J Manipulative Physiol Ther 2005 (Mar); 28 (3): 187-193 ~ FULL TEXT
Jeb McAviney, MS(Chiro), Dan Schulz, BSc, Richard Bock, MS(Chiro),
Deed E. Harrison, DC, Burt Holland, PhD
Deed Harrison, DC,
123 Second Street,
Elko, NV 89801
Objective: To investigate the presence of a “functionally normal” cervical lordosis and identify if this and the amount of forward head posture are related to neck complaints.
Methods: Using the posterior tangent method, an angle of cervical lordosis was measured from C2 through C7 vertebrae on 277 lateral cervical x-rays. Anterior weight bearing was measured as the horizontal distance of the posterior superior body of the C2 vertebra compared to a vertical line drawn superiorly from the posterior inferior body of the C7 vertebra. The measurements were sorted into 2 groups, cervical complaint and noncervical complaint groups. The data were then partitioned into age by decades, sex, and angle categories.
Results: Patients with lordosis of 20° or less were more likely to have cervicogenic symptoms (P < .001). The association between cervical pain and lordosis of 0° or less was significant (P < .0001). The odds that a patient with cervical pain had a lordosis of 0° or less was 18 times greater than for a patient with a noncervical complaint. Patients with cervical pain had less lordosis and this was consistent over all age ranges. Males had larger median cervical lordosis than females (20° vs 14°) (2-sided Mann-Whitney U test, P = .016). When partitioned by age grouping, this trend is significant only in the 40- to 49-year-old range (2-sided Mann-Whitney U test, P < .01).
Conclusion: We found a statistically significant association between cervical pain and lordosis <20° and a “clinically normal” range for cervical lordosis of 31° to 40°. Maintenance of a lordosis in the range of 31° to 40° could be a clinical goal for chiropractic treatment.
From the Full-Text Article:
Assessing “normal” for any given field of study is of value to have a reference point from which to draw comparisons. This is important in chiropractic, where assessment, diagnosis, treatment methodology, and patient education require a reference point on which to base outcomes. We examined the existence of a normal value for cervical lordosis (ARA) and AWB. Here, normal refers to the anatomic configuration least likely to predispose to cervical pain syndromes and therefore provides a definition of “clinical” relevance.
For subject inclusion, only recognized biomechanical and orthopedic causes of pain were considered. The diagnoses given for the cervical symptoms ranged from cervical facet sprain/strain and cervical disk syndromes to cervicogenic headaches. Patients who presented to the clinic with other types of primary complaints (eg, lumbar and peripheral lower limb complaints) or for spinal screening purposes were included in the noncervical complaint group if cervical or referred cervical symptoms were absent.
Our results illustrate that kyphotic (ARA < 0°) and straight (ARA = 0°) cervical curves are more likely to be associated with a cervical complaint. This range of curvatures represented 25% of the total sample and is therefore of high statistical and clinical significance (Fig 2). Our findings verify that straight and kyphotic cervical configurations are not normal variants. In support of this, only 4% of this group (straight and kyphotic) were without cervicogenic symptoms (Fig 4). Previously, Gore et al  found an incidence of 9% of segmental kyphosis in 200 asymptomatic subjects and found no subjects to have a complete kyphosis from the C2 through C7 vertebrae. Compared with the data from Gore et al  in asymptomatic subjects, our data in patients with neck pain strongly suggest that kyphosis is a risk factor for neck pain.
All subjects with an S-shaped configuration were excluded. This was done for 3 reasons. First, we only measured the ARA from the C2 through C7 vertebrae and felt it inappropriate to include subjects with local kyphotic segments in the groups with a lordotic ARA. Second, we wanted to see the clinical significance of neck pain in relation to hypolordotic, kyphotic, and straightened cervical curves. Third, the findings from Hardacker et a [l5] and Harrison et a [l7] demonstrated that segmental kyphosis is a significant risk factor in neck pain as it occurs 35% to 39% of the time in this subject population.
Concerning hypolordosis and neck pain, we identified an association between cervical pain and those patients with a cervical angle of 20° or less (P < .001). The ARA may be used as a dichotomization parameter into cervical pain and noncervical pain groups. The statistical verification of this hypothesis was performed by using the ROC analysis (Table 2). The purpose is to see how well ARA classifies noncervical pain vs cervical pain subjects. A ROC curve is a plot of sensitivity (1 - false negatives, desirable) vs specificity (false positives, undesirable) for the entire spread of cutoff values. An AUC of 1.0 is a perfect test whereas a test that is no better than flipping a coin has an AUC of 0.5. In Table 2, we found the AUC was 0.803 and this is considered to be a good classification between the groups. The optimum cutoff values were ARA of 19° and 20° depending upon which criteria were used.
Across the entire sample set, the “average” value of cervical lordosis was 14.5°, but the average symptomatic cervical spine had an ARA of 9.6°. This 9.6° average lordosis in patients with neck pain is very close to the values reported by Harrison et al in their patients with chronic neck pain undergoing treatment (ARA of 12.4° ± 11.0° and 4.2° ± 12.7°). [20-21] In contrast, the average asymptomatic cervical curve in our study had an ARA of 23.4°, which could be suggested as a minimum standard for the cervical lordosis. However, 3 subjects in this group had straightened or kyphotic necks and our data suggest this to be abnormal. By eliminating these subjects, we obtain a value of 25°. Our value of 25° is close to the average from Gore et al  of 23° (excluding the 18 subjects with segmental kyphosis) and the average ARA of 22.3° found by Owens and Hoiris. 
Aside from statistically significant findings, Fig 4 gives a clinical indication of the trends associated with lordosis and cervical complaint. In Fig 4, the ARA range with the least percentage of cervical complaints was 31° to 40°. Therefore, a clinically (functional) normal ARA range could be from 31° to 40° and is consistent with the average ARA of 34° found by Harrison et al.8 However, several of these ranges suffer from small sample sizes, especially greater than 45°.
In the current study, we found that patients with cervical pain had less lordosis across all age ranges and no trend in cervical lordosis with age was identified (Fig 2). Similarly, in 372 headache subjects matched to normal controls, Nagasawa et al  found that the cervical curves of those who have headache were straight more frequently with increasing age. Furthermore, these data contrast nicely with the findings from Gore et al,  where cervical lordosis was found to increase with age in asymptomatic subjects.
In our 277 subjects, ages ranged from 9 to 78 years but only small sample sizes were obtained for the 0 to 20 and 60+ years age groups because very few people younger than 20 years and older than 60 years attend the clinic. In addition, many x-ray films from the 60+-year-old range were excluded because of severe degenerative changes. The exclusion of subjects with cervical degenerative changes is based on the finding that degeneration is a significant risk factor for the development of neck pain at 10-year follow-up in previously asymptomatic subjects. 
The exclusion of subjects with degenerative disk disease, however, may have influenced the current results regarding neck pain and AWB. Although Fig 5 shows that the chance of neck pain increases as the amount of AWB approaches 40 mm, this result was not statistically significant. This might be because of the limited number of subjects with AWB greater than 40 mm. There is evidence to suggest that increased magnitudes of AWB increases the stresses on the vertebral body and disks and likely leads to the development of cervical disk degenerative disease.  By eliminating significant degenerative disk disease we may have artificially eliminated large AWB postures of the head. Another explanation for the limited number of subjects with AWB of more than 40 mm is the exclusion of subjects with S-shaped cervical curvatures. It is well documented that increased magnitudes of AWB of the head will cause flexion of the lower cervical spine and extension of the upper cervical spine creating an S-shaped cervical curve. [6, 23] Therefore, we may have unintentionally excluded subjects with large AWB postures.
In addition to the small sample sizes in some of the cervical ARA (>45°) and AWB (>40 mm) categories, there are other limitations to our study design. First, it would be important to include pain intensity and functional disability scores in the analysis of sagittal cervical alignment and its effect on neck pain. For example, it might be possible that certain magnitudes of AWB and cervical lordosis are associated with more severe disabling pain. Second, because of the retrospective study design, there is no way to verify that the x-ray positioning procedures were truly standardized. Some might believe that this could be a confounding factor. However, Harrison et al18 presented original data on initial and repeated lateral cervical x-rays in 3 different control groups and compared their results to 12 other manuscripts in the literature. They concluded that x-ray positioning is highly repeatable even when x-rays were taken by different examiners. Furthermore, it would be extremely unlikely to have such dramatic differences in cervical lordosis between the 2 groups that are neither age- nor sex-dependent if the difference were merely a result of x-ray positioning.
Thirdly, for complete analysis of the cervical lordosis, all segmental angles need to be measured. Harrison et al  pointed out that segmental angle analysis allows the discrimination of S-shaped curvatures and areas that have too much angulation, suggestive of damage to the connective tissues (instability). In the current study, we cannot rule out the possibility that some of the patient categories (especially ARA >40°) had hyperextension at one or more segmental levels thus causing pain.
Lastly, there is now evidence showing that the cervical lordosis can be increased with chiropractic treatment methods in patients with chronic neck pain. [20-21] However, no direct evidence exists documenting the effect on symptoms by introducing a cervical lordosis into a previously kyphotic neck. We hope that the results from our study enable the clinical application of normal values for cervical lordosis.
Patients with straight and kyphotic cervical curves were 18 times more likely to present with cervicogenic symptoms. We found a statistically significant association between cervical pain and lordosis of less than 20°. In addition, we found a “clinically normal” range for cervical lordosis between 31° and 40°, reflected by an increase in the number of cervical complaints above and below this range.
It is suggested that maintenance of a lordosis in the range of 31° to 40° might be a clinical goal for chiropractic treatment. The patients in this study demonstrated a functionally normal range of cervical lordosis of 20° to 40°, a range of 20°. Values outside this range, especially above 40° and less than 20° demonstrated a predisposition to assorted cervical symptoms. However, larger sample sizes in certain angle categories are needed for verification of this finding. Future projects should address the issues of larger sample size, pain and disability indices, and segmental angle measurements of cervical lordosis to increase understanding of the relation between sagittal cervical alignment and pain.