Natural head position and natural head orientation: basic considerations in cephalometric analysis and

research

Lundstrom", F. Lundstrom**, L. M. L. Lebret***, and C. F. A. Moorrees*** Departments of Orthodontics,

*Karolinska Institutet, Faculty of Dentistry, Huddinge, **Public Dental Health Service, l.inkopinq, Sweden,
and ***Harvard School of Dental Medicine, Boston, MA, USA

SUMMARY

Natural head position (NHP) provides the key for meaningful cephalometric analysis because an extracranial
reference line is used instead of intracranial reference lines, known to be subject to considerable biological
variation in their inclination. Although the principle of natural head position is being recognized in the
orthodontic literature, its registration may contain an element of unavoidable error that requires correction. A
new concept of natural head orientation (NHO) has been introduced to maximize the contribution of NHP to
cephalometrics. It is defined as the head orientation of the subject perceived by the clinician, based on general
experience, as the natural head position in a standing, relaxed body and head posture, when the subject is
looking at a distant point at eye level. The present study was designed to determine the accuracy and validity
of NHO, using lateral profile photographs cut in a circular shape, obtained from 27 orthodontic patients, 10-
14 years old. Findings among four investigators showed a high correlation (r= 0.82-0.96), in orientating these
profile photographs in estimated natural head position (NHP). Moreover, investigators showed good
correlation (r= 0.57-0.84) in head orientation after a 3-week interval. Mean differences for the same period
varied between 0.1 and 2.9 degrees. The validity of NHO was supported also by a panel investigation.
Clinicians as well as auxiliary personnel can be trained to make a critical judgement of the recorded natural
head position and correct head orientation, whenever indicated, to enhance the reliability of cephalometric
analysis in clinical practice and research.

INTRODUCTION

The concept of natural head position (NHP) was introduced in orthodontics in the 1950s by Downs (1956),
Bjerin (1957), and Moorrees and Kean (1958). An increasing interest in this concept has occurred from the
1970s to-date, when NHP orientation has been used in more than 40 publications during the last 8 years.
Registration of NHP has been carried out mainly in four different ways:

1. By lateral head radiographs with subjects sitting (Moorrees and Kean, 1958), or standing (Solow and
Tallgren, 1971; Cooke, 1986; Cooke and Wei, 1988a).
2. Photographically with subjects standing or sitting in NHP, (Bjerin, 1957; Lundstrom, 1982b;
McWilliam and Rausen, 1982; Lundstrom and Lundstrom, 1992).
3. Through estimation (Bass, 1991; Lundstrom et al., 1991).
4. Through a combination of radiographic registration and estimation to correct gross errors (Moorrees
and Kean, 1958).
Lundstrom and Lundstrom (1989,1992) recommended the photographic method to keep subjects undisturbed
by ear-rods of a cephalostat and to allow repeated registrations without any added roentgen dosage to the
patient. Many investigators have used mirror orientation as a means to attain NHP, with subjects looking into
their own eyes during the registration (Bjerin, 1957; Moorrees and Kean, 1958). NHP is not a fixed position,
demonstrated longitudinally by Loreille (1975) and by others through triple (Bjerin, 1957) or duplicate
registrations (Moorrees and Kean, 1958; Lundstrom 1982b; McWilliam and Rausen, 1982; Cook eand Wei,
1988a; Siersbaek-Nie1sen and Solow, 1982). Random variation of NHP registrations has been calculated to
within a range of 1.5-2 degrees (Bjerin, 1957; Moorrees and Kean, 1958; Lundstrom, 1982b; Luyk et al.,
1986; Cooke, 1988; Cooke and Wei, 1988a; Lundstrom and Lundstrom, 1992) owing to difficulties for
subjects to reproduce a mean natural head position (MNHP, Lundstrom, 1982a).

Many investigators have studied the variability of reference lines used in cephalometric analysis (Downs,
1956; Bjerin, 1957; Moorrees and Kean, 1958; Foster et al., 1981; Lundstrom, 1981; Cooke, 1986; Cooke and
Wei, 1988b; Lundstrom and Lundstrom, 1989, 1992). Results obtained confirm each other and demonstrate
large variability in the indination of intracranial reference lines related to the extracranial horizontal (HOR).
The small differences in registering natural head position is actually a limited problem in comparison with the
variation of intracranial reference lines.

The increasing interest in NHP motivates further study to elucidate how registration and estimation thereof
can enhance the utilization of this concept.

AIMS

1. To study the reproducibility of photographic NHP-registration and NHP-estimation.

2. To compare means and standard deviations for measurements of the soft tissue nasionpogonion
1inejhorizontal plane angle (N'-Pg'jHOR) for registered and estimated NHP determinations.
3. To correlate assessor differences between registered and estimated NHP determinations.
4. To compare the validity of registration and estimation of NHP. 5. To calculate confidence intervals
for the angle sella-nasion to the horizontal plane (S-NjHOR).

SAMPLE

Subjects selected for the investigation were 27 consecutive patients (14 boys and 13 girls), 10-14 years, from
the orthodontic public dental health dinic in Linkoping,

METHODS

Duplicate photographic NHP registrations, available as routine material in the orthodontic dinic, were used.
The X-ray assistant performing the registrations was experienced in instructing patients how to stand relaxed
in front of a vertical mirror on the wall, 1 m ahead of them. Patients looked into their eyes during the
registration. A fixed plumbline chain was photographed together with the patient. Enlarged copies were
produced for all subjects (18 x 24 em).
Two of the photographs (accepted by A. Lundstrom and C. F. A. Moorrees as reliable examples of NHP)
were kept for instructive purposes for the assessors before commencing NHP estimations (Fig. 1). The
remaining 25 photographs were cut in a circular form (CF: 1 to CF:25). The angle (Fig. 2) between the soft
tissue nasion-pogonion line and the plumbline was calculated after digitizing landmarks, and one upper and
one lower point on the plumbline.

A white rectangular paper (210 x 297 mm) was placed on a dark background, its right side representing the
vertical plane. CF: 1 (and so on to CF:25) was placed on the paper and turned by each assessor into a position
perceived to correspond with NHP. A ruler was placed on CF: 1, through the nasion (N') and pogonion (Pg')
landmarks. The elongation of this line was drawn with a sharp pencil on the rectangular paper below and
above CF: 1. A second estimation was performed on a new paper after a 3-week interval.

The angles (VI and V2, Fig. 2) between the elongated N'-Pg' line and the right side of the rectangular paper
were measured. These angles thus represented the estimated profile orientation in NHP on two separate
occasions. All computer measurements as well as statistical calculations were made by the senior investigator.

Four assessors participated in this study, all of whom were experienced in NHP evaluation. A test was also
performed using two undergraduate dental students, correlating their assessments with one experienced
investigator (assessor 1).

Figure 1 Two patients registered and accepted as good examples of natural head position.

In order to compare the validity of the two methods, records differing more than 4 degrees between methods
were assessed by a panel of orthodontists and orthodontic students. Observers were asked to evaluate random,
coded pairs of photographs to define the head orientation they perceived as most natural. This part of the
study was performed by 10 orthodontists and 4 orthodontic students.

The angle S-N/HOR was calculated from the angles N'-Pg'/VERT and the angle S-N/N'-Pg', as shown in Fig.
3.

Statistical analysis Means and standard deviations were determined for VI (= Vreg) and V2 (= Vest). The
standard error of single determinations (s.) were calculated (based on differences in V-angles between first
and second determinations), according to the formula:

where is the sum of squared differences and N the number of duplicate determinations. Differences between
registered and estimated head orientations (Vreg - Vest) were correlated between assessor 1, and assessors 2,
3, and 4, respectively, through calculation of correlation coefficients (Bravais-Pearson).

RESULTS

Photographic registration showed a method error (sJ of 1.6 degrees, while s, varied for estimations between
s,= 1.0 degrees and s,= 2.7 degrees for different investigators (1, 2, 3, and 4=AL, FL, LL, and eM,
respectively). The comparatively high value of 2.7 degrees for assessor 2 was due to a mean difference of 2.9
degrees ((P<0.01) between first and second assessments(table IA).

Table lA Correlation coefficients(r) showing degree of covariation between Ist and 2nd measurements of the angle N'-Pg'/VERT and the
standard error of single measurements (s;) of the same angle for photographic registrations and assessors (in degrees). N=25. Si=
±..}'Ld2/2N.

Table IB Correlation coefficients (r) showing intercorrelation between assessors for the V-angle (I st estimation). Mean correlations for
each assessor.

Figure 2 Three instances showing marked differences in head positioning for registered (left) and estimated (right) natural head
position (diff. 9, 8, and 6 degrees, respectively). Measurement of the N'-Pg'jVERT angle.

In order to compare assessors, intercorrelation coefficients were calculated between assessors (lst
estimations). Mean correlation coefficients varied between 0.52 and 0.63, indicating conformity of the
assessors for their first series of NHP estimations (Table 1B). Mean differences between registered and
estimated V-angle measurements (1st determinations) were for assessors 1 and 2 small (1.1 and 1.2 degrees,
Table 2) and not statistically significant. For assessors 3 and 4 these differences were slightly larger (2.4 and
2.4 degrees, P<0.05 and p<0.01). All differences were positive indicating NHP, somewthat extend in relation
to NHP (table 2).

Figure 3 Calculation of the S-NjHOR angle from the N'-Pg'jVERT angle (V) and the S-NjN'-Pg' angles (V3).

Table 2 Means and standard deviations of the N'-Pg'jHOR angle (in degrees) for duplicate registered and estimated NHP determinations.
N = 25.

For all assessors the standard deviation for estimated NHP (0.7-3.5 degrees) was smaller than for registered
NHP (4.1 degrees), statistically significant for assessors 1 and 4 (P<0.001; table 2).

Direction and amount of extension/flexion between registered and estimated head positions showed a strong
correlation for all assessors (r=0.82-0.96, P<O.OOI), the strongest (r= 0.96) between assessors 1 and 4
(Table 3, Fig. 4). The panel investigation showed that the estimated head position by the majority of
observers (83 per cent) was perceived to be the 'most natural' (Table 4).

In six of the seven patients (photograph 2 missing for one subject) with large differences in head orientation
(>4 degrees) between regis-tration and estimation of NHP, duplicate photographs showed similar head
positions in both registrations in five instances (Table 5).

Five individuals with differences exceeding 5 degrees were, as compared to estimations, perceived to have an
unnatural, flexed photo-graphic position of the head (Fig. 2). The cor-responding mean differences (Table 2)
agreed with this finding.

Calculation of the angle S-N/HOR was lim-ited to observers 1 and 4, who showed a remark-able covariation
in their perception of NHP. The correlation coefficient for S-N/HOR est between them was as high as r=0.93
with a non-significant mean difference of only 1 degree. The 95 per cent confidence interval range was 23.8
degrees for photographic registrations, versus 13.6 degrees by estimations for assessor 1 and 13.0 degrees for
assessor 4 (Table 6). Maximum and minimum differences in the sample are shown in Fig. 5. The standard
devi-ation of S-N/HOR(NHO) = 5.3 degrees was larger (P<O.OOI), than for S-N/HOR(NHO) (3.3 and 3.1
degrees).

DISCUSSION

Mean correspondence between assessors is, on the whole, very good. The slight extension for assessors 3
and 4 compared to assessors 1 and 2 (2-3 degrees) may depend on a slightly differ-ent perception of natural
head position between centres. The fact that assessors 3 and 4 were used to sitting patients in their NHP
evaluations

Table 3 Correlation coefficients (r) showing covari-ation between assessor I and the other assessors for differences between registered
(Vreg ) and estimated (Vest) values of the N'-Pg'jVERT angle (1st obs).

Figure 4 Scattergram showing covariation for differences between photographic registered and estimated natural head position
between assessors I and 4.

Table 4 Numbers and percentages of answers from 14 panel members, asked to state their preference among random pairs of
photographs of photoregistered (NHP) and estimated (NHP ) natural head positions, in order to choose the 'most natural' one, in seven

subjects differing more than 4 degrees between methods.

may perhaps have had some influence. According to Bjerin (1957), who compared photographic registrations
for standing and sitting subjects, the difference in NHP was on the same level (diff.= 1.9 degrees).

The strong correlation (Table 3, r= 0.82-0.96) between assessors for their deviations from NHP in estimating
natural head orientation (NHO) shows that photographic registrations must be checked to eliminate obvious
errors in head orientation, resulting in clinically misleading information.

The panel investigation (Table 4) confirms this conclusion. An experienced clinician should be able to adjust
the head orientation of the photograph to what is perceived to be the natural head orientation (NHO) (Fig.2) in
instances for which· registration has resulted in an 'unnatural' head position. The definition of NHO is, in
general, synonymous with NHP. Ifthe registered head position shows 'unnatural' flexion or extension of the
head, an adjustment should be undertaken by a person trained for this task.

Table 5 First and second of photoregistered N-PgjVERT angles in six patients with large differences (> 4 degrees) between
photoregistered (Vreg) and estimated (Vest)NHP-related angles (in degrees). A negative angle indicates that the estimated head position
is flexed in relation to the registered head position.

Table 6 Means, standard deviations, and confidence intervals (95 per cent) for the angle S-NjHOR (in degrees). The horizontal plane was
related to the head for photographic registration (NHP) and for estimation (NHO) of natural head position.

Figure 5 Minimum (1.7 degrees) and maximum (11.2 degrees) inclination of the S-N line related to the mean horizontal (HOR)
estimated by assessors 1 and 2 (first registration).
A difference between NHP and NHO is not necessarily the result of gross errors in the registration of NHP.
The strong correlation between the two photographic registrations of the V-angle (r=0.9) shows that the
difference between NHP and NHO essentially results from habitual deviations of NHP from NHO.
Individuals, who habitually keep their heads flexed or extended in a way perceived as unnatural (Fig. 2),
should not be cephalometrically analysed from the horizontal related to their registered head position. The
NHO related horizontal is then the preferred reference line.

This procedure may be questioned because of its subjective character, but for the time being, no better
alternative seems to be available.

The systematic difference of 2.9 degrees for one assessor after a 3-week interval indicates that individual
deviations from mean standards must be judged with caution. A 3-degree difference is, however, small and
should be of little significance. This problem should be less liable to give clinically misleading information if
the extracranial S-HOR line is chosen instead of a more variable intracranial reference line. According to
Lundstrom and Lundstrom (1992), the S-NjHOR showed a standard deviation of 5.3 degrees which can be
compared to 1.6 degrees head position instability in photographic registration ofNHP in the present study
(Table 1).

Differences must be expected between individuals in their ability to estimate NHP. Two undergraduate dental
students were tested in this respect on the present material. One student correlated relatively well with
assessor 1 (r= 0.67, P < 0.001), while for the other a weaker correlation was found (r=0.44, P<0.05). To what
extent training can compensate for initial difficulty in this respect, remains an interesting question.

The panel study confirms the difference between individuals in estimating NHP. Ten of 14 panel members
were very close in their evaluations (complete concordance for four of them and difference in only one case
for six), while the other four differed in two instances (one member) or three cases from the majority. It seems
probable that the first ten have a better grasp for natural head orientation than the last four.

The finding (Table 6) that variation in the S-NjHOR angle is larger for photographic registrations than for
estimations of NHP should not be overlooked (SD = 5.3 degrees, versus 3.1-3.3 degrees, P<0.001). This
difference in variability is interesting and the following three explanations may be considered:

1. The error component of the photographic registrations is larger than that for estimations.
2. Registrations and estimations do not express the same concept.
3. Assessments are biased through a tendency for assessors to evaluate the inclination of the N'-Pg' line
closer to the mean value than appropriate.

The strong correlation (r=0.90) between duplicate photographic registrations, confirmed by V-angles given in
Table 5, do not favour hypothesis no. 1. The difference in standard error for assessor 1 and 4 between
registrations and estimations (1.6 degrees, versus 1.0 and 0.6 degrees, p<0.05 and < 0.001 ). may be
interpreted as a confirmation of this hypothesis, as at least part of the explanation. Moorrees and Kean (1958)
recommend that a small mirror (10 em diameter), vertically adjustable to the pupillary level, should be used at
NHP registrations rather than a long mirror, the purpose being to ensure that the visual axis of the subject is in
the horizontal plane. Since this method was not used in the present study it is not possible to know whether
the error component might have been reduced, had that method been followed.

The validity comparison (Tables 4 and 5) of large differences between registered and estimated NHP
determinations, showed that estimated NHP gives a 'more natural' head orientation than the registered
orientation, indicating that the second hypothesis should give part of the explanation. The result of the panel
investigation confirms the corresponding study on Bjerin's sample by Lundstrom et al. (1991).

To what extent the third hypothesis may be valid is difficult to assess, but is not confirmed by the panel
examination of extremes, for which estimations indicated a 'more normal' head orientation than for
registrations. The variation of the S-N line to the horizontal is considerable when NHP is used (SD = 5.3
degrees, Table 6) and although less if NHO is employed, it is still substantial:

SD (assessor 1) 3.3 degrees, 95 per cent CI range = 13.6 degrees

SD (assessor 4) 3.1 degrees, 95 per cent CI = 13.0 degrees

These findings confirm the marked variability of the S-N line and the risk of misleading information when it
is used for clinical cephalometric analysis.

Based on the findings presented a distinction is made between the following concepts:

1. Natural head position (NHP), defined as the registered, mirror orientated head position in a relaxed
body and head posture.
2. Natural head orientation (NHO), defined as the head position estimated by a trained clinician as the
natural one.

NHO determinations will correct gross errors in registrations, as well as habitual tendencies for some
individuals to keep the head in what is perceived as a flexed or extended posture. NHO should be of interest
mainly in combination with cephalometric analysis of patients with malocclusion, especially for cases
requiring orthognathic surgery (Lundstrom et al., 1993).

Norm values based on NHP have been published adapted to mesh analysis (Moorrees and Lebret, 1962;
Moorrees et al., 1976; Moorrees, 1985), by Cooke and Wei (1981b, 5-point cephalometric analysis),
Lundstrom and Lundstrom (1989; maxillary and mandibular prognathism), Lundstrom et al., (1992, soft tissue
profile), and Viazis ( 1991), cephalometric analysis). Viazis mean values for the angles N-A/HOR and N-
B/HOR differ considerably (5.3 and 4.5 degrees) from those published by Lundstrom and Lundstrom (1989).
Since Viazis' (1991) values are not based on individually determined NHP-orientated material, but were
derived from the Bolton standards on the assumption that in all illustrations the heads were orientated in NHP,
Viazis' norm values must be confirmed by further analysis.
Norm values are needed for standard cepalometric variables based on the NHO concept as defined above.

RESEARCH

In connection with research projects, aiming at studying changes of natural head orientation over time,
uncorrected NHP must for obvious reasons be used, for instance after orthognathic surgery (Fromm and
Lundberg, 1970), treatment to relieve blocked nasal passages to attain free nasal breathing (Linder-Aronson,
1979; Solow and Greve, 1979) or with age (Tallgren and Solow, 1987). Reference can also be made to
analysis of basic differences in facial configuration between various malocclusion types (Cole, 1988;
Lundstrom and Cole, 1990). NHP registrations are also indicated in certain ethnic comparisons between
population groups (Yen, 1973; Lundstrom and Cooke, 1991).

In growth studies based on NHP, superimposition of radiographs on the initial, or best, record in a series may
be used (Moorrees et al., 1985).

CONCLUSIONS

1. Because of the large variation of intracranial lines, the extracranial reference line S-HOR should
substitute, or at least supplement, the use of intracranial reference lines for cephalometric analysis of
patients with malocclusion.
2. Photographic registration of natural head position is recommended. This registration should be
complemented with an adjustment when head orientation is perceived to be 'unnatural'.
3. Estimation of natural head orientation (NHO), performed by experienced clinicans, gives a valuable
addition to the natural head position (NHP) concept.
4. Clinicians as well as auxiliary personnel can be trained to make critical judgement of the recorded
head orientation, whenever indicated, to enhance the reliability of cephalometric analysis in clinical
practice and research.