Hopp til innhold

Selected items added to basket

Go to basket

Systematic review

Demirjian’s development stages on wisdom teeth for estimation of chronological age: a systematic review

  • Year: 03.2017
  • By: Folkehelseinstituttet
  • Authors Rolseth V, Mosdøl A, Dahlberg PS, Ding KY, Bleka Ø, Skjerven-Martinsen M, Straumann GH, Delaveris GJM, Vist GE.
  • ISBN (digital): 978-82-8082-823-1
Aldersbestemmelse tenner web Forside.jpg

This systematic review summarizes the scientific documentation about how chronological age is distributed according to Demirjian’s stages from A to H for wisdom teeth.

Downloadable. In Norwegian. English summary.



Key message

Demirjian’s development stages of teeth are used to estimate the age of children and youth. This systematic review summarizes the scientific documentation about how chronological age is distributed according to Demirjian’s stages from A to H for wisdom teeth.

We found 18 relevant cross-sectional studies, all published after 2005. They were from 13 different countries from all continents except Australia. The studies presented the mean age of participants in each stage.

The majority of these studies were conducted in such a way that we do not have confidence in the results.  The number of participants in each age group and the total age range of the study influence the results. If studies do not consider this, a bias called age mimicry occurs.

We considered that only a few studies were performed in such a way that they provide an adequate description of the method’s ability to estimate age.

The studies do not provide sufficient evidence to conclude whether there are differences in the development of wisdom teeth between populations from different regions.



Every year, young asylum seekers who are unaware of their age or cannot document their age arrive in Norway. To ensure that children receive their rights and that adults are not treated as children, it is necessary to estimate a chronological age.  Evaluation of skeletal maturation of the hand and dental development has been used for age estimation of asylum seekers in Norway. These methods have been widely criticized for being imprecise, but currently no better methods have been put forward. In 2016, the Norwegian Institute of Public Health was assigned a national professional responsibility to evaluate and improve the methods (from 1.1. 2017 transferred with the Department of Forensic Sciences to Oslo University Hospital). The forensic expert group, in collaboration with the Knowledge Centre for the Health Services, conducted a systematic review of the evidence behind several of the medical age assessment methods.

The objective of this systematic review is to assess the scientific evidence on how mean chronological age is distributed in Demirjian’s stages on wisdom teeth (third molar), and, if possible, to elucidate any variations between different populations.

We chose to concentrate on Demirjian’s developmental stages in molars as this method has the strongest scientific documentation for estimating age in children and young adults. In parallel, we have also conducted a systematic review of age estimation based on maturation stages of hand from the Greulich and Pyle atlas


We searched for studies in the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase and Google Scholar. End search date was May 2016. The literature search covered studies that use X-rays of teeth or hand, and CT or MRI of the clavicle, knee and ankle for age assessment of children and young people between 10 to 25 years. Two persons independently read title and summary of 10059 references. Of these, we found 589 potentially relevant publications for age estimation based on dental X-ray and two people independently considered these in full texts. We included 18 studies with results for children and young people that covered the whole or parts of the specified age range with data as mean chronological age with variance within each of Demirjian’s stages on third molar. Two people independently considered the risk of bias in the studies based on the QUADAS-2 checklist and summarized this jointly.


We found 18 cross-sectional studies using Demirjian’s stages for molars to assess development of wisdom teeth, which presented results as mean age and standard deviation (variance) in tooth stages and provided information about the number of participants. The studies included from 209 to 3510 people with known chronological age. All studies included both boys and girls. Four studies were from China, two studies each from South Africa and Turkey, and one study each from Austria, Brazil, Britain, Canada, France, Iran, Malaysia, Malta, South Korea and Spain respectively. The age composition of the different studies varied considerably. The most narrow age range was from 15 years to 23 years (8 age cohorts) and the widest age range was from 4.1 years to 26.9 years (23 age cohorts). We considered that the majority of the studies had a high risk of a particular form of selection bias. Otherwise, the studies generally achieved low or unclear risk of bias based on our QUADAS-2 assessment.

Average chronological age for the different tooth stages varied considerably from study to study. We found that results from each study were highly influenced by the included age range and the number of individuals in each age group. This form of bias has previously been described as age mimicry. Age mimicry may strongly influence the observed mean age and standard deviation in each stage depending on the number of participants in each age group and the age span of the participants. Only a few of the 18 studies included an even number of individuals in the different age groups and an age range that was appropriate for all eight tooth stages.

We can highlight one study by Lee et al (2009) as an example of how the participants can be selected to describe the mean and standard deviation of age for tooth development stages of wisdom teeth in an adequate way. With this study, we demonstrate how one can use such data to describe the distribution of chronological age within each of the various stages for the purpose of age estimation. We also calculated prediction intervals (based on the mean age, standard deviation and the number of individuals for each stage), which gives a range of where the chronological age for an individual falls within, with a defined probability. We calculated such prediction intervals based on the results from Lee et al (2009) as an example. The calculations showed that the width of the 95 % prediction intervals for the various development stages ranged from 4.7 years to 6.8 years. For example, the mean age of individuals in stage E was 16.2 years and the estimated limits of the 95% prediction interval were 12.7 and 19.6 years.


We found that the majority of the 18 included studies examining the mean age of Demirjian’s stages on wisdom teeth, did include an uneven number of individuals in the different age cohorts. If you aim to describe how chronological age distributes within these stages, it is important to include an approximately equal number of people in each age group. In addition, the age range in such a study should be broad enough to cover all probable ages for the described stages. There is a particular problem with the stage H, as it is the final stage and also an important stage to describe whether a person is over or under 18 years. The upper ages of the enrolled participants in a study will strongly influence the mean age and the variance you observe for this stage. The higher the age included the higher the mean age and the standard deviation will be for stage H. There are several proposed strategies to solve this problem.      

Variation observed between studies using Demirjian’s stages on wisdom teeth has often been interpreted as differences in tooth development between populations and ethnicities. However, we consider age mimicry to be such a dominant source of variation in these studies that the scientific evidence is insufficient to conclude whether such differences exist.


The majority of studies we have summarized were conducted in such a way that we do not have confidence in the results. The number of participants in each age group and the total age range in the studies affect the results. If studies do not consider this, a bias called age mimicry occurs. Only one of the 18 studies prevented this phenomenon in their study design and could thus describe all eight Demirjian stages of wisdom teeth in an appropriate way. Based on this study, we calculated 95% prediction intervals for age in each development stage and found that these ranged from 4.7 to 6.8 years. Only when more studies with study designs that prevent bias due to age mimicry and with individuals from different regions, will it be possible to describe the age distribution in tooth stages in an appropriate way and conclude whether regional differences is of importance for tooth development.