A stepwise approach to determine periodontal attachment loss in longitudinal studies

Modelling changes in clinical attachment loss to classify periodontal disease progression

AIM

The goal of this study was to identify progressing periodontal sites by applying linear mixed models (LMM) to longitudinal measurements of clinical attachment loss (CAL).

METHODS

Ninety-three periodontally healthy and 236 periodontitis subjects had their CAL measured bi-monthly for 12 months. The proportions of sites demonstrating increases in CAL from baseline above specified thresholds were calculated for each visit. The proportions of sites reversing from the progressing state were also computed. LMM were fitted for each tooth site and the predicted CAL levels used to categorize sites regarding progression or regression. The threshold for progression was established based on the model-estimated error in predictions.

RESULTS

Over 12 months, 21.2%, 2.8% and 0.3% of sites progressed, according to thresholds of 1, 2 and 3 mm of CAL increase. However, on average, 42.0%, 64.4% and 77.7% of progressing sites for the different thresholds reversed in subsequent visits. Conversely, 97.1%, 76.9% and 23.1% of sites classified as progressing using LMM had observed CAL increases above 1, 2 and 3 mm after 12 months, whereas mean rates of reversal were 10.6%, 30.2% and 53.0% respectively.

CONCLUSION

LMM accounted for several sources of error in longitudinal CAL measurement, providing an improved method for classifying progressing sites.

Outcome variables in periodontal research: means and threshold-based site changes

BACKGROUND

The exclusive use of mean measurements in periodontal research might at times be misleading, as changes in different sites in the same individual might nullify each other. The purpose of the present study was to compare disease progression and response to periodontal therapy using both individual site activity with thresholds and mean patient changes.

METHODS

Seventy-nine (79) subjects with established periodontitis were monitored for 1 year (no treatment [NTx] group); 108 subjects who received scaling, root planing, and quarterly prophylaxis were observed in a similar time interval (treatment [Tx] group). Probing depth (PD), attachment level (AL) and alveolar crestal height (ACH) were measured at baseline and 1 year using pressure-sensitive probes and computer-assisted image analysis of radiographs.

RESULTS

Mean reduction in PD (0.50 mm) was observed in the Tx group compared to a small increase (-0.04 mm) in the NTx group (P= 0.0001). Treatment resulted in mean AL gain (0.44 mm) compared to net AL loss (-0.21 mm) in the NTx group (P= 0.0001). Subjects in the Tx group had twice as many sites with AL gain (beyond the threshold) compared to NTx subjects (16.64+/-1.07% versus 8.11+/-0.68%) which was highly significant (P = 0.0001, Student t test). Similarly, although in the opposite direction, percentage of sites with AL loss beyond threshold was 6.21+/-0.47% (Tx group) and 14.02+/-1.15% (NTx group) which was also highly significant (P = 0.0001, Student t test). Change in ACH was minimal for the Tx group (-0.07+/-0.03 mm) while NTx subjects experienced greater bone loss throughout the year (-0.16+/-0.02 mm); percentage of sites with ACH loss was similar for both groups, while the Tx group had 3 times the number of sites with ACH gain (11.76% versus 3.42%), suggesting that the reduction in mean bone loss in the Tx group did not result from a reduction in losing sites, but rather from an increase in gaining sites. The use of both means and binary data seems to supplement the information regarding the nature and mechanism of disease progression and arrest.

CONCLUSIONS

The characteristics of periodontal disease with its multifactorial patient, local, and site-related etiology support the use of both mean and site-based changes in clinical studies of periodontal disease and treatment.

Option-4 algorithm for third generation disc probe: agreement of selected site-specific relative attachment level measurements and detection of longitudinal site-specific attachment level change

BACKGROUND

Longitudinal site-specific attachment level change (SSAC), identified from serial relative attachment level measurements (RAL), is the principal indicator of progression/regression of periodontal diseases. Many variables confound RAL reproducibility and affect measurement error. The Option-4 algorithm was designed to reduce measurement error and improve accuracy and sensitivity of SSAC detection. The study aimed to evaluate the performance of the Option-4 algorithm.

METHODS

A precalibrated clinician recorded full mouth RAL with a third generation disc probe on 4 occasions over 6 months in 16 subjects (mean age 48.1 years) with moderately advanced chronic adult periodontitis (2,312 sites). Option-4 allowed up to 4 RAL recordings per site per visit until 2 values had differences < or =1.0 mm and their mean was < or =1.0 mm from the previous visit mean: the clinician made the selection if these criteria were unfulfilled.

RESULTS

Within-visit agreement < or =1.0 mm was > or =99.6%: all within-visit correlation coefficients = 0.98 (P<0.001). At each visit, mean difference in Option-4 values was < 0.05 mm, mean absolute difference (ignoring direction) was < or =.34 mm. Mean site-specific variances ranged from 0.092 mm2 to 0.097 mm2 across all visits. Subject thresholds for site-specific attachment level change (from estimated 95% confidence limits of visit 1 data) ranged from 0.52 mm to 0.67 mm. Linear SSAC (by linear regression) and between-visit patterns of SSAC were investigated. SSAC was detected in 100% subjects and at 51.0% measured sites. Linear SSAC (R2 > or =0.90: P < or =0.05) occurred at 105 sites (4.5%): 32 sites (1.4%) deteriorated, 73 sites (3.1%) improved. Between-visit SSAC occurred at 1,074 sites (46.5%): 391 sites (16.9%) deteriorated, 295 sites (12.8%) improved, and 388 sites (16.8%) showed exacerbation/remission patterns.

CONCLUSIONS

The Option-4 algorithm produced high RAL agreement. Site-specific attachment level change was detected in both directions in 100% subjects and at 51.0% measured sites.

The effect of different diagnostic thresholds on incidence of disease progression

Detection of periodontal disease progression occurs when a predetermined threshold of attachment loss is exceeded during longitudinal monitoring. The incidence of disease progression in a population may be dependent on the method and threshold utilized to identify significant changes in attachment level measurements. The aim of this study was to investigate the effect of utilizing different methods and thresholds on the incidence of disease progression in an untreated periodontitis population. The relationship between baseline clinical parameters and disease progression was also examined. A total of 411 interproximal sites in 46 individuals were monitored monthly over a 6-month period. Disease progression was determined by the cumulative sum (CUSUM) method and by the absolute change in relative attachment level between months 0 and 6 utilizing 3 different thresholds for attachment level change (0.58 mm, 1.16 mm, and 1.74 mm) based upon examiner repeatability using an automated probe. Utilizing the CUSUM method, 49 of 411 sites (11.9%) demonstrated attachment loss over the 6-month observation period. When attachment level changes > or = 0.58 mm, > or = 1.16 mm, and > or = 1.74 mm were used to identify disease progression, the percentage of sites exhibiting deterioration were 19.5%, 8.8%, and 2.9%, respectively. These results demonstrate that the apparent incidence of disease progression was dependent on the method and threshold utilized to detect progressive sites. When utilizing the CUSUM and 0.58 mm thresholds a significant (P < 0.05), but weak relationship (r = -0.26) was observed between baseline relative attachment level measurements and sites exhibiting disease progression. This finding suggests that sites with significant but relatively less attachment loss may be more likely to experience further breakdown compared to sites with a history of greater periodontal destruction.

Is the chemical prevention of gingivitis necessary to prevent severe periodontitis?

Gingival inflammation seldom causes discomfort, social embarrassment or loss of function. As most sites with gingival inflammation do not progress to severe periodontal disease, gingivitis should not be considered a public health problem. Periodontitis is always preceded by gingivitis. But most gingivitis remains stable for years without progressing to periodontitis. The number of gingivitis sites that do convert is small. The levels of oral cleanliness achieved by the majority of populations in industrialized countries are below the threshold for severe destructive periodontal disease of personal and public health concern. Because methods of measuring the progression of periodontal disease are unreliable, definitive answers regarding conversion of gingivitis to severe periodontitis are lacking. Gingival inflammation frequently remains contained; most gingivitis remains stable for years without progressing to periodontitis. Decreasing gingivitis does reduce shallow pocketing, but the effect on severe periodontitis is not clear. Although the underlying justification for the reduction of plaque is to reduce gingival inflammation to prevent or reduce severe periodontitis and tooth loss, the basis for the approach is equivocal. A reasonably high level of plaque appears to be compatible with acceptably low levels of periodontal disease. Reducing nonspecific plaque levels to such levels is therefore a rational goal. The conventional methods of controlling periodontal disease involve mechanical removal of plaque and calculus. A complimentary ecological approach, using chemicals, would be to alter the environment of the pocket to prevent growth of putative pathogens. Any ecological approach should be sensitive to the dangers of disrupting the natural ecology of dental plaque. Some antimicrobial and antimetabolic agents such as fluoride, chlorhexidine and triclosan and zinc citrate can selectively suppress certain organisms or inhibit bacterial proteases implicated in tissue damage. The uncertainties about factors that convert gingival inflammation into periodontitis and periodontitis into severe periodontitis coupled with insufficient data from controlled clinical trials on the effectiveness of chemical reduction of gingivitis to prevent severe periodontitis leads one to conclude that more research is required before the need for the chemical prevention of gingivitis to prevent severe periodontitis can be justified.

Option-4 algorithm for automated disc probe: reduction in the variance of site-specific relative attachment level measurements

Physical periodontal measurement is plagued by many confounders which result in aberrant values. Replicate measurements can reduce the number of aberrant values, the measurement error, and the variance of site-specific measurements. This study aimed to reduce the variance of site-specific measurements by using a new clinical algorithm (the Option-4 algorithm) for an automated disk probe. A single clinician recorded full-mouth relative attachment levels (RAL) at one visit in 32 patients (mean age 45.5 years) with moderately advanced chronic adult periodontitis. RAL was recorded over two passes at six sites per tooth (4,675 sites). The algorithm accepted the first and second pass RALs (RAL1 and RAL2) if their difference was < or = 1 mm, otherwise a maximum of two further RALs (RAL3 and RAL4) were recorded until the difference between any two RALs was < or = 1 mm (SAL1 and SAL2): 4,048 sites (86.6%) required two recordings, 580 sites (12.4%) required three recordings and 47 sites (1%) required four recordings. Correlation coefficients for RAL1 and RAL2 and SAL1 and SAL2 (4,675 sites) were both > or = 0.91 (P = 0.00). Site-specific variances were calculated for RAL1 and RAL2 and SAL1 and SAL2. The mean of the RAL1/RAL2 site-specific variances (A) was 0.45 mm2 (range 0.00 mm2 to 35.28 mm2) whilst the mean of the SAL1/SAL2 variances (B) was 0.09 mm2 (range 0.00 mm2 to 0.5 mm2): the respective medians were 0.08 mm2 and 0.02 mm2. The study demonstrated high intra-examiner RAL reproducibility. The Option-4 algorithm produced an 80% reduction in the mean site-specific variance of RAL1/RAL2 (Y) and a 75% reduction in the median site-specific variance of RAL1/RAL2 (y = [(A - B)/A] x 100).

Cytokines modulate routes of collagen breakdown. Review with special emphasis on mechanisms of collagen degradation in the periodontium and the burst hypothesis of periodontal disease progression

In this paper, we review recent work on collagen degradation, 2 main routes of breakdown are described and their relevance during healthy and inflammatory conditions of the periodontium is discussed. Special attention is paid to the possible role of cytokines, in particular interleukin 1 (IL-1) and transforming growth factor beta (TGF-beta), on the modulation of collagen phagocytosis and metalloproteinase production. IL-1 has been shown to have a dual function in collagen digestion. It inhibits the intracellular phagocytic pathway, but at the same time, it strongly promotes extracellular digestion by inducing the release of collagenolytic enzymes like collagenase. TGF-beta has an opposite effect on both pathways and antagonizes IL-1. Collagenase is released in an inactive form, and a considerable fraction of the proenzyme may become incorporated in the extracellular matrix. This reservoir of latent enzyme can be activated (for instance by plasmin), leading to a sudden and extensive breakdown of the collagenous fibre meshwork. It is suggested that this phenomenon may also take place during progressive periodontitis and could explain an episodic nature of collagenolysis, clinically resulting in bursts of attachment loss (burst hypothesis).

Reproducibility of attachment level recordings using an electronic and a conventional probe

This study describes the variations in the reproducibility of attachment level recordings for different subjects at different examinations. Twenty patients with different degrees of periodontal disease were recruited and examined bi-monthly for their attachment levels using an electronic probe in two quadrants and a conventional probe in the other two quadrants. At each of the 7 examinations attachment level recordings were repeated with the appropriate probe after approximately 30 minutes. Results demonstrate that the reproducibility of attachment level recordings was slightly better for the conventional probe than for the electronic probe. Attachment level recording reproducibility varied considerably between subjects and between examinations for the same subject. Aggregation of attachment level recordings in the form of mean tooth values or mean individual values reduced the range of the differences, but considerable between-subject and between-examination variation was still seen. These results indicate that no single uniformly valid estimate of attachment level reproducibility exists which can be used to set diagnostic thresholds such that a minimum number of incorrect diagnoses are made.

Sources of error for periodontal probing measurements

This study determined the relative contribution to probing measurement reliability of several factors, compared with that of random error. Probing measurements were performed by examiners properly trained and calibrated. A total of 5771 pairs of replicate pocket depth (PD) and relative attachment level (RAL) measurements were performed with the Florida Probe. A total of 1488 replicate clinical attachment level (CAL) measurements were performed with the North Carolina 15 mm probe. In addition, longitudinal replicate measurements of RAL were performed at 0 and 12 months on 816 sites in 11 patients utilizing the Florida Probe 20 mm disk probe. Measurement reliability with the Florida Probe resulted in mean intraexaminer variances of 0.21 and 0.33, for PD and RAL, respectively (s.e.m. 0.46 mm for PD and 0.57 mm for RAL). Measurement reliability with the conventional probe resulted in mean intra-examiner variances of 0.19 for PD and 0.32 for CAL (s.e.m. 0.44 mm and 0.56 mm). Pocket depth contributed to = 5% of the variability of the intra-examiner variance with both probes with other contributing factors being the individual patient, tooth and site location. Mean intra-examiner reproducibility for duplicate RAL measurement performed at 0 and 12 months was 0.24 and 0.19, respectively (s.e.m. 0.49 mm and 0.43 mm). In conclusion, a mean intra-examiner variance of < or = 0.24 can be achieved for replicate measurements with both electronic and conventional probes for moderate and severe periodontitis patients. Individual examiner, individual patient and site location contribute up to 10% to the overall variance. Hence, the pattern of variability for intra-examiner variance of probing measurements performed with either electronic or conventional probes by trained and calibrated examiners is mostly random error.

Effect of model fitting artifacts on the stepwise approach to identifying patterns of attachment loss

The stepwise approach to the determination of periodontal attachment loss involves fitting linear, logarithmic, and exponential models to individual site data and concluding that the form of loss is consistent with the model that has the greatest r-value, provided that the model predicts loss in excess of a site-specific threshold. Logarithmic and exponential fits are considered to define early and late bursts, respectively, while linear fit describes loss at a constant rate. In a recently published study, the stepwise approach was applied to 6,935 sites in patients with established periodontitis and, of 581 loss detections, 195 (33.6%) were linear, 224 (38.6%) logarithmic, and 162 (27.9%) exponential. However, curvilinear patterns may occur by chance and regression algorithms that can fit such curvature may have an advantage unrelated to the true mechanism(s) [correction of mchanism(s)] of periodontal destruction. To investigate the implications of this possibility, proportions of linear, logarithmic, and exponential fit were estimated by simulation under four different conditions. These conditions incorporated random or random plus linear change, but no nonlinear effects. The relative proportions of model fits described in the published study were approximated in all of these conditions. It would appear that the observed proportions are ubiquitous to the modelling approach itself, and do not constitute evidence of a causal non-linear biological mechanism. The stepwise approach may be useful for detecting change but relevance to causal processes seems problematic.

The effect of post-surgical flap placement on probing depth and attachment level: a 2-year longitudinal study

Postsurgical flap placement might affect the outcome of the operative procedure. Modified Widman flap surgery with primary closure and flap approximation (usually away from the bone crest) and apically positioned flap surgery with near crestal bone positioning are both widely used in surgical periodontal treatment. Several comparative investigations have studied these modalities, however, none have been able to show conclusively that either is superior to the other. The purpose of this longitudinal study was to explore the optimal postsurgical flap placement in respect to final probing depth and changes in clinical attachment level. Following routine hygienic phase of treatment, 12 subjects (186 teeth) with adult periodontitis received surgical periodontal treatment. Prior to the flap surgery, probing depth and clinical attachment level were recorded. Sounding depth measurements were taken to record postoperative flap placement. Patients were placed on a 3-month maintenance program. Probing depth and clinical attachment level were again measured at 2 years postoperatively and compared to baseline measurements. An overall positive correlation (R = 0.43; P = 0.0248) was found between immediate postoperative sounding measurements and probing depth after 2 years. Conversely, attachment level changes over the 2-year period showed only weak inverse correlation (R = 0.27; P = 0.0121) with sounding depth measurement immediately postsurgically. Sites where postoperative sounding depth were < or = 3 mm had a mean probing depth (2.52 mm) which was significantly (P < 0.001) smaller compared to sites with sound depth > or = 4 mm (3.58 mm). Changes in clinical attachment level varied between sites and sounding depth groups; however, none of these differences were statistically significant. Based on our findings it is suggested that following periodontal flap surgery, in those cases where minimal probing depth is desired, the flap be secured to the underlying structures at or slightly coronally to the bone crest (< or = 3 mm).(ABSTRACT TRUNCATED AT 250 WORDS)