Saliva detects periodontitis in smokers too: MMP-8 with an AUC of 0.855
Salivary MMP-8 distinguished stage III periodontitis from a healthy periodontium with an area under the curve of 0.855, and in heavy smokers, in whom tobacco suppresses bleeding on probing, it retained a value of 0.831.
A marker that does not depend on bleeding
Bleeding on probing is a fundamental indicator of inflammation, but in heavy smokers it falls short: vasoconstriction suppresses gingival bleeding, so the gingiva looks calmer than it is. The team of authors led by Bozaba tested whether a salivary marker can circumvent that pitfall.
Eighty systemically healthy adults were included, 20 in each of four groups: periodontally healthy non-smokers, stage III grade B in non-smokers, periodontally healthy heavy smokers and stage III grade C in heavy smokers. Alongside a full periodontal status, which covered probing depth, bleeding on probing, clinical attachment level, plaque index and radiographic bone loss, MMP-8, MMP-9 and HIF-1α were measured in unstimulated saliva by ELISA, and diagnostic accuracy was assessed by ROC analysis with a bootstrap of 2,000 iterations.
MMP-8 distinguished periodontitis from a healthy periodontium with an area under the curve of 0.855 (95% CI 0.756 to 0.934), with a sensitivity of 82.5% and a specificity of 85.0%. In the subgroup of smokers the accuracy remained high, with a value of 0.831, a sensitivity of 75.0% and a specificity of 90.0%, and this despite the fact that bleeding in them was lower than in non-smokers with grade B (30.9 ± 16.5% versus 44.6 ± 25.3%). MMP-9 lagged behind with a value of 0.702 (difference 0.154, p = 0.005). HIF-1α had no diagnostic value (0.453), but smoking suppressed it independently (p = 0.007).
Limitations
A cross-sectional design with 20 subjects per group and strictly defined phenotypes gives a clean contrast, but says nothing about how the marker behaves during therapy or in the mixed population of a practice. The difference in bleeding between the groups did not survive Bonferroni correction (p = 0.593).
For your practice
- In heavy smokers, do not rely on bleeding as a measure of inflammatory burden; a low value may be a consequence of vasoconstriction rather than of quiescent disease.
- Where a chairside MMP-8 test is available, it has the greatest added value in smokers, while in non-smokers the conventional indices still carry the diagnosis.
- Do not interpret HIF-1α as a periodontal marker; in this sample it tracked smoking, not the condition of the periodontium.