14
Back Pain and Sagittal Alignment
Pedro Berjano
■ Introduction
Every degenerative case is a deformity case too. In the past, degenerative conditions in the lumbosacral spine were attributed to nerve compression, motion in worn or swollen joints, or segmental instability; these were considered the only causes of pain and disability. Alignment received little attention as a potential means of reducing pain and disability. Sagittal alignment has been demonstrated to be an independent predictor of pain and disability in adult patients with spinal deformity.1–3 Moreover, adjacent-segment degeneration has been attributed mainly to stiffness of the fused segments, with consequent overload at the boundaries of surgical fusion. Adjacent-segment failure following spinal fusion is frequently observed in patients treated for lumbar degenerative segment disease, and it is often expressed initially as hyperextension at the upper adjacent level, leading to the intuitive idea that compensation for lower lumbar hypolordosis could be the mechanism for accelerated degeneration of the adjacent level. Several published studies provide some support for this idea. More recent studies have shown that improving sagittal alignment in surgery for degenerative conditions can increase the clinical benefit that patients perceive. When treating degenerative conditions, the surgeon should always consider spinal alignment and implement a reasonable strategy to restore it.
This chapter discusses degenerative conditions of the lumbar spine in relation to issues of alignment. Basic concepts of sagittal alignment, findings from the literature on the impact of alignment restoration, and technical tips to improve alignment when performing spinal fusion are presented.
■ Sagittal Balance
A person whose spinal alignment results in minimal requirements of energy to sustain the upright position is considered to be sagittally balanced. Alternatively, a person whose trunk’s shape requires exerting great effort to remain upright is considered to be imbalanced. Therefore, sagittal balance is in effect the composite product of spine shape, body mass distribution, and the forces from the stabilizing muscles. Individuals with good alignment require little effort to stand and to walk. For slight deviations from the best possible alignment, standard muscle activity enables an effective, comfortable upright position and walking. But greater deviations in alignment require increased muscle effort, in some instances resulting in decreased function and pain. Individuals with better muscular functional reserve (in terms of muscle power and endurance) can compensate for higher degrees of misalignment with little or no impairment of function and with little or no increase in pain. In contrast, those with lower muscle reserve have poorer function and increased pain for the same degree of misalignment.
A Constant Battle: Loss of Lordosis and Compensatory Efforts
Life events can have kyphosing effects on the spine, such as disk degeneration, muscle deterioration, and changes in bone shape (i.e., due to fractures). These effects reduce lumbar lordosis and increase thoracic kyphosis progressively as the person ages.
A person with sagittal misalignment stands upright by compensating for the misalignment, which entails assuming an unnatural position. When some lordosis at one or a few lumbar segments is lost, the first adjustment usually consists of hyperextension of the adjacent levels. Increasing deformity, for example with loss of lordosis along the entire lumbar spine, can be compensated for by extension of the intact thoracic spine. But these modalities of active extension require increased muscle effort. An additional contribution to compensation can be obtained by posterior rotation of the pelvis around the hip joints (pelvic retroversion), which can be measured by the increase of pelvic tilt. In patients with extreme sagittal deformity, the most energy-consuming compensatory mechanism of knee flexion may be elicited, greatly decreasing the endurance of the patient (Fig. 14.1 ).4
How Can Sagittal Balance and Compensation Be Measured?
The knowledge we have gained about sagittal balance and its influence on clinical outcomes has been derived by evaluating the sagittal alignment of the spine in full-spine standing lateral radiographs that included the pelvis. The position of the arms has an influence on how the alignment of the spine presents in radiographs. It has been shown that positions of the fingers to the clavicle position or the hands crossed in front of the pelvis introduce little variation in the alignment of the spine, and thus they are preferred during radiographic evaluation.5
C7 Plumbline and the Sagittal Vertical Axis
Although some earlier work has demonstrated that the degree of forward displacement of the C7 plumbline correlates with health-related quality of life (HRQoL) (see Why Is Sagittal Balance Important?, below), a milestone was represented by the 2010 paper by Schwab et al,3 which confirmed that three main sagittal parameter measurements correlate with outcomes in adult deformity patients: sagittal vertical axis (SVA), pelvic tilt (PT), and lumbar lordosis-pelvic incidence mismatch (LL-PI).
The SVA is the length of displacement from the C7 plumbline to the posterior angle of S1. Greater anterior displacement of the C7 plumbline correlates with more severe sagittal misalignment. Adolescents frequently have a negative SVA (a posterior C7 plumbline), young adults usually have a neutral SVA, and elderly people tend to have a positive (anterior) SVA. However, the more positive the SVA, the greater the patient’s pain and disability and the poorer the function. The SVA threshold to distinguish between patients with better and worse HRQoL is around 50 mm. SVA is a measurement of the degree of efficiency of the global upright trunk position that a patient can achieve.3
Pelvic Tilt
Another important parameter to examine is pelvic tilt (PT), which measures the retroversion of the pelvis in the standing position. The PT is related to the pelvic incidence (PI), which measures the inclination at which the sacrum articulates to the pelvis (Fig. 14.2). This PI angle can vary greatly among persons with balanced spines, from 30 to 90 degrees. PI has a positive correlation with the curvature of the spine. In general, a balanced spine with a larger PI exhibits both greater lumbar lordosis (LL) and greater thoracic kyphosis. Conversely, a balanced spine with a smaller PI has a smaller LL and smaller thoracic kyphosis.6
In spite of the striking variability in PI, the pelvis tends to align itself economically to achieve a rather vertical “magic line” (see below).
The angle of the PT is defined by the intersection of a vertical line with a second line that links the midpoint of the sacral end plate and the midpoint of the bifemoral line (the line drawn through the center of both femoral heads). We now abbreviate the latter line (from the midsacrum end plate to the middle of the bi-coxofemoral axis) and simply refer to it as the “magic line.” Of course this nickname is not based on any scientific property or characteristic, but it conveys the idea that this line can “magically” provide intuitive information on the amount of PI and pelvic retroversion present in a specific patient’s X-rays. A rather vertical magic line demonstrates that the patient has little or no pelvic retroversion, whereas a more inclined magic line is an expression of increasing pelvic retroversion. Similarly, the angle formed by the magic line and the S1 end plate implies the amount of pelvic incidence. Patients with a very high pelvic incidence tend to have a magic line nearly parallel to the S1 end plate, whereas in patients with decreasing values of PI, the line is more perpendicular to the S1 end plate.
Experimental work in asymptomatic individuals has shown how the PT tends to increase in normal individuals as the PI increases.6 But although the PI can vary between 30 and 90 degrees in asymptomatic persons, the PT increases correspondingly from a few degrees to 21 degrees, and in very infrequent cases with the highest values of PI reaching as great as 24 to 25 degrees. Thus, asymptomatic individuals very seldom have a PT greater than 21 degrees.
An increase of PT is accompanied by an increase in the pelvic torque, which is the torsion moment of the body weight, mainly applied to the pelvis at the sacrum, and the ground reaction force, mainly applied to the pelvis at the acetabula (Fig. 14.3). Therefore, as PT increases, an increase in compensatory muscle effort ensues to stabilize the pelvis and keep the person upright. More balanced spines exhibit a smaller PT (with a more vertical magic line and a smaller pelvic torque).
Fig. 14.2 (a) Low pelvic incidence due to less forward inclination of the sacrum in the pelvis. (b) The sacrum has more forward inclination within the pelvis, corresponding to high pelvic incidence.
Pelvic Incidence-Lumbar Lordosis Mismatch
Asymptomatic individuals have a substantial correlation between the PI and the total lordosis,6 and the total LL usually equals or exceeds the PI. Patients with too few degrees of LL compared with PI have poor clinical outcomes in terms of pain and disability.3 One recent study suggests that patients undergoing surgery to fix sagittal malalignment obtained good postoperative alignment when their LL exceeded PI by 10 degrees. In the same study, patients with an adequate distribution of LL (i.e., with two thirds of LL between L4 and S1 and no thoracolumbar junction hyperkyphosis) obtained satisfactory postoperative alignment if their LL was superior to the PI.7 It is the author’s experience that accepting a postoperative value of LL that is inferior to the PI significantly increases the risk of sagittal malalignment, except for patients with PI > 70 degrees, in which case LL > PI – 10 degrees can be sufficient.
Distribution of Lordosis
Lordosis is not distributed equally all along the lumbar spine. We know that in normal subjects, approximately two thirds of LL is located between the superior end plate of L4 and S1. Some individuals with a loss of lordosis between L4 and S1, due to degeneration of the lower lumbar disks, compensate with increased lordosis in the upper lumbar spine, resulting in a normal amount of global lordosis. However, this normal lumbar lordosis is the result of active extension that requires increased muscle effort and can potentially lead to increased pain and decreased function. Although the two-thirds rule has clear clinical utility and is widely used, a more refined approach is available. In a study of 160 volunteers, Roussouly et al8 defined four types of sagittal alignment, based on the variables of pelvic incidence, length of lordosis, sacral slope, and distribution of lordosis. Their classification can be useful to help determine what could have been the original distribution of lordosis in a given patient and to interpret the patient’s adaptation to pathological changes. One recent study suggests that in patients with higher values of PI, the contribution of L4-S1 to total lordosis tends to be lower.9
Fig. 14.3 (a) With a low pelvic tilt angle (relatively vertical “magic line,” in red) the axis of the trunk weight load on the sacrum and that of the ground reaction force on the hip joint are close to each other, resulting in low pelvic torque. (b) Increasing pelvic tilt (relatively horizontal “magic line,” in red) makes the distance between the forces increase, resulting in higher pelvic torque. Between these two drawings the torque increases by about fourfold.
■ Why Is Sagittal Balance Important?
Sagittal Alignment and Clinical Outcomes
There is substantial evidence in the literature regarding the correlation of sagittal alignment with HRQoL Mac-Thiong et al10 demonstrated that the displacement of the C7 plumbline is an important indicator of HRQoL by showing a very strong correlation between increasing anterior displacement of the C7 plumbline and increasing spinal malfunctioning measured by Oswestry Disability Index (ODI) scores. In adult scoliosis patients with and without prior surgery, Glassman et al11 found that pain and HRQoL measures from the Scoliosis Research Society’s SRS-22 questionnaire, the Short Form SF-12 questionnaire, and ODI assessments all correlated with changes in SVA, whereas coronal imbalance (lateral displacement of the C7 plumbline from the center of the sacrum) did not correlate with variations in HRQoL scores. Furthermore, the correlation persisted when controlled for by age. HRQoL scores are also improved with sagittal plane correction in patients without scoliosis.12 Lastly, Schwab et al3 established groundbreaking correlations among SVA, PT, PI-LL, and both pain and ODI scores.
The above-mentioned evidence is derived from studies of patients with deformity. It can be demonstrated that similar, albeit more subtle, correlations can also be found in patients with degenerative spine conditions. There is also anecdotal evidence indicating that if sagittal alignment is not considered when planning and performing surgery, it can lead to poorer results and earlier failure. Kim and colleagues13 provide an excellent example in patients with degenerative spondylolisthesis and stenosis. In this study, sagittal alignment was not considered in the surgical planning before performing posterior interbody fusion, making the resulting changes in PT random. Although all patients’ Visual Analogue Scale (VAS) and ODI outcome scores improved, a greater margin of improvement was seen in patients whose PT was decreased, reflecting postoperative improvement of lumbar alignment with a decrease of pelvic compensation.
Sagittal Alignment and Adjacent-Segment Degeneration
Patient Studies
In a study by Heo et al,14 378 patients whose spondylolisthesis at the L4-L5 or L4-S1 level was treated with fusion were followed for a minimum of 2 years. The authors found that nearly 25% of the patients exhibited clinically symptomatic adjacent-segment degeneration (ASD) between 5 and 10 years postprocedure, although only 33 patients (8.7%) underwent a subsequent fusion extension surgery on the adjacent L3-L4 segment in this period. They also found that two of the most important risk factors for the incidence of clinically symptomatic ASD following fusion were low overall lordosis and low lordosis in the instrumented segments.
Immediately postoperatively, the amount of lumbar lordosis was greater in patients who did not develop clinically symptomatic ASD. A third parameter found to correlate with a low risk of clinically symptomatic ASD was having interbody fusion (IBF) rather than posterolateral fusion (PLF). This decreased risk can be explained by the fact that the IBF technique enables the surgeon to better maintain lumbar lordosis and can therefore achieve better sagittal alignment than with PLF.14
Kumar et al15 measured the C7 sagittal plumbline and sacral slope immediately postoperatively in a cohort of patients who underwent fusion for lumbar degenerative conditions. Patients with normal sagittal alignment immediately postoperative had the lowest incidence of ASD when compared with patients having poor postoperative sagittal alignment. Furthermore, the authors recommended following patients with postoperative sagittal malalignment for at least 5 years because of the enhanced risk of ASD.
Jackson et al16 conducted a radiographic comparative study of 100 asymptomatic volunteers and 100 patients suffering from low back pain (LBP). The authors found smaller amounts of overall lumbar lordosis in the LBP patients, as well as a tendency to have greater proximal segment lordosis and less distal segmental lordosis, indicating the occurrence of proximal compensation for distal degenerative loss of lordosis.
Park et al17 studied patients with isthmic spondylolisthesis and found that patients with a better (smaller) postoperative PT had a lower incidence of ASD, again confirming the importance of sagittal alignment in patient outcomes.
Sagittal alignment has been shown to influence the incidence of some distinct modalities of spinal pathology. In a retrospective study by Barrey et al,18 patients with higher PI were found to be more predisposed to developing degenerative spondylolisthesis (DS). The phenomenon can be explained by higher amount of lordosis in these patients, causing increased compressive stress and accelerated degeneration in their lower lumbar facet joints and higher shear forces in the lower lumbar disks, finally resulting in wear and vertebral slip.
In a prospective controlled study, Aono et al19 arrived at a similar conclusion by correlating greater PI with a greater likelihood of DS.
Lastly, a radiological study by Lazennec and collaborators20 that focused on the importance of considering the pelvis when planning for surgery of the lumbar spine found that sacral tilt (ST) was strongly correlated with postfusion pain.
Experimental Studies
Umehara et al21 first demonstrated how the intraoperative kneeling position with 90 degrees of hip and knee flexion results in hypolordosis of the fused segments. They then attempted to shed light on the biomechanical effects of a hypolordotic lumbar spine in a cadaveric study in which they mimicked the hypolordosis created by the approach that entails 90 degrees of hip and knee flexion. They found that decreased lordosis in the instrumented segments significantly increased the load across the posterior column and increased the lamina strain in the adjacent levels, which could explain the greater propensity for ASD to manifest in patients with a loss of lordosis.
In another cadaveric study, Akamuru and colleagues22 aimed to understand how different sagittal alignments affect the movement of adjacent segments. They found that when compared with a baseline specimen, hypolordotic fusion caused greater flexion-extension movement in the cranially adjacent segments and that fusion in a nonanatomic hyperlordotic conformation induced greater flexion-extension motion in the caudally adjacent segments.
An experimental in vivo study by Oda et al23 aimed to learn more about the effects of kyphosis on the biomechanics of adjacent segments. The authors found that fusion in kyphosis was able to induce ASD. They reported that kyphosis increased the lamina strain and produced significant degeneration in the superior adjacent segments, and they therefore deduced that hypolordosis of the lumbar spine may cause facet joint arthritis, leading to ASD in the upper adjacent segments.23
■ How to Improve Sagittal Alignment
Preoperative Planning
In the previous sections we presented DATA supporting the importance of alignment in patients undergoing spinal fusion. Surgical planning and execution in patients undergoing spinal fusion for degenerative conditions should include an attempt to optimize sagittal alignment to increase the clinical benefit from the procedure and to minimize the risk of ASD.
As each patient’s spinopelvic parameters are unique, planning the surgical intervention should ideally include full-spine, standing lateral radiographs. Nonetheless, standing lumbar radiographs that include the hip joints have been shown to be effective in measuring PI, PT, LL, and lower LL, which are some of the most important parameters required for evaluating sagittal alignment and in screening for sagittal malalignment.24 Planning should include an assessment of PI, LL (globally and for each segment), as well as a calculation of the ideal LL (again, globally and for each segment). No universal agreement exists on the best formula to predict ideal LL. One suggested calculation25 (derived from studies on normal subjects) is LL = 0.54 PI + 32.56. An approximation to this formula is yielded by the rule LL = PI + 10 degrees, which is fairly accurate for patients in the midrange of PI; subjects with very low PI may need slightly more lordosis than PI + 10 degrees, whereas patients with the highest values of PI need a lesser amount of lordosis than PI + 10 degrees. Regarding the distribution of lordosis along the lumbar spine, the clinician should classify the patient’s morphology according to Roussouly et al’s8 types and thus decide the ideal extension of the lordosis and its distribution; again, a practical rule, valid for most patients in the midrange of PI, is to place two thirds of the total lordosis between L4 and S1.
Patient Positioning
After determining the LL goals for the patient, surgical positioning should be considered, as it will affect the potential results one can achieve. Different modalities of prone positioning result in different degrees of lordosis in the patient. As demonstrated by Umehara et al,21 different types of kneeling positions (Fig. 14.4a), although effective in reducing abdominal pressure and epidural bleeding, put the spine into flexion and should be avoided in surgical interventions that aim to attain fusion.
One study by Benfanti and Geissele26 found that as little as 20 degrees of hip flexion on the operative table on a Wilson frame reduced the LL by 25%. Conversely, with extension of the hips, the preoperative LL was maintained (Fig. 14.4b,c).
Interbody Cages and Surgical Approach
Long posterior instrumentation for sagittal deformity can provide and maintain substantial correction until fusion is achieved. Multiple anchors and a long lever arm of the instrumentation explain the good ability of long constructs to provide correction. As a general rule, posterior instrumentation from the lower thoracic spine to the sacrum/pelvis can provide a correction similar to that obtained preoperatively in a fulcrum-extension supine radiograph. Facetectomies or posterior column osteotomies can increase the correction by some degrees per osteotomy level. In case of severe spinal stiffness or correction goals much superior to that observed in fulcrum extension films, three column osteotomies27 (or more recently anterior release techniques28) can be used.
