2005D-0240 Guidance for Industry on Development and Evaluation of Drugs for Treatment or Prevention of Gingivitis; Availability
FDA Comment Number : EC12
Submitter : Dr. Athena Papas Date & Time: 11/03/2005 10:11:07
Organization : Tufts School of Dental Medicine
Category : Health Professional
Issue Areas/Comments
In response to FDA"s draft guidance document governing the conduct of clinical study design for trials examining the anti-gingivitis effects of oral care products, I am gravely concerned with the delegation of clinical endpoints of bleeding to secondary endpoint status. The guidance as written would perclude the use of indicies based solely on bleeding (PBI,SBI,etc) and dichotomous derived bleeding scores from combined indices such as LS GI from being employed as primary clinical endpoints for the measurement of gingivitis. In my clinical investigations and that of my colleauges at Tufts University School of Dental Medicine, bleeding indices relate most closely to disease and identify patients at greatest risk.

In contrast, a pure visual clinical scale suchh as MGI, could be used as a primary outcome in the absence of measuring a bleeding effect. I do not believe that there is a rational biological basis for this point of view. I also beleive that the guidelines as written, will substantially affect the development of anti-gingivitis products. Please consider the technical argument outlined below in support of gingival bleeding as a primary outcome for the testing of anti-gingivitis products.

It has been accepted in the dental community that plaque-associated gingivitis is identified clinically by gingival bleeding, redness, edema causing loss of tissue form, and gingival tenderness since the early work of Loe in 1965, and later by Suzuki in 1988. Gingival bleeding upon stimulation is internationally accepted as a clinical sign of gingival inflammation. Since Muhleman's paper in 1971, gingival bleeding has been reported in clinical and histological studies to be an earlier and more sensitive sign of gingival inflammation relative to visual alterations, such as redness and edema. ( Korman 1987, Greenstein 1981). In some indices, such SBI, bleeding precedes color changes while other combined indices such as, LS GI, color change proceeds bleeding. The SBI and PBI indices have both been shown to be statistically siginificant and positively correlate to number of inflammatory cells in gingival connective tissue, upon histomorphometric analysis. (Engelbeger 1983)

Additionally Oliver (1969) found that the percentage of inflamed area of total connective tissue is weakly correlated to LS GI scores, although minimal differences in area of inflamed tissue was observed between biopsies with GI=1 and GI=2 scores. The studies demonstrate that gingival bleeding is an early sign of gingivitis and my occur prior to or at the same time as color change and edema. There is little data supporting that color changes are an earlier event than bleeding, or that ordinal combined gingival indices like LS GI accurately represent the underlying biology.

Gingival bleeding is a more relevant and visible endpoint for patients. Examination of gingival bleeding points is a routine part of standard oral exams and a common question that dentists ask as part of a dental history. In screening for periodontal studies, asking about bleeding is frequently the best way to determine eligibility. Patients are aware of bleeding: it is frequently the reason they seek care. Reduction in the number of bleeding sites is a more interpretable result than a reduction observed in an index score based on color. It is both objective and ordinal in nature and predictive of future disease risk as measured by periodontal attachment loss. Absence of gingival bleeding over a 4 year maintenance period ensured periodontal health in 98.5% of sites. In contrast, sites that repetitively bled following probing had significant attachment loss in 30% of the instances.(Lang 1986)

In calibration exercises in multicenter studies, bleeding indices are easier to control through standardization than visual combined indices. We and others doing both, have found it very difficult to calibrate to visual combined indices due to examiner subjectivity. (McClanahan 2001)