| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
J Am Dent Assoc, Vol 139, No 10, 1318-1327.
© 2008 American Dental Association |  |
COVER STORY |
A Computer-Based Survey Study
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS, METHODS AND SUBJECTSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods. The authors created a survey by using a digital image editing software package, which enabled raters to manipulate intraoral photographs featured in the survey. They altered smile characteristics in photos of a sex-neutral face showing nasal tip to mentolabial fold. The authors administered an electronic survey in Boston (n = 78); Columbus, Ohio (n = 81); and Seattle (n = 84). An interactive interface allowed raters to select the ideal for each smile characteristic presented and identify the range of acceptability for the variables.
Results. Raters were reliable (
= 0.34–0.88). Survey location was not significant except that raters from the West accepted a broader smile than did those from the Midwest and the East. Raters identified ideals and thresholds for the following smile characteristics: smile arc; buccal corridor; gingival display; canine and posterior crown torque, ideal and large corridor; maxillary midline to face; maxillary to mandibular midline; overbite; maxillary central incisor gingival height discrepancy; maxillary lateral incisor gingival height discrepancy; maxillary central to lateral incisal step; maxillary central incisor crown width-to-height ratio; maxillary central-to-lateral incisal ratio; and occlusal cant. Generally, the values for ideal paralleled existing data, and new guidelines for some variables emerged. The ranges of acceptability were large.
Conclusions. The ideal and an acceptable range for each smile characteristic can be identified reliably.
Practice implications. Laypeople can reliably identify ideal smile characteristics. The ranges of acceptable deviations for smile characteristics are large, and practitioners should avoid unnecessarily sensitizing patients to minor discrepancies.
Key Words: Smile esthetics; smile arc; buccal corridor; attractiveness; tooth proportion; gingival display; orthodontics; dental esthetics
The contemporary practice of dentistry must include management of the dentition and soft tissues to create an esthetic smile by means of tooth movement, soft-tissue modification, restorative procedures or some combination of these techniques. The development of an esthetic smile should be based on as much objective evidence as possible. Until recently, however, the refereed literature had not clarified the importance of smile characteristics.1–9 Investigators conducting studies in restorative dentistry and orthodontics historically have identified esthetic issues1,4,6–37 and demonstrated that dental professionals and laypeople can identify smile characteristics that both enhance and detract from a smile.
Researchers in some studies have applied computer-based methods to alter dental morphology,2,18,26,27,31 and this computerized alteration appears to be an effective method of exploring esthetics owing to its consistency of variable manipulation and controlled presentation. Kokich and colleagues,2 to our knowledge, were the first to use computer-based image modification in attempting to quantify smile characteristic acceptability by using images of female lips and teeth. They found that orthodontists, general dentists and laypeople were able to detect discrepancies in smile characteristics at differing levels and that, for many variables, laypeople were less discriminating than were practitioners.
However, excessive cropping, skin tone differences, lipstick application and tooth shape can affect the perception of smile characteristics. Other authors6,9 have advocated the inclusion of more facial features, such as a full-face view, to demonstrate the interaction between all smile-related tissues. Furthermore, presentation of large incremental differences between images (for example, differences of 1 millimeter or more) may have obscured the true threshold of acceptability. It may be possible to refine, confirm and expand the results of previous studies using more advanced digital imaging methods and survey techniques.
A number of variables have been suggested to influence the attractiveness of a smile. Rosenstiel and colleagues26,27 suggested that the ideal width-to-height ratio for the maxillary anterior teeth is between 70 and 80 percent. The ideal smile arc should mimic the curvature of the lower lip from central incisor to canine.1,30 Buccal corridor minimization is a critical smile feature.1,6,8,12,25 Increased torque in posterior teeth is a way to improve the esthetics of narrow smiles.37,38 Excessive gingival display does not appear to be well-tolerated by raters.2 The highly variable permanent maxillary lateral incisor can be a challenging tooth to manage orthodontically and restoratively.39 Maxillary midline deviations can upset the balance of an otherwise esthetic smile.40–42 A maxillary-to-mandibular midline discrepancy alters anterior esthetics and indicates how posterior teeth will occlude.43 Overbite generally is characterized as ideal when its value is between zero and 2 mm.43 The vertical position of the lateral incisor affects the continuity of the smile arc.44 Kokich and colleagues2 found an occlusal plane cant to be an overwhelmingly displeasing smile characteristic to health professionals and laypeople. The location, shape and contour of the gingiva in the maxillary anterior region affect smile esthetics.29 Clearly, there are numerous characteristics that make up a smile, but these must be disaggregated and systematically evaluated to determine their effect.
Bearing these considerations in mind, we undertook a study to identify definitively the ideal and acceptable range of several smile characteristics through presentation of a standardized sex-ambiguous circumoral view including the lower face (a context that provided facial cues to symmetry) using raters from three regions of the United States. Most importantly, the raters were able to manipulate smile characteristics on a visually continuous scale so they could appreciate the realm of possibilities before they chose what they found most appealing or at the threshold of acceptability.
Model selection and image manipulation.
We examined the digital archive of the The Ohio State University College of Dentistry’s Division of Orthodontics for initial images following the protocol described by Parekh and colleagues.1 We used Adobe Photoshop CS2 (Adobe Systems, San Jose, Calif.) to crop the facial images to show only the lips, nasal tip and mentolabial fold to reduce distractions. To create a "hollow" lip set, we erased the teeth and periodontium from the lower face. We used a single intraoral frontal photograph of an ideally treated dentition for all smile-characteristic manipulation and inserted it inside the lip set after alteration. While conducting the survey, we projected these facial images on the computer screen at a size comparable to that of a face at typical conversational distance and standardized the resolution at 1,024 x 768 pixels.
Survey.
We designed the survey to encompass a range of values for many smile characteristics. To make the raters’ time requirement reasonable, we constructed two surveys; the same raters did not answer both surveys.
We used FormArtist Professional (Quask, Campbell, Calif.), a survey administration software allowing questions to be linked to images in both surveys. We used the images in this study to display continuously modifiable smiles across a predefined physiological range (Figure A, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]). The changes made during the initial image manipulation in Photoshop produced a visually continuous scale of possible choices when the slider bar coupled with the image was manipulated (Figure B, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]).
The study was provided to 300 participants older than 18 years who did not work in the dental profession. We used voluntary-response questions to gather demographic data, including sex, ethnicity and sociodemographic status. Any previous professional dental affiliation disqualified respondents from participation. Raters evaluated 28 (survey 1) or 26 (survey 2) image-based questions. Each question was presented randomly with one of two statements:
Raters completed the surveys in Boston; Columbus, Ohio; and Seattle. They used identically configured laptop computers, and all responses were anonymous.
Survey 1 content.
Below, we describe the image manipulation for all Survey 1 variables. Table 1
MATERIALS, METHODS AND SUBJECTS
TOP
ABSTRACT
MATERIALS, METHODS AND SUBJECTS
RESULTS
DISCUSSION
CONCLUSIONS
REFERENCES
This study and its procedures were approved by the institutional review board of The Ohio State University, Columbus.
presents a summary of the range of possible values and how each variable was measured.
|
Survey 2 content.
For all Survey 2 variables, Table 1
presents a summary of the image manipulation, as well as the range of possible values and how each variable was measured.
Statistical analysis.
We assessed raters’ reliability for each variable by means of the weighted statistic and using statistical software (SAS, Version 3.1, SAS Institute, Cary, N.C.). We analyzed differences in attractiveness ratings by means of descriptive statistics, including median and 95 percent confidence intervals. To investigate for regional differences, we conducted multiple Mann-Whitney-Wilcoxon tests with a Bonferroni-Holm correction. We used the signed rank test to evaluate the possibility that raters prefer different amounts of crown torque in canines and posterior teeth when the size of the buccal corridor varies. We set the level of significance at 
= .05 for all analyses.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS, METHODS AND SUBJECTSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Reliability.
Table 2 presents weighted statistics.
|
Effect of buccal corridor size on torque preference. There was no difference in raters’ perception and preference of buccal crown torque in canines only or in posterior teeth related to the size of the buccal corrditors (Table A, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]).
Defining ideal and acceptable smile characteristics.
The numeric value and the associated image for each variable selected by the raters as ideal and the threshold of acceptability are reported for each smile characteristic by using medians. These summary statistics and images for the ideal smile characteristics are displayed in Figures 1 through 10 and in Figures C, D, E, F, G and H (available as supplemental data to the online version of this article [found at "http://jada.ada.org"]).
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS, METHODS AND SUBJECTSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In this study, we surveyed laypeople, because they are the primary consumers of orthodontic services, instead of practitioners, who are providers of care. The possibility of regional differences among the East, West and Midwest had not been considered in previous research, to our knowledge and, ultimately, we found that lay raters in different locations were not significantly different in their assessment of individual smile characteristics. The penetration of mass media and popular culture may be responsible for this finding. The one difference, that raters on the West Coast accepted a broader smile, probably is not clinically significant, given the magnitude of the difference.
It still is unclear how laypeople evaluate smile esthetics. There are many potential distracters and interactions among different smile characteristics. Numerous facial or dental characteristics can be distracters that supersede interpretation of smile characteristic deviation.15 Standardization of image presentation in a realistic context should allow for true assessment of lay raters’ preferences across numerous smile characteristics.
The observation that orthodontic treatment flattens the smile arc13,45 is less troubling given the findings of this study. Lay raters preferred a consonant smile but accepted a smile with minimal curvature as well (Figure 1). Parekh and colleagues1 found flat smile arcs to be extremely objectionable, but it appears that there are increments flatter than ideal that are acceptable. We found that the addition of more upward curvature beyond what follows the lower lip did not rate well. Clearly, there is a wide and clinically significant difference between the upper and lower thresholds, but, generally, smiles should follow the curvature of the lower lip.
To enable the smile to follow the curvature of the lower lip, clinicians customarily increase the overbite. Our findings (Figure 2) suggest that lay raters are tolerant of a deeper bite, which enables the creation of a congruent smile arc. In addition, it is common orthodontic practice to incorporate a modest step between maxillary central and lateral incisors despite the absence of evidence that this is a desirable technique. Our findings (Figure 3) support a lateral step up beyond the recommended value (0.5 mm), which allows establishment of a smile arc without excursive interferences. These three aspects of the smile—smile arc, overbite and lateral step—all can work harmoniously and contribute to a more esthetic smile. However, many raters preferred even incisal edges, emphasizing that individual preference should be assessed during finishing.
Previous reports of ideal buccal corridor size vary from 2 percent6 to 19 percent.46 Our ideal buccal corridor size (Figure 4) was 16 percent, and our acceptability range was 8 to 22 percent. It appears that raters prefer the appearance of a buccal corridor approaching the 19 percent preferred by untreated patients in the study by Ritter and colleagues.46 Facial perspective or inclusion of more than the circumoral area, however, may make a difference.
Buccal corridor ratings indicate that visibility of the buccal segments may have an esthetic effect. Crown torque in the posterior segments may be visible, but claims that increased torque improves the esthetics of a narrow smile37,38 have not been substantiated. In our study, laypeople tolerated nearly every image presented to them and unexpectedly preferred negative torque in narrow smiles (Figures C, D, E and F, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]). Therefore, torque of canine and posterior teeth probably should be controlled primarily for purposes of functional occlusion.
Gingival display has been investigated extensively.2,13,17,18,28,46–51 Kokich and colleagues2 first reported that 4.0 mm of gingival display represented the threshold of acceptability but more recently18 used smaller increments and found it to be 3.0 mm. Our data (Figure 5) indicate the ideal value for gingival display to be 2.1 mm incisor coverage and the acceptable range essentially ± 4 mm. Vertical lip changes occur with aging52–56 and may make maintenance of gingival display difficult, but finishing orthodontic treatment within the acceptability range should be possible.
The location, shape and contour of the maxillary anterior gingiva are important smile characteristics,29 especially in the context of a smile with some amount of gingival display. Crown length discrepancies are most common when one maxillary central incisor is shorter than the contralateral incisor16 because of uneven wear of one central incisor combined with active incisor eruption. In recent years, investigators have found that laypeople did not detect asymmetric crown length unless one crown was 1.5 to 2.0 mm shorter than the other.18,57 Our study results corroborated that 2.0 mm is the limit of acceptability for this variable (Figure 6); however, one-third of the raters did not find the discrepancy unacceptable until the heights had a difference of 4.0 mm.
Kokich and colleagues2 noted that gingival discrepancies between the maxillary central and lateral incisors were not obvious to laypeople. Our study results confirmed the broad range of acceptability for this variable and demonstrated it was acceptable even when the lateral gingival margin was superior to the central gingival margin (Figure 7).
Kokich and colleagues2 also found that laypeople needed a 2.0-mm deviation of the ideal crown length to classify the central incisors as noticeably less esthetic. In their research, Kokich and colleagues defined the ideal central incisor crown width-to-height ratio at approximately 0.77, and the 2.00-mm deviation resulted in a width-to-height ratio of approximately 0.90. In a study in which they reviewed anatomic crown width-to-length ratios, Magne and colleagues58 found that unworn central incisors had a ratio of 0.78. Results of other studies have shown similar values.4,27,59 Our findings confirmed these values (Figure G, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]).
Proportional size of the maxillary lateral incisor is an interesting variable because of the tooth’s variability in size. Its most frequent anomaly is a peg shape, in which the tooth’s width is grossly decreased in comparison with the height. Kokich and colleagues2 found that the threshold for acceptability was 4.0 mm narrower than the ideal width of the lateral incisor. At the ideal, the lateral incisor’s width was 78 percent of the central incisor’s width, whereas at the threshold it was 45 percent. Our findings were similar: 72 percent for the ideal and a threshold value of 53 percent (Figure H, available as supplemental data to the online version of this article [found at "http://jada.ada.org"]). We found that lateral incisors can be as wide as 76 percent of the central incisor before becoming unacceptable.
The maxillary dental midline often is compared with the facial midline using the center of the philtrum22,23,28,60 and soft-tissue nasion.23,28 Some authors11,14,61,62 have demonstrated that maxillary midline discrepancies of more than 2.0 mm were likely to be noticed by laypeople, whereas others2,57 found that laypeople could not perceive a 4.0-mm deviation. Our findings established the maximum acceptable value to be 2.9 mm (Figure 8), although one-third of our respondents accepted a deviation of 4.3 mm.
Maxillary and mandibular midlines are noncoincident in three-fourths of the population,22 and small deviations do not cause any detriment to smile esthetics.43 The contribution of the mandibular midline to esthetics may be diminished owing to the narrow width and uniform size of mandibular incisors.63 We found that mandibular midline deviation was acceptable until it exceeded 2.1 mm (Figure 9) and one-third of the respondents accepted the maximal deviation of 2.9 mm. This demonstrates that many respondents found this deviation to be acceptable when more than one-half of the mandibular incisor deviated from the maxillary midline. This finding adequately accommodates patients who have a missing or extracted lower incisor.
Asymmetry, even among esthetically pleasing faces, is a typical finding.64 An occlusal cant is a form of asymmetry that is apparent when a person smiles but is not perceived on intraoral images or study casts.28 Kokich and colleagues2 found that laypeople did not detect this type of asymmetry until it reached 3.0 mm (equivalent to 4 degrees). Results of other studies have showed that deviations in cant are not noticeable unless they exceed 2 degrees,65 3 degrees64 or 4 degrees.24 We found that our lay raters accepted cants of as much as 4 degrees (Figure 10), but one-third of the respondents accepted cants at the maximum deviation of 6 degrees.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMATERIALS, METHODS AND SUBJECTSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Our most important finding probably is the range of acceptability. Lay raters tolerated a wide range of variability for most characteristics, and clinicians’ knowledge of what is ideal does not make it appropriate for them to ignore this range. It probably is sound for the clinician to use care in identifying the ideal for patients when that knowledge could sensitize them to unrealistic or unattainable goals. In their naiveté, they are more accepting than practitioners might expect. Remembering that our values as clinicians should not be imposed except near the margins of accept-ability probably is the best course.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS, METHODS AND SUBJECTSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. Hickman, A. R. Firestone, F. M. Beck, and S. Speer Eye fixations when viewing faces J Am Dent Assoc, January 1, 2010; 141(1): 40 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Relevant research from non-orthodontic journals J. Orthod., June 1, 2009; 36(2): 130 - 133. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
