has been inferred, but the use of TS and CL as a descriptor of primary cervical deformity has not been well established.^8,15

The purpose of this study was to investigate the use of TS-CL as a marker of cervical sagittal deformity (CSD) and to evaluate the development of CSD following thoracolumbar three-column osteotomy. We expect TS-CL to correlate with the variety of existing descriptive measures of CSD and to more reliably identify primary CSD especially in the setting of concurrent thoracolumbar deformity.

MATERIALS AND METHODS

Radiographic Parameters

Preoperative and 1-year postoperative, sagittal cervical, thoracolumbar, and spinopelvic parameters were measured using Spineview software (Laboratory of Biomechanics, ENSAM ParisTech, Paris, France),^17,18 a validated and computer-based tool that enables quantitative measures of the spine and pelvis. In addition, preoperative thoracolumbar deformity severity was quantified using the Scoliosis Research Society-Schwab Classification of adult spinal deformity.¹⁹

• Cervical radiographic parameters (Figure 1) included C2 slope (C2S), CPL, C2-C7 lordosis (CL), T1 slope (TS), and T1 slope minus cervical lordosis (TS-CL).

• Thoracolumbar and spinopelvic radiographic parameters included T2-T12 thoracic kyphosis (TK_T2), T4-T12 thoracic kyphosis (TK_T4), L1-S1 lumbar lordosis (LL), pelvic incidence (PI), Pelvic tilt, SS, and PI-LL.

• Global radiographic parameters (Figure 2) included SVA, cervico-thoracic pelvic angle (CTPA), C2 pelvic angle, and T1 pelvic angle (TPA).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Cervical radiographic parameters measured including the cervical lordosis (CL), the C2 slope, the cervical plumb line (CPL), and the T1 slope.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Global radiographic parameters measured including the cervico-thoracic pelvic angle (CTPA), the C2 pelvic angle (CPA), and the T1 pelvic angle (TPA). CTPA is a global angular measure of cervical sagittal alignment and a correlate of the C2-C7 plumb line. T1 pelvic angle is a measure of global sagittal alignment and a correlate of the C7 sagittal vertical axis. CPA is the angle of a line from C2 centroid to the femoral heads (FH) and a line from the FH to the middle of the S1 endplate. T1 pelvic angle is the angle of a line from the center of T1 to the FH and a line from the FH to the center of the S1 endplate. CTPA is the angle of a line from C2 centroid to the FH and a line from the FH to the center of T1. CTPA is the result of subtracting T1 pelvic angle from CPA.

Statistical Analysis

Analysis of total population radiographic and demographic parameters was conducted. Correlation of existing and novel radiographic parameters was reviewed. Because global sagittal correction is known to affect cervical sagittal malalignment, postoperative values were used in linear regression analysis to determine the threshold of TS-CL corresponding to a CPL of 4 cm. This cervical alignment threshold was then used to classify patients into uncompensated (TS-CL above threshold) and compensated (TS-CL below threshold) groups (Figures 3 and 4). Radiographic parameters between groups were compared using unpaired t test analysis. Radiographic correlations were reviewed within CSD and Reciprocal Sagittal Alignment groups using Pearson correlations. Level of significance was set to .05.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

(a) Preoperative x-rays of a patient with thoracolumbar deformity from the uncompensated cervical sagittal deformity (CSD) group. Though there is lordotic alignment of the cervical spine, there is a mismatch in T1 slope and cervical lordosis (TS-CL > 17°). Cervical plumb line, TS-CL, and CTPA are elevated. (b) Postoperative x-rays of the same patient from the uncompensated group. There has been good global correction, but there has been progression of the cervical deformity with kyphotic alignment of the cervical spine, an increase in the cervical plumb line, a progression in the mismatch of T1 slope and cervical lordosis, and an increase in the cervico-thoracic pelvic angle. Abbreviations: CL, cervical lordosis; TS, T1 slope; TPA, T1 pelvic angle; CTPA, cervico-thoracic pelvic angle; C2-C7 PL, cervical plumb line; SVA, sagittal vertical axis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

(a) Preoperative x-rays of a thoracolumbar deformity patient from the compensated group (no cervical deformity). The magnitude of the cervical lordosis matches that of the T1 slope (TS-CL < 17°), demonstrating that there is good cervical compensation for the large global deformity allowing for horizontal gaze. (b) Postoperative x-rays of the same patient from the compensated group. Good global correction has been achieved and reciprocal change has occurred, with a decrease in the cervical sagittal alignment. The magnitude of the cervical lordosis remains in balance with the T1 slope. Abbreviations: CL, cervical lordosis; TS, T1 slope; TPA, T1 pelvic angle; CTPA, cervico-thoracic pelvic angle; C2-C7 PL, cervical plumb line; SVA, sagittal vertical axis.

Logistic regression modeling was used to determine predictors of postoperative CSD (TS-CL above threshold). Potential predictors were identified using univariate analysis and expert opinion. The model was built using a combination of backward elimination and bootstrap selection. Our final model had 20 events per variable. Model fit was assessed using the Hosmer-Lemeshow test and the C statistic/receiver operating characteristic curve. Internal validation of the model was performed using 10-fold cross-validation.

RESULTS

Radiographic Correlations

CTPA correlated with CPL as an angular analog to this linear measure of cervical alignment with preoperative r = .85, and postoperative r = .699, (all P < .001; Figure 5). C2 pelvic angle correlated with C2-S1 plumb line with preoperative r = .756 and postoperative r = .683 (all P < .001). TS-CL correlated with CTPA with preoperative r = .52, and postoperative r = .37, and TS-CL correlated with CPL with preoperative r = .57 and postoperative r = .37 (all P < .001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Patient with thoracolumbar deformity after pedicle subtraction osteotomy with a long fusion to the pelvis; patient developed cervical sagittal deformity. T1 slope minus cervical lordosis (TS-CL) is 72°.

DISCUSSION

Sagittal spinal alignment has been well established as a primary driver of disability in patients with thoracolumbar deformity.^20,21 It is becoming increasingly more apparent that cervical sagittal alignment is similarly an important determinant of health-related quality of life measures.^7,22,23 Tang et al⁷ studied 113 patients undergoing posterior cervical fusion, and positive cervical sagittal balance was associated with greater disability by validated outcome measures including the neck disability index and 36-item short form health survey. Villavicencio et al²² demonstrated that if cervical sagittal alignment was maintained or improved, SF-36 and NDI outcomes were improved among 122 patients undergoing anterior cervical discectomy and fusion.

However, CSD is difficult to define, especially when there is concurrent thoracolumbar deformity. Several authors have noted that an underlying thoracolumbar sagittal malalignment causes positive cervical sagittal balance.^1,23–26 This reciprocal cervical malalignment reverses with correction of the underlying thoracolumbar deformity. Therefore, a positive cervical sagittal alignment in the setting of thoracolumbar sagittal malalignment is not necessarily indicative of primary cervical deformity.

Likewise, cervical kyphosis is considered to be a pathologic indicator of cervical deformity. However, many have noted that cervical kyphosis in a standing individual can be a flexible physiologic state that depends on global standing balance and the underlying spinopelvic morphology.^10,25 The presence of cervical kyphosis is, therefore, not necessarily indicative of a primary cervical deformity.

We used the relationship between TS and CL to determine patients who have concurrent primary CSD and thoracolumbar deformity. The balance between pelvic incidence and lumbar lordosis has been shown to be an important relationship in the determination of lumbar flatback deformity.^11–13,26 Similarly, TS-CL is a relationship that can define primary CSD (Figure 1).^8,14,15 The relationship between TS and CL is analogous to that of SS and LL. SS has been shown to be highly correlated to LL in asymptomatic patients.^2,14 In fact the correlation between SS and LL was stronger than that for pelvic incidence and LL.¹⁴

In a retrospective study, Hyun et al²⁷ showed that there is a significant positive correlation between C2-C7 and SVA and TS-CL, with the latter measurement significantly affecting cervical alignment. The study revealed that a mismatch greater than 26.1° corresponded to positive cervical sagittal malalignment, defined as C2-C7 SVA greater than 50 mm.²⁷

In our study, we used a linear regression analysis to determine the postoperative TS-CL that corresponds to a CPL of 4 cm. Tang et al⁷ previously demonstrated that a CPL of greater than 4 cm is associated with poor health measures in patients with primary cervical deformity. We correlated the postoperative values of TS-CL and CPL because correction of the underlying thoracolumbar deformities in our patient population should lead to a reciprocal correction of any cervical malalignment. Following the thoracolumbar deformity correction, any residual positive cervical sagittal alignment is more likely to be the result of an underlying primary cervical deformity. By determining that a TS-CL threshold of 17° correlates to a CPL of 4 cm, we used this threshold to categorize the patients into the compensated and uncompensated groups.

We found it interesting that the compensated group (TS-CL < 17°) had improved cervical alignment following correction of the underlying thoracolumbar deformities. Conversely, the postoperative alignment of the uncompensated cohort (TS-CL > 17°) did not demonstrate any reciprocal change in CL and there was a paradoxical increase in the CPL and an increase in the cervico-thoracic pelvic angle (CTPA), which is a novel global measure of cervical sagittal alignment (Figures 2 through 5).²⁸

The TS-CL measure identified 36 patients with CSD (TS-CL > 17°) that did not meet criteria for deformity by established measures (CPL > 4 cm, or CL < 0°). In these 36 patients, the mean CPL worsened preoperatively to postoperatively (from 29.07 ± 9.01 mm to 36.61 ± 16.7 mm) and the mean CL worsened from 15.36° ± 11.33° to 1.03° ± 12.6°. Conversely, there were only 2 patients from the compensated group with deformities by the established criteria. In both patients the CPL and CL improved preoperatively to postoperatively. This suggests that the TS-CL measure can be useful in defining CSD and in predicting those patients who may have a tendency to worsen in postoperative alignment.

A linear regression analysis was used to determine risk factors for the development of cervical sagittal malalignment (TS-CL above threshold). Potential predictors were identified using univariate analysis, and a statistical model was built using a combination of backward elimination and bootstrap selection. The model fit was assessed using the Hosmer-Lemeshow test and the C statistic/receiver operating characteristic curve; internal validation of the model was performed using 10-fold cross-validation. Positive predictors of postoperative CSD included baseline TS-CL > 17° (OR, 2.78; P = .007), longer fusion (OR, 1.3; P = .033), and baseline CTPA (OR, 1.24 per degree; P = .029). This suggests that patients with preoperative cervical deformities, as defined by a larger baseline TS-CL mismatch and high CTPA, were more likely to maintain postoperative cervical deformities and unlikely to undergo reciprocal change of cervical alignment. Moreover, this analysis shows that those with longer fusions were more likely to have postoperative cervical deformities. This may be due to diminished capacity for adjacent cervical regional compensation or the result of disruption of the cervical paraspinal muscle attachments and the capsuloligamentous complex at the cervicothoracic junction during exposure and instrumentation.

A primary limitation of the study is that this data set did not include health-related quality of life measures and so the threshold of CSD (CPL > 4 cm) had to be extrapolated from the literature.⁶ However, it is important to note that a questionnaire specific to cervical sagittal alignment does not exist. Another limitation of this study is the retrospective design, although the analyses were based on prospectively collected data. The strengths of the study include the contribution of cases from multiple spinal-deformity centers and the use of multiple standardized measures of regional and global sagittal spinal alignment.

This study demonstrates the utility of TS-CL to determine when primary CSD is present concurrently with thoracolumbar deformity. Further research is needed to establish the balance between TS and CL (TS-CL) as an integral CSD parameter, correlated with health-related quality of life measures.