Maintenance of gingival health – a measure based on clinical indices

Abstract

Objectives: To introduce a variable which directly measures maintenance of gingival health. Design: The maintenance of gingival health index (MGHI) is based on the change in score at each tooth site compared to a reference visit for a clinical index (eg MGI or BI). For a subject the number of sites that (1) improve-NDEC, (2) show no change-NNOCHG or (3) worsen-NINC in score are determined. Then the MGHI for a subject is defined as (NDEC + NNOCHNG)/NINC. Results: This method was applied to two clinical studies and both studies demonstrated significantly better maintenance of gingival health in those subjects who received the experimental 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice compared to the 0.204% w/w sodium fluoride control dentifrice based on a gingival and a bleeding clinical index. Conclusion: The MGHI is a useful measure to assess the maintenance of gingival health. It is very applicable to the maintenance of gingival health study designs as it directly looks at a measurement of maintaining the health of tooth sites. It also provides a meaningful measure of the relative effectiveness of dentifrice products to the dental practitioner.

INTRODUCTION

Oral Healthcare Clinical indices have been developed and available for use in clinical trials for many years. The first dental index is reputed to have been introduced as far back as 19491., 2.. Commonly, many sites around the whole mouth (maxillary and mandibular arches, facial and lingual sites) are used. For each individual subject, prior to statistical analysis, a whole mouth score (typically the mean) is calculated based on all available sites. Although all sites are taken into account, a large amount of information (and clinicians’ time) goes into this summary measure, for example, for the Modified Gingival Index (MGI) and Bleeding Index (BI), 108 sites will contribute to the whole mouth mean.

Numerous studies have indicated that clinical dental hygiene services have limited long range probability of success if the subject does not cooperate in the daily procedures of care. Educational and clinical services therefore are mutually dependent in the total dental hygiene care of the subject. Clinical studies have repeatedly demonstrated that removing and subsequently preventing the reformation of plaque are important elements in maintaining health and controlling periodontal disease. Even after this, there may be potential to improve gingival health. Using the maintenance of gingival health model3., 4. it may be possible to bring the subjects to their most optimal gingival health, although this may not be zero MGI score in all cases, and then investigate dentifrice products that can maintain this level of optimal gingival health.

As a measure to look at the ability of dentifrice products to maintain this healthy level, one could look at the change in the dental index from a ‘baseline reference’ level, and if the score remained the same or decreased this would be an indication of a maintained level.

This paper investigates a method to measure maintenance of gingival health and is based on the score at each tooth site assessed and if it improves; shows no change; or worsens in score compared to a reference visit, e.g., baseline. A maintenance of gingival health index (MGHI) is constructed for each subject based on all sites assessed in the whole mouth based on these three outcomes. The MGHI can be calculated for an individual treatment and compared between treatments using conventional statistical techniques.

This methodology has been used in two, 12-week clinical studies recently conducted by GSKCH3., 4., and the application of this methodology in these two studies is described later in this paper.

METHODS

The two clinical indices to be discussed in this paper are the Modified Gingival Index (MGI) and the Bleeding Index (BI). These two indices are described below.

Modified Gingival Index (MGI)

The MGI⁵ was assessed on the facial and lingual surfaces at two sites on each tooth (papillae and margin). These assessments were performed on all evaluable teeth. The scoring of the MGI was performed under dental office conditions using a standard dental light for illuminating the oral cavity. Compressed air, water and mouth mirrors were available to each examiner.

This procedure was performed by a single examiner to avoid inter-examiner variability. The MGI is a non-invasive evaluation of early visual changes in severity and extent of gingivitis.

The MGI scoring system is as follows:

0 = absence of inflammation

1 = mild inflammation; slight change in color, little change in color; little change in texture of any portion of the marginal or papillary gingival unit

2 = mild inflammation; criteria as above but involving the entire marginal or papillary gingival unit

3 = moderate inflammation; glazing, redness, edema, and/or hypertrophy of the marginal or papillary gingival unit

4 = severe inflammation; marked redness, edema and/or hypertrophy of the marginal or papillary gingival unit, spontaneous bleeding, congestion, or ulceration.

Bleeding Index (BI)

The Bleeding Index (modified Index of Saxton⁶), was performed by a single examiner using a dental probe.

The BI scoring system is as follows:

0 = no bleeding after 30 seconds

1 = bleeding upon probing after 30 seconds

2 = immediate bleeding observed

The probe was engaged approximately 1 millimetre (mm) into the gingival crevice. A moderate pressure was used whilst sweeping from interproximal to interproximal along the sulcular epithelium.

Tooth site level measures

As an alternative to the whole mouth measures of MGI and BI, changes at the tooth site level can be determined. Summary measures across all evaluable tooth sites within a subject can include the number or proportion of sites that decrease, increase or show no change in their score compared to a reference visit. Since the number of tooth sites for each subject is a fixed quantity, these three outcomes are inter-related. A single measure which encompasses all three outcomes, the Maintenance of Gingival Health Index (MGHI), is introduced. The derivation of this index and how it can be analysed is described below.

Maintenance of Gingival Health Index (MGHI)

The MGHI was calculated for both MGI and BI separately, and was derived from tooth site level changes compared to a reference visit.

MGHI was defined as: MGHI = (NDEC + NNOCHNG)/NINC

Where:

• •NDEC = The number of sites that show a decrease in their index score (Visit score - reference visit score < 0). This is seen as an improvement in score
• •NNOCHNG = The number of sites that do not change in their index score (Visit score - reference visit score = 0)
• •NINC = The number of sites that increase in their index score (Visit score - reference visit score > 0). This is seen as a worsening in score.

This index is analogous to an odds of the maintenance of gingival health such that if the MGHI is > 1 this means that there is a greater number of sites that do not increase compared to the number of sites that do increase, i.e. a greater maintenance of health. The ratio of MGHI for treatment A/MGHI for treatment B is defined as the MGHI ratio which in essence is an odds ratio. If the MGHI ratio is >1, this translates to the odds of maintaining gingival health being greater on treatment A compared to treatment B.

If the number of sites that worsen is zero the calculation of the MGHI would not be possible due to division by zero. To avoid this, the MGHI can be set to the total number of sites assessed in the subjects’ mouth and this would still maintain the ordered nature of the measure. If a subject had 108 evaluable sites and only one worsened, then the MGHI would be 107/1 = 107. If zero sites had worsened then we would set the MGHI to the number of sites assessed, which would be 108. This is numerically larger than its nearest neighbour but more meaningful than if infinity (108/0) is used.

MGHI can be calculated separately for MGI and BI scores. So for each subject, an MGHI value is constructed which ranges from 0 to 108. This lends itself to be analysed using an analysis of covariance (ANCOVA). The statistical model used for the two example studies included factors for treatment group and the pre-prophylaxis baseline and randomisation baseline levels of MGI and BI (whole mouth) as covariates. For study I³, where assessments were performed at six and 12 weeks, a repeated measures model was used, whereby in addition to those stated above, terms were included for time, treatment by time and baseline by time interactions. Since the MGHI is bounded at the lower end by zero, it tends to have a skewed distribution. The log MGHI can take any positive or negative value (Figure 3) and so has an approximately normal distribution⁷. Therefore the MGHI was transformed prior to analysis using the log₁₀ of the MGHI (a small value e.g. 0.001 can be added to MGHI to avoid taking logs of 0). Various plots were produced to investigate assumptions of normality and constancy of variance. These were upheld with the log₁₀ transformation, therefore ANCOVA was appropriate for this variable in both studies.

Figure 3.

NDEC = The number of sites that show a decrease in index score. NNOCHNG = The number of sites that do not change in index score. NINC = The number of sites that increase in index score. NSITES = The number of tooth sites. Plot to show MGHI against the proportion maintained.

RESULTS

The tooth site level changes and MGHI results after 12 weeks of twice daily brushing, from the two clinical studies3., 4. are presented in Tables 1 and 2, and Figures 1 and 2. Figures 1 and 2 show the mean possible changes (increase, decrease and no change) for each tooth index by treatment group. The results presented in Table 1 demonstrate that the MGHI (based on MGI) in those subjects who received the 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice is 0.97 (i.e. approximately half the sites decreased in score or showed no change at week 12 compared to baseline, and approximately half the sites increased in score). For those subjects who received the control 0.204% w/w sodium fluoride dentifrice, the MGHI was 0.09 (i.e. approximately 10% of the sites decreased in score or showed no change at week 12 compared to baseline, and approximately 90% of the sites increased in score). The ratio of 0.97/0.09 converts to a greater than 10 fold increase in the MGHI based on the MGI for the 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice compared to the control 0.204% w/w sodium fluoride dentifrice. For the result based on BI, the MGHI was significantly higher for the 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice. The corresponding result for the second study is that there was a 2.9 and 2.3 fold increase in the MGHI based on MGI and BI respectively for the 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice compared to the control 0.204% w/w sodium fluoride dentifrice. Although there were differences in the size of the treatment effect from the two clinical studies, both of them demonstrated significantly better maintenance of gingival health in those subjects who received the 0.1% w/w o-cymen-5-ol/ 0.6% w/w ZnCl₂ dentifrice compared to the control 0.204% w/w sodium fluoride dentifrice, whether based on MGI or BI.

Table 1.

Study I³ Maintenance of Gingival Health Index (MGHI)

Variable	Maintenance of Gingival Health		MGHI ratio (95% CI)	P-value*
	0.1% w/w o-cymen-5-ol/0.6% w/w ZnCl2 (N = 104)	0.204% w/w Sodium fluoride (N = 101)
MGI	0.97	0.09	10.68 (7.59, 15.14)	<0.0001
BI	51.66	20.68	2.50 (2.04, 3.09)	<0.0001

^*P Value for difference between treatment groups in log₁₀ MGHI from a repeated measures ANCOVA.

Table 2.

Study II⁴ Maintenance of Gingival Health Index (MGHI)

Variable	Maintenance of Gingival Health		MGHI ratio (95% CI)	P-value*
	0.1% w/w o-cymen-5-ol/0.6% w/w ZnCl2 (N = 100)	0.204% w/w Sodium fluoride (N = 105)
MGI	6.95	2.38	2.92 (2.25, 3.79)	<0.0001
BI	7.27	3.15	2.31 (1.79, 2.98)	<0.0001

^*P Value for difference between treatment groups in log₁₀ MGHI from ANCOVA.

Figure 1.

Study I³ Tooth site level changes in MGI and BI.

Figure 2.

Study II⁴ Tooth site level changes in MGI and BI.

DISCUSSION

Figure 3 shows how the MGHI and the log of MGHI changes with the proportion of sites maintained. The figure also shows that the MGHI equals 1.0 when 50% of the sites are maintained. As soon as more than 50% of the sites are maintained the MGHI rises quite sharply compared to when <50% of the sites are maintained. The plot also shows the linearity of the function when it is log transformed.

One of the big problems with dental indices (e.g. MGI and BI) is that they are a subjective assessment. However, the dental examiners are experienced in using these indices. A repeatability assessment on the MGI was performed during both studies on a small number of subjects. This is where the small number of subjects have a repeat MGI in addition to their initial main MGI. The repeat MGI is carried out shortly after the main initial exam. The level of agreement between the initial and repeat MGI was calculated using a Kappa coefficient. The intra-examiner kappa scores were 0.96 (95% CI 0.95, 0.96) for Study I³ and 0.95 (95%CI 0.94, 0.96) for Study II⁴. These two results demonstrate excellent⁸ repeatability and consistency of the examiners scoring. An auxiliary method could use photographs as a reference but these are two dimensional and miss the contour changes which are critical in the MGI. Also photographic techniques are not as developed for consistent posterior and lingual photographs. The methodology described in the Contour study⁹ could be used but the measurement is on a continuous scale and one would need to establish a range ordered categorical measure to be able to establish a meaningful measurement index.

This MGHI described in this paper is very applicable to the maintenance of gingival health study design3., 4. rather than just a simple single prophylaxis model as even after this single deep clean it may be possible to improve in MGI/BI score especially under clinical conditions where use of the study treatment is monitored in some way.

The methodology described in this paper could be applied to other indices but care may need to be taken in what is used as a reference visit score, e.g. for plaque it is common for a prophylaxis to be performed before the subjects begin to use their assigned treatment. This prophylaxis is assumed to remove all plaque and thus scores would be zero at this point for every site. The dilemma is then whether to compare back to the plaque score prior to the prophylaxis or to a post prophylaxis score of zero. If the zero score is used, many sites will worsen in index score. If the pre-prophylaxis score is used, it could be that many sites may still be below this score due to the effect of the prophylaxis as well as the treatment intervention. Some indices will have more than one measure across the tooth, e.g. Lobene Stain Index and Modified Turesky Modification of the Plaque index. For these indices, a dilemma exists to whether the change at a site on the tooth should be used or whether an overall measure for the tooth would be better. For the MGI and BI, the sites are located across the tooth margin and the papillae between the teeth. It is meaningful to break out into facial and lingual areas as lingual areas are harder to reach and inflammation can exist on either side of the tooth or papillae.

The MGHI is better than just looking at, for example, the proportion maintained as it encompasses all aspects of the change in index score at the site in that the score is maintained or better or it worsens in score. It is also for this same reason that the MGHI is more advantageous than just looking at the mean score, as a single site change can have a greater impact on the score than on an overall mean. This can be seen from Figure 3 as the MGHI changes quite dramatically the closer to 100% of the sites are maintained. The whole mouth mean would be constrained between a score of 0 to 4 for the MGI. This single site impact can be important from a subject’s viewpoint as they could suffer an oral health problem at a single site which might require a visit to a dental practice. This increased sensitivity at this end of the scale is an advantage of the MGHI over an overall whole mouth mean score. It is also better than looking at the overall mean as increases and decreases can cancel each other out. If 50% of the scores increase by 1 point and 50% of the scores decrease by 1 point this would lead to an overall mean which is the same. Looking at no change in the whole mouth means would indicate that close to 100% of sites are maintained. In contrast using the individual site changes would be closer to 50%.

There is no evidence of a measure to look directly at maintenance of gingival health in the public domain. Many papers have used no change in the mean score to attempt to measure maintenance10., 11., 12., 13., 14., 15.. One paper does set a criterion of ≤ 5 bleeding sites to achieve gingival health¹⁶. The MGHI does not rely on a cut off criteria and the problems of using mean levels have been identified previously.

CONCLUSION

The MGHI is a useful measure to assess the maintenance of gingival health. It is very applicable to the maintenance of gingival health study designs3., 4. as it directly looks at a measurement of maintaining the health of tooth sites. It also provides a meaningful measure of the relative effectiveness of dentifrice products to the dental practitioner.

CONFLICT OF INTEREST AND SOURCE OF FUNDING

The work described in this manuscript was funded by GlaxoSmithKline Consumer Healthcare. The authors are employed by GSK but confirm no potential conflicts of interest.