Cervical Alignment Variations in Different Postures and Predictors of Normal Cervical Kyphosis

A New Understanding

Hwee Weng Dennis Hey, MBBS (Sing), MRCS (Ire), MMED (Orth), MCI (Sing), FRCSEd (Orth), FAMS (Orth); Eugene Tze-Chun Lau, MB BChir (Cantab); Gordon Chengyuan Wong, MBBS (Sing); Kimberly-Anne Tan, MBBS (Aus), BSc (Med) Hons; Gabriel Ka-Po Liu, MBBCh (Ire), MSc (Ire), FRCS (Ire), FRCSEd (Orth); Hee-Kit Wong, MBBS (Sing), MMED (Surg), FRCS (Glas), MCh (Orth) Liv, FAMS (Orth)


Spine. 2017;42(21):1614-1621. 

In This Article


Conventional teaching describes normal cervical spine alignment as being lordotic, with loss of "natural" lordosis often implying a pathological cause.[4,5] This concept occurred during an era in which cervical x-rays were the only investigation required in cases of suspected cervical pathology, and thus preventing this concept from being challenged. The present rapidly growing field of adult spinal deformity and the increasingly recognized importance of sagittal spinal profile restoration[19–21] has led to greater use of full body radioimaging.[16] Understanding of cervical spine alignment is currently changing as we begin to realize that patients without cervical pathology may exhibit nonlordotic alignments.[1,3]

The varied morphology of the cervical spine ranges from lordosis to kyphosis, and at times involves complex forms such as S-shaped or inverted S-shaped curves.[1,3] This could be a result of primary pathologies involving the cervical spine,[22,23] or adaptive changes secondary to pathologies in the caudal spine.[7,8,11,24] Deformity surgeons strive to achieve lordotic corrections of the cervical spine on the principles that cervical spine alignment correlates with clinical symptoms,[5,25] and that lordotic correction of the cervical spine leads to positive outcomes.[26–28] Insufficient restoration of lordosis has also been shown to lead to poor outcomes.[29] Loss of lordosis has been associated with neck pain[5,29] and neurovascular deficits including myelopathy.[30,31] When correcting the cervical alignment of these patients to a more lordotic profile, improvement in neck disability index scores are observed.[26] Although many surgeons strive to adhere to the recommendations of the aforementioned studies to restore CL, there is mounting evidence, which shows that asymptomatic healthy individuals have nonlordotic cervical spine alignments,[1,2,32,33] and therefore generating lordosis may not always necessarily be physiological or ideal.

Our study showed similar findings when compared with other studies, confirming the varying morphology of the normal cervical spine in standing posture.[1,2] In fact, only 27.0% of the patients demonstrated CL from C2-C7 with a mean Cobb angle of −0.6° ± SD 11.1°. In the natural sitting posture, all patients had a lordotic profile (mean C2-C7 Cobb −17.2° ± SD 12.1°) regardless of their standing profile, implying that cervical spine alignment is heavily influenced by posture. Greater lordosis occurs in the erect sitting posture, and even more so in the natural sitting posture due to an increase in T1 slope (from a mean of 17.4° during standing to 30.2° during natural sitting; P < 0.001), and an increase in SVA (from −8.8 mm during standing to 65.2 mm during natural sitting; P < 0.001). Our previous study demonstrated that erect sitting has more radiographic parameters predictive of cervical alignment than standing,[3] a finding likely secondary to adaptive changes corresponding with longer hours spent sitting than standing.[12] The present study did not elucidate any predictive radiographic parameters in the natural sitting posture. This could be attributable to the varied postures individuals assume to "naturally" sit,[34,35] precluding meaningful conclusions from being made.

The present study demonstrates that the highly variable profile of the cervical spine is dependent on T1 slope and SVA. Because both these parameters determine the angle and translation of the C7 vertebra, it is reasonable to infer that they will in turn influence cervical spine alignment. Patients with an SVA less than 10 mm require less lordosis in the cervical spine because they are already translated posteriorly. Instead, such patients tend to have cervical kyphosis, which allows forward translation as a form of compensation for the backward shift in center of gravity. Greater effect is observed in LCA (C4-C7; P < 0.001) than global cervical alignment (C2-C7; P < 0.02). This observation is likely due to the relative proximity of C4-C7 to the thoracic spine, it being lower, allowing greater magnitude of translational compensation.

Similarly, a low T1 slope of less than 20°, which may or may not be a result of negative sagittal balance, also promotes cervical kyphosis. This is due to the need to flex the cervical spine, to generate horizontal gaze. T1 slope of less than 20° is a more significant predictor of global cervical kyphosis as compared to SVA less than 10 mm (OR 17.33, P = 0.001 vs OR 11.67, P = 0.02, respectively). This is unsurprising as the most direct and biomechanically advantageous way in which the body can compensate for angular changes (such as loss of T1 slope) is through opposing angular changes (i.e., increasing cervical kyphosis). SVA is a less direct, albeit important, predictor of cervical kyphosis. Employing the SVA to correct angular changes would inevitably result in multiple noneconomical alignment changes throughout the whole body. The above concepts, which demonstrate SVA and T1 slope as key predictors of kyphotic cervical alignment can also be observed in patients with adolescent idiopathic scoliosis.[8] These patients often have a kyphotic cervical spine due to hypokyphosis of the thoracic spine[7,8,24] secondary to vertebral rotation[8] and negative SVA.[24]

The present study is the first to investigate the effects of natural sitting on cervical spinal alignment while comparing it with conventional standing and erect sitting postures. It has previously been shown that a higher T1 slope in healthy individuals predisposes to greater CL,[36] similar to how a pathological hyperkyphotic thoracic spine does.[11] We were able to reiterate this concept by studying individuals in natural sitting posture (Figure 1), a position associated with high T1 slope and positive SVA.

Figure 1.

EOS images of two representative patients with changes in cervical alignment from standing to erect sitting to natural sitting. Lordosis increases particularly in the lower cervical spine when transitioning from standing to erect sitting to natural sitting due to forward shift in sagittal vertical axis (SVA) and increase in T1 slope. Comparing between both patients, a negative SVA and low T1 slope in standing predisposes patient 2 to a kyphotic cervical alignment.

The study also challenges conventional thinking that the normal cervical alignment is necessarily lordotic. We confirmed the existence of a spectrum of cervical spine morphologies during standing (namely lordosis, straight, S-shaped and kyphosis) and our findings suggest that these are perhaps the consequence of individual variations in SVA and T1 slope. In particular, physiological kyphosis of the cervical spine could occur in the context of low T1 slope and negative SVA (Figure 2). Although T1 slope has already been established as a strong predictor of cervical alignment, our findings highlight the need to consider as well the effect of a negative SVA, so as to better appreciate the full spectrum of normal spinal alignment. Although previous studies have shown that the creation of CL generally leads to good outcomes,[5,9,29] a tailored approach in which both T1 slope and SVA are considered may prove useful in providing even more optimal deformity correction targets.

Figure 2.

Diagram of standing cervical profile based on SVA and T1-slope. A negative SVA and low T1-slope predisposes to a kyphotic cervical spine. A higher T1 slope and positive SVA predisposes to a more lordotic cervical spine. In the natural sitting posture, both SVA and T1 slope increase drastically. This posture generates the largest cervical lordosis. SVA indicates sagittal vertical axis.

Although the present study has a relatively small number of subjects, power analysis ensured that the numbers are sufficient to prove a difference between the three x-rays performed in the same individual using a paired statistical test. Large SDs are a reflection of the large variations in cervical morphology, which has been evidently shown in the literature. This should be appreciated when studying radiographic alignments of the cervical spine. The predominantly male cohort is a potential limitation of this study due to hormonal differences between gender which may indirectly affect spinal alignment. However, there is currently no evidence in the literature suggesting gender as a confounding factor of sagittal alignment.

In conclusion, this study adds to the growing body of evidence describing the cervical spine as a highly variable structure with a wide spectrum of morphology. Key determinants of its alignment include SVA and T1 slope, even in patients with non-pathological cervical spines. Physiological postures such as natural sitting increase T1 slope, resulting in CL. Conversely, low T1 slope and negative SVA lead to cervical kyphosis. Therefore, routine lordotic correction of the cervical spine may not necessarily be physiological or beneficial in certain patients. Deformity correction targets should thus be individualized and determined with consideration of these parameters. Future studies should be performed to understand the value of alignment corrective surgeries in patients undergoing cervical fusion for radiculopathy or myelopathy.