Comparison between standardizedperiapical and bitewing radiographsin assessing alveolar bone loss

Jasim M. Albandar, Dhia K. Abbas,

Marthe Waerhaug and Per GJermoDepartment of Periodontology and Institute otCommunity Dentistry, University ot Oslo, Oslo.Norway

Albandar JM, Abbas DK, Waerhaug M, Gjermo P: Comparison between standard-ized periapical and bitewing radiographs in assessing alveolar bone loss. CotnmunityDent Oral Epidemiol 1985; !3: 222-5.

Abstract - Readability and alveolar bone level were cotnpared in 455 sets ofstandardized radiographs each consisting of four periapical (PA) and two bitewing(BW) films. BW radiographs showed significantly higher proportions of unreadablesites in the canines and the mesial surface of first premolars (/'<0,0001). while PAradiographs exhibited significanlly higher proportions of unreadable sites in thedistal surface of second molars (/ '< 0.0001). Alveolar bone level was assessable in atotal of 10016 sites, and the comparison showed identical readings in 82% of thesites. In the tnaxilla, a significantly higher proportion of sites exhibited larger honeloss measurements (/»< 0.0001) and more sites with significantly higher mean values(/^< 0,005) in BW than in PA films. Mandibular sites showed the opposite trend.The majority of the sites revealed a Pearson's correlation coefficient (r) rangingbetween 0.60 and 0.82. No significant difference between mean bone loss measure-ments in the two types of radiographs for all the sites was found. It was concludedthat the two radiographic techniques may be used interchangeably in epidemiologicstudies.

Key words: bone level; epidemiology, oral:periodontal disease: radiography, oral

Jasim M, Albandar, Department ofPeriodontology. Dental Faculty. Geitmyrsvn,69. 0455 Oslo 4. Norway

Accepted for publication 10 December 1984

Dental radiographs are important in thedetection and assessment of the amountof periodontal tissue breakdown, a valu-able adjunct in treatment planning andin the determination of prognosis (1). Su-OMi. PLUMBO & BARBANO (2) could not

find a clearcut superiority to either radio-graphic or pocket measurements with re-gard to their ability to assess alveolarbone destruction. Comparison of radio-graphic and direct measurements in thealveolar crest region of dried mandibleshas shown that a dimensionally accurateradiographic image could be obtained(3).

Bitewing technique is characterized hyits simplicity and reduction of the radi-ation exposure since fewer films areneeded than when the periapical tech-nique is used. Many dentists take twobitewing films routinely for patients atrecall visits (4), Such radiographs haveformed the material ofa number of epid-emiologic and clinical studies which havedealt with the assessment of bone loss(5-8). Few studies have been conductedto compare alveolar bone level measure-ments assessed on periapical and bite-

wing radiographs (9, 10). REEt> & POLSON(9) concluded Ihat the two methods couldnot be used interchangeably on an indi-vidual basis.

The aim of this study was to compareperiapical and bitewing radiographs withregard to readability and bone lossmeastiretnents.

Material and methods

The material of the present study con-sisted of four periapical (PA) and twobitewing (BW) radiographs from the lat-eral Jaw segments of each of 455 subjectsranging in age from 17 to 68 yr (mean46.3).

Kodak Ultraspeed intraoral films No.2 were used and the exposure time was0.5 s, A mobile X-ray unit with long cone(Siemens Heliodent. 50 kVp. 7 mA) wasemployed. The periapical radiographswere taken with an Eggen film holder forintraoral standardized paralleling tech-nique (11), and bitewing radiographswith the Kwikbite film holder. The radio-graphs were developed under standard-ized conditions using an automatic roent-

gen film processor (Durr Periomat).Each radiograph was magnified 10

times using an ADA-Realist radiographviewer (12), A transparent millimeter pa-per was magnified at the same time anda drawing compass was employed tomeasure the distance from the ccmento-enamel junction (CEJ) to the alveolarcrest (AC). The most apical margin of ametal crown or proximal restoration wasused as a reference point if CEJ couldnot be identified. A distance exceeding 2mm was considered to represent boneloss. The measurements were made to thenearest millimeter and the area from themesial surface of the canine to the distalof the second molar in each quadrantwas scored. If bone loss could not bemeasured according to the criteria de-scribed above, the site was regarded asunreadable.

The proportion of unreadable siteswas calculated by dividing the number ofunreadable sites by the total number ofsites. For the purpose of comparingbone loss measurements, a site which wasunreadable in one or both filtii types wasexcluded. A minimum of 2 weeks elapsed

Bone loss in periapical and bitewing radiographs 223

between examining periapical and bite-wing radiographs of the stime subject. Allmeasurements were performed by oneexaminer.

Analyses of data - The differences be-tween the proportions of unreadable sitesin the two techniques were tested by thetest of difference between proportions(13). This test was also used to testwhether the proportions of sites showingthe greatest bone loss measurements inthe two techniques was different. Thetnean values of the two techniques foreach site were compared using the paired/-test. A low a-level (/*< 0.005) was cho-sen due to the high number of sites whichwere included in each test.

The relationship betw^n each pair ofmeasurements was calculated by meansof Pearson's correlation coefficient (r).

Intra-examiner reliability - PA andBW radiographs from 50 randomlyselected subjects were examined for boneloss twice by the same examiner with aninterval of 2 wk. The mean bone lossvalue for each subject was calculated andthe association between the means of thefirst and the second readings was com-puted using Pearson's correlation. Theresulting correlation coefficient (r) for PAand BW radiographs was 0.987 and0.989. respectively. When individual siteswere compared the result showed that in95.4% of Ihe sites the two measurementswere identical.

Results

A total of 5406 sites were excluded dueto unreadability in one or both radio-graphic techniques. They constituted35% of the total sample examined. Dueto the high unreadability in the mesialsurfaces of canines, these surfaces wereexcluded from the comparisons of boneloss measurements, thus a total of 10016pairs of surfaces could be read; the meanfor each of the 36 sites (from the distalsurface of canines to the distal of theseeond molars) was 278.

Unreadability was most frequently en-countered in canines and the mesial sur-face of first premolars where BW radio-graphs showed significantly higher pro-portions of unreadable sites than did PAfilms (P < 0.0001). PA radiographsshowed significantly greater unreadabil-ity in the distal surface of the secondmolars (/'<0.000!) than the rest ofthesites where the differences in unreadabil-

ity between the two techniques appearedto be less demarcated (Fig. I).

The results showed that in 82% of thesites bone loss measurements were iden-tical in both techniques (Fig. 2). Thenumber of sites showing differencesgreater than 2 mm was small. In themaxilla, more sites exhibited larger CEJ-AC distance on BW than on PA films.

1G0^ N-15522 • PA

Fig. I. Distribution of unreadable sites in per-iapical (PA) and bitewing (BW) radiographsaccording to site.

lOQN-IDOIE

0.3 U

•2mm 2mtn 1mm

PA>BW

Imm 2 mm >2mm

PA<BWFig. 2. Comparison between bone lossmeasurements in periapical (PA) and bitewing(BW) radiographs showing frequencies of dif-ferences in millimeters (mm).

whereas the opposite was observed formandibular teeth (Fig. 3). and the differ-ence between the two measurements wassignificant on both occasions( f < 0.0001).

The distribution of mean bone lossmeasurements in both techniques accord-ing to site is shown in Table 1. Of the ninemaxillary sites, five showed significantlyhigher values in BW than in PA films,while the difference was not statisticallysignificant in three of the sites. In themandible, three sites displayed sig-nificantly higher mean values in PA thanin BW (Table I).

The majority of correlations betweenthe two bone loss measurements showedr-value.s ranging between 0.60 and 0.82(Table 2). Mesial surfaces of first pre-molars and second mandibular pre-

100

50

MAXILLA

MANDIBLE

Fig.

PA - BW PA < BW PA ^ BW

3. Comparison between bone lossmeasurements in pcriapicat (PA) and bitewing(BW) radiographs in maxillary and mandibu-lar sites.

Table I. Mean alveolar bone loss (mm) in periapical (PA) and bitewing (BW) radiographsaccording to site

MaxillaSecond molarsFirst molarsSecond premolarsFirst premoiarsCanines

MandibleSecond molarsFirst molarsSecond premolarsFirst premolarsCanines

PA

0.360.260.190.15

0.160.160.240.33

MesialBW

0.380.330.290.23

0.21O.IO0.130.08

NS*

*

NS•

PA

0.390.260.240.300.25

0.190.180.380.340.27

DistalBW

0.320.370.300.340.31

0.160.180.350.320..10

*

NSNS

NSNSNSNSNS

NS, not significant. * / '<0.005. •* / '<0.0005.

224 ALBANDAR F.T AL.

Table 2. Relationship between bone lossmeasurements in periapicai and bitewingradiographs on each site as computed by Pear-son's correlation coefficient (r)

with our results in which such compari-son of measurements in mandibular sitesresulted in identical values in 84.5% ofthe comparisons (Fig. 3).

The distribution of the differences be-tween PA and BW measurements did notseem to be different on the left and theright sides of the jaws which permittedpooling ofthe results from the two sides.Out of the 18 sites which were tested,paired /-tests revealed that 50% of thesites did not have significantly differentmeasurements on PA and BW radio-graphs when the mean value for eachsite was used (Table 1). Although a /-test showed significant differences in the

Table 3. Mean bone loss (mm) in periapicai (PA) and bitewing (BW) radiographs and relationshipbetween bone loss measurements Tor all sites

Bone loss (tnm)PA BW

n Mean (SD) mean (SD) P r

MaxillaSecond molarsFirst molarsSecond prcmolarsFirst premolarsCanines

MandibleSecond molarsFirst molarsSecond premolarsFirst premolarsCanines

Mesial

0.670.630.680.46

0.710.600.490.25

Distal

0.690.650.800.750.61

0.650.610.820.700.62

AU sites 10016 0.26(0.69) 0.27(0.67) NS 0.67

P, Probability. NS, not significant, r, Pearson's correlation coeflicieni.

molars exhibited the lowest values. Themean bone loss values for PA and BWradiographs for the total of 10016 siteswere 0.26 and 0.27 mm, respectively (Ta-ble 3). The difference between the twovalues was not statistically significant.

Discussion

PA radiographs exhibited higher read-ability than BW films in the area of ca-nines and the mesial surface of first pre-molars. and lower readability at the dis-tal surface of the second molars (Fig.I). Because of the limited ability of BWradiographs to cover extensive bone loss,the difference between the readability ofthe two techniques may be expected tobe more evident in populations with ad-vanced periodontal disease.

The site by site comparisons of boneloss measurements that were made onPA and BW radiographs resulted in ahigh degree of agreement between thetwo values (Figs. 2 and 3). HAKANXSON.

BJORN & JoNSSON (14) employed a rela-tive method for assessing alveolar boneheight on BW radiographs in the man-dible. Using such a device their resultsshowed agreement of measurementsmade on PA and BW radiographs in85% of the cases. This is in agreement

other 50% of sites, the clinical impor-tance of such differences is small. Onlyone site showed a difference of more than0.11 mm. REED & POLSON (9) compared

the alveolar bone levels in PA and BWradiographs and noticed that in 50% ofthe sites no significant differences couldbe detected, but they reported larger dif-ferences between the two measurementsin the other 50% which ranged between0.2 and 0.6 mm. The use of a wholemillimeter as the measuring unit for as-sessing bone loss in the present studymay have contributed to the disparitywith the results of REED & POLSON (9)

who used 1/5.9 mm as a unit.Radiographs have been shown gener-

ally to underestimate alveolar bone loss(3. 15). Thus, a difference between corre-sponding bone loss measurements on two

radiographs would mean larger under-estimation o( bone loss on the type ofradiograph exhibiting the shortest CEJ-AC distance (16). In the maxilla our datashowed that BW radiographs reproducethe alveolar bone level better than PAradiographs (Fig, 3 and Table I). This isin accordance with REF.D & POLSON (9)

who found that BW radiographs showedsignificantly higher mean alveolar bonelevels than PA radiographs in 9 of 10sites. A similar result was also shown bySEWERIN (10). Six out of the nine man-

dibular sites showed no significant differ-ence, whereas the rest showed sig-nificantly higher means of bone lossmeasurements in PA than in BW radio-graphs (Table 1).

The low bone loss values in our ma-terial may be ascribed partly to the utiliz-ation or proximal filling margins as acriterion instead of CEJ when the lattercould not be used. However, this has noinfluence on the comparison of the tworadiographic techniques, which was theaim of the study.

When site is the observational unit,interchangeable use of bone lossmeasurements assessed on PA and BWradiographs is possible only if the Pear-son's correlation coefficient (r) betweenthe two measurements approaches I.O(17). In clinical studies, sites are oftenemployed as a basis for comparison, andour data indicated that in such trials thetwo techniques cannot be used inter-changeably (Table 2) because of the rela-tively low correlations at individual sites,However, the main objective of epidemi-ologic studies is to provide an estimate ofbone loss by aggregating the individualscores in terms of group means. The re-sult from the present study revealed nosignificant difference between mean boneloss values from the two techniques (Ta-ble 3). and it is concluded that both PAand BW radiographs may be used inter-changeably in epidemiologic studies.

References1. LANG N P . IIUX RW. Radiographs in periodontics. J Clin Periodotuol 1977; 4: 16-28.2. SuoMi JD, PLUMBO J. BARBANO JP. A comparative study of radiographs and pocket measure-

ments in periodontal disease evaluation. J Periotioniol 1968: J9: 311 5,3. STONER J E . An investigation into the accuracy of measurements made on radiographs of

the alveolar crests of dried mandibles. J Feriodoniol 1972; 43: 699-701.4. MATTESON S R , MORRISON W S . STANEK E J . PHILLIPS C . A survey of radiographs obtained

at the initial dental examination and patient selection criteria for bitewings at recall. J AmDem Assoc 1983; 107: 586-90.

5. STONER JE. The value ofthe bitewing radiograph in the diagnosis of dental disease in youngsubjects. J Dent Child 1974; 41: 190-3.

Bone loss in periapicai and bitewing radiographs 225

6. SELiKOWrrz H-S. SHI:IHAM A, ALBERT D. WILLIAMS JM. Retrospective longitudinal studyof the rate of alveolar bone loss in humans using bite-wing radiographs. J Clin Perintlonlol1981; 5.'431-8.

7. GJERMO P. BELLLNI H T , SANTOS VP, MARTINS J G , FERRACYOLI JR. Prevalence of bone loss

in a group of Brazilian teenagers assessed on bite-wing radiographs. J Clin Periodontol 1984;/ / . 104-13.

8. HANSEN BF. GJERMO P, BERGWITZ-LARSEN K R . Periodontal bone loss in 15-year-oldNorwegians. J Clin Periodontol 1984; //.• 125-31.

9. REED BE, POLSON AM. Relationships between bitewing and periapicai radiographs inassessing crestal alveolar bone levels. J Periodontol 1984; 55: 22-7.

10. SEWERIN I. SamnienJigning af marginale parodontale knogletabsmSl afljest p4 pcriapikaleog bitewing-optagetser. Timdlaegebladei 1983; 87: 733-1.

11. EGGEN S. Standardized projection technique in intraoral roentgenography. Sver Tandldkar-forb Tid 1969; 61: 867-72.

12. ALBANDAR JM, GJERMO P. Associations between gingivitis, pocket depths, and radiographicbone loss in inlerproximal areas in young adults. Scand J Dcni Re.s 1983; 9}: 371-5.

13. HILL AH. Principles of medical stalisiics. London; The Lancet. 1971; 133-42.14. HAKANSSON J . BJORN A - L , JONSSON B G . Assessment of the proximal periodontal bone

height from radiographs with partial reproduction of the teeth. Swed Dem J 1981: 5.- 147-53.15. THEILADE J. An evaluation of the reliability of radiographs in the measurement of bone loss

in periodontal disease. J Periodontol 1960; 31: 143-53.16. SEWERIN L Undervurdcring af marginalt parodontali knogleiab pS periapikale og bitewing-

optagelser. Tandhegebladet 1983; 87: 779-83.17. NUNNALLY JC. Psychometric theory. New York; McGraw-Hill, 1978; 245-6.