Clinical trials with crystal growth conditioning as an alternative to acid-etch enamel pretreatment

Jon htun and Sven Bergland Oslo, Norway

It has been claimed that different ion solutions containing sulfate induce crystal growth and might be a better alternative than conventional acid etching for enamel pretreatment in bracket bonding. It should thus combine optimal bond strength with easy and quick debonding. Two clinical experiments were performed to test this hypothesis. The first experiment dealt with the debonding procedure. Following conditioning with dilute sulfuric acid which contained sodium sulfate (to be termed solution A) on one side of the mouth and etching with 37% phosphoric acid on the other, brackets were bonded on the maxillary and mandibular incisors of twenty dental nurse students. Debracketing and a subsequent cleanup procedure were performed after 2 days. The mode of loosening was mainly between the enamel surface and adhesive on the crystal-growth-conditioned teeth and between the bracket mesh and adhesive on the teeth etched with phosphoric acid. This difference in mode of loosening was statistically significant (P < 0.001). In the second experiment, which dealt with the clinical bond strength, 250 brackets were bonded in forty patients. One side served as a control and was conventionally acid etched. On the experimental side conditioning was done with solution A in thirty patients. In ten patients, 10% phosphoric acid was added to the dilute sulfuric acid used (to be termed solution B). Failure rates and modes of failure were recorded for a &month period. It was found that the failure rates were significantly higher after enamel conditioning with solution A than after conditioning with solution B (P < 0.001) and after conditioning with solution B than after phosphoric acid etching (P < 0.05). Subsequent to solution A conditioning, nearly all the brackets came loose during the first 2 weeks. When solution B was used, the failures occurred at a later point of time.

Key words: Enamel conditioning, crystal growth, debonding, adhesiveremnants, failure rates

onventional etching with phosphoric acid is known to produce dissolution of the outermost enamel layer and provide mechanical attachment for bonded orthodontic brackets when suitable acrylic or diacrylate resins are used .lm3 However, the debonding and subsequent cleanup procedure at the end of treatment may be rather time-consuming. Also, care must be taken not to induce iatrogenic effects, including cracks, scratches, and removal of pieces of enamel.- All adhesive remnants should be removed at the time of debonding, as abrasive wear of most orthodontic adhesives apparently is minimal .* Several in vitro bond-strength tests indicate that bond failure with metal brackets is likely to take place in the interface between the adhesive resin and the mesh., +-11 This might imply that the mechanical interlocking that occurs when the adhesive penetrates into etched enamel to a depth of up to 50 pm* l3 may be stronger than is really necessary. If this holds true, a
From the Department of Orthodontics, University of Oslo.

weaker connection between enamel and adhesive would allow easier and less time-consuming cleanup, without higher failure rates. It has been claimed by Smith14 that polyacIylic acid cements display adhesive bonding to dental enamel and that the interaction of the aqueous polyacrylic acid component of the cement with enamel varies according to the formulation. Purified polyacrylic acid produces only slight etching of the enamel surface. However, polyacrylic acid solutions which contain residual sulfate ion produce, in addition to slight etching, a crystalline deposit that bonds firmly to the enamel surface and resists mechanical removal.15 The crystals were shown to be calcium sulfate dihydrate (gypsum). The crystal formation depends mainly on the sulfate ion concentration in the polyacrylic acid solution and is independent of the molecular weight or concentration of the so1ution.l The potential value of this crystalline interface as mechanical interlocking for orthodontic bracket bonding has been tested in vitro by Maijer and Smith.j
333

334

&-run

and

Bergland

Am. J. Orthod. April 1984

Table

I. Total number of brackets bonded in each experimental group, divided into tooth groups
Group I II III IV Total Incisors 36 32 16 28 112 Canines 20 12 12 16 60 Premolar-s 30 8 18 14 70 side were contralateral Molars 8 Total 94 52 46 58 250 2).

Table
Group I II III IV

II. Study design on the test side in each experimental group (experiment 2)
Conditioning Solution Solution Solution Solution agent A A A B Bonding Concise Concise Crystal Concise adhesive Saliva evacuator

Bond

Conventional High-speed High-speed High-speed

8 (experiment

The test side and the control

They found the maximum density of long, needleshaped crystals growing on the enamel surface after conditioning for 4 minutes with 40% polyacrylic acid containing the appropriate amount (3.8%) of sulfate ions. The crystalline interface produced a tensile bond strength equivalent to that of a conventionally acidetched surface. At debonding the fractures occurred mostly in the crystal/resin interface when the brackets were pulled off the teeth. However, the remnants could be removed easily with an ultrasonic scaler and pumice, apparently leaving the fluoride-rich outer enamel surface smooth and intact .lfi It has since been discovered that crystal growth can be achieved with other ion solutions. After extensive laboratory investigation, it was found that a sulfuric acid anion source gave more uniform and reliable growth patterns than the polyacrylic acid while reducing application time to 90 seconds. The purpose of the present study was to test the applicability of two crystal bonding agents in routine clinical orthodontic practice, using two test solutions containing sulfuric acid distributed by Ormco/Sybron for clinical evaluation. It was of particular interest: (1) to find out if debracketing and subsequent adhesive cleanup were easier and quicker with this method than after etching with phosphoric acid and (2) to determine clinically whether the two enamel-conditioning methods result in similar or different failure rates in terms of the number of loose brackets.
MATERIAL AND METHODS

Two experiments were performed, the first one on young adult volunteers who were to receive no orthodontic treatment and the second on patients who were going to start routine orthodontic treatment.
Experiment 1

from 17 to 28 years (mean, 22; SD, 15). The usual prophylaxis was performed. The two conditioning methods were distributed at random. On one side, the maxillary and mandibular incisors were etched with 37% phosphoric acid* for 60 seconds. On the other side, the maxillary and mandibular incisors were treated with an ion solution containing 1% sulfuric acid buffered to pH 1.5 and 15% anhydrous sodium sulfate? (to be termed solution A) for 90 seconds as recommended in the manufacturers instructions to achieve crystal growth. Following rinsing with water spray for 30 seconds and thorough drying,, 2, the enamel surfaces were examined carefully for any difference in appearance between the two sides. Standard meshbacked edgewise brackets+ were bonded with prediluted Concise as described elsewhere.16 After 2 days, the brackets were removed by two different commonly used debonding techniques. On nine individuals, the tie wings were grasped mesiodistally with Weingart utility plierss and gently squeezed together, thus peeling the edges of the bonding base away from the adhesive (technique S). On the remaining nine, the peel force was transmitted at the enamel-bracket base junction with bracket-removal pliers11 used to gently squeeze the edges of the base (technique P). An attempt was made to subjectively register the force necessary to remove the brackets. An Adhesive Remnant Index (ARI) system was used to evaluate the amount of adhesive left on the tooth after debracketing. This index system was developed on the basis of a pilot study on twenty extracted teeth, and the criteria are as follows: Score 0 = No adhesive left on the tooth. Score 1 = Less than half of the adhesive left on the tooth. Score 2 = More than half of the adhesive left on the tooth. Score 3 = All adhesive left on the tooth, with distinct impression of the bracket mesh.
*Concise etching liquid, 3M Company, St. Paul, Minn. tCrysta1 Bond Gel Conditioner, OmxolSybron, Glendora, $Chmesh, Ormco/Sybmn, Glendora, Calif. OIIIWJ 801-0062, Ormco/Sybron, Glendora, Calif. /IOrmco 801-0056, OmcoiSybron, Glendora, Calif.

The first part of the study was performed on the maxillary and mandibular central and lateral incisors of eighteen student dental nurses at the School of Dentistry, University of Oslo. The students ranged in age

Calif

Volume 85 Number 4

Crystal

growth

conditioning

335

Attempts were made to scrape the adhesive remnants off gently with a Schure scaler,* and the ARI system was used to evaluate the amount that could be removed in this manner. The residual adhesive was removed with a tungsten carbide bur according to the technique described earlier.* 6, l8 Students t test was conducted to test statistically significant differences in appearance of the adhesive remnants between the two conditioning methods and the two debracketing methods. After conditioning with 37% phosphoric acid or crystal growth gel conditioner as described above, four young premolars subsequent to extraction for orthodontic reasons were dehydrated and given an electronically conducting coat of carbon and gold in a vacuum evaporator. The surfaces were then examined in a scanning electron microscope, usually operated at 25 kv. Photographs were taken at magnifications of X 100 and X 1,000.
Experiment 2

Fig. 1. Clinical appearance after conditioning with dilute sulfuric acid that contained sodium sulfate and etching with 37% phosphoric acid. Careful examination reveals a slight difference. The crystal-growth-conditioned teeth are not so evenly white and frosty as the teeth etched with phosphoric acid. AE, Enamel etched with phosphoric acid. CG, Crystal-growth-conditioned enamel. SE, Saliva ejector.

The second part of the study was performed by one of us (S.B.) on forty persons, 24 girls and 16 boys, 11 to 15 years of age (meaty, 13 years; SD, 1.9), scheduled for routine orthodontic treatment of various types of malocclusion. The material was divided into four groups of ten persons. The patients were in a beginning stage of treatment. The number of brackets to be bonded initially varied from patient to patient according to the clinical situation. This resulted in different-sized groups in terms of number of brackets bonded (Table I). However, the teeth were bonded symmetrically, the test side and the control side being contralateral and thus comparable. Standard mesh-backed brackets were used,? and the test side and the control side were chosen in a random manner. On all control sides etching was performed with 37% phosphoric acid. Brackets were direct-bonded with prediluted Concise according to the routine described earlier.16 The study design on the test side is outlined in Table II. Crystals were grown on the enamel surfaces for 90 seconds, as recommended in the manufacturers instructions with two different ion solutions. In Groups I, II, and-111solution A was used. According to the manufacturer, the sulfuric acid in this solution etched the enamel to a depth of only about 4 to 6 prn.19 In Group IV, an ion solution containing 1% sulfuric acid buffered to pH 1.2, 15% anhydrous sulfate, and 10% phosphoric acid (to be termed solution B) was
*Schure 349, Rocky Mountain, Denver, Cola tAmerian, fine mesh, American Orthodontics,

used. This acid combination etched the enamel down to a depth of approximately 8 to 10 prn.17 After rinsing with water spray for 30 seconds and thorough drying, 3 * * brackets were bonded with Concise according to usual routines8 in Groups I, II, and IV on the test side as, well as on the control side. In Group III, an adhesive containing submicron-sized tiller particles, recommended by the manufacturer of the conditioning gel,* was used. In Group I only a conventional saliva ejector was used. In contrast, a high-speed evacuatort was used in Groups II, III, and IV. The patients were recalled and examined every 4 weeks. Bond failures were recorded whenever they occurred and at 6 months of observation. The enamel surface under loose brackets was scored according to the ARI system (see above). A test for comparison of two proportions based on normal approximation was conducted to determine statistically significant differences in bond failure rates between the groups.
RESULTS Experiment 1

Clinically, the difference between a tooth conditioned with dilute sulfuric acid containing sodium sulfate and a tooth etched with 37% phosphoric acid was not very obvious. However, careful examination revealed a slight difference, the crystal-growth-conditioned teeth not being so evenly white and frosty as the phosphoric-acid-etched teeth (Fig. 1). In the SEM, the pattern differed significantly between the two conditioning methods (Fig. 2). After conditioning with
*Crystal Bond, OrmcolSybron, Glendora, Calif. tDiirr Omsuc Mini, Diin Dental, Bietigheim, West Germany.

Sheboygan,

Wis.

336

htun

and Bergland

Am. J. Orrhod. April 1984

Fig. 2. SEM appearance after the two conditioning methods. A, After phosphoric acid etching. (Magnification, x 100.) B, After crystal growth conditioning. (Magnification, x 100.) C, After acid etching. (Magnification, x 1,000.) D, After crystal growth conditioning. (Magnification, x 1,000.) CD, Crystal deposits. EE, Etched enamel between crystals.

crystal growth conditioner, crystal deposits were seen to spread evenly over the enamel surface. Between the crystals, acid-etched enamel areas could be seen (Fig. 2, B and 0). After the 2 days, one maxillary central incisor bracket was dislodged on the experimental side. Debracketing was easier and better controlled with technique S than with technique P, especially in difficultto-reach areas such as the mandibular incisors. The force needed to debracket teeth treated with crystal growth conditioner apparently was much weaker than that needed to debracket teeth conditioned with phosphoric acid. Where crystals were grown, nearly all the adhesive followed the bracket base, and the little bit left on the enamel surface was easily removed after gentle scraping with a scaler. Where phosphoric acid was used,

most adhesive remained on the enamel surface after debracketing, and only minor amounts of the adhesive remnants could easily be scraped away. The difference is demonstrated in the typical clinical examples shown in Fig. 3, A and Fig. 4. The ARI difference shown in Table III was statistically significant between the two conditioning groups (P < 0.001). On the acid-etched side, there was a tendency for more adhesive to remain on the enamel when the bracket wings were squeezed together (technique S) than when a peel force was applied to the pad (technique P). However, this difference was not statistically significant.
Experiment 2

Table IV gives an overview of all loose brackets on the test and control sides in each experimental group. The failure rates in terms of loose brackets were ex-

Volume 85 Number 4

Crystal

growth

conditioning

337

Fig. 3. A, Typical

clinical appearance after debracketing of crystal-growth-conditioned phoric-acid-etched teeth. 8, Clinical appearance after removal of residual adhesive crystal-growth-conditioned teeth and with tungsten carbide burs on phosphoric-acid-etched Acid-etched enamel. CG, Crystal-growth-conditioned enamel.

teeth and phoswith scaler on teeth. AE,

Table

III. Adhesive remnants on the teeth scored according to the AR1 system after debracketing with technique S (squeezing bracket wings with Weingart utility pliers) and technique P (applying a peel force to bracket pads) and subsequent scaling (experiment 1) I
Lg. 4. A, Typical clinical appearance after debracketing of crystal-growth-conditioned teeth. B, Typical clinical appearance after debracketing of acid-etched teeth. AR, Adhesive remnants with distinct impression of bracket mesh.

Technique Scaling Technique Scaling

S P

30 35 20 36

5 16 4

1 9 8 14

5 21 17 12

30 6 11 6

One upper central incisor was dislodged on the side conditioned with crystal growth conditioner. No adhesive remnants on the tooth,

tremely high after conditioning with solution A. There was no difference in failure rates whether the brackets were bonded with Concise or the Crystal Bond adhesive recommended by the manufacturer. Out of a total of 96 brackets bonded after conditioning with solution A, as many as 93 (97%) came loose. Several brackets had already come loose during the ligation of the first arch wire, and the majority of the failures occurred during the first month (Table V and Fig. 5, A and B). According to the ARI system, 91 enamel surfaces were given score 0 and 2 were given score 1. After conditioning with solution B, 8 of 22 brackets (27%) came loose, and these failures happened at a later time (Table V). According to the ARI system, 3 surfaces were given score 0, 4 were given score 1, and 1 was given score 3. The failure rates were significantly higher after conditioning with solution A than after conditioning with solution B (P < 0.001); they were also significantly higher after conditioning with solution B than after phosphoric acid etching (P < 0.05). On the control side in experimental Group I, where

a conventional saliva ejector was used, 4 out of 47 brackets (8.5%) came loose. According to the ARI system, 2 were given score 2, 1 was given score 1 and 1 was given score 0. In experimental Groups II, III, and IV, where a high-speed evacuator was used, 3 out of 78 brackets (3.8%) came loose. According to the ARI system, 2 were given score 1 and 1 was given score 0. The difference in failure rates was not statistically significant .
DISCUSSION

Since the works performed by Smith, Cartz, and Maijer,14-16 it has been found possible to achieve crystal growth not only with polyacrylic acid but with many different ion solutions.17 The calcium sulfate crystals must enucleate from bound calcium in the enamel; to achieve proper crystal enucleation, some enamel etching may be required. A direct correlation between enamel acid solubility and crystal enucleation has been found. The mechanical attachment for bonding after crystal growth conditioning is created around the crystalline interface and the superficially etched enamel between the crystals (Fig. 2). The bond strength depends upon a combination of the strength of the chemical bond between the crystal deposit and the enamel sur-

338

A&n

and Bergland

Am. J. Orthod. April 1984

Fig. 5. A and B, Typical clinical appearance etched teeth. CG, Crystal-growth-conditioned

a few weeks after teeth with solution

mounting of the appliance. A. LB, Loose brackets.

AE, Acid-

Table

IV. Number of failures and failure rates on the test side and the control side in each experimental group, divided into tooth groups (experiment 2)
Tooth Incisors Group I Side No. % No. Canines 70 No. group
MOhS

Premolars % No.

Total 70 No. 5%

II III IV

Test Control Test Control Test Control Test Control

18 0 I6 0 8 0 3 0

100 0 100 0 100 0 21 0

IO 0 6 I 6 2 2 0

100 0 100 16.7 100 33 25 0

13 2 3 0 8 0 3 0

86.7 13 100 0 88.9 0 42 0

4 2

100 50

45 4 26 I 22 2 8 0

95.1 8.5 100 3.8 95.7 8 27.6 0

V. Number of loosened brackets on the test side in each experimental group related to time of loosening (experiment 2)
Table

Days
0

Group

Group

II

Group

111

Group

IV

I-10 1 l-20 21-30 31-40 41-

9 17 11 9 2

1 10 6 6 2 I

2 8 II 2

1 I 2 I 3

face and the strength of the mechanical bond achieved when the resin penetrates into the superficially etched enamel between the crystal deposits. The only published information on the applicability of the crystal growth principle in orthodontics is the work performed by Maijer and Smith. On the basis of laboratory test results, they reported that conditioning the enamel with sulfated polyacrylic acid for 4 minutes produced a tensile bond strength with an acrylic resin equivalent to that of phosphoric acid. Furthermore, they found that the bond fracture occurred mostly within the crystal adhesive resin layer when brackets were pulled off the teeth, indicating that the crystals

were bonded strongly to the enamel surface. However, the bonding was not too strong to allow easy cleaning of the enamel with an ultrasonic scaler and pumice. In our opinion, laboratory bond-strength tests have significant weaknesses. The forces can be standardized in only two directions (shear and tensile) and many factors, such as the surface curvature of the enamel, may influence the results. However, with the work by Maijer and Smith16 as a reference, the first part of this study undoubtedly confirmed that debonding of orthodontic adhesives is easy after crystal growth conditioning. The minor amounts of adhesive left on the enamel surface after debracketing (Figs. 3, A and 4, A) were easily scraped off with a scaler (Fig. 3, B). After phosphoric acid etching, on the other hand, especially after debracketing with technique S, nearly all the adhesive was left on the tooth and distinct impressions of the bracket mesh were generally seen (Fig. 4, B). One might argue that the proposed Adhesive Remnant Index (ARI) system is largely a subjective evaluation which cannot be used to calculate adhesive remnants. In our opinion, however, the system was of value and quite similar indices have been used by other investigators when attacking clinical problems.6. 3 ? The second part of this study did not confirm the

Volume 85 Number 4

Crystul

growth

conditioning

339

results of the previous in vitro bond-strength tests. Rather, it tended to confirm a clinical impression that what is easy to debond and remove may also have inadequate strength. However, it must be stressed that Maijer and Smith used polyacrylic acid to produce crystal growth, while in this investigation sulfuric acid was used. A close look at the photomicrographs in this investigation (Fig. 2, B and D), reveals that the crystals are not as long and needlelike as those reported by Maijer and Smith but are somewhat rounder and flatter. This difference in crystal morphology may be one explanation for the high failure rates. However, comparing the varying surface pattern seen after conventional acid etching, some differences in crystal morphology after crystal growth conditioning may be expected. The test population in this investigation had known histories of fluoride supplements. The extremely high failure rates after conditioning with solution A could be a result of reduced crystal growth due to reduced acid solubility of the fluoride containing hydroxylapatite in the enamel. To overcome the variation in acid solubility , 10% phosphoric acid was added in solution B This change improved the results significantly, but not to the extent of clinical applicability. Further increase in etchant concentration would be of advantage with respect to high concentrations of fluoride in the hydroxylapatite, but for the majority of patients this would merely approach a conventional etching process and negate any potential benefits of the crystal growth. Residual sodium sulfate left behind on the enamel surface as a result of improper washing may be another explanation for the high failure rates. It may be that prolonged rinsing with water spray would increase retention. It is also possible that the material used for buffering may interfere with proper bonding. The failure rates experienced in this investigation are considered clinically unacceptable. However, through further research in the areas just discussed, it would appear possible to arrive at more optimal conditioning agents. Since several different ion solutions may produce crystal growth, it seems important to make further clinical testing of other different crystal bonding agents to see if the results are comparable with those presented in this study for two such agents. Also, the problem of enamel variation in acid solubility has to be overcome. If crystal growth could produce enough retention for the brackets in the clinic, the advantages might be (1) easier and quicker debonding with little damage to the enamel surface, (2) minimal effect on the outer, fluoride-containing enamel surface, (3) no resin tags left behind in the enamel surface, and (4) possibility in

the future of incorporating fluoride in the crystal interface to provide an anticariogenic action. However, if it became necessary to make a choice between easier debonding and optimal strength, the latter would certainly be the choice of most clinicians. The time consumed in rebonding several loose brackets would probably far exceed the time saved during debonding. Besides, when removing adhesive remnants with careful use of selected tungsten carbide burs operated at speeds of around 25,000 rpm (and no water coolant so that adequate contrast to enamel is obtained), the risk of inducing iatrogenic damage is minimal. The technique is relatively quick, and the enamel surface may look smooth and sound, as shown in Fig. 3, B.
REFERENCES 1. Zachrisson BU: A posttreatment evaluation of direct bonding in orthodontics. AM J ORTHOD 71: 173-189, 1977. 2. Zachrisson BU, Brobalcken BO: Clinical comparison of direct versus indirect bonding with different bracket types and adhesives. AM J ORTHOD 74: 62-78, 1978. 3. Buzzitta VAJ, Hallgren SE, Powers JM: Bond strength of orthodontic direct-bonding cement-bracket systems as studied in vitro. AM J ORTHOD 81: 87-92, 1982. 4. Gwinnett AJ, Gorelick L: Microscopic evaluation of enamel after debonding. AM J ORTHOD 71: 651-665, 1977. 5. Pus MD, Way DC: Enamel loss due to orthodontic bonding with filled and unfilled resins using various clean-up techniques. AM J ORTHOD 77: 269-283, 1980. 6. Zachrisson BU, &tun J: Enamel surface appearance after various debonding techniques. AM J ORTHOD 75: 121-137, 1979. I. Zachrisson BU, Skogan 0, Hqymyhr S: Enamel cracks in debonded, debanded, and orthodontically untreated teeth. AM J ORTHOD 77: 307-319, 1980. 8. Brobakken BO, Zachrisson BU: Abrasive wear of bonding adhesives: Studies during treatment and after bracket removal. AM J ORTHOD 79: 134-147, 1981. 9. Alexander P, Young J, Bowman D: Bond strength of three orthodontic adhesives. AM J ORTHOD 79: 653-660, 1981. 10. Maijer R, Smith DC: Variables influencing the bond strength of metal orthodontic bracket bases. AM J ORTHOD 79: 20-34, 1981. 11. Reynolds IR, von Fraunhofer JA: Direct bonding of orthodontic attachments to the teeth: The relation of adhesive bond strength to gauze mesh size. Br J Orthod 3: 91-95, 1976. 12. Silverstone LM: The acid etch technique: In vitro studies with special reference to the enamel surface and the enamel-resin interface. In Silverstone LM, Dogan L (editors): Proceedings of an international symposium on the acid etch technique, St. Paul, 1975, North Central Publishing Company, pp. 13-39. 13. Diedrich P: Enamel alterations from bracket bonding and debonding: A study with the scanning electron microscope. AM J ORTHOD 79: 500-522, 1981. 14. Smith DC: A new dental cement. Br Dent J 125: 381-384.1968. 15. Smith DC, Cartz L: Crystalline interface formed by polyacrylic acid and tooth enamel. J Dent Res 52: 1155, 1973. 16. Maijer R, Smith DC: A new surface treatment for bonding. J Biomed Mater Res 13: 975-985, 1979. 17. Swartz M (Ormco/Sybron, 1322 South Lone Hill Ave., Glendora, CA 91740): Personal communication, 1983. 18. &tun J, Zachrisson BU: Improving handling properties of a

340

h-tun

and Bergland

Am. J. Orthod. April 1984

composite
1982.

resin

for direct

bonding.

AM J ORTHOD

81: 269-276,

22.

Fehr FR vd, Lij, H, Theilade Caries Res 4: 131-148. 1970.

E: Experimental

caries

in man.

19. Crystal Bond TM system: Fact sheet, Glendora, Calif., 1981, Ormco/Sybron 20. Dunn OJ: Basic statistics: A primer for the biomedical sciences, third printing, New York, 1967, John Wiley & Sons. 21. Tronstad L, Leidal TI: New instrument for finishing of embrasure margins of Class II cavities. J Am Dent Assoc 93: 94-97,
1976.

Reprint requests to: Dr. Jon Artun Department of Orthodontics University of Oslo Geitmyrsvn 71 Oslo 4, Norway