J Manipulative Physiol Ther 2002 (Nov); 25 (9): 550–555 ~ FULL TEXT
Kristjansson E, Jonsson H Jr
PhD candidate in Health Science at the Faculty of Medicine,
University of Iceland,
Chief Orthopaedic Surgeon at Landspitalinn,
SEE ALSO: Sagittal Alignment of Cervical Flexion and Extension: Lateral Radiographic Analysis
OBJECTIVE: To reveal whether women with chronic whiplash-associated disorder (WAD) symptoms, grade I-II, demonstrate regional and/or segmental radiographic signs of altered cervical lordosis.
DESIGN: Case-control study.
SETTING: Radiography department at a university hospital.
PARTICIPANTS: Three age-balanced groups comprising 120 women. The case group included women with chronic whiplash syndrome (n = 41), and the control group included women with chronic insidious onset neck pain (n = 39) and an asymptomatic group (n = 40), who were given baseline data. The sample was referred from informed doctors and physiotherapists.
INTERVENTION: The women sat in a standardized sitting position and radiographs were taken in a lateral position with fluoroscopic control for alignment.
OUTCOME MEASURES: Two distinct measurements were taken; 1 of the angles of the upper and lower cervical curvatures, respectively, and 1 of the angles between the inferior borders of each pair of vertebrae in the lower cervical spine. The 3 groups were compared on the ratio of the lower to upper cervical spine angles and on the mean angular values for each segment in the cervical spine.
RESULTS: The whiplash group showed a decreased ratio between the lower versus upper cervical spine but comparisons between groups were not statistically significant. The whiplash group was in a significantly more flexed position at the C4-C5 level compared with the asymptomatic group (P =.007). The reliability measures have to be strengthened to render these results definitely conclusive.
CONCLUSION: The whiplash group exhibited a different configuration of cervical lordosis. This is clinically important and needs to be studied more closely.
From the Full-Text Article:
This study investigated whether women with chronic WAD of grades I-II had radiographic signs of altered configuration of the cervical lordosis compared with a group of women with insidious onset neck pain and an asymptomatic group. Figure 3 shows that in comparison with the other 2 groups, the group with whiplash exhibited a decreased ratio of lower to upper cervical lordosis. This difference was not statistically significant but may have clinical importance. The statistically significant difference for the C4-C5 segment between the whiplash group and the asymptomatic group, a mean difference of 3° (95% CI, 0.8–5.2) is considered clinically important.
According to Bland and Altman,  plotting the difference between 2 methods, testers, or repeated measurements against their mean reveals their agreement. The aforementioned between-group differences at the C4-C5 segment are within the 0.38° ± 4.01 limit of agreement for that level (Table 1). The statistical significance and the practically important difference at the C4-C5 segment between the group with whiplash and the asymptomatic group show that an improvement of the measurement method is needed. For the angles shown, the limits of agreement ranged from 2.84° to 4.21°. However, this method was considered better than the unit-free intraclass correlation coefficients, which have several disadvantages when documenting the reliability of clinical measurements. [27, 28] To improve the agreement between measurements, assessing the whole radiographic procedure will be important.
The seated position has several limitations when documenting the sagittal alignment of the cervical spine because the cervical curvature is sensitive to the position of the pelvis, the trunk-thigh angle, and the inclination of the backrest. [29-31] There is evidence from several studies that a standing position is more reliable. [32, 33] In addition, the Harrison posterior tangent method has been found to more accurately depict cervical curvature than the more commonly applied Cobb method.  Therefore, the standing position and the Harrison method are recommended for future studies.
According to our knowledge, the relation between lower and upper cervical spine lordosis in different patient groups has not been investigated before. Measuring each segmental level's contribution to total cervical lordosis and comparing these values across different groups has also not been a common practice in most research of cervical lordosis. The results of this study show the preponderance of the angle between the C1-C2 vertebrae to the total cervical lordosis (Fig 4).
Comparing the segmental angular values for the asymptomatic subjects in this study with the segmental angular values obtained in a study of 100 asymptomatic subjects conducted by Hardacker et al  is of great interest. In this latter study, higher values were obtained for the C1 level but lower values were obtained for the C2 level compared with the corresponding values in our study. These discrepancies obviously were caused by a different reference point on the posterior aspect of the C2 vertebrae between the 2 studies. The Cobb method used in our study was modified for the C2 vertebra to reflect the contribution of the C2 vertebra to the lordosis more efficiently. A tangent line between the anterior hook-shaped corner and the posterior corner on the inferior body of C2 was found to underestimate the lower cervical lordosis and overestimate the upper cervical lordosis. When we marked the caudal posterior point of the C2 vertebral body on 30 radiograms, as Hardacker et al  had done, instead of the caudal posterior point of the spinal canal, we obtained similar values to theirs. The results of the 2 studies are therefore comparable, despite the different radiographic assessment positions, and show that the lower cervical spine below the C2 vertebra contributes far less to total cervical lordosis (Fig 4). The fact that the atlas contributes most to cervical lordosis is consistent with its weight-bearing function.
Many symptoms in late whiplash syndrome remain obscure. [36, 37] Preventing the development and persistence of these symptoms is therefore difficult. Altered cervical curvature may play a role in the symptomatology of some whiplash subjects. In a whiplash-type movement, the passive integrity of a cervical motion segment may be threatened and thereby its biomechanic  and/or neurophysiologic stability.  Correct segmental alignment of the spine depends on adequate function of the deep local muscles to provide a stable base for efficient limb and spinal movements. [40, 41] The change in the ratio of lower to upper cervical spine lordosis observed in the whiplash subjects in this study may indicate dysfunction of the deep flexors in the upper cervical spine and of the deep extensors in the lower cervical spine. This hypothesis warrants further investigation.
There are much data to support the view that the function of the cervical spine is best preserved and in the least strenuous way by maintaining physiological lordosis. [42-46] Apart from intrinsic factors, the cervical curvature depends on the head on trunk position  and the position of the trunk under the head, including the shape of the thoracic kyphosis.  The inclination of the sacrum and the configuration of lumbar lordosis may also play important roles in the size of the cervical lordosis.  The muscles must be activated more in subjects where the head's line of gravity falls more anteriorly as in forward head posture  or when the thrust line through the cervical spine (cervical gravity line) falls outside the arc formed by the anterior segment (discs and vertebrae) of the column.  Therefore, the misalignment at the C4-C5 level observed in the whiplash group may have clinical importance because the load-bearing capacity of the cervical spine will be greatly reduced. 
This study indicates that the cervical lordosis of patients with whiplash may be differently configured. The group with whiplash showed a decreased ratio of lower to upper cervical spine lordosis. The between-group differences for this ratio were not statistically significant but may be clinically important. The whiplash group was in a significantly more flexed position at the C4-C5 level compared with the asymptomatic group. Future studies should use a standing position and the posterior tangent method to enhance reliability when measuring sagittal alignment of the cervical spine.