Craniofacial characteristics of subjects with normal

and postnormal occlusions-A longitudinal study

W. John S. Kerr, M.D.S., D. Orth., F.D.S., F.F.D.,* and David Hirst, B.Sc.**
Glasgow, Scotland, United Kingdom

Lateral skull radiographs of 85 growth study children taken at 5, 10, and 15 years of age, with
normal/Class I occlusion (normal) and Class II, Division l/Division 2 occlusion (postnormal) were
digitized and analyzed by means of stepwise discriminant analysis. The cranial base angle proved to
be the best discriminator between the two groups, the value at age 5 years being an accurate
predictor of the occlusal type at 15 years in 73% of subjects. Although the majority of subjects grew
predictably, their craniofacial characteristics being compatible with their ultimate occlusal type,
17% showed a growth trend from postnormal to normal and 9% a trend from normal to postnormal.
By 15 years of age, only 8% of subjects possessed an occlusion at variance with their facial type.
The cranial base angle is suggested as the fundamental determinant of jaw relation, but in some
subjects this may be compensated by differential jaw growth manifested by a change in angle ANB.
(AM J ORTHOD D~NTOFAC ORTHOP 1987;92:207-12.)

C
ross-sectional cephalometric studies have
attempted to define the craniofacial differences between
normal/Class I subjects and Class II subjects of both
divisions; the general consensus’.2.3 has been that it is
the position, rather than the size of the mandible, that
produces the Class II relationship.
Bjork“ was one of the first to draw attention to the
relationship between cranial base angle and the position
of the glenoid fossa; Houston’ showed a trend for a
larger cranial base angle in Class II groups, which might
explain the distal positioning of the mandible, although
increased cranial base length also found in this study
might have produced the same effect. Anderson and
Popovich’ have shown a correlation in longitudinal ma-
terial between cranial base angle and occlusal relation-
ship, with Class II malocclusion more commonly found
in subjects with increased angles.
In opposition to this view, some workers’,6 have
found no difference in cranial base angle between the
groups; others’ have noted that the mandible is smaller
in Class II malocclusions.
There is considerable controversy as to how the
cranial base angle should be measured, particularly with
regard to its distal point. Articulare, Bolton point, po-
rion, and basion have all been advocated, making a
*Department of Orthodontics, University of Glasgow
School.
**Department of Statistics, University of Glasgow.
comparison of results difficult. Varjanne and Kosk?
have denigrated the use of articulare because of its
remoteness from the cranial base and advocate the use
of basion because of its anatomic significance in spite
of difficulties in locating it.
To clarify the points of difference between these
two major malocclusion groupings and to ascertain how
early and with what degree of accuracy typical facial
characteristics could be identified, the following study
was undertaken.
MATERIAL (Table I)
The material was that used in a previous study’ and
consisted of lateral skull radiographs of 85 children who
were part of the Belfast Growth Study and of whom
films were available at 5, 10, and 15 years of age, taken
with their teeth in occlusion.
The subjects, none of whom had undergone ortho-
dontic treatment, were divided on the basis of the 15-
year-old study models into (1) normal occlusion, (2)
Class I malocclusion, (3) Class II, Division 1 maloc-
clusion, and (4) Class II, Division 2 malocclusion after
the method described by Angle.
METHOD
Each of the films was traced on two occasions and
the tracings digitized. The 16 points shown in Fig. 1
Dental Hospital and
were plotted and from these the following linear and
angular variables derived:

207
208 Kerr and Hirst Am. J. Orthod. Dentofac. Orthop.
September 1987

Table I. Distribution of material by occlusion

group and age (N = 85)
Normal (n = 51) Postnormal (n = 34)

Normal Class I Class II’ Class IS

s N
. Boys 7 18 10 9
Girls 9 17 11 4
g fi
Mean age {years) Range (years)

Film
Ar 1 5.4 5-6
Ba - 2 10.4 IO-11
3 15.4 14.3-16

SN line, a perpendicular was dropped from it to the

original SN line. The process was repeated for the sec-
ond and third films in each series. The SN line, there-
fore, bore a constant relationship to the cranial base.
The magnification factor for the cephalostat was
0.92 and the mean of each double determination was
used in calculating the results.
Stepwise discriminant analysis was used to separate
Fig. 1. Points digitized. the groups and locate the major areas of difference
between them. The calculations were performed by
means of the BMDP Statistical Software program
1. S-N: Anterior cranial base P7M.” As has been described previously,‘2 in the first
2. N-Ba: Total cranial base step of a stepwise discriminant analysis, the single vari-
3. ANS-PNS: Maxillary length able that best discriminates between the groups is se-
4. Ar-T,: Mandibular ramus height
lected. In subsequent steps variables are selected that
5. T,-Pog: Mandibular body length
best discriminate between the groups conditionally on
6. Ar-Pog: Total mandibular length
7. N-M: Total facial height the variables already entered in the process. Thus, a
8. N-ANS: Upper facial height variable that individually might be a useful discrimi-
9. Ba-S-N: Cranial base angle nator could be omitted if another variable with which
10. Ar-T,/T,-M: GoniaI angle it is highly correlated is already included. At each step
11. SNA: Maxillary prognathism all variables currently in the process were tested to see
12. SNB: Mandibular prognathism if any might be removed given the other variables al-
13. SNPog: Chin-point prognathism ready entered. The procedure continues until no vari-
14. SN/ANS-PNS: Anterior cranial base/maxillaryplane able meets predefined criteria for entry or removal. On
15. SN/T,-M: Anterior cranial base/mandibular plane the basis of the variables selected, a rule is devised that
16. 4 Max. Pl.: Maxillary incisor inclination can be used to classify new cases into a group. A clas-
17. 1 Max. PI.: Mandibular incisor inclination
sification matrix indicates the success rate of the rule
The original SN line was transferred from the age- when applied to new cases from each group. The ob-
5-year-old film to subsequent films and nasion corrected jective in this study was to see how closely the clas-
to this line to eliminate vertical movement in relation sification based on craniofacial measurements corre-
to the cranial base.” This was achieved using the Ad- sponded to the occlusal classification and which cra-
am’s blink comparator* to view the first two films in niofacial measurements best discriminated between the
each series, simultaneously superimposing on de Cos- occlusal groups.
ter’s line. The points on each were traced and the SN
RESULTS
line from the first tracing transferred to the second. If
nasion as the second tracing did not lie on the original Initial attempts to separate the four groups by step-
wise discriminant analysis were disappointing because
*G.S. Ross Ltd., Macclesfield, Cheshire, England of the marked overlap between the normal occlusion
Volumr 92 Craniofacial characteristics of subjects with normal and postnormal occlusions 209
Number 3

Table II. Means and standard deviations for variables by occlusion group and age
I
5 Years IO Years 15 Years

Variable Normal Postnormal Normal Postnormal Normal Postnormal

S-N 64.2 2.9 64.6 2.8 68.5 3.2 69.0 3.1 72.2 3.8 72.7 4.0
N-Ba 93.1 4.2 94.5 3.4 101.2 4.5 102.8 4.2 108.2 5.4 109.5 5.1
ANS-PNS 44.7 2.1 45.5 2.2 41.8 3.1 50. I 2.3 52.2 3.2 55.2 2.9
Ar-T, 42.6 3.2 42.0 2.5 48.3 3.7 46.8 1.9 56.1 4.6 53.8 3.4
Trf’og 66.7 3.0 65.5 3.1 77.2 3.9 75.8 3.7 85.6 4.1 83.1 4.6
Ar-Pog 88.0 3.8 86.1 3.3 99.1 4.8 97.9 3.4 1 Il.6 5.7 107.9 4.8
N-M 99.2 5.6 98.9 4.0 Ill.3 6.5 110.1 5.6 124.9 9.3 122.4 7.5
N-ANS 45.3 3.1 45.7 2.5 51.4 3.4 52.1 2.8 57.9 5.0 58.2 4.0
Ba-S-N 127.4 4.6 132.5 4.3 126.7 4.5 131.7 4.6 127.6 5.3 132.3 4.5
Ar-T2 IT,-M 129.3 4.9 129.0 5.9 125.6 5.6 125.0 5.7 123.8 5.9 122.3 6.4
SNA 81.4 3.5 80.2 2.9 81.4 3.6 80.5 3.1 81.4 3.6 80.9 3.4
SNB 76.9 2.9 74.3 3.1 11.9 2.9 75.1 3.0 78.5 3.3 75.1 3.5
SNPog 76.0 2.9 73.8 2.8 78.1 2.1 75.8 3.0 79.3 3.4 16.4 3.5
SNIANS-PNS 7. I 2.6 7.5 2.6 7.2 2.5 x.3 2.8 8.4 3.0 9.7 3.2
SN/T,-M 36.0 4.9 36.5 4.6 34.5 5.1 34.6 5.6 33.8 6.4 34.0 7.0

Table Ill. Classification-Age 5 years Table IV. Classification-Age 10 years

8 Correct Normal Postnormal % Correct Normal f’ostnormal

Normal 70.6 36 I5 Normal 78.2 40 11

Postnormal 76.5 8 26 Postnormal 82.4 6 28
TOTAL 12.9 TOTAL 80.0

Variable Variable

Step Entered Removed Step Entered Removed

1 Cranial base angle - 1 Cranial base angle -

2 Maxillary length -
3 Total mandibular length -

and Class I malocclusion groups on the one hand and

the Class II groups on the other.
In addition, paired t tests between normal and
Class I groups showed only one significant difference
(P < 0.05), the anterior cranial base/maxillary plane
angle being larger in the normal occlusion group at ages
10 and 15 years. Adams,13 with material drawn from
the same source, could find no consistently significant
differences between normal occlusion and Class I mal-
occlusion subjects in respect to sagittal craniofacial di-
mensions, thus confirming that only small random dif-
ferences occur between these two occlusal groupings.
In comparing the Class II groups, apart from highly
significant differences in upper incisor inclination at all
ages (P < O.OOl), the only differences were a smaller
gonial angle in the Division 2 subjects at ages 5 and
10 years (P < 0.05), and a larger mandibular body
length in Division 2 subjects at age 5 years (P < 0.05).
On this evidence it was considered justifiable to
amalgamate the normal and Class I groups (normal)
and the two Class II groups (postnormal) to ascertain
the fundamental areas of difference between them. In
this process the incisor inclinations (variables 16 and
17) were deleted from the analysis because they were
the basis for the groupings and an average value for the
postnormal groups would have had little meaning.
The discriminant analysis was run with the sexes
analyzed both separately and together. The differences
between the occlusion groups at all ages were found to
be greater than those between the sexes to the extent
that discrimination was not greatly improved by sepa-
ration.
The combined means and standard deviations for
each variable are shown in Table II, and stepwise dis-
criminant analysis at ages 5, 10, and 15 years in Tables
III, IV, and V, respectively.
The analysis shows discrimination between the two
groups improving steadily from 73% to 80% to 92%
210 Kerr and Hirst
Table V. Classification-Age 15 years
% Correct Normal Postnormal
Normal 92.2 41
Postnonnal 91.2 3 31
TOTAL 91.8
Variable
Step Entered Removed
1 SNB -
2 SNA -
3 Total facial height - present at 15 years. Eight subjects exhibited normal
4 Maxillary length -
- characteristics at 5 years, six of whom developed ad-
5 Cranial base angle
6 SN/Pog - versely developing postnormal characteristics by 15
between ages 5 and 15 years as more variables accu-
rately define the differences.
JackknifingI the classification to eliminate bias did
not produce any alteration in the percentages correctly
classified.
Initially, the best discriminator was the cranial base
angle, which was larger in the postnormal group; to
this were added maxillary length and total mandibular
length at age 10 years, the former being larger and the
latter smaller in the postnormal group. By 15 years
prognathism of the jaws was the best discriminator be-
tween the groups, both SNB and SNA being smaller
in the postnormal group as was total facial height.
Of particular interest were the subjects whose cra-
niofacial classifications changed between ages 5 and 15
years or were contrary to their occlusal classifications
throughout. The progression of classification for each
occlusal grouping between ages 5 and 15 years is shown
in Table VI.
Of the 5 1 subjects who developed normal occlusions
by 15 years, 34 exhibited compatible craniofacial char-
acteristics at 5 years that were still present at 15 years.
Fifteen such subjects exhibited postnormal craniofacial
characteristics at 5 years; 13 of these matured favorably
to exhibit compatible craniofacial characteristics at 15
years and two subjects retained their postnormal char-
acteristics, but dentoalveolar compensations neverthe-
less facilitated the accommodation of normal occlu-
sions. Two subjects who exhibited normal character-
istics at age 5 years grew adversely so as to manifest
postnormal characteristics at 15 years, but again ac-
commodated normal occlusions.
Of the 34 subjects who developed postnormal oc-
clusions by age 15 years, 25 exhibited compatible cra-
niofacial characteristics at age 5 years that were still
Am. J. Ortbod. Dent&c. Orthop.
September 1987
Table VI. Craniofacial type for each occlusion
group-Progression from ages 5 to 15 years
Normal occlusion (n = 51) ( Postnormal occlusion (n = 34)
4
5 years
N 34 2
PN w 13 2
N PN 15 years N PN 15 years
years that accommodated postnormal occlusions; two
retained their normal characteristics at 15 years but
supported postnormal occlusions. One subject who ex-
hibited postnormal characteristics at 5 years grew fa-
vorably to show normal characteristics at 15 years, but
nevertheless supported a postnormal occlusion.
The results of this study show that the craniofacial
characteristics of subjects with normal and postnormal
occlusions become more defined with advancing age.
Whereas at 5 years of age almost 30% of subjects
who developed a normal occlusion exhibited charac-
teristics consistent with postnormal subjects and almost
24% of subjects who developed a postnormal occlusion
exhibited characteristics consistent with normal sub-
jects, by 15 years of age this ambiguity had all but
disappeared. Nevertheless, even then approximately 8%
of each occlusion group exhibited facial characteristics
consistent with the other.
The majority of subjects, therefore, grew predict-
ably from 5 to 15 years of age and supported the type
of occlusion that was consistent with their craniofacial
characteristics. A considerable number of subjects who
exhibited postnormal characteristics at age 5 years de-
veloped a normal occlusion and normal facial charac-
teristics by age 15 years. A smaller number moved in
the opposite direction. By 15 years of age, a small
number of subjects supported occlusions at variance
with their facial characteristics due to what must have
been dentoalveolar and soft-tissue compensations.
Stepwise discriminant analysis showed the cranial
base angle at 5 years of age to be a good predictor of
the ultimate occlusal type at 15 years in 73% of subjects.
Global figures for mean cranial base angle of 129.4” (5
years), 128.8” (10 years), and 129.5” (15 years) with
a grand mean of 129.2” (mean age, 10 years) compare
favorably with the 129.6” of Ricketts” for a wide age
Volume 92 Craniofacial characteristics
Number 3
range of subjects with a mean age of 9 years. As pre-
viously shown,” on average this angle varies little be-
tween 5 and 15 years of age and can be considered one
of the few craniofacial constants during this period of
growth; however, GeorgeI has shown a mean decrease
in this angle of about 12” from birth to 5 years 9 months
with a mean value of 129.6”. The results of this study
showed a divergence from this mean when normal and
postnormal subjects were considered separately, with a
value of 127” being typical for the former group and
132” for the latter.
More detailed investigation of the ‘incorrectly’ clas-
sified subjects showed that, of the 15 normal occlusion
subjects classified as postnormal at 5 years of age, all
had increased cranial base angles; of the four subjects
‘incorrectly’ classified at 15 years of age, two had in-
creased cranial base angles; two who were classified as
postnormal for the first time at age 15 years had normal
angles but were included on the basis of other variables.
The eight postnormal occlusion subjects classified
as normal at age 5 years all had reduced cranial base
angles; of the three subjects classified as normal at age
15 years, two had reduced angles, and one who was
classified as normal for the first time on the basis of
other variables had an average postnormal value for this
angle.
This would suggest that, even if the cranial base
angle is the fundamental skeletal variable that deter-
mines the relationship of the maxilla to the mandible,
skeletal compensation can be effected to minimize its
importance, which from the orthodontic viewpoint can
be either favorable or unfavorable.
Although it is difficult to comprehend the subtleties
of skeletal modification that compensate for a funda-
mentally adverse cranial base angle yet produce normal
facial characteristics at age 15 years, analysis of the
normal occlusion subjects classified as postnormal at 5
years but who matured to normal facial characteristics
showed that in all cases a favorable change in angle
ANB occurred (that is, a reduction). On the other hand,
those postnormal occlusion subjects classified as normal
at 5 years and who matured to postnormal facial char-
acteristics all showed an increase in angle ANB.
Therefore, the cranial base angle, being relatively
constant, appeared to determine the fundamental jaw
relationship in these two occlusal groupings. In the
majority of subjects, this relationship remained un-
changed, but in approximately 26% of subjects, dif-
ferential growth of the maxilla and mandible could com-
pensate for the underlying relationship, leading either
to an improvement or a deterioration of facial harmony.
It would appear from the mean values for mandib-
of subjects with normal and postnormal occlusions 211
ular variables in this investigation that the mandible
was on average both distally positioned and smaller in
size in postnormal occlusions as compared with normal
occlusions. Mandibular size, however, was a poorer
discriminator than mandibular position between the
groups in the fully fledged malocclusion at age 15 years.
It is hoped that analysis of the study models of this
material will give further insight into the dynamics of
occlusion that accompany the craniofacial development
described here.
CONCLUSIONS
1, Stepwise discriminant analysis showed the best
discriminator at 5 years of age between subjects with
potential normal and postnormal occlusions to be the
cranial base angle, which was an accurate predictor in
approximately 73% of cases.
2. As a predictor of ultimate facial type, cranial
base angle was slightly less accurate-that is correct
in 67% of normal occlusion subjects and 74% of post-
normal occlusion subjects (mean 69%).
3. Mean values for cranial base angle for normal
and postnormal subjects were 127” and 132”, respec-
tively, these values being relatively constant between
5 and 15 years of age.
4. By 15 years only 8% of subjects possessed an
occlusion at variance with their facial characteristics.
5. The subjects whose classification changed be-
tween 5 and 15 years of age did so by differential
skeletal growth, commonly by a change in angle ANB,
which may be favorable or unfavorable, but which com-
pensates for a cranial base angle that is close to the
mean of the opposite occlusal grouping.
6. Seventeen percent of all subjects showed a
growth trend from postnormal to normal and 9% a trend
from normal to postnormal.
REFERENCES
1. Renfrce EW. A study of the facial patterns associated with
Class I, Class II, Division 1, and Class II, Division 2 maloc-
clusions. Angle Orthod 1948;18:12-15.
2. Houston WJB. Cephalometric analysis of Angle’s Class II, Di-
vision 2 malocclusion in the mixed dentition. Dent Pratt
1967;17:45-50.
3. Fischer-Brandies H, Fischer-Brandies E, Konig A. A cephalo-
metric comparison between Angle’s Class II, division 2 mal-
occlusion and normal occlusion in adults. Br J 0rthod 1985;
12:158-62.
4. Bjork A. Cranial base development. AM J ORTHOD 1955;41:198-
225.
5. Anderson D, Popovich F. Relation of cranial base flexure to
cranial form and mandibular position. Am J Phys Antbropol
1983;61:181-7.
6. Menezes DM. Comparison of craniofacial featums of English
212 Kerr and Hint Am. J. Orthod. Dentofac. Orthop.
September 1987

children with Angle Class II division 1 and Angle Class I oc- occlusion and with crowding of the teeth, Angle Class I mal-
clusions. J Dent 1974;2:250-4. occlusion, in respect of face size and shape and tooth size. Swed
I. Gilmore WA. Morphology of the adult mandible in Class II Dent J [Suppl] 1982;15:1 l-26.
Division 1 malocclusion and in excellent occlusion. Angle Or- 14. Lachenbmch P, Mickey RM. Estimation of error rates in dis-
thod 1950;20:137-46. criminant analysis. Technometrics 1968;lO: l-l 1.
8. Varjanne I, Koski K. Cranial base, sagittal jaw relationship and 15. Ricketts RM. A foundation for cephalometric communication.
occlusion. A radiological-craniometric appraisal. Proc Finn Dent AM J ORTHOD1960;46:330-57.
Sot 1982;78: 179-83. 16. George SL. A longitudinal and cross-sectional analysis of the
9. Kerr WJS. A longitudinal cephalomettic study of dentofacial growth of the postnatal cranial base angle. Am J Phys Anthropol
growth from 5-15 years. Br J Orthod 1979;6:115-21. 1978;49:171-8.
10. Kerr WJS. A method of superimposing lateral cephalometric Reprint requests IO:
films for the purpose of comparison: a preliminary report. Br J Dr. W. J. S. Kerr
Orthod 1978;5:51-3. University of Glasgow
11. BMDP Statistical Software.. Berkeley, California: University of Dental Hospital and School
California Press, 1983. 378 Sauchiehall St.
12. Kerr WJS, Ford I. A comparison of facial form in three western Glasgow G2 352, Scotland, United Kingdom
European male groups. Eur J Orthod 1986;8:106-11.
13. Adams CP. A comparison of 15 year old children with excellent