The age distribution of facial metrics in two large Korean populations

Growth and alterations in craniofacial morphology have attracted interest in many fields of science, especially physical anthropology, genetics and forensic sciences. We performed an analysis of craniofacial morphology alterations by gender and ageing stage in Korean populations. We studied 15 facial metrics using two large Korean populations (1,926 samples from the Korea Medicine Data Center cohort and 5,643 samples from the Ansan-Ansung cohort). Among the 15 metrics, 12 showed gender differences and tended to change with age. In both of the independent populations, brow ridge height, upper lip height, nasal tip height, and profile nasal length tended to increase with age, whereas outer canthal width, right palpebral fissure height, left palpebral fissure height, right upper lip thickness, left upper lip thickness, nasal tip protrusion, facial base width, and lower facial width tended to decrease. In conclusion, our findings suggest that ageing (past 40 years of age) might affect eye size, nose length, upper lip thickness, and facial width, possibly due to loss of elasticity in the face. Therefore, these facial metric changes could be applied to individual age prediction and aesthetic facial care.

ethnic comparison of the human orbital region. Facial metrics in the orbital region, including PFH, IW, and OW, were compared between Koreans and other ethnicities (Fig. 2). RPFH and LPFH were relatively low among non-Asians relative to Asians ( Fig. 2A). The Korean subjects had a large OW, ranking between those of Japanese and Indian individuals, whereas the IW was of Koreans was smaller than those of other Asian populations, such as Japanese and Chinese (Fig. 2B,C). Interestingly, the gender difference in IW was no more prominent in Koreans than in other ethnic groups.

Discussion
In this study, we analysed 15 facial metrics for gender differences and for ageing-related changes in two large Korean cohorts. Among these 15 metrics, 12 showed gender differences and tendencies toward age-related change. Increasing tendencies were observed in BRH, ULH, NTH, and PNL, and decreasing tendencies were observed in OW, RPFH, LPFH, RULT, LULT, NTP, FBW, and LFW. The decreasing tendencies were not changed even when adjusted for body mass index, indicating that some other structures under the skin (such as collagen and muscle) were reduced in the aged population.
In the orbital region, BRH showed a tendency to increase with age in both the KDC and Ansan-Ansung cohorts. The BRH is the nodule or crest of bone situated on the frontal bone of the skull 10 ; this ridge separates the forehead itself from the tops of the eye sockets. Normally, in humans, the ridges arch over each eye, offering mechanical protection 11 . Typically, the arches are more prominent in men than in women and vary between different ethnic groups 12 .
"RPFH" and "LPFH" are abbreviations for the right and left palpebral fissure height. The palpebral fissure is the elliptic space between the medial and lateral canthi of the two open eyelids. There are many studies of palpebral fissure height; this metric is approximately 10 mm in adults and is smaller in East Asian populations than in white populations 13,14 . The gender differences in palpebral fissure height have been well summarized in reports on a South Indian population 14 . The bilateral orbital region, which is part of the upper face, acts as a key determinant of the perception of facial attractiveness, youthfulness and health 14 . Our comparison of palpebral fissure height implies that a very small PFH is a typical Korean facial characteristic ( Fig. 2A). IW did not show any alteration with age, but OW showed a prominent decreasing pattern. The ethnic comparison of Fig. 2B,C showed that our study group had a shorter IW than Japanese or Chinese individuals, but the gender difference was no more prominent in Koreans than in other ethnic groups. In contrast, the OW of our study population was relatively large and was similar to that of the Japanese population. Therefore, the characteristics of the orbital region of Koreans are a small palpebral height and a wide canthal region. In the nasal area, the NTH slightly increased with age, and the NTP slightly decreased with age, but the NBH did not show any tendency to change. Additionally, the profile nasal length increased slightly with ageing only in males. The nasal tip is a challenging part of the nose for plastic surgeons because of variations in the anatomy of the lower lateral cartilage 15 . Therefore, our results help to clarify the direction of surgical correction for nasal tip deformities.
In the orthodontic region, the ULH slightly increased with ageing, and the RULT and LULT showed slight decreases with ageing. The upper lip is an important attractive point for the smile. According to Hulsey, the "smile is one of the most effective means by which people convey their emotions" 16 . The ULH is affected by the gingival margin of the upper central incisors and influences the attractiveness of a smile 16 . Generally, a shorter ULH indicates a higher smile line. On average, the smile line was found to be 1.5 mm higher in women than in men 17 ; similarly, in the current study, the ULH was greater in males than in women, indicating a shorter smile line in women (see Table 1g). Although the concept of beauty has changed throughout the centuries, the thickness of the upper lip has always been a subject of interest and has importance in every culture 18 . These results may be beneficial to forensic anthropologists, plastic and reconstructive surgeons, and orthodontists.
Regarding overall facial width, both FBW and LFW slightly decreased with ageing, and the tendency was more prominent in male than female subjects. Facial width is reportedly associated with testosterone levels in males 19 . Therefore, the decrease in testosterone in elderly males may decrease the facial width.

conclusion
Our study found that 12 of 15 facial metrics changed with age in subjects of both genders. The changes in eye size, nose length, upper lip thickness, and facial width might be general trends in Koreans because of the consistent tendencies in two large independent cohorts. Changes in facial metrics could serve as indicators to predict age from photographs and would be helpful for aesthetic facial care. To our knowledge, there are very few studies on the alteration of facial metrics with ageing, particularly in groups over 40 years old. Moreover, our study did not analyse the facial ageing of other ethnic populations. Therefore, we hope that the currents results will be validated in other ethnic groups and applied to diverse fields.  Table 2). The subjects were photographed with a neutral expression in both frontal and profile views under the following standard conditions: the hair should be pulled back with a hair band; the centre points of the two pupils should be horizontally aligned, as should the upper auricular perimeters; and a ruler should be  www.nature.com/scientificreports www.nature.com/scientificreports/ placed approximately 10 mm below the chin to convert pixels into millimetres. All participants provided written informed consent to participate in the study. This protocol was approved by the Korea Institute of Oriental Medicine Institutional Review Board (I-0910/02-001), and all research was performed in accordance with the relevant guidelines/regulations. craniofacial measurements. Detailed descriptions of candidate feature variables and the corresponding measurement methods have been provided in a previous report 6,9 . Briefly, the facial variables were limited to those that could be easily quantified. Facial feature points in frontal and lateral images were automatically extracted by detecting and analysing the face, eyes, nose, mouth, and contours via an in-house program in Visual Studio C++ using OpenCV (Open Source Computer Vision Library). The positions of the extracted points were confirmed by a well-trained operator (accuracy: 98.8% on average). Fifteen facial metrics, each defined by the distance between two facial points, were derived from the photographs by converting pixels into millimetres using MATLAB software 21 .

Study participants.
For statistical analysis, 5 severely skewed facial variables, including PNL, NTH(V), NTP(H), RULT, and LULT, were ln transformed. We removed the outliers, which were defined using the first and third quartiles and the interquartile range of each facial variable. In each facial variable, measurements below the first quartile -2.0 × interquartile range or above the third quartile + 2.0 × interquartile range were defined as outliers and excluded.
Statistical analysis. For each facial variable, the mean length was compared between age groups (40 s vs 50 s, 60 s or 70+) using Student's t-test. Additionally, the mean lengths of male and female facial variables were plotted ( Fig. 1) for both the KDC and Ansan-Ansung cohorts. We used the following criteria to identify phenotypes that tended to change with age: the mean underwent a significant gradual increase or decrease (p < 0.05), and the pattern appeared to be similar between the KDC and Ansan-Ansung cohorts.