A Quantitative Analysis of Mineral Loss and Shrinkage of in vitro Demineralized Human Root Surfaces J.M. TEN CATE, B. NYVAD', Y.M. VAN DE PLASSCHE-SIMONS, and 0. FEJERSKOV' Department of Cariology & Endodontology, Academic Centre for Dentistry Amsterdam (ACTA), Louwesweg 1, 1066 EA Amsterdam,The Netherlands; and 'Department of Oral Anatomy, Dental Pathology, and Operative Dentistry, Royal Dental College Aarhus, Aarhus, Denmark

Demineralization of dentin specimens proceeds at a faster rate than that ofenamel. Although this is generally accepted, a quantification of the rate offormation of root lesions is hampered by the shrinkage of the lesions when these are dried prior to microradiographic analysis.' This leads to a significant underestimation of the lesion depth and total mineral loss. The aim ofthis paper was to quantitate the rate of mineral loss during root lesion formation in vitro and to determine the shrinkage of root specimens as a result of drying. Unerupted roots ofhuman teeth were subjected to a demineralizing system of 0.1 mol/L lactate buffer (pH = 4.8) with 0.2 mmol/L methanehydroxydiphosphonate during four, 11, 22, and 44 days. The root lesions were assessed by quantitative microradiography. The demineralizing solutions were analyzed to determine the amounts of root tissue dissolved. A comparison of these two sets of data showed that, with the demineralizing system used, root lesions may shrink up to 62%. Fixation of the specimens in fixative did not affect this shrinkage. Chemical analysis showed that mineral loss proceeded linearly with time. From the data-sets of this study, a model was developed to compensate for the shrinkage in the dentin specimens. In this way, it was possible to calculate the lesion depth at four demineralization times as being 130, 220, 320, and 530 PM, respectively. These values were in agreement with a microscopic determination of the lesion depth.

month intra-oral period. Moreover, Nyvad et al. (1989) suggested that the values determined for the lesion depth underestimated the true extent of the lesion, since the specimens had undergone considerable shrinkage prior to and duringmicroradiographic analysis. The aim ofthe present study was to obtain quantitative data on the magnitude of shrinkage by correlating the microradiogaphic and chemical data of root lesions created under chemical conditions in the laboratory.

Materials and methods.

Specimen preparation. -Specimens were obtained from mesial or distal root surfaces of 48 partially-erupted human third molars that had completed root formation. After extraction, the teeth were stored on cotton wool saturated with thymol-H20. Before the experiments, remnants of the periodontal ligament were removed by immersing the teeth in 5% (v/v) NaOCl for three h. This procedure was followed by ultrasonic cleaning for 10 min (Branson 2200' ultrasonic cleaner) and rinsing in distilled water for three days under constant stirring. Finally, a "window" of standardized size (10 mm2) was created in the cervical part of the root cementum of each tooth by the use of nail varnish. Demineralization.-Artificial lesions were produced by exposing the teeth to a demineralizing solution (0.1 mol/L lactate buffer at pH J Dent Res 70(10):1371-1374, October, 1991 4.8, 0.2 mmol/L MHDP) for four, 11, 22, and 44 days at 370C, respectively. The number of teeth from each group is given in the Introduction. Table. The demineralizing solution was changed once every week. This time scheme allowed accurate quantification ofthe amount of Recent studies of root surface caries have focused on the behavior of the root mineral dissolved, without the solution becoming saturated exposed dentin to de- and remineralization in the oral cavity. to a significant degree. Hoppenbrouwers et al. (1986, 1987) studied the thermodynamic In each demineralizing solution, the calcium concentration was properties of dentin and showed dentin to be more soluble than measured with an atomic absorption spectrophotometer (Perkin dental enamel under conditions simulating cariogenic challenges in Elmer 372), with phosphate interference being suppressed by the the oral environment. Several authors have examined the kinetic addition of 1.56% (w/v) La(NO)2 .6H20 in 50 mmol/L HCl. This aspects of lesion formation in dentin by different methodologies. calcium value gives the chemicalinformation regarding the formation The conclusions derived by the groups differ. For undersaturated of the root lesion. calcium phosphate solutions and for lactate/methyleneLesion analysis.-After demineralization, 12 of the 48 teeth hydroxydiphosphonate (MHDP) buffers, Featherstone et al. (1987) (groups 2b and 3b) were fixed in Karnovsky (1965) fixative for 24 h, and Almqvist et al. (1988) found the lesion depth and mineral loss whereas the remaining teeth were not treated further. to be linearly related to the square root of demineralization time, Upon termination of the demineralization period, the specimens indicating a diffusion-controlled demineralization process. Arends were sectioned in a Leitz sectioning machine with a water-cooled et al. (1987) studied lesion formation in dentin in a pH 5 gel system. diamond-coated wheel and ground to a uniform thickness of about Their data for lesion depth fit equally well to linear and to square 80 jim. Microradiographs were produced using Cu-Kcx radiation root of time relationships. (Philips X-ray generator 1729) filtered by a Ni-filter, on high For experimental root caries in situ, Ogaard et al. (1988) re- resolution plates (Kodak Type 1A). The dentin specimens and an ported very fast initial demineralization of dentin in stagnant areas aluminum step-wedge (covering 25-200 gum and prepared from under orthodontic bands, followed by a slower linear increase in ultrapure aluminum) were simultaneously exposed to make the lesion depth for periods beyond two weeks. The lesions observed by contact microradiograph. The plates were processed under stanthese authors were characterized by the lack ofa surface layer. This dardized conditions and analyzed in a microscope (Zeiss) equipped observation is in contrast to results from in situ studies by Nyvad et with a CCD videocamera (Panasonic) and computer with a frameal. (1989), who found that root caries lesions with free access to grabber and processor (Data Translation). Image analysis software saliva were subsurface in nature and that mineral loss and lesion (Imagepro) and dedicated software were used to transform the depth progressed linearly with time throughout the entire three- image-scan to mineral-content-depth profiles by Angmar's formula (Angmaret al., 1963). The mineral profiles were measured at three to six locations on each section (depending on the degree ofinterscan Received for publication December 7, 1990 variation). After averagingper section andper specimen, an overall Accepted for publication May 15, 1991 Downloaded from jdr.sagepub.com at UNIV OF PITTSBURGH on March 13, 2015 For personal use only. No other uses without permission.

1371

TEN CATE et al.

1372

Can~ ~ +

J Dent Res October 1991

volume % mineral 2

50

E 0 E

4-a 06

40

._

-a0

30

0

20

10

40

30

0

group 1

+

group 2a

So

group 2b

A

goup 3a

o

g0ou

V

group 4

tb

50

time (days) Fig. 2-Mineral loss during dentin lesion formation in vitro determined at different time intervals from the individual demineralizing solutions. Group numbers as indicated in the Table or Fig. 1.

20

mineral 1088 (pmol Ca/mm2)

c

10

4.50

E

0

0

3.00

0 CA 0

15

L

0

100

200

300

400

depth (pm) Fig. 1-Average mineral content profiles for each of the six experimental groups. Demineralization times varied from four, 11, 22, to 44 days (groups 1, 2a/b, 3a/b, 4, respectively). For the intermediate time periods, lesions were sectioned either with (2b, 3b) or without (2a, 3a) fixation in Karnovsky fixative after demineralization.

group average and standard deviations were calculated at 1-pmdepth intervals. From each profile, the lesion depth and the amount of mineral removed (A MC) were calculated. These values constitute the microradiographic data of the root lesions. In addition to the microradiographic assessment, the overall lesion depth (including the shrinkage) was measured directly under a microscope. A thin coverglass was carefully moved over the microradiograph in such a way as to create a line connecting the anatomical surfaces of the sound tissue at both sides of the lesion. For non-curved roots, this procedure allowed an estimation of the location of the original surface at the site of the lesion. Then, at high magnification, the lesion depth-being the total ofthe depth of the lesion observed in the microradiograph and the depth caused by specimen shrinkage-was determined using a measuring grid.

Results. The data obtained from the chemical and microradiographic analyses of the demineralizing solution and of the root lesions, respectively, are given in Figs. 1 and 2 and in the Table. Fig. 1 depicts the mineral profiles for the six experimental groups as well as for a group of untreated root specimens. The values for the sound dentin were normalized at 50% volume of mineral. The mineral profiles for the different demineralization periods showed a similar pattern with no surface layer and fairly constant values for mineral content in the lesion body. At the lesion front, the mineral content increased sharply to reach the sound dentin values within about 50 pm. Statistical analysis of the lesion parameters (integrated min-

E0.0

.50I

E < 0.00

a

+

1.50

X

3.00

data points l/ine

of unity

v~~~~~~~regression line 4.50

6.00

chemical assessment Fig. 3-Comparison of mineral loss (in pnol CaImm2) determined from microradiographs (y-axis) and from chemical analyses ofthe demineralizing solutions (x-axis). The dotted line shows the theoretical relationship, whereas the solid line denotes the regression line of the data points from this study.

eral loss and lesion depth) showed no systematic pattern of significant difference between the groups with and without fixation. The mineral loss with time-calculated from the change in the individual

demineralizing solutions and averagedper groupper demin-

eralization time-is shown in Fig. 2. The lesion parameters and the chemical analysis of the demineralizing solutions are given schematically in the Table. Finally, a direct comparison between the chemical and microradiographic assessment of the lesion formation showed that microradiographic analysis resulted in a significant underestimation ofthe extent ofthe lesion (Fig. 3), the slope ofthe regression line being 0.41.

Discussion. The results of this study show the vulnerability of root surfaces to relatively mild demineralizing conditions. Also, they demonstrate the rather extensive shrinkage experienced by the specimens when dried prior to microradiographic analysis (Phankosol et al., 1985; Mellberg and Sanchez, 1986; Wefel et al., 1987). This leads to a significant underestimation of the lesion depth and mineral loss of lesions when these are determined by microradiography. Root specimens do not shrink to the same extent when they are analyzed in an aqueous medium. This would favor examination by polarized

Downloaded from jdr.sagepub.com at UNIV OF PITTSBURGH on March 13, 2015 For personal use only. No other uses without permission.

1373

MINERAL LOSS AND SHRINKAGE OF ROOT SURFACE LESIONS

Vol. 70 No. 1 0

11 days dwmineralizatlon

4 days domlneraNzatlon

I-

0

0 L.

2 .E

L.

2

E0

0

.2

0

0

0

600

a~w*eh spookam

100

200

300

400

500

depth (microns)

depth (microns)

22 days dInalzatlon

44 days danlralzatlon

a t. 0

L.

0

C

600

40

._

E

30 /

0 20

0

2

__

_

20

2

20

:0

100

0

200

300

400

500

10

-,

100

0

600

{

-

200

depth (microns)

-

300

400

I

I

500

600

depth (microns)

Fig. 4 (a-d)--Mineral content profiles, re-calculated by mathematic expansion of the specimens. The dotted verticle line marks the location of the shrunken surface.

light microscopy in various watery media, but this technique does not give information regarding mineral content patterns in tissues. Fixation of the specimens prior to handling did not affect the degree of shrinkage. Fixation was employed in an attempt to stabilize the collagen matrix remaining during demineralization (Hayat, 1970), but apparently this treatment did not influence the results when microradiography was used as the analytical tool. In this context, it should be appreciated that fixation and subsequent drying of specimens prior to examination by, for example, SEM in and of itself causes substantial dimensional changes in the tissue (Boyde, 1978), the extent of shrinkage of course being strongly dependent on the type of tissue examined. Comparison of the mineral-loss data calculated from chemical and microradiographic analysis (Table, Fig. 3) revealed that the latter technique underestimated the mineral loss by values up to 62%. Data from this study were used to quantitate the shrinkage and to correct the mineral profiles for this artefact. The reasoning used and the mathematical procedure are described below. We assume that the local shrinkage is (linearly) related to the local degree of porosity:

(shrinkage)

=

f.

=

shrinkage factor; mineral content at depth x, and;

shrinkage at depth x.

To compensate for the shrinkage, each depth layer (x) is expanded (mathematically) proportional to the degree of mineral loss of that layer: (MC)sounld - (MC)x

(Ax)x~expanded =(x)xmeasured

[1

+ f

]

(MC)sound

Because the total amount of mineral for the expanded layer of dentin must remain constant, its mineral content decreases proportional to the degree of expansion: (Ax) x~esred

(MC)xexpanded

=(MC)xmeasured

.

Axexpanded

The total amount of mineral loss for the expanded dentin was calculated by summation over all the (hypothetical) depth layers of the root lesion.

lesion depth (A

(M)sound

=

=

(MC)sound (MC)x (MC),Ud

with f (MC)X

(shrinkage)x

MC)ta=

I x = 0

[(MC)sound

(MC)xexpanded]

(AX)x ,expanded

AX,.....

Since the mineral removed during demineralization [i.e., (A MC),Ot] is measured chemically, the shrinkage factor f and the "expanded"

Downloaded from jdr.sagepub.com at UNIV OF PITTSBURGH on March 13, 2015 For personal use only. No other uses without permission.

1374

TEN CATE et al.

Time of Number of Demineralization Group Specimens (days) 4 1 8

J Dent Res October 1991

TABLE DEMINERALIZATION CONDITIONS AND RESULTS Chemical Results Microradiographic Results (MRG) Shrinkage Observed Karnovsky Mineral Removed Lesion Depth Mineral Removed Microscopically: Fixation (pmol Ca/mm2) MRG/Chemical (pmol Ca/mm2) (pm) Range (pm) 0.81 (0.14) 113 (7) 0.53 (0.18) 15-50 0.65

2a

10

11

-

1.90 (0.09)

141 (5)

0.72 (0.19)

0.38

n.d.

2b

6

11

+

1.99 (0.05)

160 (15)

0.91 (0.17)

0.46

n.d.

3a

10

22

-

2.83 (0.14)

211 (11)

1.58 (0.29)

0.56

n.d.

3b

6

22

+

3.19 (0.14)

220 (19)

1.32 (0.23)

0.41

n.d.

4

8

44

-

5.40 (0.07)

322 (13)

2.36 (0.24)

0.44

150-300

mineral profiles can be calculated using simple numerical mathematical methods (Fig. 4). The shrinkage amounted to about 20, 75, 80, and 210 pum for the four demineralization periods. These values are in good agreement with the amount of shrinkage measured microscopically for the two extreme demineralization periods. We therefore conclude that this simple model may be used in future studies to compensate numerically for the shrinkage of dentin specimens, in particular in those cases where quantitative estimation ofmineral loss is required and shrinkage cannot be prevented. For the model to be used quantitatively, it seems necessary to validate the current findings for different demineralizing systems and for partially remineralized lesions. An interesting observation was the relatively constant value for the mineral content in the lesion body. Based on the expanded profiles, about 10 volume % mineral was not dissolved during the demineralization. We assume that this is a fraction of the mineral that was either bound to the collagen matrix or present within the collagen fibers. The chemical data for demineralization showed that the mineral dissolution process during demineralization was linear with time (Fig. 2). This finding is consistent with some other reports in which lesions were formed under various conditions in vivo and in vitro (Featherstone et al., 1987; Arends et al., 1987; Almqvist et al., 1988; Ogaard et al., 1988; Nyvad et al., 1989), and showed that the demineralization process was not governed by diffusion restrictions butbythe rate ofdissolution atthe crystallite surface. The apparent high permeability ofthe tissue to acid explains the high vulnerabilty of the dentin to caries-like challenges in vitro as well as in situ. Moreover, the larger surface area of the small crystallites (as compared with that of enamel) may enhance the rate at which dissolution takes place. The suggested higher solubility product of dentin mineral may add to this (Hoppenbrouwers et al., 1987). As would be expected, microradiography was not as sensitive and reliable for the detection of mineral loss from dentin as was chemical analysis. Our data showed that, on average, the microradiographic technique was able to record about half of the mineral that was actually lost. However, there was a constant linear relationship between the results from the two methods employed. The pattern of demineralization ofroot surfaces described in this laboratory experiment may elucidate certain basic physico-chemical principles governing dissolution ofthis type of tissue. However, Fig. 1 demonstrates that the pattern of mineral distribution in the artificially developing lesion was different from that obtained during controlledroot-surface cariesdevelopmentinsitu (Nyvadetal., 1989). In the latter case, a pronounced surface layer and a more gradual increase in mineral content from lesion body to sound tissue were observed. Therefore, to improve our understanding of the processes

that control root surface caries development, we suggest that a comparison between in situ and in vitro studies be carried out. REFERENCES ALMQVIST, H.; WEFEL, J.S.; LAGERLOF, F.; EKSTRAND, J.; and HENRIKSON, C.O. (1988): In vitro Root Caries Progression Measured by 125I Absorptiometry: Comparison with Chemical Analysis, JDent Res 67:1217-1220. ANGMAR, B.; CARLSTROM, P.; and GLAS, J.E. (1963): Studies on the Ultrastructure of Dental Enamel, IV. The Mineralization of Normal Human Enamel, J Ultrastruct Res 8:12-23. ARENDS,J.; CHRISTOFFERSEN,J.;RUBEN,J.;andCHRISTOFFERSEN, M.R. (1987): Lesion Progress in Dentine and the Role of Fluoride. In: Dentine and Dentine Reactions in the Oral Cavity, A. Thylstrup, S.A. Leach, and V. Qvist, Eds., Oxford: IRL Press, pp. 117-125. BOYDE, A. (1978): Pros and Cons of Critical Point Drying for SEM, Scan Electron Microsc 11:303-314. FEATHERSTONE, J.D.B.; McINTYRE, J.M.; and FU, J. (1987): Physicochemical Aspects of Root Caries Progression. In: Dentine and Dentine Reactions in the Oral Cavity, A. Thylstrup, S.A. Leach, and V. Qvist, Eds., Oxford: IRL Press, pp. 127-137. HAYAT, M.A. (1970): Principles and Techniques of Electron Microscopy, Biological Applications, Vol. 1, New York: Van Nostrand Reinhold Co., pp. 65-81. HOPPENBROUWERS,P.M.M.;DRIESSENS,F.C.M.; andBORGGREVEN, J.M.P.M. (1986): The Vulnerability of the Unexposed Human Dental Roots to Demineralization, J Dent Res 65:955-958. HOPPENBROUWERS, P.M.M.; DRIESSENS, F.C.M.; andBORGGREVEN, J.M.P.M. (1987): The Mineral Solubility of Human Tooth Roots, Arch Oral Biol 32:319-322. KARNOVSKY, M.J. (1965): A Formaldehyde-Glutaraldehyde Fixative offHigh Osmolality for Use in Electron Microscopy, J Cell Biol 26:137A-138A. MELLBERG, J.R. and SANCHEZ, M. (1986): Remineralization by a Monofluorophosphate Dentifrice in vitro of Root Dentin Softened by Artificial Caries, J Dent Res 65:959-962. NYVAD, B.; TEN CATE, J.M.; and FEJERSKOV,O. (1989): Microradiography of Experimental Root Surface Caries in Man, Caries Res 23:218-224. OGAARD, B.; R0LLA, G.; and ARENDS, J. (1988): In vivo Progress of Enamel and Root Surface Lesions under Plaque as a Function of Time, Caries Res 22:302-305. PHANKOSOL, P.; ETTINGER, R.L.; HICKS, M.J.; andWEFEL, J.S. (1985): Histopathology ofthe Initial Lesion ofthe Root: An in vitro Study,JDent Res 64:804-809. WEFEL, J.S.; CLARKSON, B.H.; and HEILMAN, J.R. (1987): Histology of Root Surface Caries. In: Dentine and Dentine Reactions in the Oral Cavity, A. Thylstrup, S.A. Leach, and V. Qvist, Eds., Oxford: IRL Press, pp. 139-146.

Downloaded from jdr.sagepub.com at UNIV OF PITTSBURGH on March 13, 2015 For personal use only. No other uses without permission.