Curso Clínico de la Periodontitis Crónica: Factores predisponentes en Enfermedad Periodontal

8/9/2019 Curso Clnico de la Periodontitis Crnica: Factores predisponentes en Enfermedad Periodontal

1/8

The clinical course of chronicperiodontitis: V. Predictive factors

in periodontal disease

Schatzle M, Faddy MJ, Cullinan MP, Seymour GJ, Lang NP, Bu rgin W, Anerud A,Boysen H, Loe H. The clinical course of chronic periodontitis: V. Predictive factorsin periodontal disease. J Clin Periodontol 2009; 36: 365371. doi: 10.1111/j.1600-051X.2009.01391.x.

Abstract

Background: The factors associated with initial periodontitis are not well understoodand cannot be identified by cross-sectional studies.

Aim: To identify the factors associated with the initiation of chronic periodontitisusing ante-dependence modelling.

Material and Methods: A 26-year longitudinal study of the natural history ofperiodontitis served as the basis for the study. In 1969, 565 Norwegian men aged1634 years were surveyed. Subsequent surveys were performed in 1971, 1973, 1975,1981, 1988 and finally in 1995, with 223 remaining subjects. Plaque (PlI), gingival(GI) and calculus indices (CI) and loss of attachment (LoA) were recorded. Ante-dependence modelling using a Markov chain enabled the results of this sequence ofexaminations to be analysed longitudinally, taking into account serial dependence,describing temporal changes in patients levels of disease and allowing for both

progression and regression between disease categories.Results: With age, the rate of disease regression decreased. Increasing calculusaccumulation and smoking increased the rate of disease progression, while increasingGI increased the rate of regression.

Conclusions: Increased mean CI and smoking were significant predictive covariatesfor progression, while increased mean GI and younger age predicted regression ofinitial periodontitis.

Key words: age; ante-dependence model;calculus; disease progression; gingivitis; initialperiodontitis; prediction; smoking

Accepted for publication 26 January 2009

Although it is generally believed that, atsome point, periodontitis must be precededby gingivitis (Loe & Morrison 1986, Page& Kornman 1997), it is also true that not

all gingivitis sites progress to periodontitis.

Marc Schatzle1

, Malcolm J. Faddy2

,Mary P. Cullinan3, Gregory J.Seymour3, Niklaus P. Lang4,Walter Burgin5, Age Anerud6,Hans Boysen6 and Harald Loe1Universitiy of Zurich, Zurich, Switzerland;2Queensland University of Technology,

Brisbane, Australia; 3University of Otago,

Dunedin, New Zealand; 4University of Hong

Kong, Pok Fu Lam, Hong Kong; 5University of

Berne, Berne, Switzerland; 6Private practice,

Kongsvinger, Norway and Vejle, Denmark

Conflict of interest and source offunding statement

The authors declare that they have no con-flict of interests.This study has been partially supported bythe Clinical Research Foundation (CRF) forthe Promotion of Oral Health, University ofBerne, Switzerland. No other external fund-ing, apart from the support of the authorsacademic institutions, was provided for thisstudy.

The role of gingivitis in the pathogen-esis of chronic periodontitis has beenelucidated previously (Schatzle et al.2003a) in a longitudinal investigation

of the initiation and progression of nat-ural periodontal disease in a randomizedgroup of middle-class Norwegian men.This study showed that before 40 yearsof age, only slight increases in perio-dontal attachment loss due to pocketformation occurred, but after this, thefrequency increased significantly. Lossof attachment (LoA) due to gingivalrecession was rarely observed. As menapproached 60 years of age, gingivalsites that consistently bled on probingover the 26 years had approximately70% more attachment loss than sites

that were consistently non-inflamed[gingival index (GI)5 0].

The fact that sites with non-inflamedgingivae also exhibited some LoA and

pocket formation was thought to be dueto fluctuations in disease expressionduring long observational intervals pos-sibly combined with the presence ofsubclinical inflammation (Schatzle et al.2003a).

A subsequent analysis of the samecohort of well-maintained and dentallyaware Norwegian men (Heitz-Mayfieldet al. 2003) revealed that 50% of the16 year olds exhibited initial LoA onthe buccal surfaces of molars and pre-molars in both the jaws, most of whichwere gingival recession. These lesions

Dedication:

This manuscript is dedicated to the memory of

our co-author, Harald Loe, friend, colleague and

mentor, who died on 9 August 2008 during the

period of the revision of this paper. Without

Haralds foresight in carrying out the surveys on

the natural history of periodontal disease, this

paper would not have eventuated.

J Clin Periodontol 2009; 36: 365371 doi: 10.1111/j.1600-051X.2009.01391.x

365r 2009 John Wiley & Sons A/SJournal compilation r 2009 John Wiley & Sons A/S


2/8

progressed at a relatively slow rate(0.1 mm/year) during their twenties andthirties. At 30 years of age, the meanindividual cumulative LoA waso1mm.As the subjects approached 40 years ofage, the mean individual loss was slightlyabove 1.5 mm (Loe et al. 1978b). Gen-

erally, the incidence of incipient perio-dontal destruction increased with age,with the highest rate occurring between50 and 60 years. Moreover, while gingi-val recession was the predominant lesionbefore age 40, periodontal pocketing wasthe principal mode of destruction bet-ween 50 and 60 years of age (Heitz-Mayfield et al. 2003).

The rate of attachment loss duringvarious stages of adult life was furtherassessed in a third analysis of the cohort(Schatzle et al. 2003b). This analysisrevealed that the annual rates and the

annualized risks of periodontal attach-ment loss vary throughout adult life.The annual mean rate and the meanannualized risk of initial attachmentloss were the highest between 16 and34 years of age. However, most of thiswas due to recession.

Finally, the fourth study on this cohortof Norwegian men (Schatzle et al. 2004)showed that different severities of gingi-vitis yielded different risks for tooth loss.Teeth surrounded with healthy gingivaltissues were maintained for a tooth age of51 years, while teeth consistently sur-

rounded with inflamed gingivae yieldeda 46 times higher risk of loss. Only two-thirds of such teeth were maintainedthroughout the 26-year observation peri-od. Based on this observation, gingivalinflammation was thought to be a riskfactor for tooth loss.

Cross-sectional statistics, however,make it difficult to analyse fluctuationsbetween gingivitis and initial perio-dontitis. Furthermore, the temporalassociation between the various factorsinfluencing the initiation of progressionand regression between disease states

cannot be identified. Owing to the cycli-cal nature of the disease process, cross-sectional statistics cannot provide a rateat which the patients progress or regressfrom health (gingivitis) to initial perio-dontitis and back again.

It is understood that attachment loss orprobing depth at a particular time may notbe representative of the activity of alesion, in terms of loss and/or gain ofclinical attachment that has occurred inthe preceding interval. This is especiallyso in studies of longer duration withlengthy periods between examinations.

Ante-dependence modelling using aMarkov chain enables the results of asequence of periodontal examinations tobe analysed longitudinally taking intoaccount serial dependence. This modeldescribes temporal changes in patientslevels of disease in terms of transition

probabilities, which allow for both pro-gression and regression of the diseasefrom health/gingivitis to initial perio-dontitis, and then back again, from initialperiodontitis to health/gingivitis. To date,there has only been one 3-year longitu-dinal study (Faddy et al. 2000) demon-strating how ante-dependence modellingof longitudinal data can reveal effects thatmay not be immediately apparent fromthe data, such as smoking and increasingage, inhibiting the healing processrather than promoting disease progres-sion.

The purpose of this analysis was touse ante-dependence modelling to iden-tify the factors associated with chronicperiodontitis in a 26-year longitudinalstudy. Chronic periodontitis was definedat a patient level according to the cate-gorization proposed by the 5th EuropeanWorkshop on Periodontology in 2005(Tonetti & Claffey 2005).

Material and Methods

Sources of data

The information presented in this paperis based on a 26-year longitudinalstudy of the initiation and progressionof periodontal disease in well-educatedmiddle-class men in Norway. As pre-viously alluded to, this cohort receivedstate-of-the-art dental care from theage of 3 years and reportedly performedan oral home care programme on a dailybasis. The study population has beendescribed earlier (Loe et al. 1978a, b,c,1986, Anerud et al. 1991, Heitz-Mayfield et al. 2003, Schatzle et al.2003a, b, 2004). The initial examination

in 1969 included 565 individuals agedbetween 16 and 34 years. Subsequentsurveys took place in 1971, 1973, 1975,1981, 1988 and 1995. Of the 565 subjects examined in 1969, 223 attendedthe seventh examination, 26 years later.

Starting shortly after World War I,the City of Oslo launched a comprehen-sive oral health care programme for theimprovement of oral health in its chil-dren. From 1936 onwards (Gythfeldt1937), all children were entitled to com-prehensive examinations and treatmenton the basis of an annual recall, and by

1946, every school child was offeredsystematic dental care including preven-tive, restorative, endodontic, orthodon-tic and surgical therapy, if needed. Overtime, other programmes were added toinclude both preschool children and uni-versity students. Thus, the dental care

programme covered the age span from3 to 23 years.

Clinical parameters

The examinations were performed inwell-equipped clinical facilities at theFaculty of Odontology, University ofOslo, and included assessments of theperiodontal tissues. At each appointment,the participants answered questionsregarding their personal dental care andsmoking habits. Also, at each examina-tion throughout the study, the same oral

indices were scored by the same twoinvestigators, both of whom who wereexperienced periodontists, and well stan-dardized and repeatedly calibrated invarious disease levels (H. B., A. A.).

The following oral indices or measure-ments were recorded (Loe et al. 1978a):

GI (Loe & Silness 1963). LoA in millimetres (Glavind & Loe

1967). Plaque index (PlI) (Silness & Loe

1964). Retention index (RI) (Loe 1967).

From the survey in 1973 and there-after, recession of the marginal gingivawas measured on all mesial andbuccal surfaces of all teeth as the dis-tance in millimetres from the cementenamel junction (CEJ) to the gingivalmargin, whenever located apically tothe CEJ. Pocket depth was calculatedfrom the measurements of the attach-ment level and the gingival recessionat each site. In the survey in 1981 andin all subsequent examinations, thedistal and lingual surfaces were also

included in the examinations. Thirdmolars were not included in the evalua-tion at any time.

Subjects were also stratified accordingto their smoking history into self-reportedsmokers and non-smokers. The non-smoking cohort was made up of indivi-duals who, at each examination, reportedthat they had never smoked. The smokinggroup consisted of all subjects, who, atevery survey in which they participated,reported smoking two or more cigarettesper day. However, as the informationrelating to smokers and non-smokers

366 Schatzle et al.

r 2009 John Wiley & Sons A/SJournal compilation r 2009 John Wiley & Sons A/S


3/8

collected at the examination in 1995 waslost, it was assumed that participants whowere smoking throughout the study (foralmost 20 years, respectively) continuedto smoke up to 1995.

Data analyses

Based on the consensus paper at the 5thEuropean Workshop of Periodontology(2005) (Tonetti & Claffey 2005), threestages of periodontal disease corre-sponding to different levels of LoAwere defined as:

Level 0 individuals with a healthyperiodontium: upto one proximalsite with LoA X3mm.Level 1 presence of proximalattachment loss ofX3mm in X2non-adjacent teeth.Level 2 presence of proximalattachment loss of X5 mm inX30% of teeth present.

This categorization was measured fora number of subjects at seven examina-tions carried out over the 26-year period.As in most longitudinal studies of thissize and length, a number of the parti-cipants dropped out and hence could notbe followed. Other subjects missed oneor more examinations, but attended thelast survey. The aim of the presentanalysis was to model the progressionand regression (healing) between dis-

ease levels or states over the period ofthe study (19691995), in terms of thecovariates determined in the previoussurvey, i.e. age (years), smoking/non-smoking status (binary 0/1), mean PlIover all available sites, mean GI over allavailable sites and mean calculus index(CI) (a component of the RI) over allavailable sites. Consequently, those withcontiguous sequences of examinationsstarting at baseline comprised the datafor this analysis.

Modelling

In view of the very small number ofsubjects whose periodontal disease everprogressed to Level 2, modelling anddata analyses were restricted to Levels 0and 1, i.e. a transition from the categoryof health to that of initial periodontitisand from initial periodontitis back tohealth. The Markov model used hasbeen described earlier by Faddy et al.(2000) and may be summarized brieflyas follows:

Disease is assumed to progress betweenstages or Levels 0 and 1 at a rate a per unit

time, and regress or reverse betweenLevels 1 and 0 at a rate b per unit timeaccording to a two-state Markov chain incontinuous time. This leads to probabil-ities pij(t) of transitions between levelsi and j in time t given by the formulae:

p00t ba expfa btg=a b

p01t aa expfa btg=a b

p10t bb expfa btg=a b

p11t ab expfabtg=ab:

The rates a and b were log-linearlydependent on the covariates accordingto the formulae:

loga a0 a1 age a2

smoking status a3

mean plaque index a4

mean gingival index a5

mean calculus index;

and

logb b0 b1 age b2

smoking status b3

mean plaque index b4

mean gingival index b5

mean calculus index:

To allow for possible changes inthese rates of disease progression andregression during the course of thestudy, some variation in the parametersa0 and b0 between examinations wasincluded. All the resulting parametersa0s, a1, a2, a3, a4 and a5, b0s, b1, b2, b3,b4 and b5 were estimated from theobserved data by maximum likelihoodwith backwards elimination used toremove any non-significant covariateeffects, using a significance level of5%. Any interactive effects of smokingstatus with mean PlI, mean GI and meanCI were estimated as additional terms inthe above forms of log(a) and log(b),and included in the model if they weresignificant at the 5% level.

Results

Of the initial 565 individuals, 380 werepresent at contiguous examinationsstarting at baseline (i.e. examinations1, 2, . . . , n before dropping out atexamination n11, for some nX2) andwere, therefore, included in this analy-sis. The average times (t) betweenexaminations were 25 months (surveys12), 24 months (surveys 23), 27months (surveys 34), 76 months (sur-

veys 45), 77 months (surveys 56) and85 months (surveys 67), respectively.The proportions of transitions betweendisease levels between surveys areshown in Table 1.

The baseline demographics of these380 individuals are shown in Table 2.Over the 26 years of investigation, thesubjects ages ranged from a minimumof 16 years (survey 1) to a maximum of59 years (survey 7), and the proportionsof subjects smoking (smk51) rangedfrom 29% at the beginning of the study

to 9% at the end. The mean PlI rangedfrom 0.35 to 1.98; the mean GI rangedfrom 0.018 to 1.84; and the mean CIranged from 0 to 1.40.

Some changes in the rates of diseaseprogression and regression were appar-ent during the course of the study (cfTable 1), and estimates of these rateswere as follows:

i between surveys 1 and 4 :

loga6:21 0:85 smk1:62 CI

logb 21:39 0:22 age 1:57 GI

Table 1. The proportions of transitionsbetween disease levels from survey to survey

Visits 14 To: 0 1 2

From0 87.0% 5.3% 0%1 4.7% 3.0% 0%2 0% 0% 0%Visits 45 To: 0 1 2From0 85.4% 7.8% 0%1 1.0% 5.8% 0%2 0% 0% 0%Visits 56 To: 0 1 2From0 63.8% 24.6% 0%1 1.5% 10.1% 0%2 0% 0% 0%Visits 67 To: 0 1 2From0 20.4% 50.0% 0%1 1.8% 24.1% 3.7%2 0% 0% 0%

Table 2. Baseline demographics

N 380Age 1635 yearsProportion of patientswith Level-1 disease

6.84%

Smoking 109 (28.7%)PlI mean (range) 1.2 (0.501.9)GI mean (range) 0.89 (0.051.8)CI mean (range) 0.16 (0.001.1)

Level 15X2 mesial sites with LOA X3mm.

CI, calculus index; GI, gingival index; PlI,

plaque index.

Ante-dependence modelling and initial periodontitis 367



4/8

ii between surveys 4 and 5 :

loga 7:10 0:85 smk 1:62 CI

logb 0:42 0:22 age 1:57 GI

iii between surveys 5 and 6 :

loga 5:96 0:85 smk 1:62 CI

logb0:42 0:22 age 1:57 GI

iv between surveys 6 and 7 :

loga 4:59 0:85 smk 1:62 CI

logb 0:42 0:22 age 1:57 GI

The time periods involved here (sur-veys 14, surveys 45, surveys 56 andsurveys 67) were all of some 67-yearduration. The rates of disease progres-sion show some reduction after survey4, but after surveys 5 and 6 there are

increases, towards higher levels thanthey were initially, and the rates ofdisease regression show a reductionafter survey 4; these changes are statis-tically significant with p-value o0.001.Any further changes in the rate of dis-ease regression over the course of thestudy (after survey 5) were not signifi-cant (p-value 40.2). To some extent,these changes in the rates of progressionand regression as the study proceededreflect the increasing ages of the sub-jects as there was no additional effect ofage on the rates of disease progression

(p-value 40.2), but there was a furthernegative effect of increasing age, redu-cing the rates of regression (p-valueo0.001). The other effects, indicatedabove, of smoking and higher mean CIincreasing the rates of disease progres-sion, and of a higher mean GI increasingthe rates of regression are all significant(p-valueso0.01), whereas the effects ofmean GI on the rate of progression,smoking and mean CI on the rate ofregression, and mean PlI on both therates, were not significant (p-values40.2). There were no significant addi-

tional interactive effects of smokingwith mean GI, mean CI and mean PlIon the rates of either progression orregression of disease (p-values 40.2).

In this population, smokers had abouttwice the rate of disease progressionas non-smokers for this particular defi-nition of disease category (Level 1;Tonetti & Claffey 2005). The effect ofcalculus on the rates of disease progres-sion was such that these were higherby a factor of approximately 5 for CI51 compared with CI5 0, and the effectof mean GI on the rates of regression

was such that these were also higher bya factor of approximately 5 for GI5 1.5compared with GI5 0.5. Tables 3 and 4show the estimated proportions of subjects experiencing some disease pro-gression and some disease regression,respectively, between surveys, for dif-ferent values of these predictive factors.

In Figs 1 and 2, there are illustrativeplots of (a) the estimated rates of diseaseprogression and regression with increas-ing age, and (b) the corresponding pro-

portions of the subjects experiencingsome disease progression, and somedisease regression, during the periodbetween surveys 1 and 4, and betweensurveys 6 and 7.

Figures 1(a) and 2(a) indicate thatafter certain ages, the declining ratesof disease regression have fallen belowthe rates of disease progression, anddisease is more likely to take hold aftersuch critical ages. All of the subjects insurvey 1 were younger than these ages,but by survey 6 none of the subjects hadrates of disease progression less than

rates of disease regression. An increasein the mean GI from 0.5 to 1.5 results inan increase of this critical age (i.e.disease tends to take hold later) bysome 78 years, as does a decrease inthe mean CI from 1 to 0. And forsmokers, this age is lower (i.e. diseasetends to take hold earlier) than that fornon-smokers by some 34 years.

Discussion

The Norwegian population analysed inthe present study has had the benefit of acomprehensive oral health care pro-gramme from an early age (3 years). Allthe subjects participating in this study hadbeen in the City of Oslos Dental Pro-gram and subsequently reported to haveseen their private dentists on a regularannual basis. There must be very fewother population groups in the world,which in 1995 and at ages up to almost60 years had documented an exposure tosystematic dental care similar to that of

Table 3. Estimated proportions of the subjectsexperiencing some disease progressionbetween surveys

Surveys Non-smokers Smokers

14 CI 0 0.14 0.301 0.54 0.84

45 CI 0 0.061 0.14

1 0.27 0.5256 CI 0 0.18 0.37

1 0.63 0.9067 CI 0 0.58 0.87

1 0.987 0.9999

CI, calculus index.

Table 4. Estimated proportions of the subjectsexperiencing some disease regression betweensurveys

Surveys Age

14 30 years 50 years

GI 0.5 0.58 0.0101.5 0.985 0.049

Surveys Age45 30 years 50 years

GI 0.5 0.13 0.00171.5 0.50 0.0081


GI 0.5 0.13 0.00171.5 0.50 0.0082


GI 0.5 0.15 0.00191.5 0.54 0.0091

GI, gingival index.

15 20 25 30 35 40 45 50 55 600

0.01

0.02

0.03

0.04

0.05

a

b

age (years)

rate

15 20 25 30 35 40 45 50 55 600

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

age (years)

proportion

Fig. 1. (a) Estimated rates of disease progres-sion for non-smoking subjects with meancalculus indices 0 ( ) and 1 ( ),and rates of disease regression for non-smok-ing subjects with mean gingival indices 0.5( ) and 1.5 ( ), for the periodbetween surveys 1 and 4. (b) Estimatedproportions of subjects experiencing somedisease progression for non-smokers withmean calculus indices 0 ( ) and 1( ), and proportions of subjects experi-encing some disease regression for non-smo-kers with mean gingival indices 0.5 ( )and 1.5 ( ), during the period between

surveys 1 and 4.

368 Schatzle et al.



5/8

those participating in this study. In thissense, the patient cohort followed for 26years in the present study represents auniquely maintained middle-class malepopulation of Caucasian ethnicity. Hence,it is reasonable to expect that diseaseprogression would be scarce in thiscohort. Indeed, only two of the subjectswith contiguous examinations (n5380)progressed from Level 1 to Level 2 usingthe 5th European Workshop on Perio-

dontology definitions of disease and hadadvanced LoA at multiple sites at the lastexamination in 1995.

In order to evaluate the factors influ-encing disease progression to initialperiodontitis, data on transitions fromLevel 0, i.e. zero LoA with at themost one inter-proximal site with LoAX3 mm, to Level 1, i.e. some perio-dontal lesions including the presence ofproximal attachment loss ofX3 mm inX2 non-adjacent teeth, were analysed.

The definitions of Levels 0, 1 or 2patients may be open to debate. At the

5th European Workshop on Perio-dontology in 2005, the proposed criteriafor a two-level definition of a perio-dontal case were discussed at length.Deliberately, the Level 1 definitionrepresents a sensitive case definitionrepresenting an initial stage of perio-

dontitis, and the Level 2 definitionallows a more specific case definitionfor patients with substantial extent andseverity of periodontitis. The proposedcriteria were recommended for the iden-tification of risk factors (Tonetti &Claffey 2005). In the light of the den-tally aware and well-maintained malepopulation of the present study, itseemed entirely reasonable to apply theLevel 1 case definition for the modellingprocess when using the ante-dependentMarkov chain model.

At all examinations, LoA was assessed

relative to probing depth to the CEJ.Hence, LoA was the primary determinantof disease progression. Probing depthswere available in examination 3 in 1973and thereafter when recession was speci-fically addressed. It is obvious that prob-ing depth reflects the severity of thedisease, while LoA may be a moredefinite assessment of past disease.

Because both the assessment of LoAand the probing depth depend on thepenetration of a periodontal probe fol-lowing an applied force, it is recognizedthat these variables yield multiple

sources of error (Mombelli 2005). It iscrucial that the dimensions and theapplied forces are standardized and con-trolled (Mombelli et al. 1992, 1997).Angulation of the probe and its incre-mental markings may also lead to varia-bility during assessment. Last, but notleast, the conditions of the gingivaltissues (healthy versus inflamed) havea profound effect on the penetrationdepth of the probe (Armitage et al.1977, Fowler et al. 1982). As sponta-neous healing of periodontal lesionswith histological gain of attachment is

unlikely to occur (Caton et al. 1980), itis logical to assume that regression toLevel 0 may be due to resolution of theinflammation rather than to true attach-ment gain, although the latter cannot beexcluded completely.

Generally, the presence of X2 teethwith LoA of 3 mm or more over time is adefinition applied to express disease pro-gression (Tonetti & Claffey 2005). Thethreshold of two teeth is set to minimizethe risk of including cases of progressionarising due to reasons other than perio-dontitis. The threshold of a longitudinal

LoA of 3 mm or more over time is basedon evidence extensively documented inthe periodontal literature.

An increasing rate of disease progres-sion and a decreasing rate of diseaseregression during the course of the studymight have been expected due to the fact

that the subjects were getting older. Thedecrease in the rate of disease progressionthat was apparent after survey 4 wastherefore rather unexpected (althoughthere were subsequent increases aftersurveys 5 and 6) and may be due to aninitial improvement in the subjects den-tal care as a consequence of their takingpart in the study, although this can onlybe a conjecture.

A higher mean CI substantially influ-enced the progression of disease, irre-spective of the smoking status. A highermean CI resulted in net disease progres-

sion occurring some 78 years earlierthan that for a low mean CI. This pointsto the importance of calculus as a pla-que-retaining and -promoting factor(Waerhaug 1956), the regular removalof which appears to be of utmost impor-tance even in this dentally aware andwell-maintained patient cohort.

It is interesting to note that increasingCalculus rather than PlI were predictivefor progression of disease. This highlightsthe difficulty in evaluating subgingivalplaque as well as accumulation of supra-gingival plaque on the teeth in longitudi-

nal studies, where the intervals betweenobservations are too long to enable aprecise representation of the oral hygienestatus of the dentition over time.Obviously, the assessment of the calcifiedplaque indirectly reflected the amount ofdeposits in a more reliable way.

The results of the present analysisalso showed that, in this Norwegianpopulation, smoking led to a doublingof the rate of disease progression, withnet disease progression (i.e. the rate ofprogression exceeding the rate of regres-sion) occurring some 34 years earlier

compared with that in the non-smokers.Increasing mean GI significantly

increased the rate of regression of dis-ease by a factor of approximately 5 (fora unit increase). This is an interestingfinding that may indicate that the inflam-matory response of the gingiva affec-ted the healing process rather than thedisease progression rate and hencemaintained the lesions in a state ofhomeostasis. Those patients (e.g. smo-kers) who do not have a strong inflam-matory response to plaque mayrepresent a susceptible population. This

15 20 25 30 35 40 45 50 55 600

0.01

0.02

0.03

0.04

0.05

age (years)

rate

15 20 25 30 35 40 45 50 55 600

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

age (years)

proportion

a

b

Fig.2. (a) Estimated rates of disease progres-sion for non-smoking subjects with meancalculus indices 0 ( ) and 1 ( ),and rates of disease regression for non-smok-ing subjects with mean gingival indices 0.5( ) and 1.5 ( ), for the periodbetween surveys 6 and 7. (b) Estimatedproportions of subjects experiencing somedisease progression for non-smokers withmean calculus indices 0 ( ) and 1( ), and proportions of subjects experi-encing some disease regression for non-smo-kers with mean gingival indices 0.5 ( )and 1.5 ( ), during the period between

surveys 6 and 7.




6/8

is supported by the findings of the pre-sent study. However, for over 50 yearolds, there seems to be little diseaseregression whatever the value of GI.In the early stages of the disease pro-cess, i.e. in the transition from health/gingivitis to initial/early periodontitis,

the apparent loss and gain of attachmentis likely to reflect a dynamic processthat in turn reflects the homeostaticmechanisms of inflammation in res-ponse to the presence of dental plaque.It is only at older ages that these fail anddisease progresses. In subjects withoutinflammation, the clinical situationappears to present with LoA severalyears earlier.

This agrees with the previous analysisof these data (Schatzle et al. 2003b),which demonstrated that the effect ofgingivitis was in the over 40 year olds

and as they approached 60 years of age.Age and smoking status were the twosystemic variables influencing diseaseprogression and regression. It has to bekept in mind that in the present study,the self-reported smoking habitsincluded the consumption of as little astwo cigarettes per day. Increasing agelowered the rate of disease regression,while smoking increased the rate ofdisease progression. In contrast to pre-vious studies (Bergstrom 2006), theanalysis of the present study has usedlongitudinal data to elucidate the role of

these covariates in predicting initialperiodontitis.In a previous study on the natural

history of periodontal disease of a short-er duration, with more frequent exam-inations and shorter intervals betweenexaminations (Faddy et al. 2000), smok-ing was recognized as having a signifi-cant negative effect on diseaseregression, a phenomenon that couldnot be detected in the present study,possibly due to the extended periodsbetween examinations varying from 2to 7 years.

From a clinical point of view, apply-ing a Markov chain ante-dependencemodel to this 26-year longitudinal studyof dentally aware and well-maintainedNorwegian males has identified threemajor factors, calculus, smoking andage, affecting the transition from zeroLoA to initial periodontitis. Increasedmean CI and smoking were signifi-cant predictors of disease progression.

Further, gingivitis and age influencedthe regression from initial periodontitisback to health. An increased mean GIincreased the regression rate from initialdisease back to level zero, whileincreased age had the opposite effect.

References

Anerud, A., Loe, H. & Boysen, H. (1991) The

natural history and clinical course of calculus

formation in man. Journal of Clinical Perio-

dontology 18, 160170.

Armitage, G. C., Svanberg, G. K. & Loe, H.

(1977) Microscopic evaluation of clinical

measurements of connective tissue attach-

ment levels. Journal of Clinical Perio-


Bergstrom, J. (2006) Periodontitis and smoking:

an evidence-based appraisal. Journal of Evi-

dence Based Dental Practice 6, 3341.

Caton, J., Nyman, S. & Zander, H. (1980)

Histometric evaluation of periodontal sur-gery. II. Connective tissue attachment levels

after four regenerative procedures. Journal of

Clinical Periodontology 7, 224231.

Faddy, M. J., Cullinan, M. P., Palmer, J. E.,

Westerman, B. & Seymour, G. J. (2000)

Ante-dependence modeling in a longitudinal

study of periodontal disease: the effect of age,

gender and smoking status. Journal of Perio-


Fowler, C., Garrett, S., Crigger, M. & Egelberg, J.

(1982) Histologic probe position in treated and

untreated human periodontal tissues. Journal

of Clinical Periodontology 9, 373385.

Glavind, L. & Loe, H. (1967) Errors in the

clinical assessment of periodontal destruction.

Journal of Periodontal Research 2, 180184.

Gythfeldt, T. (1937) Oslo kommunale tannkli-

nikker, Oslo.

Heitz-Mayfield, L. J., Schatzle, M., Loe, H.,

Burgin, W., Anerud, A., Boysen, H. & Lang,

N. P. (2003) Clinical course of chronic

periodontitis. II. Incidence, characteristics

and time of occurrence of the initial perio-

dontal lesion. Journal of Clinical Perio-


Loe, H. (1967) The gingival index, the plaque

index and the retention index systems. Jour-

nal of Periodontology 36, 610616.

Loe, H., Anerud, A., Boysen, H. & Smith, M.

(1978a) The natural history of periodontal

disease in man. Study design and baselinedata. Journal of Periodontal Research 13,

550562.


(1978b) The natural history of periodontal

disease in man. Tooth mortality rates before

40 years of age. Journal of Periodontal

Research 13, 563572.


(1978c) The natural history of periodontal

disease in man: the rate of periodontal

destruction before 40 years of age. Journal

of Periodontology 49, 607620.

Loe, H. & Morrison, E. (1986) Periodontal

health and disaese in young people: screening

for priority care. International Dental Jour-

nal 36, 162167.

Loe, H. & Silness, J. (1963) Periodontal disease

in pregnancy. I Prevalence and severity. Acta

Odontologica Scandinavia 21, 533551.Mombelli, A. (2005) Clinical parameters: bio-

logical validity and clinical utility. Perio-

dontology 2000 39, 3039.

Mombelli, A., Muhle, T., Bragger, U., Lang, N.

P. & Burgin, W. B. (1997) Comparison of

periodontal and peri-implant probing by

depth-force pattern analysis. Clinical Oral

Implants Research 8, 448454.

Mombelli, A., Muhle, T. & Frigg, R. (1992)

Depth-force patterns of periodontal probing.

Attachment-gain in relation to probing force. Journal of Clinical Periodontology 19,

295300.

Page, R. C. & Kornman, K. S. (1997) The

pathogenesis of human periodontitis: an

introduction. Periodontology 2000 14, 911.Schatzle, M., Loe, H., Burgin, W., Anerud, A.,

Boysen, H. & Lang, N. P. (2003a) The

clinical course of chronic periodontitis: I.

The role of gingivitis. Journal of Clinical

Periodontology 30, 887901.

Schatzle, M., Loe, H., Lang, N. P., Burgin, W.,

Anerud, A. & Boysen, H. (2004) The clinical

course of chronic periodontitis: IV. Gingival

inflammation as a risk factor for tooth mor-

tality. Journal of Clinical Periodontology 31,

11221127.

Schatzle, M., Loe, H., Lang, N. P., Heitz-

Mayfield, L. J. A., Burgin, W., Anerud, A.

& Boysen, H. (2003b) The clinical course of

chronic periodontitis: III. Patterns, variationsand risks of attachment loss. Journal of

Clinical Periodontology 30, 902908.

Silness, J. & Loe, H. (1964) Periodontal disease

in pregnancy. II. Correlation between oral

hygiene and periodontal condition. Acta

Odontologica Scandinavica 22, 121135.

Tonetti, M. S. & Claffey, N. (2005) Advances in

the progression of periodontitis and proposal of

definitions of a periodontitis case and disease

progression for use in risk factor research:

group C consensus report of the 5th European

workshop in periodontology. Journal of Clin-

ical Periodontology 32 (Suppl. 6), 210213.

Waerhaug, J. (1956) Effect of rough surfaces

upon gingival tissue. Journal of Dental

Research 35, 323325.

Address:Niklaus P. Lang

Implant Dentistry

The University of Hong Kong

34, Hospital Road

Sai Ying Pun

Hong Kong SAR

E-mail: [email protected]

370 Schatzle et al.



7/8

Clinical Relevance

Scientific rationale for the study:Because cross-sectional studies can-not identify factors influencing initia-tion, progression and regression ofdisease, ante-dependence modellingusing a Markov chain has allowed

the results of a sequence of perio-dontal examinations to be analysedlongitudinally to describe temporalchanges in patients levels of disease.

Principal findings: In analysing theparameters of a 26-year longitudinalcohort with seven consecutive exam-inations, the ante-dependence model-ling revealed increasing mean CI andsmoking as significant predictive fac-tors for initiation of chronic perio-

dontitis, while increasing mean GIand lower age were associated withregression of initial periodontitisback to a healthy state.

Practical implications: In a middle-class, dentally aware cohort like theNorwegian males in this study,increased calculus deposits andsmoking were the determining fac-tors for progression from health toinitial periodontitis as defined by

the 5th European Workshop onPeriodontology, while older ageincreased the likelihood of perio-dontitis persistence.




8/8

Curso Clínico de la Periodontitis Crónica: Factores predisponentes en Enfermedad Periodontal

Documents

Transcript of Curso Clínico de la Periodontitis Crónica: Factores predisponentes en Enfermedad Periodontal