utility of Ramfjord index teeth toassess periodontal diseaseprogression in longitudinalstudies

Thomas E. SJacqueline Oler̂ ,Max A. Listgarten^and Jergen Siots^^Department of Periodontics, University ofPennsylvania Schooi of Dentai Medicine,Philadeiphia, PA USA; 'Department ofQuantitative Methods, Drexei University,Philadeiphia, PA USA; ^Department ofPeriodontoiogy, University of SouthernCaiifornia School of Dentistry, Los Angeles,CA USA

Rams TE, Oler J, Listgarten MA and Slots J: Utility of Ramfjord index teeth toassess periodontal disease progres.sion in longitudinal studies. J Clin Periodontol1993; 20: 147-150. © Munksgaard 1993,

Abstraet. The feasibility of using the Ramfjord index teeth to estimate whole-mouth periodontal disease activity was investigated. Whole-mouth examinationswere carried out semi-annaally over a 36-month period in 98 maintenance patientspreviously treated for adult periodontitis. Recurrent periodontitis was definedas either a 3-mm or greater probing depth increase from baseline, or a 2-mm orgreater probing depth increase together with 2-mm or greater of relative attach-ment loss measured from a reference stent. Whole-mouth disease activity wascompared to Ramfjord index teeth data, with and without adjustment. Adjustmentwas made by multiplying disease activity rates on Ramfjord index teeth by theratio formed from the sum of all teeth present over the sum of all Ramfjord indexteeih in the study population. Without adjustment, Ramfjord index teeth mark-edly underestimated subjects with recurrent disease activity when compared towhole-mouth findings. However, with adjustment, the hypothesis that upper andlower limits on whole-mouth incidence of recurrent periodontitis could be estimatedfrom 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 activesubjects as detennined from whole-mouth evaluations was below the upperlimit for disease incidence calculated, with 95% confidence, from point estimatesderived from adjusted Ramfjord index teeth data. Partial-mouth examinations withappropriate adjustment of Ramfjord index teeth data may be useful for assessingperiodontal disease progression in longitudinal population studies of human perio-dontitis.

Key words: Ramfjord index teeth; periodontaidiagnosis; periodontitis disease activity;partiai recording; Bonferroni inequality;epidemioiogic and longitudinal studies.

Accepted for publication 27 February 1992

Ramfjord (1959) designated 6 indexteeth for epidemioiogic studies of hu-man periodontai diseases. Teeth selectedwere the maxillary right and mandibu-lar left first molars, maxillary left andmandibular right first premolars, andmaxillary left and mandibular right cen-tral incisors. Cross-sectional studieshave found a good correlation betweenmean values from the Ramfjord indexteeth and whole-mouth scores for dentalplaque (Alexander 1970, Mills et al.1975, Gettinger et al, 1983, Goldberg etal, f985, Silness & Roynstrand 1988),supragingival and subgingival caiculus(Alexander 1970, Gettinger et al, 1983,Ainamo & Ainamo 1985), gingival in-flammation (Alexander 1970, Downer1972, Chiiton et al. 1978, Gettinger etai. 1983, Goldberg etal, 1985, Siiness &

Roynstrand 1988, Ainamo & Ainamo1985), radiographic bone loss (Berg etal. 1984), probing depth (Downer 1972,Mills et al. 1975, Berg et al, 1984, Aina-mo & Ainamo 1985, Silness & Royn-strand 1988), chnicaliy determinedattachment loss (Gettinger et al. 1983),and the periodontai disease index (Jami-son 1963).

The usefulness of the Ramfjord indexteeth in longitudinal population studiesremains to be delineated. The reliabilityof utilizing the Ramfjord index teethfor assessing the rate of periodontitisprogression in longitudinal studies mustbe resolved, particularly since this pro-gression may occur episodically and ina site-specific fashion (Socransky et al.1984). The present study assessed thediagnostic value of employing the

Ramfjord index teeth as estimators ofwhole-mouth incidence of recurrentperiodontitis in an adult population onmaintenance care.

Material and MethodsPatients

The study population, the experimentaidesign, examination techniques, andbaseline clinicai and microbiologicalfindings are described in detail else-where (Listgarten et al, 1989). In brief,98 adults previously treated for moder-ate to advanced periodontitis weremaintained on a 3-month recall pro-gram. This report analyzes clinical dataon site-specific disease recurrence de-tected semi-annuaily over a 36-monthperiod.

148 Rams et al.

Clinical examinations

2 clinical examiners were cahbrated atthe beginning of the study and there-after 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 aprobing force of approximately 50ponds was utilized. The length of theprobe tip at the bottom of the pocketto the gingival margin was recorded tothe nearest mm. At each tooth, meas-urements were taken from the mesial,distal, vestibular (buccal, labial), andmid-oral (lingual, palatal) tooth sur-faces. At each tooth surface, the deepestsite constituted the recorded probingdepth.

Relative attachment level. Relativeattachment level was determined at thesame time as probing depth as the dis-tance from the probe tip to a fixed refer-ence consisting of a trimmed thermo-plastic occlusal stem. The same sites asdescribed above were examined.

Recurrent periodontitis

Recurrent disease activity was defmedas any site, identified by the 2 indepen-dent examiners, which exhibited eithera probing depth increase of 3 mm ormore from baseiine, or a probing depthincrease of 2 mm or more together witha loss of clinical attachment of 2 mm ormore.

Data Analysis

The hypothesis that the whole-mouthrate of recurrent periodontitis in apopulation (n) can be inferred from re-current periodontitis incidence on theRamfjord index teeth (j.?^) in a samplefrom that population was examined asfollows.

Subjects were categorized YES/NOfor each 6-month period with respectto recurrent periodontitis activity on (i)one or more of the Ramfjord indexteeth, and (ii) one or more of all teethpresent. For each 6-month period, dis-ease activity rates on Ramfjord indexteeth were adjusted the ratio r, formedfrom the sum of all teeth present overthe sum of all Ramijord index teeth inthe study population,

A point estimate (P^) of ^n upperlimit for n was given by the formula:

where:Pa — the observed percentage of subjects

• with recurrent periodontitis on atleast one Ramfjord index tooth,

r =the ratio of the sum of all teethpresent over the sum of allRamfjord index teeth in the exam-ined population, recalculated foreach 6-month period.

In this study, TI, the observed % ofsubjects with recurrent disease activityover a given 6-month period on at leastone tooth in the entire dentition, wasavailable from whole-mouth examin-ations. Each JT was examined to deter-mine if it supported the hypothesis thata 95% confidence upper limit for whole-mouth disease incidence {n) can be cal-culated from Ramfjord index teeth data.

An upper bound for TZ {n^) was ob-tained in an application of the first Bon-ferroni inequality (Morrison 1983),which established that the probabilityfor a union of K random events, V/i,,K / ^ A : , is bounded about by SPU,) ,The set of disease-active subjects wasrepresented as the union of (not necess-arily disjoint subsets oO subjects iden-tified as disease-active based onRamfjord index teeth data (A^) or non-Ramijord index teeth data (^R). Fromthe first Bonferroni inequality, an upperlimit for disease incidence is given by:

tivity rates, were calculated as:

A point estimate for the right-handside of the equation was made using theadjusted Ramijord index teeth disease-activity rate, P̂ ,̂ — p^ x r. which assumesequal likehhoods of disease activity inequal sized subsets of Ramfjord andnon-Ramijord index teeth.

The unadjusted Ramfjord index teethdisease-activity rate (p^j provides apoor estimate of the lower limit onwhole-mouth disease-activity incidence(;rL). since subjects disease-active atRamfjord index teeth constitute a sub-set of subjects disease-active anywherein the mouth. An improved estimatorof 71, would require the probabihty ofthe joint event that a subject is disease-active at both Ramfjord and non-Ramfjord index teeth, which would notbe available from partial-mouth exam-inations involving only Ramfjord indexteeth.

From the central limit theorem,bounds below which the whole-mouthdisease-activity incidence would be ex-pected to occur in the target population,estimated with 95% confidence from ad-justed Ramfjord index teeth disease-ac-

where:/7 = number of subjects examined for

disease-activity at Ramfjord indexteeth (number of subjects on whichthe disease-activity incidence esti-mate was based).

As in any application of the centrallimit theorem, the sampling error, ormargin for error estimation, is inverselyproportional to the sample size, and de-creases as the study population size in-creases.

Results

Table 1 describes the percentage of sub-jects with recurrent periodontitis iden-tified by the Ramfjord index teeth (/?|Oand whole-mouth {iz) evaluations atsemi-annual intervals through 36months, Ramfjord index teeth by them-selves without adjustment (/JR, column3) underestimated subjects with recur-rent disease activity when comparedwith whole-mouth examination findings(Table 1, column 3 versus column 8). Aconsiderably improved correlation wasseen by adjusting disease-activity rateson Ramfjord index teeth by r (Column5), the ratio of total teeth to Ramfjordindex teeth in the population. Interest-ingly, the ratio r remained constant overtime (Table 1), due to a generally equaltemporal distribution of tooth loss be-tween Ramfjord index teeth and non-Ramfjord index teeth over the 36-month period (data not provided).

At 5 of 6 semi-annual examinations,the %) of disease-active subjects as ob-served in whole-mouth evaluations (ii)fell below 95% confidence upper limitestimates constructed from point esti-mates (PR) based on adjusted Ramfjordindex teeth data. At 6 months, the nullhypothesis that whole-mouth perio-dontitis-activity (;;) was less than orequa! to 20.2 (Row 1, Column 7) couidnot be rejected (P>0,20, z-test) with theobserved ji = 23.5.

Discussion

The present study examined the possi-bility of using partial-mouth examin-ations to detect recurrent periodontitisin adult patients on trimonthly mainten-ance care. Clinical measurements overtime were made on Ramfjord index

Ramfjord teeth and disease activity 149

Tahle J. Whok-mouth versus Ramfjord index teeth analysis of recurrent periodontitis for subjects by posl-baseline examinalion period

Exam

6 months12 months18 months24 months30 months36 months

No, ofsubjects

989893918911

WithoutRamfjord

adjustment ofindex leeth data

No, (%) of subjects exhibitingrecurrent periodonlitis

Based onRamijord teeth Based on all

(unadjusted)

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

(/)«) teeth (ii)

23 (23,5)16(16,3110(10,8)15(16,5)15(17,1)15(19,4)

r

4,644,644,644,644,644,64

Pointestimate

(-Pw)

14,428,314,925,515,818,1

Without adjustment of Ramfjord

95'Mi confidenceupper limit onwhole-mouthincidence (n._^)

20.235,821,033,022,225,3

Observedwhole-mouth

incidenceill)

23,5!6,310,816.517.119.4

index leeth data^'

Predictive efficacy of adjusiedRamfjord teeth point estimate

(P\<j) of upper limit forwhole-mouth disease activity [it)

h exceeds 95% upper bound*"n below 95% upper boundTl beiow 95% upper bound7c below 95% upper boundn below 95%> upper boundn below 95% upper bound

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 tookinto account the proportional represen-tation of index teeth among all teethpresent in the patient population. Itmust be pointed out that the reportedcorrelations between partial and whole-mouth disease activity are valid only ona population level, and are not appli-cable in the clinical management of indi-vidual patients.

The present study findings agree withFleiss et al, (1987) that Ramfjord itidexteeth data by themselves markedlyunderestimate the whole-mouth inci-dence of periodontitis progression.However. Fleiss et al, (1987) did notassess changes in ciinical attachmentlevel over time, but rather, evaluated theoceurrenee of moderate to deep perio-dontal pocketing by recording only thedeepest probing depth per looth perexamination in patients with relativelyhealthy periodontal conditions.

Improved correlations were found byadjusting the Ramijord index teeth datato provide an estimation of an upperlimit for whole-mouth disease subjectincidence for recurrent periodontitis. Ateach of the 6 semi-annual examinations,the observed whole-mouth recurrentperiodontitis rate supported the hypo-thesis that 95% confidence upperbounds on disease-activity incidencecan be projected frorn appropriate ad-justment of Ramfjord index teeth data.

It is important to emphasize that theadjustment employed on the Ramfjordindex teeth data provided otily upperlimit whole-mouth incidence level esti-mates, in accordance with the ftrst Bon-ferroni inequality (Morrison 1983),

sinee disease activity rates at non-Ramfjord index teeth is not known orobserved in the partial mouth examin-ations. Disease activity on Ramfjordindex teeth by themselves without ad-justment provide a crude lower limit onwhole-mouth disease incidetice. Moreprecise lower limit estimates, followingthe second Bonferroni inequality (Mor-rison 1983), requires the hkelihood ofthe joint event to be determined, i.e.,that a subject experienced disease activ-ity both at one or more Ramfjord indexteeth and at one or more non-Ramfjordindex teeth.

No attempt was made in the presentstudy, as previously suggested (Mills etal. 1975, Fleiss et al, 1987), to improveefficacy of partial-mouth scoring bysubstituting other teeth if one or moreRamfjord teeth was missing in specificpatients. Since disease activity was as-sumed to be equally distributed withindentitions and exhibiting the same pat-tern among all patients, the usefulnessof the proposed partial examinationtechnique may be limited for some pa-tient groups, such as persons wilh local-ized juvenile periodontitis or exhibitingsevere tooth loss. However, for adultperiodontitis, as indicated in the presentstudy at least on a group basis, use ofadjusted Ramijord teeth data provides95% upper limit estimations on the inci-dence of whole-rnouth periodontitis dis-ease-activity.

Examiner fatigue as well as the ex-tended time required for full-mouth,site-specific clinical measurements ham-per assessment of periodontitis diseaseactivity in epidemioiogic studies and

large-scale clinical trials. An advantageof the Ramfjord index leeth is thatexaminations can be carried out fasterand more econornically than evaluationof the entire dentition. Partial-mouthscoring in conjunction with newly-de-veloped electronic periodonlal probes(Jeffcoat et al, 1986, Birek et al, i987.Gibbs et al, 1988) are likely to facilitatestudies of periodontitis disease activityinvolving large numbers of subjects. Cli-nical data needed to utilize the proposedmethod would be (i) the disease-activityrate on Ramfjord index teeth, (ii) thenumber of Ramfjord index teeth, and(iii) the total number of all teeth in thestudy population. The relatively large95% estimation errors observed for thecalculated upper limit on periodontitisdisease-aetivity incidence suggests themethod is most suitable where largesample sizes are involved.

In summary, parliai-mouth examin-ations and appropriate adjustment ofRamfjord index teeth data appear use-ful in estimating, wilh 95%. confidence,the upper limit of whole-mouth perio-dontitis activity in longitudinal popula-tion studies. Further evaluation of theseand other partial-mouth assessmenttechniques is indicated.

Acknowledgments

This study was supported by GrantsRO1-DE06085 and RR 01224/00040from the National Institute of DentalResearch, National Institutes of Heaith,Bethesda, MD, USA.

150 Rams et al.

Zusammenfassung

Der Nutzen der Ramfjord hidex-Zahne inLangzeitstiidien zur Bestimmung der Progres-sion einer ParodontalerkrankungDie Brauchbarkeit der Rarafjord Iiidex-Zah-ne zur EJnschatzung der parodontalen Er-krankungsaktivitat des gesamlen Gebisseswurde untersuchl. Bei 98 Recall-Patienten,die vorher wegen einer Erwachsenenpar-odontilis behandelt wurden, hai man fur einePeriode von 36 Monaten halbjahrige Unler-suchungen des gesamten Gebisses durchge-fuhrt. Wiederkehreiide Parodontitis wurdeentweder ais eine 3 oder mehr mm groBereSondierungstiefe im Vergleich zur Ausgangs-untersuchung oder eine 2 oder mehr mm gro-Bere Sondierungstiefe zusammen mit einem 2oder mehr mm groBeren Attachmentverlustim Vergleich zu einer Referenzschiene defi-niert. Die Erkrankungsaktivitiit des gesamtenGebisses wurde mit den angepaBten odernicht angepaBten Daten aus den RamijordIndex-Zahnen vergliehen. Die Anpassung er-folgte iiber eine Multiphkation der Erkran-kungsaktivitiitsraten der Ramfjord Index-Zahne mit dem Verhaltnis aus der Summealier vorhandenen Zahne zur Summe allerRamfjord Index-Zahne in dieser Studienpo-pulation, Im Vergieich mit den Ergebnissendes gesamten Gebisses unterschatzten dieRamfjord Index-Zahne ohne Anpassung diePersonen mit wiederkehrender Erkrankungs-aktivitat betrachtlich, Mit Anpassung jedociiwurde die Hypothese, daB die oberen undunteren Grenzen der Inzidenz einer wieder-kehrenden Parodontitis durch die Erkran-kungsaktivitatsraten der Ramfjord Index-Zahne eingeschatzt werden kann zu jedemSechsmonats-Intervall. nicht verworfen {p>0.20, r-Test), In 5 der 6 Untersuchungen lagder Prozentsatz der erkrankungsaktiven Per-sonen bei Bestimmung iiber die Evaluationdes gesamten Gebisses unter der oberenGrenze der Erkrankungsinzidenz. die anhandder angepaBten Daten der Ramfjord Index-Zahne mit 95% Konfidenz berechnet woirde.Untersuchungen eines Teils des Gebisseskonnten bei geeigneter Anpassung der Datender Ramfjord Index-Zahne zur Messung derProgression einer Parodontalerkrankung inLangzeitstudien der menschlichen Parodonti-tis nijtzlich sein.

Resume

Vliliie des dents indices de Ramfjord pour eva-luer la progression de la maladie parodontaledam les etudes longitudinalesLa faisabilite de I'emploi des dents indices deRamfjord pour estimer I'activite de la mala-die parodontaie de l'ensemble de la bouche aete etudiee, Des examens de toute ia boucheont ete pratiques 2 fois par an pendant uneperiode de 36 mois chez 9S patients traitesanterieurement pour parodontite de j'adulteet actueliement en phase de maintenance. Ona defini comme recidive de ia parodontite

soit une augmentation de ia profondeur ausondage de 3 mm ou plus depuis le debut del'etude, soit une augmentation de 2 mm oupius de la profondeur au sondage en memetemps qu'une augmentation de 2 mm ou plusde la perte d'attache par rapport a une gout-tiere occlusale de reference, L'activite de lamaladie dans l'ensemble de la bouche a etecomparee aux donnees concernant les dentsiiidiees de Ramfjord, avec et sans ajustement.Cet ajustement consistait a multiplier les tauxd'activite de la tnaladie au niveau des dentsindices de RamQord par le rapport: sommede toutes ies dents presentes sur somme detoutes les dents indices de Ramfjord dans lapopulation etudiee. En l'absence d'ajuste-ment. les dents indices de Ramfjord sous-estimaient fortement les sujets ayant une reci-dive de la maladie par comparaison avec lesdonnees concernant l'ensemble de la bouche,Cependant. avee ajustement, on n'a a aycundes intervailes de 6 mois rejete l'hypotheseque les limites superieures et inferieures del'incidence des recidives de la parodontitepour l'ensemble de la bouche pouvaient etreestimees en se basant sur les taux d'activitedes denis indices de Ramfjord (P>0.20, testr). Lors de 5 des 6 examens. la proportiondes sujets presentant une maladie en activite,a en juger par les evaluations de l'ensemblede la bouche, etait en dessous de la limitesuperieure pour l'incidence, calculee. a un ni-veau de confiance de 95%, a partir d'estima-tions de points derivees des donnees ajusteespour les dents indices de Ramfjord,

Les examens buccaux partiels avec ajuste-ment adequat des donnees sur les dents indi-ces de Ramijord peuvent etre uliles pour eta-blir la progression de la maladie parodontaledans les etudes longitudinales de populationconcernant la parodontite humaine.

References

Ainamo, J.& Ainamo, A. {19S5) Partial indi-ces as indicators of the severity and preva-lence of periodontal disease. InternationalDental Journal 35, 322-326,

Alexander, A, G, (1970) Partial mouth re-cording of gingivitis, plaque and calculusin epidemiological surveys. Journal ofPeriodontal Research 5, 141-147,

Berg, L.. Becker, W. & Becker, B, E, (1984)The use of index teeth to predict the heahhstatus of the balance of the mouth. Interna-tional Journal of Periodontics and Restora-tive Dentistry 4, 47-53,

Birek. P, McCulloch. C, A, G, & Hardy, V.(1987) Gingival attachment level measure-ments with an automated periodontalprobe. Journal of Clinical Periodontology14, 472-477.

Chilton, N, W., Eertig, J, W, & Talbott, K.(3978) Partial and fuil mouth recording ofgingivitis scores. Pharmacology and Thera-peutics in Dentistry 3, 39-44,

Downer, M, C, (1972) The relative efficiencesof some periodontal partial recordingselections. Journal of Periodontal Re.'ieareh7, 334-340,

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 epidemiologicstudies of gingivitis and periodontitis.Community Dentistry and Oral Epidemi-ology 15, 221-224.

Gettinger. G., Patters, M. R,, Testa, M, A.,Loe, H,, Anerud, A., Boysen, H. & Rob-ertson, P, B, (1983) The use of six selectedteeth in population measures of perio-dontal status. Journal of Periodoniologv54, 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) De-scription and clinical evaluation of a newcomputerized periodontal probe - the Elo-rida probe. Journal of Clinica! Periodonlol-ogy 15. 137-144.

Goldberg, P.. Matsson. L, & Anderson. H,(1985) Partial recording of gingivitis anddental plaque in children of different agesand in young adults. Community Dentistryand Oral Epidemiology 13, 44-46,

Jamison, H, C. (1963) Some comparisons oftwo methods of assessing periodontal dis-ease. American Journal of Public Health53, 1102-1106.

Jeffcoat, M, K., Jeffcoat. R, L,, Jens, S, C, &Captain, K. (1986) A new periodontaiprobe with automated cemento-enameljunction detection. Journal of ClinicalPeriodontology 13, 276-280,

Listgarten, M, A,. Slots, J.. Rosenberg, J,,Nitkin, L,, Sullivan, P & OJer. J, (1989)Clinical and microbiological character-istics of treated periodontitis patients onmaintenance. Journal of Peridontologv 60,452-459,

Mills, W, H., Thompson, G, W, & Beagrie,G, S, (1975) Partial-mouth recording ofplaque and periodontal pockets. Journal ofPeriodontal Re.'ieareh 10, 36—43,

Morrison, D, E, (1983) Applied linear statisti-cal models. Prentice-Hall, New Jersey.

Ramfjord, S. P. (1959) Indices for prevalenceand incidence of periodontal disease. Jour-nal of Periodontology 30, 51-59,

Siiness, J, & Roynstrand, T, (1988) Partialmouth recording of plaque, gingivitis andprobing depth in adolescents. Journal ofClinical Periodontology 15, 189-192,

Socransky, S, S,, Haffajee, A. D,, Goodson,J, M, & Lindhe, J, (1984) New concepts ofdestructive periodonta] disease. Journal ofClinical Periodontology 11, 21-32,

Address:

Thomas E. RamsDepartment of PeriodonticsUniversity of PennsylvaniaSchool of Dental Medicine4001 Spruge StreetPhiladelphia. PA ]9!04USA