utility of Ramfjord index teeth to assess periodontal disease progression in longitudinal studies

Rams TE, Oler J, Listgarten MA and Slots J: Utility of Ramfjord index teeth to assess periodontal disease progres.sion in longitudinal studies. J Clin Periodontol 1993; 20: 147-150. Munksgaard 1993, Abstraet. The feasibility of using the Ramfjord index teeth to estimate wholemouth periodontal disease activity was investigated. Whole-mouth examinations were carried out semi-annaally over a 36-month period in 98 maintenance patients previously treated for adult periodontitis. Recurrent periodontitis was defined as either a 3-mm or greater probing depth increase from baseline, or a 2-mm or greater probing depth increase together with 2-mm or greater of relative attachment loss measured from a reference stent. Whole-mouth disease activity was compared to Ramfjord index teeth data, with and without adjustment. Adjustment was made by multiplying disease activity rates on Ramfjord index teeth by the ratio formed from the sum of all teeth present over the sum of all Ramfjord index teeih in the study population. Without adjustment, Ramfjord index teeth markedly underestimated subjects with recurrent disease activity when compared to whole-mouth findings. However, with adjustment, the hypothesis that upper and lower limits on whole-mouth incidence of recurrent periodontitis could be estimated from Ramfjord index teeth disease-activity rates were not rejected (P>0,20, z-test) at any 6-month interval. At 5 of 6 examinations, the % of disease active subjects as detennined from whole-mouth evaluations was below the upper limit for disease incidence calculated, with 95% confidence, from point estimates derived from adjusted Ramfjord index teeth data. Partial-mouth examinations with appropriate adjustment of Ramfjord index teeth data may be useful for assessing periodontal disease progression in longitudinal population studies of human periodontitis.

Thomas E. S Jacqueline Oler^, Max A. Listgarten^ and Jergen Siots^

^Department of Periodontics, University of Pennsylvania Schooi of Dentai Medicine, Philadeiphia, PA USA; 'Department of Quantitative Methods, Drexei University, Philadeiphia, PA USA; ^Department of Periodontoiogy, University of Southern Caiifornia School of Dentistry, Los Angeles, CA USA

Key words: Ramfjord index teeth; periodontai diagnosis; periodontitis disease activity; partiai recording; Bonferroni inequality; epidemioiogic and longitudinal studies. Accepted for publication 27 February 1992

Ramfjord (1959) designated 6 index teeth for epidemioiogic studies of human periodontai diseases. Teeth selected were the maxillary right and mandibular left first molars, maxillary left and mandibular right first premolars, and maxillary left and mandibular right central incisors. Cross-sectional studies have found a good correlation between mean values from the Ramfjord index teeth and whole-mouth scores for dental plaque (Alexander 1970, Mills et al. 1975, Gettinger et al, 1983, Goldberg et al, f985, Silness & Roynstrand 1988), supragingival and subgingival caiculus (Alexander 1970, Gettinger et al, 1983, Ainamo & Ainamo 1985), gingival inflammation (Alexander 1970, Downer 1972, Chiiton et al. 1978, Gettinger et ai. 1983, Goldberg etal, 1985, Siiness &

Roynstrand 1988, Ainamo & Ainamo 1985), radiographic bone loss (Berg et al. 1984), probing depth (Downer 1972, Mills et al. 1975, Berg et al, 1984, Ainamo & Ainamo 1985, Silness & Roynstrand 1988), chnicaliy determined attachment loss (Gettinger et al. 1983), and the periodontai disease index (Jamison 1963). The usefulness of the Ramfjord index teeth in longitudinal population studies remains to be delineated. The reliability of utilizing the Ramfjord index teeth for assessing the rate of periodontitis progression in longitudinal studies must be resolved, particularly since this progression may occur episodically and in a site-specific fashion (Socransky et al. 1984). The present study assessed the diagnostic value of employing the

Ramfjord index teeth as estimators of whole-mouth incidence of recurrent periodontitis in an adult population on maintenance care.
Material and Methods Patients

The study population, the experimentai design, examination techniques, and baseline clinicai and microbiological findings are described in detail elsewhere (Listgarten et al, 1989). In brief, 98 adults previously treated for moderate to advanced periodontitis were maintained on a 3-month recall program. This report analyzes clinical data on site-specific disease recurrence detected semi-annuaily over a 36-month period.

148

Rams et al. where: Pa the observed percentage of subjects with recurrent periodontitis on at least one Ramfjord index tooth, r =the ratio of the sum of all teeth present over the sum of all Ramfjord index teeth in the examined population, recalculated for each 6-month period. In this study, TI, the observed % of subjects with recurrent disease activity over a given 6-month period on at least one tooth in the entire dentition, was available from whole-mouth examinations. Each J T was examined to determine if it supported the hypothesis that a 95% confidence upper limit for wholemouth disease incidence {n) can be calculated from Ramfjord index teeth data. An upper bound for TZ {n^) was obtained in an application of the first Bonferroni inequality (Morrison 1983), which established that the probability for a union of K random events, V/i,, K / ^ A : , is bounded about by S P U , ) , The set of disease-active subjects was represented as the union of (not necessarily disjoint subsets oO subjects identified as disease-active based on Ramfjord index teeth data (A^) or nonRamijord index teeth data (^R). From the first Bonferroni inequality, an upper limit for disease incidence is given by: tivity rates, were calculated as:

Clinical examinations

2 clinical examiners were cahbrated at the beginning of the study and thereafter once a year. Probing depth. A Hu-Friedy (Cat. =^LL20, Hu-Friedy Co., Chicago, IL) probe calibrated in mm with specia! markings at 5, 10. \5 and 20 mm and a probing force of approximately 50 ponds was utilized. The length of the probe tip at the bottom of the pocket to the gingival margin was recorded to the nearest mm. At each tooth, measurements were taken from the mesial, distal, vestibular (buccal, labial), and mid-oral (lingual, palatal) tooth surfaces. At each tooth surface, the deepest site constituted the recorded probing depth. Relative attachment level. Relative attachment level was determined at the same time as probing depth as the distance from the probe tip to a fixed reference consisting of a trimmed thermoplastic occlusal stem. The same sites as described above were examined.

where: /7 = number of subjects examined for disease-activity at Ramfjord index teeth (number of subjects on which the disease-activity incidence estimate was based). As in any application of the central limit theorem, the sampling error, or margin for error estimation, is inversely proportional to the sample size, and decreases as the study population size increases. Results

Table 1 describes the percentage of subjects with recurrent periodontitis identified by the Ramfjord index teeth (/?|O and whole-mouth {iz) evaluations at semi-annual intervals through 36 months, Ramfjord index teeth by themselves without adjustment (/JR, column 3) underestimated subjects with recurrent disease activity when compared Recurrent periodontitis with whole-mouth examination findings (Table 1, column 3 versus column 8). A Recurrent disease activity was defmed considerably improved correlation was as any site, identified by the 2 indepenseen by adjusting disease-activity rates dent examiners, which exhibited either on Ramfjord index teeth by r (Column a probing depth increase of 3 mm or 5), the ratio of total teeth to Ramfjord more from baseiine, or a probing depth index teeth in the population. Interestincrease of 2 mm or more together with A point estimate for the right-hand a loss of clinical attachment of 2 mm or side of the equation was made using the ingly, the ratio r remained constant over time (Table 1), due to a generally equal more. adjusted Ramijord index teeth disease- temporal distribution of tooth loss beactivity rate, P^^, p^ x r. which assumes tween Ramfjord index teeth and nonequal likehhoods of disease activity in Ramfjord index teeth over the 36Data Analysis equal sized subsets of Ramfjord and month period (data not provided). The hypothesis that the whole-mouth non-Ramijord index teeth. rate of recurrent periodontitis in a At 5 of 6 semi-annual examinations, The unadjusted Ramfjord index teeth population (n) can be inferred from re- disease-activity rate (p^j provides a the %) of disease-active subjects as obcurrent periodontitis incidence on the poor estimate of the lower limit on served in whole-mouth evaluations (ii) Ramfjord index teeth (j.?^) in a sample whole-mouth disease-activity incidence fell below 95% confidence upper limit from that population was examined as (;rL). since subjects disease-active at estimates constructed from point estifollows. Ramfjord index teeth constitute a sub- mates (PR) based on adjusted Ramfjord Subjects were categorized YES/NO set of subjects disease-active anywhere index teeth data. At 6 months, the null in the mouth. An improved estimator hypothesis that whole-mouth periofor each 6-month period with respect to recurrent periodontitis activity on (i) of 71, would require the probabihty of dontitis-activity (;;) was less than or one or more of the Ramfjord index the joint event that a subject is disease- equa! to 20.2 (Row 1, Column 7) couid teeth, and (ii) one or more of all teeth active at both Ramfjord and non- not be rejected (P>0,20, z-test) with the Ramfjord index teeth, which would not observed ji = 23.5. present. For each 6-month period, disbe available from partial-mouth examease activity rates on Ramfjord index inations involving only Ramfjord index teeth were adjusted the ratio r, formed Discussion from the sum of all teeth present over teeth. the sum of all Ramijord index teeth in From the central limit theorem, The present study examined the possithe study population, bounds below which the whole-mouth bility of using partial-mouth examinA point estimate (P^) of ^n upper disease-activity incidence would be ex- ations to detect recurrent periodontitis limit for n was given by the formula: pected to occur in the target population, in adult patients on trimonthly maintenestimated with 95% confidence from ad- ance care. Clinical measurements over justed Ramfjord index teeth disease-ac- time were made on Ramfjord index

Ramfjord teeth and disease activity

Tahle J. Whok-mouth versus Ramfjord index teeth analysis of recurrent periodontitis for subjects by posl-baseline examinalion period Without adjustment of Ramfjord index leeth data No, (%) of subjects exhibiting recurrent periodonlitis Based on No, of Ramijord teeth Based on all subjects (unadjusted) (/)) teeth (ii) Point estimate (-Pw) 14,4 28,3 14,9 25,5 15,8 18,1 Without adjustment of Ramfjord index leeth data^'

149

Exam 6 months 12 months 18 months 24 months 30 months 36 months

r 4,64 4,64 4,64 4,64 4,64 4,64

95'Mi confidence Observed upper limit on whole-mouth whole-mouth incidence incidence (n._^) ill) 20.2 35,8 21,0 33,0 22,2 25,3 23,5 !6,3 10,8 16.5 17.1 19.4

Predictive efficacy of adjusied Ramfjord teeth point estimate (P\<j) of upper limit for whole-mouth disease activity [it) h exceeds 95% upper bound*" n below 95% upper bound Tl beiow 95% upper bound 7 c below 95% upper bound n below 95%> upper bound n below 95% upper bound

98 98 93 91
89 11

3(3,1) 6(6,1) 3 (3,2) 5(5,5) 3 (3,4) 3(3,9)

23 (23,5) 16(16,31 10(10,8) 15(16,5) 15(17,1) 15(19,4)

r = ratio of lotal teeth to Ramfjord teeth in study subjects, n = "/u of subjects with disease activily as observed from whole-mo nth examinations, p^^.^ point estimate of upper limit of Ji-based on adjusted Ramfjord index teeth data, "' Disease activity rates on Ramfjord index teeth adjusted by multiplying by r. "' Do not reject //: :Jr^20.2 (F>0,20, r-test).

teeth and the entire dentition, and took into account the proportional representation of index teeth among all teeth present in the patient population. It must be pointed out that the reported correlations between partial and wholemouth disease activity are valid only on a population level, and are not applicable in the clinical management of individual patients. The present study findings agree with Fleiss et al, (1987) that Ramfjord itidex teeth data by themselves markedly underestimate the whole-mouth incidence of periodontitis progression. However. Fleiss et al, (1987) did not assess changes in ciinical attachment level over time, but rather, evaluated the oceurrenee of moderate to deep periodontal pocketing by recording only the deepest probing depth per looth per examination in patients with relatively healthy periodontal conditions. Improved correlations were found by adjusting the Ramijord index teeth data to provide an estimation of an upper limit for whole-mouth disease subject incidence for recurrent periodontitis. At each of the 6 semi-annual examinations, the observed whole-mouth recurrent periodontitis rate supported the hypothesis that 95% confidence upper bounds on disease-activity incidence can be projected frorn appropriate adjustment of Ramfjord index teeth data. It is important to emphasize that the adjustment employed on the Ramfjord index teeth data provided otily upper limit whole-mouth incidence level estimates, in accordance with the ftrst Bonferroni inequality (Morrison 1983),

sinee disease activity rates at nonRamfjord index teeth is not known or observed in the partial mouth examinations. Disease activity on Ramfjord index teeth by themselves without adjustment provide a crude lower limit on whole-mouth disease incidetice. More precise lower limit estimates, following the second Bonferroni inequality (Morrison 1983), requires the hkelihood of the joint event to be determined, i.e., that a subject experienced disease activity both at one or more Ramfjord index teeth and at one or more non-Ramfjord index teeth. No attempt was made in the present study, as previously suggested (Mills et al. 1975, Fleiss et al, 1987), to improve efficacy of partial-mouth scoring by substituting other teeth if one or more Ramfjord teeth was missing in specific patients. Since disease activity was assumed to be equally distributed within dentitions and exhibiting the same pattern among all patients, the usefulness of the proposed partial examination technique may be limited for some patient groups, such as persons wilh localized juvenile periodontitis or exhibiting severe tooth loss. However, for adult periodontitis, as indicated in the present study at least on a group basis, use of adjusted Ramijord teeth data provides 95% upper limit estimations on the incidence of whole-rnouth periodontitis disease-activity. Examiner fatigue as well as the extended time required for full-mouth, site-specific clinical measurements hamper assessment of periodontitis disease activity in epidemioiogic studies and

large-scale clinical trials. An advantage of the Ramfjord index leeth is that examinations can be carried out faster and more econornically than evaluation of the entire dentition. Partial-mouth scoring in conjunction with newly-developed electronic periodonlal probes (Jeffcoat et al, 1986, Birek et al, i987. Gibbs et al, 1988) are likely to facilitate studies of periodontitis disease activity involving large numbers of subjects. Clinical data needed to utilize the proposed method would be (i) the disease-activity rate on Ramfjord index teeth, (ii) the number of Ramfjord index teeth, and (iii) the total number of all teeth in the study population. The relatively large 95% estimation errors observed for the calculated upper limit on periodontitis disease-aetivity incidence suggests the method is most suitable where large sample sizes are involved. In summary, parliai-mouth examinations and appropriate adjustment of Ramfjord index teeth data appear useful in estimating, wilh 95%. confidence, the upper limit of whole-mouth periodontitis activity in longitudinal population studies. Further evaluation of these and other partial-mouth assessment techniques is indicated.

Acknowledgments This study was supported by Grants RO1-DE06085 and RR 01224/00040 from the National Institute of Dental Research, National Institutes of Heaith, Bethesda, MD, USA.

150

Rams et al.
soit une augmentation de ia profondeur au sondage de 3 mm ou plus depuis le debut de l'etude, soit une augmentation de 2 mm ou pius de la profondeur au sondage en meme temps qu'une augmentation de 2 mm ou plus de la perte d'attache par rapport a une gouttiere occlusale de reference, L'activite de la maladie dans l'ensemble de la bouche a ete comparee aux donnees concernant les dents iiidiees de Ramfjord, avec et sans ajustement. Cet ajustement consistait a multiplier les taux d'activite de la tnaladie au niveau des dents indices de RamQord par le rapport: somme de toutes ies dents presentes sur somme de toutes les dents indices de Ramfjord dans la population etudiee. En l'absence d'ajustement. les dents indices de Ramfjord sousestimaient fortement les sujets ayant une recidive de la maladie par comparaison avec les donnees concernant l'ensemble de la bouche, Cependant. avee ajustement, on n'a a aycun des intervailes de 6 mois rejete l'hypothese que les limites superieures et inferieures de l'incidence des recidives de la parodontite pour l'ensemble de la bouche pouvaient etre estimees en se basant sur les taux d'activite des denis indices de Ramfjord (P>0.20, test r). Lors de 5 des 6 examens. la proportion des sujets presentant une maladie en activite, a en juger par les evaluations de l'ensemble de la bouche, etait en dessous de la limite superieure pour l'incidence, calculee. a un niveau de confiance de 95%, a partir d'estimations de points derivees des donnees ajustees pour les dents indices de Ramfjord, Les examens buccaux partiels avec ajustement adequat des donnees sur les dents indices de Ramijord peuvent etre uliles pour etablir la progression de la maladie parodontale dans les etudes longitudinales de population concernant la parodontite humaine. Fleiss, J, L,. Park, M. H., Chilton, N, W., Alman, J, E., Feldman, R. S, & Chauncey, H, H, (1987) Representativeness of the "Ramfjord teeth" for epidemiologic studies of gingivitis and periodontitis. Community Dentistry and Oral Epidemiology 15, 221-224. Gettinger. G., Patters, M. R,, Testa, M, A., Loe, H,, Anerud, A., Boysen, H. & Robertson, P, B, (1983) The use of six selected teeth in population measures of periodontal status. Journal of Periodoniologv 54, 155-159. Gibbs, C, H,, Hirschfeld, J. W., Lee, J, G,. Low. S, B,, Magnusson, 1,, Thousand, R, R,. Yemeni, P & Clark, W. B, (1988) Description and clinical evaluation of a new computerized periodontal probe - the Elorida probe. Journal of Clinica! Periodonlology 15. 137-144. Goldberg, P.. Matsson. L, & Anderson. H, (1985) Partial recording of gingivitis and dental plaque in children of different ages and in young adults. Community Dentistry and Oral Epidemiology 13, 44-46, Jamison, H, C. (1963) Some comparisons of two methods of assessing periodontal disease. American Journal of Public Health 53, 1102-1106. Jeffcoat, M, K., Jeffcoat. R, L,, Jens, S, C, & Captain, K. (1986) A new periodontai probe with automated cemento-enamel junction detection. Journal of Clinical Periodontology 13, 276-280, Listgarten, M, A,. Slots, J.. Rosenberg, J,, Nitkin, L,, Sullivan, P & OJer. J, (1989) Clinical and microbiological characteristics of treated periodontitis patients on maintenance. Journal of Peridontologv 60, 452-459, Mills, W, H., Thompson, G, W, & Beagrie, G, S, (1975) Partial-mouth recording of plaque and periodontal pockets. Journal of Periodontal Re.'ieareh 10, 3643, Morrison, D, E, (1983) Applied linear statistical models. Prentice-Hall, New Jersey. Ramfjord, S. P. (1959) Indices for prevalence and incidence of periodontal disease. Journal of Periodontology 30, 51-59, Siiness, J, & Roynstrand, T, (1988) Partial mouth recording of plaque, gingivitis and probing depth in adolescents. Journal of Clinical Periodontology 15, 189-192, Socransky, S, S,, Haffajee, A. D,, Goodson, J, M, & Lindhe, J, (1984) New concepts of destructive periodonta] disease. Journal of Clinical Periodontology 11, 21-32,

Zusammenfassung
Der Nutzen der Ramfjord hidex-Zahne in Langzeitstiidien zur Bestimmung der Progression einer Parodontalerkrankung Die Brauchbarkeit der Rarafjord Iiidex-Zahne zur EJnschatzung der parodontalen Erkrankungsaktivitat des gesamlen Gebisses wurde untersuchl. Bei 98 Recall-Patienten, die vorher wegen einer Erwachsenenparodontilis behandelt wurden, hai man fur eine Periode von 36 Monaten halbjahrige Unlersuchungen des gesamten Gebisses durchgefuhrt. Wiederkehreiide Parodontitis wurde entweder ais eine 3 oder mehr mm groBere Sondierungstiefe im Vergleich zur Ausgangsuntersuchung oder eine 2 oder mehr mm groBere Sondierungstiefe zusammen mit einem 2 oder mehr mm groBeren Attachmentverlust im Vergleich zu einer Referenzschiene definiert. Die Erkrankungsaktivitiit des gesamten Gebisses wurde mit den angepaBten oder nicht angepaBten Daten aus den Ramijord Index-Zahnen vergliehen. Die Anpassung erfolgte iiber eine Multiphkation der Erkrankungsaktivitiitsraten der Ramfjord IndexZahne mit dem Verhaltnis aus der Summe alier vorhandenen Zahne zur Summe aller Ramfjord Index-Zahne in dieser Studienpopulation, Im Vergieich mit den Ergebnissen des gesamten Gebisses unterschatzten die Ramfjord Index-Zahne ohne Anpassung die Personen mit wiederkehrender Erkrankungsaktivitat betrachtlich, Mit Anpassung jedocii wurde die Hypothese, daB die oberen und unteren Grenzen der Inzidenz einer wiederkehrenden Parodontitis durch die Erkrankungsaktivitatsraten der Ramfjord IndexZahne eingeschatzt werden kann zu jedem Sechsmonats-Intervall. nicht verworfen {p> 0.20, r-Test), In 5 der 6 Untersuchungen lag der Prozentsatz der erkrankungsaktiven Personen bei Bestimmung iiber die Evaluation des gesamten Gebisses unter der oberen Grenze der Erkrankungsinzidenz. die anhand der angepaBten Daten der Ramfjord IndexZahne mit 95% Konfidenz berechnet woirde. Untersuchungen eines Teils des Gebisses konnten bei geeigneter Anpassung der Daten der Ramfjord Index-Zahne zur Messung der Progression einer Parodontalerkrankung in Langzeitstudien der menschlichen Parodontitis nijtzlich sein.

References
Ainamo, J.& Ainamo, A. {19S5) Partial indices as indicators of the severity and prevalence of periodontal disease. International Dental Journal 35, 322-326, Alexander, A, G, (1970) Partial mouth recording of gingivitis, plaque and calculus in epidemiological surveys. Journal of Periodontal Research 5, 141-147, Berg, L.. Becker, W. & Becker, B, E, (1984) The use of index teeth to predict the heahh status of the balance of the mouth. International Journal of Periodontics and Restorative Dentistry 4, 47-53, Birek. P, McCulloch. C, A, G, & Hardy, V. (1987) Gingival attachment level measurements with an automated periodontal probe. Journal of Clinical Periodontology 14, 472-477. Chilton, N, W., Eertig, J, W, & Talbott, K. (3978) Partial and fuil mouth recording of gingivitis scores. Pharmacology and Therapeutics in Dentistry 3, 39-44, Downer, M, C, (1972) The relative efficiences of some periodontal partial recording selections. Journal of Periodontal Re.'ieareh 7, 334-340,

Resume
Vliliie des dents indices de Ramfjord pour evaluer la progression de la maladie parodontale dam les etudes longitudinales La faisabilite de I'emploi des dents indices de Ramfjord pour estimer I'activite de la maladie parodontaie de l'ensemble de la bouche a ete etudiee, Des examens de toute ia bouche ont ete pratiques 2 fois par an pendant une periode de 36 mois chez 9S patients traites anterieurement pour parodontite de j'adulte et actueliement en phase de maintenance. On a defini comme recidive de ia parodontite

Address: Thomas E. Rams Department of Periodontics University of Pennsylvania School of Dental Medicine 4001 Spruge Street Philadelphia. PA ]9!04 USA