The American Association of Orthodontists Foundation (AAOF) Craniofacial Growth Legacy Collection is a digital repository of records from 9 craniofacial growth study collections in the United States and Canada. The purposes of this article were to describe the use of materials from the AAOF Craniofacial Growth Legacy Collection in the orthodontic literature in comparative and follow-up studies, and to analyze trends before and after the project’s launch in 2009.
An electronic search without date or language restriction was conducted in the following databases: PubMed, Embase, Evidence-Based Medicine Reviews, and CINAHL. Grey literature resources and the bibliographies of the selected studies were also consulted. Three independent reviewers assessed the studies for inclusion. The criteria were human subjects of any age, sex, and ethnicity; at least 1 of the 9 AAOF legacy collections used as either the main sample population or the comparison or control; and orthodontic outcomes assessed. Data were analyzed using STATA software (version 14.2; StataCorp, College Station, Tex).
A total of 199 studies (127 follow-up, 72 comparative) were included. The most commonly used collection in comparative studies was the Michigan Growth sample. The number of published studies more than doubled after the AAOF Legacy Collection project testing and launch in 2009. The increase continued through 2010 to 2014, during which there was a trend to use multiple collections. The Burlington Growth collection was the most commonly used collection for follow-up studies. The overall use of the legacy collection showed a small increase in published studies after 2009.
The overall numbers of published studies in the comparative and follow-up categories increased after 2009, reflecting the efforts of the AAOF team and collection curators to make the records available worldwide. Further research should consider studying each collection to identify utilization predictors.
The numbers of studies using collection material increased after the AAOF Legacy Project launch (2009).
The Michigan sample was the most-used sample for comparative studies.
The Burlington sample was the most-used sample for follow-up studies.
Further research should consider studying each collection to identify utilization predictors.
Between 1930 and 1985, researchers in the United States and Canada collected longitudinal growth data on children with malocclusions. The data included radiographic and nonradiographic records that were amassed in individual, independent collections. Preservation of these collections is absolutely critical because these records were accumulated at considerable human and economic costs. Because of the radiation exposure risk, these studies are unlikely to be repeated in view of ethical considerations. Therefore, these records are likely to remain invaluable for decades to come.
Due to concerns about loss of the physical records, the American Association of Orthodontists Foundation (AAOF) supported the development of a centralized database to preserve the records in a digital format, allowing them to be accessible from a single Web site. The AAOF Craniofacial Growth Legacy Collection is the result. It currently holds 9 of the 11 known collections of longitudinal craniofacial growth study records. The 9 collections, comprising a total sample size of 762 subjects with different malocclusions, are the Bolton-Brush Growth, Burlington Growth, Denver Growth, Fels Longitudinal, Forsyth Twins, Iowa Facial Growth, Mathews Growth, Michigan Growth, and Oregon Growth studies.
Each collection had its own sampling technique, study follow-up duration, data collection method, and types of records. Collectively, however, a wealth of information can be gathered from the data, including physical growth, craniofacial growth and development, skeletal maturation, and aging of the craniofacial complex. This rich array of knowledge is invaluable. Records collected annually, semiannually, or quarterly as part of the original growth studies include dental casts, cephalometrics, twin studies, family studies, implant records, hand-wrist films, facial photographs, demographic records, and other records of importance. Some studies followed children from infancy to adulthood, others only during adolescence. See Table I for a full description of each collection. The data provided by these records is the segue to treatment of growing and nongrowing patients in contemporary orthodontics.
|Legacy Collection||Study duration||Location||Type of study||Total n||M/F Ratio||Age range (y)||Ethnicity ∗||Types of records †||Frequency of records ‡||RG images (n)||LC enlargement factor (%)||Total on AAOF Web site (n)||Type of malocclusion on AAOF Web site|
|Bolton-Brush||Bolton: 1929-1959 §
|Case Western Reserve University||Longitudinal||4309||2021/2288||<3-18+||90.7% white, 9.2% black, 0.1% other||0, 1, 2, 3, 4, 5, 6, 7, 9, 11||<1 y: every 3 mo; 1-5 y: every 6 mo; > 5 y: annual||250,000+||Bolton: 12.9
|102||I: 25; II: 55; III: 22|
|Burlington||1952-1971 §||University of Toronto||Longitudinal||1698||‖||3-21||Northern European white||0, 1, 2, 3, 4, 5, 6, 7, 9, 11||Annual||48,365||9.84||100||I: 38; II:56; III: 6|
|Denver||1927-1967||University of Oklahoma||Longitudinal||300||150/150||1-20+||European white||0, 1, 2, 3 4, 7, 11||Annual||3000+||4||94||I: 58; II:22; III: 1; Unknown: 13|
Skeletal/physical growth: 1929-ongoing
|Wright State University||Cross-sectional; Longitudinal||1379||‖||0.1-59+||98% European white; 2% black||0, 1, 7, 11||0-6 y: every 6 mo; >6 y: annual||9500+||9.09||102||I: 44; II: 3; III: 1; Unknown: 54|
|Forsyth Twin||1959-1975||Forsyth Institute||Longitudinal||533 families||‖||6-10; 10-16; 6-16+||European||0, 2, 3, 5, 6, 7, 8, 11||Annual||‖||6||10||I: 6; II; 4|
|Iowa||1946-1960 §||University of Iowa||Longitudinal||183||92/91||2-56||European||0, 1, 2, 5, 6, 11||< 5 y: every 3 mo; 5-12 y: every 6 mo; 12-18 y: annual||3546||8.7-9.4||100||I: 80; II:16; III: 2; Unknown: 2|
|Mathews||1967-1979||University of the Pacific||Longitudinal||36||14/22||7-18||European white||0, 2, 3, 10||Annual||∼1400||8.3||35||I: 19; II: 16|
|Michigan||1935-1970 §||University of Michigan||Longitudinal||720||‖||5-18||European and Asian||0, 1, 2, 3, 9||Annual||2400+||12.7||102||I: 80; II: 20; III: 1; Unknown: 1|
|Oregon||1952-1974 §||Oregon Health Sciences University||Mixed longitudinal||409||188/221||3-18||Northern European||0, 1, 2, 4, 5, 6, 7, 9, 11||Annual||‖||Varied||107||I: 67; II: 34; III: 4; Unknown: 2|
∗ Data sources: Burlington, Fels, Forsyth, Mathews, and Oregon, from Hunter et al ; Denver, from a personal communication via e-mail with curators; Iowa and Michigan, from the AAOF Web site ; Bolton-Brush, from Behrent and Broadbent.
† 0 , Lateral ceph; 1 , dental cast/model; 2 , frontal ceph; 3 , lateral oblique radiograph; 4 , photograph; 5 , standing height; 6 , weight; 7 , hand-wrist radiograph; 8 , twin study; 9 , family study; 10 , implant studies; 11 , other.
Use of the AAOF Legacy Collection and trends in the use of the individual collections are not clear. It has been claimed that the AAOF Legacy Collection project significantly improved the prospects of orthodontic research over the 8 years since initiation of the project.
The aims of this systematic review were to describe the use of the AAOF Legacy Collection in the orthodontic literature focusing on both comparative and follow-up studies and to analyze trends of use before and after the Legacy project launch. As a secondary objective, we hoped to highlight the importance of the AAOF Legacy Collection for future research opportunities.
Material and methods
Protocol and registration
This review followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) as closely as possible. The protocol for this systematic review was registered on PROSPERO: international prospective register of systematic reviews (PROSPERO 2016: CRD42016038395). Details of the protocol can be accessed at www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42016038395 .
Studies were eligible for inclusion based on the following criteria.
Primary studies that used an experimental or observational study design (randomized controlled trial, cohort/longitudinal study, case-control study, cross-sectional study).
At least 1 of the 9 collections in the AAOF Legacy Collection was used as either the main sample population or as a comparison/control group within the study.
The study included human subjects of any age, sex, and ethnicity.
The study measured orthodontic appliance outcomes, longitudinal craniofacial growth and development, malocclusions, or any other orthodontically related outcome.
Narrative reviews, case reports, and case series studies were excluded from review. Animal studies were also excluded, because the goal of this systematic review was to analyze trends in the use of the AAOF Legacy Collection in the orthodontic literature as it pertains to human subjects. Studies were also excluded if they did not measure at least 1 orthodontic outcome (eg, anthropological studies).
Information sources, search strategy, and study selection
An electronic search was conducted in March 2016 in the following databases: PubMed (1946-March 2016), Embase (1966-March 2016), CINAHL (1937-March 2016), and Evidence-Based Medicine Reviews (1991-March 2016). Evidence-Based Medicine Reviews is a database available through the Ovid platform and comprises 7 databases, which include the Cochrane Register of Controlled Trials, Cochrane Database of Systematic Reviews, Cochrane Methodology Register, ACP Journal Club, Database of Abstracts of Reviews of Effects, Health Technology Assessment, and NHS Economic Evaluation Database. The search was developed and performed by a health sciences librarian (E.S.). The search included 2 main components: the 9 components of the AAOF Legacy Collection, and orthodontic keyword and MeSH terms to locate relevant orthodontic studies. No language restrictions were used. The complete search strategy used in PubMed can be found in Appendix 1 . Additionally, a search for grey literature was conducted on ClinicalTrials.gov ( clinicaltrials.gov/ ), OpenGrey.eu ( opengrey.eu/ ), and Grey Literature Report ( greylit.org/ ) by searching each AAOF Legacy Collection name, AAOF Legacy, or American Association of Orthodontists Foundation. The ProQuest Dissertations and Theses Global database were also searched according to each collection name, AAOF Legacy, or American Association of Orthodontists combined with “orthodontics,” “craniofacial growth,” or “dental.” The bibliographies of the included studies were also used to identify additional studies for possible inclusion.
Studies were screened with the previously stated inclusion criteria at the title and abstract level by 2 independent investigators (T.A.J. and R.C.) to reduce bias. A high level of agreement was obtained at the 2 stages. Due to the magnitude of studies that needed to be reviewed at the full-text level, the results were divided among 3 investigators to be reviewed without overlap (T.A.J., R.C., and L.L.). A fourth independent external member reviewed all studies to confirm inclusion or exclusion. Any disagreement was resolved by discussion among all authors.
Data items and collection
Two customized data abstraction forms were used to extract data from each study. For comparative studies, the following variables were recorded: study authors, origin, study design, sample sizes in the treatment and control groups, age categories at baseline, malocclusion classifications for treatment and control groups, type of intervention, treatment or observation duration in months, skeletal maturational age, name of the particular craniofacial growth collection used, type of orthodontic records, and controls matching criteria. Similar variables were collected for the follow-up studies on the main follow-up sample. Because of the lack of control samples or interventions in the follow-up studies, this information was not collected.
Data were analyzed using STATA software (version 14.2; StataCorp, College Station, Tex). Descriptive statistics were conducted for all categorical and continuous variables in the comparative and follow-up studies separately. Chi-square tests were used to determine publication trends of each craniofacial legacy collection at the 2009 cutoff, which represented the testing and launch of the AAOF Legacy Collection project. A time series graph was used to denote overall trends of Legacy Collection utilization over time.
The search identified 3188 articles ( Fig 1 ). After exclusion by title and abstract, 304 studies underwent full-text reviews, and 105 articles were excluded for various reasons ( Supplementary Table ). A total of 199 studies met the inclusion and exclusion criteria of this study. Of these, 127 were follow-up studies using a legacy collection as the main sample, and 72 were comparative studies with a legacy collection used as a control group for contemporary treatment groups ( Appendices 2 and 3 ).
All 9 collections in the AAOF Legacy Collection were used in studies included in this systematic review.
Of the 72 comparative studies, 63 (87.5%) were published articles, and the remaining were theses or dissertations. The countries of origin where the studies were conducted were the United States (55; 76.4%), 1 international country (13; 18.1%), and more than 1 country (4; 5.6%). All studies were retrospective comparative studies except for 2 (2.7%) that were prospective controlled studies.
The total median sample size of the treatment groups was 43 subjects, compared with 32 in the control groups. Table II presents the sample distribution by sex in each group. The most common age category at baseline among the treatment and control groups was 10 to 14 years ( Table III ).
|Comparative studies (n = 72)||Follow-up studies (n = 127)|
|Treatment groups (n)||Control groups (n)||Study samples (n)|
|Total||Females||Males||Total ∗||Females||Males||Total ∗||Females||Males|
|Variable||Comparative studies||Follow-up studies|
|Treatment group||Control group||Study sample|
|n (%)||n (%)||n (%)|
|Age at baseline (y)|
|1-4||2 (2.8)||1 (1.5)||22 (22)|
|5-9||19 (27.1)||21 (31.8)||51 (51)|
|10-14||46 (65.7)||34 (51.5)||13 (13)|
|15-19||3 (4.3)||6 (9.1)||10 (10)|
|>19||0 (0)||4 (6.1)||4 (4)|
|Type of malocclusion|
|Normal occlusion||1 (1.7)||3 (5.1)||9 (12.5)|
|Class I||3 (5.2)||11 (18.6)||22 (30.6)|
|Class II||32 (55.2)||26 (44.1)||7 (9.7)|
|Class III||6 (10.3)||5 (8.5)||2 (2.8)|
|Other||16 (27.6)||14 (20.3)||32 (44.4)|
More than half of the treatment group samples had Class II malocclusion followed by other malocclusion conditions. The other category included studies on multiple malocclusion classifications or those with other conditions such as skeletal and dental anterior open bite, deepbite, large cranial base angle, transverse maxillary deficiency, and cleft lip and palate. Similarly, 44.1% of subjects in the control samples had Class II malocclusion followed by other types of malocclusion (14; 20.3%). Of the 14 studies, 35.5% assessed 2 or more malocclusion categories, and 64.5% included other forms such as anterior open bite, deepbite, small cranial base angle, transverse maxillary deficiency, and noncleft patients.
The types of interventions varied among the studies. They included appliances such as the mandibular anterior repositioning appliance, AdvanSync, Herbst, Bionator, Activator, Twin-block, Fränkel-2, Fränkel-3, maxillary expanders (bonded, Haas, modified Haas, quad helix, hyrax, acrylic splint), facemask, Xbow, Forsus, chincup, Schwartz, mandibular lingual holding arch, spring-loaded module, removable maxillary plate with anterior bite plane, headgear with anterior bite plane, Tweed and standard edgewise appliances, preangulated and torqued “Siamese” edgewise appliances, passive self-ligating system (Damon 3MX), intermaxillary Class II, or intra-arch elastics on round archwires. They also included treatment procedures such as premolar extraction, LeFort I osteotomy, midfacial advancement with distraction osteogenesis, and 1-stage closure of complete unilateral cleft lip and palate. Additionally, the study groups included patients with craniofacial anomalies or syndromes such as complete unilateral cleft lip and palate, Down syndrome, Pierre Robin sequence, and craniosynostosis. The median total treatment or follow-up duration among the treatment groups was 28 months (interquartile range, 21-42; minimum, 0.7; maximum, 132), whereas the median control total observation duration was 36 months (interquartile range, 15-66; minimum, 4; maximum, 132).
Figure 2 presents the percentages of utilization of the Legacy Collection in the comparative and follow-up studies. Of the 9 collections, the most commonly used growth study in the comparative studies was the Michigan growth sample (31.9%), followed by multiple collections (22.2%), and then the Burlington collection (18.1%). The Denver growth study was used in combination with other databases in 13 studies, and the most common combination was the Michigan and Denver samples (9 studies). No included comparative study used the Fels, Forsyth Twin, Oregon, or Mathews collections.