POSTOPERATIVE BLEEDING TENDENCY AS A RISK FACTOR IN ACTINOBACILLUS-ACTINOMYCETEMCOMITANS-ASSOCIATED PERIODONTITIS

Frequent bleeding on probing (BOP) has been considered a risk factor f or recurrence of periodontitis. In the present study, 29 patients with Actinobacillus actinomycetemcomitans-associated periodontitis were en rolled in a carefully performed recall system. At 6 sites per tooth, p eriodontal probing depth (PPD), gingival index (GI), plaque index (PlI ) and BOP was assessed 6 weeks, 6 months, 1 and 2 years after comprehe nsive therapy. Professional toothcleaning and subgingival scaling at s ites with PPD greater-than-or-equal-to 5 mm and BOP was carried out ev ery 2nd or 3rd month. Subgingival samples from 2 sites, a pooled subgi ngival sample, cheek mucosa, saliva and tongue samples were selectivel y cultivated for A. actinomycetemcomitans after 2 years. Following act ive therapy, 8 % sites had a PPD of greater-than-or-equal-to 4 mm, whe reas 21 % sites bled on probing. After 2 years, respective figures wer e 12 and 27 %. During maintenance, frequent BOP (greater-than-or-equal -to 3 times at 4 visits) had a predictive value of 0.133 to indicate a n increase in PPD of greater-than-or-equal-to 2 mm and a negative pred ictive value of 0.947. The predictive value of no bleeding to indicate a stable site was 0.972, the negative predictive value 0.078. There w as evidence for heterogeneity of associations between increase in PPD of greater-than-or-equal-to 2 mm and greater-than-or-equal-to 3 times BOP among patients (chi2(28) = 41.45, p < 0.05). Significant sources f or the variation of weighted ln-transformed estimates of individual od ds ratios (range -0.83 to 6.21, median 1.52) were relative numbers of A. actinomycetemcomitans-positive samples 2 years after therapy, age, and mean % of PlI 2 (R2 = 0.439, p < 0.001). No association between in crease in PPD and BOP was found in patients where A. actinomycetemcomi tans was not recovered from any sample (chi2MH = 1.96), but A. actinom ycetemcomitans-positive subjects still had inconsistent associations ( chi2(17) = 37.65, p < 0.01). Ignoring patient characteristics may be m isleading in the search for risk factors for recurrence of the disease .