Pararadicular cementum deposition as a criterion for age estimation in human beings

Thomas John Stein, MEd, DMD, MSa and John F. Corcoran, DDS, MSh Ann Arbor, Mich. VETERANS ADMINISTRATION MEDICAL CENTER AND UNIVERSITY OF MICHIGAN

The purpose of this article is to ascertain if pararadicular cementum can be used as a reliable criterion for age

estimation in human beings. Fifty-two nonrestorable teeth were extracted from 42 patients at the Veterans Administration Medical Center at Ann Arbor, Mich. The specimens were prepared to a thickness of 500 pm with a Buehler lsomet bone saw (Bronweil Scientific, Inc., Rochester, N.Y.). Longitudinal sections were cleaned of artifacts in an ultrasonic cleaner and stained with 1% alizarin red. Photomicrographs were taken of each prepared section. The cementum was composed of multiple light and dark bands that were counted on the photograph and added to the average eruption time of the individual tooth. There was an overall Pearson’s product-moment correlation coefficient of r = 0.93 between the patient’s predicted age with the use

of cementum annulations as compared with the actual chronologic age of the person. Predicted age counts showed greater divergence from actual age in persons older than 55 years. A formula is presented to adjust for this discrepancy. The data indicate that quantitation of cementum annuli is a moderately reliable means for age estimation in humans. (ORAL SLJRC ORAL MED

ORAL PATHOL 1994;77z266-70)

The hard tissues of the human dentition are able to resist decay and degradation long after other tissues are lost. This resistance has made teeth useful indica- tors for assessing variation in diet, expressions of metabolic diseases, and calculation of age at the time of death. The enamel, dentin, and cementum that comprise teeth have been used to estimate the chro- nologic age of unknown persons. Rough estimation of age has been proposed from the color of the root sur- faces of teeth.’ Gustafson2 found many masculatory aspects of the dentin and enamel useful for gross age estimation in human beings including faceting, mul- tiple wear patterns, and generalized attrition.

Because of its position, cementum has not been used to the extent of enamel and dentin. However, the counting of cemental annulations may offer a more accurate method for age estimation in human beings. Zander and Hurzeler3 state that cementum is poten- tially a better age-estimating tissue because it is pro- tected from the ravagement of dentin and enamel. This is due to its unique location in the alveolar pro- cess. Cementum exhibits progressive growth through- out the life of the human dentition. Recent studies43 5 found that cementum growth is approximately linear

The views expressed in this article are strictly those of the authors and do not necessarily reflect the policy of the Veterans Adminis- tration. aEndodontist, Ann Arbor Veterans Administration. bChairman, Department of Endodontics, School of Dentistry, Uni- versity of Michigan. Copyright @ 1994 by Mosby-Year Book, Inc. QO30-4220/94/$3.00+ 0 7/14/52562

266

and that cementum thickness is proportional to the age of the tooth. Linhart,6 Fog1 and Mosby,7 and De- Ricqules8 suggested that this linear growth may be the result of regular apposition of cementum in bands similar to those bands exhibited by lower animal spe- cies.

Attempts have been made93 lo to estimate the chro- nological age of preserved animal species by counting cementum annulation. Annulation counts have been used as a game-managing tool for many modern spe- cies of mammals, including bear,” caribou,’ 2 moose, 1 3 elk,t4 deer,t5 bisonI red fox,17 and two species of primates.18 The annulations appear as alternating light and dark bands throughout the cementum layer. It has been inferred from animal studies19 that intense seasonal variations, such as hibernation, wet and dry seasons,20 or prolonged fasting2’ are the cause of this banding. However, banding appears in human beings and counting annulations appears to approximate the chronologic age of human beings. Applied to hu- mans, counting annulations would enable a forensic pathologist to use a single tooth to estimate the age of an unknown person that died many years or centuries before from a single tooth. Gustafson’s2 estimations have been used in forensic medicine for many years, but their accuracy is questionable. The purpose of this article is to ascertain if counting cementinal annula- tions is a reliable method for age estimation in humans.

MATERIAL AND METHODS Fifty-two nonrestorable teeth from 42 patients un-

dergoing dental treatment at the Veterans Adminis-

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Stein and Corcoran 267 Volume 77. Number 3

Fig. 1. Longitudinal section of specimen 37 of lower right bicuspid of 64-year-old man. Den = dentin; Cem = cementum; Pm = periodontal membrane. Each line represents annulation for a total of 51 annula- tions that are added to the eruption factor of 9 + 1 for a total count of 60 + 1.

tration Medical Center in Ann Arbor, Mich. were se- lected for the study. Consent and history forms were obtained for each patient. The teeth selected for ex- traction were either lower bicuspids or central inci- sors.

After extraction, the teeth were placed in a solution of 10% formalin. The specimens were dehydrated in a solution of 95% ethyl alcohol, which was changed six times in 48 hours with equal parts acetone and mod- ified Spurr’s low-viscosity embedding medium (Beu- hler, Ltd., Evanston, Ill.). The tissue was embedded in a fresh change of 100% modified Spurr’s medium in plastic molds and polymerized for 24 to 48 hours at 60” C. After curing the plastic, longitudinal sec- tions were cut on a Buehler Isomet 1 l- 1180 low-speed saw (Bronwell Scientific, Inc., Rochester, N.Y.) with a precision diamond wafering blade (No. H-4242, Buehler, Ltd., Evanston, Ill.). The sections were washed in an ultrasonic bath with distilled water and stained with 1% alizarin red stain for 2 minutes. Dis- tilled water was used to remove excess stain. The stained sections were rinsed with 95% alcohol and covered with a slip mounted with Permount (Fisher Scientific, Fair Lawn, N.H.). Sections were photo- graphed with the use of a green filter and Kodak Technical Panfilm (Eastman Kodak Co., Rochester, N.Y.) with an Axiophot Zeiss Microscope (Carl Zeiss, Inc., Thornwood, N.Y.).

Annulation counts were made on a x10 photomi- crograph (Fig. 1). An annulation is composed of both a light band and a dark band. A band is represented by either a light or dark region. The annulation count

was added to the average age of eruption in years for each tooth-as presented in Gray’s Anatomy.22 This number was then compared with the known ages of the person. All statistics were done with a Biological Science Statistical Package (Eastern Michigan Uni- versity, Ypsilanti, Mich.).

RESULTS Calculated and known ages of persons are shown in

Table I. Fig. 2 represents the scattergram of uncorre- lated data for 52 specimens for ages from 27 to 84 years with a mean age of 61 years. The Pearson’s product-moment correlation coefficient is r = 0.93. This indicates a fairly strong positive correlation (JJ < 0.05) between the two variables of estimated age and actual age. Regression equations were used to compute the predicted age or y. The equation for pre- dicted age is y = 1.47 (x) -18.46. Fig. 3 represents a subgroup of persons 55 years of age and younger, which consisted of 17 specimens with a mean age of 43 years. The correlation coefficient is r = 0.98 for this subgroup. The equation for the predicted age is y = 1.08 (x) -1.84.

Fig. 4 represents an older subgroup of persons over 55 years, which consisted of 35 specimens with a mean age of 70 years. The correlation is r = 0.85. The pre- dicted age formula is y = 1.40 (x) - 13.03. A line of best fit was calculated for each scattergram.

DISCUSSION The results of this article are less supportive than

the Stott et a1.23 direct correlation of estimated age to

268 Stein and Corcoran ORAL SUKCERY ORAL MEDICINE ORAL PATHOLOGY Mnrch I994

Table I. Counts of cemental annulations with eruption time estimates compared with known age of persons (in years)

Specimen number

Tooth number annulations

1 45 31 11 2 21 90 7 3 41 43 7 4 21 44 I

5 25 46 10

6 45 21 II I 42 36 7 8 22 25 8 9 23 42 II

10 15 45 10

11 44 43 9

12 45 42 10

13 24 32 9

14 34 45 IO 15 33 46 I I

16 44 52 9

17 21 55 7

18 11 56 7 19 14 32 9

20 24 18 IO

21 15 51 10

22 25 53 IO

23 33 40 1 1

24 35 41 IO

25 24 45 9

26 11 41 7

27 21 48 7

28 21 110 I

29 23 53 I I

30s 25 42 10

31 14 52 9

32 33 46 11

33 32 50 x

34 24 32 10

35 15 30 IO

36 13 28 II

37 44 51 9

38 23 53 II

39* 11 49 7

40 21 60 7

41 32 57 8

42 35 62 IO

43 31 58 7

44 24 51 9

45 25 54 I 0

46 13 42 11

47 11 49 I 1

48 21 57 7

49 22 49 8

50 22 47 8

51* 24 46 9

52* 22 48 8

* = Female

actual age. This study, however, has a higher corre- patients under 30 years and r = 0.001 for patients lation than the results of Lipsinic et a1.24 The corre- over 30 years. In the Lipsinic study, after a specimen lation of Lipsinic’s study for his 42 specimens is that varied more than three standard deviations was r = 0.51, with subgroup correlations of r = 0.93 for separated from the group over 30 years of age, the

42 40 91s 2148 5 2 50 t 2/25 27

Sl 59

56 75

32 29

43 45

33 30

53 56

55 61

52 55

s2 55

41 43

55 67 57 67

0 I 67

62 7x

63 78

II 43

28 30

6 1 77

63 79

51 52

51 52

54 60

48 60

55 60 ll7+2/58 6X

64 84

52 53

61 75

51 66

58 66

42 42

40 42

39 42

60 64

64 75

56 74

61 82

65 79

72 79

65 19

60 79

64 79

53 55

60 68

64 12

57 65

55 58

55 65

56 65

Prcdicircl irge i\”

L- .~.--- __~._

43 .y 0 2.5 5x 6S 29 45

31

60

63 59

59

43

64

61

12

74

75

43

24

12 75

51

57

62

53

63

6X

71

59 72

66

6X

44

41

40

70

16

65

80

78

88

78 71

16

60

71 77

66 63 63

65

Stein and Corcoran 269 ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 77, Number 3

Actual Age Actual Age

30 40 50 60

Estimated Age

Fig. 2. Computer-generated scattergram of 52 specimens. Line-of-best-fit plotted, r = 0.93.

correlation improved to r = 0.64. The lower correla- tion in older persons in the Lipsinic et al. study and the present study may be due to both a small sample size (n = 11) and a decreased apposition of cementum in persons over 60 years as discussed in Solheim’s25 article. Solheim had a correlation of r = 0.63 between thickness of cementum and age in his sample of over 1000 specimens. Solheim showed that cementum ap- position diminishes by one third after the age of 60. The rate of deposition decreased, becoming noncon- cordant with age. That is to say that cementum appo- sition thickness is greater up to age 60 and then ap- pears to diminish after that age.

Another factor relevant to discrepancies in results is tooth selection. Whereas Lipsinic et al. chose to use maxillary bicuspids, others such as Solheim25 state that mandibular second bicuspids and central incisors have the best correlation for annulation counting.

There was a doubling of the counts in specimens 2, 3, and 28 (Table I). If the higher counts are divided in half, then the estimates are very close to the actual chronologic age. This doubling is mentioned by Kay et a1.26 Doubling may explain why the dark/light in- terface appears to be less distinct in higher magnifi- cations. This lack of distinction can lead to problems with annulation counts as expressed by Miller et a1.27

The sectioning method to be used has also not been a topic of agreement. Many authors prefer the sections to be longitudinal, whereas others prefer cross sections. Both methods seem to have advantages and disadvantages. Although the longitudinal sec- tions allow viewing the whole root surface such as ad- vocated by Klevezal and Kleinenberg,28 others such as Stott et a1.23 prefer a cross section that allows a series of observations. It would appear that cross sections would be easier to perform on a routine basis. In ad-

.? 2 . 407 . . 30 .

L/ 1 I , 30 40 50

Estimated Age

Fig. 3. Computer-generated scattergram of 17 specimens 55 years of age and younger with line-of-best-fit, r = 0.98.

Actual Age

60

70 -

/

. * 60 -

. . . * . . I/, .

50 I I 50 60 70

Estimated Age

Fig. 4. Computer-generated scattergram of 35 specimens, over 55 years of age with line-of-best-fit, r = 0.85.

dition, cross sectioning would remove a possible problem with longitudinal sectioning performed tan- gentially to the long axis of the tooth with resultant distortions.

Annulation counts for those persons 55 years of age and younger appear to be very close to the actual age as expressed by the simplified regression equation, y = x - 1.8, where x is equal to the sum of the annu- lation count and average eruption age and y is equal to the newly expressed predicted age.

Annulation for those persons over 55 years show an increased tendency to be unrepresentative. This can be seen by the regression equation for the older age group, where y = 1.4x - 13. In samples of actual ages of 60 years (specimen 25), 68 years (specimen 47), and 79 years (specimen 43) we see calculated age es- timates of 54,60, and 65 respectively. This represents

270 Stein and Corcoran ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY March I994

undercounts of 6, 8, and 14 years. As age increases, the estimates are at a greater divergence for an aver- age undercount of 9 years. Applying the formula for adjusted predicted age y = 1.4x - 13, an overcount of 2 years (specimen 25), 3 years (specimen 47), and an undercount of 1 year (specimen 43) can be seen. The data indicate that cementum annuli are a moderately reliable means to estimate the age in human beings.

The predictive power of regressive analysis can make age estimation more accurate. This predictive power is greater in younger persons than in older per- sons. The calculated regression numbers appear at the right of Table I. There is not a linear 1: 1 relationship between cementum annulations and age in years. The regression slope of the scattergram differs from 1. The actual annulation counts cannot be used as age pre- dictors in themselves, but they can become relative values to be used as input in the appropriate regres- sion equations. This generates age estimates or pre- dictions of actual age.

We thank George Harrison for his excellent slide prep- aration and Shirley Tate for her assistance in the prepara- tion of this article. We also thank the Hospital Administra- tion and staff of the Veterans Administration in Ann Arbor, Mich. for their assistance, patience, and flexibility in allow- ing us to pursue this endeavor. The triad of patient care, education, and research that is the mission of the Veterans Administration has allowed many health providers an op- portunity to help patients and to strive to provide a better health delivery system.

REFERENCES

5.

6.

I.

8.

Tencate A. The estimation of age of skeletal remains from the color of roots of teeth. J Can Dent Assoc 1977;2:83-6. Gustafson G. Age determinations on teeth. J Am Dent Assoc 1950;41:45-54. Zander HA, Hiirzeler B. Continuous cementum apposition. J Dent Res 1958;37:1035-44. Nitzan D. The effect of aging on tooth morphology: a study on impacted teeth. ORAL SURG ORAL MED ORAL PATHOL 1986; 61:54-60. Stein TJ, Corcoran J. Anatomy of the root apex and its histo- logic changes with age. ORAL SURG ORAL MED ORAL PATHOL 1990;69:238-42. Linhart SB. Age determination and occurrence of incremen- tal growth lines in the dental cementum of the common vam- pire bat (Desmodus rotundus). J Mammal 1973;54:493-6. Fog1 JG, Mosby HS. Aging gray squirrels by cementum annuli in razor-sectioned teeth. J Wildlife Management 1978; 42: 444-8. deRicqlis AJ. Evolution of endothermy: histological evidence. Evol Theory 1974;1:51-80.

9

10

I I,

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Peabody FE. Annual growth Lanes tn living and iossrl \*ertc- brates, .I Morph01 l96l;lOX:l I-h?.. Kock PL, Fisher DC, Dettman D. Oxygen isotope variatmn in the tusks of extinct Proboscideans: a measure of season of death and seasonability. Geology I989:17:515-9. Sauer PR, Free S. Browne S. Age determination in black bears from canine tooth sections. NY Fish Game J 1966:33:125-39. McEwan E. Seasonal annuli in the cementum of the teeth of barren ground caribou. Can .I Zoo 1963;41:1 l-l 13, Gasaway WC, Harkness DB, Rausch RA. Accuracy of moose age determinations from incisor cementum layers. .I Wildlife Management 1978;42:558-63. Keiss RE. Comparison of eruption-wear patterns and cemen- turn annuli as age criteria in elk. .I Wildlife Management 1969;33:175-80. Lochard GR. Further studies of dental annuli for aging white- tailed deer. J Wildlife Management 1972;36:46-55. Novakowski M. Cemental deposition as an age criterion in bi- son, and the relation of incisor wear, eye-lens weight, and dressed bison carcass weight to age. Can J Zoo 1965;43: 173-8. Monson RA. Stone WB, Parks E. Aging red foxes (Vulpes falva) by counting the annular cementum rings of their teeth. NY Fish Game J 1973;20:54-61. Wada K, Ohtaishi N, Hachiya N. Determination of age in the Japanese monkey from growth layers in the dental cementum. Primates 1978;19:775-84. Free S, Sauer I’. Cementum layering in canine tooth of a 28- year-old captive black bear. NY Fish Game J 1966;13:233-4. Charles DK. Condon K, Cheverud JM, Buikstra J. Cementum annulation and age determination in homosapiens: tooth vari- ability and observer error. Am J Phys Anthropol 1986;71:31 I- 20. Mansfield AW, Fisher HD. Age determination in the harbor seal, Photo vitulina L. Nature 1960;186:92-3. Gray’s Anatomy, 35th ed., Warwick R, and Williams PL. eds.. Philadelphia: WB Saunders Co., 1973:1184-5. Stott GG. Sis RF, Levy BM. Cemental annulation as an age criterion in forsenic dentistry. J Dental Res 1982;61:814-7. Lipsinic FE, Paunovich E, Houston GD, Robison SF. Corre- lation of age and incremental lines in the cementum of human teeth. J Forensic Sci 1986;31:982-9. Solheim T. Dental cementum apposition as an indicator of age. Stand J Dent Res 1990;98:510-9. Kay RF, Rasmussen DT, Beard K. Cementum annulus counts provide a means for age determination in Macaca Mulatta (Primates, Anthropoidea). Folia Primatol (Basel) 1984;42:85- 95. Miller CS, Dove SB, Cottone J. Failure of use of cemental an- nulations in teeth to determine the age of humans. J Forensic Sci 1988;33:137-43. Klevezal GA, Kleinenberg SE. Age determination of mammals from annual layers in teeth and bones [translated from Rus- sian by J. Salkind, 19691 Springfield, Virginia: U.S. Depart- ment of Commerce, 1967.

Reprint requests. Thomas J. Stein, MEd, DMD, MS Endodontics Department Veterans Administration Medical Center Ann Arbor, MI 48105