Methods: Thirty-six periodontitis patients were monitored during supportive therapy in three month intervals after conclusion of non-surgical therapy. GCF samples were collected at one site/patient. PD, clinical attachment level (CAL) and BOP were assessed with an electronic constant-force periodontal probe. Calprotectin levels in GCF were determined with an ELISA kit. Sites presenting a PD increase > 0.5 mm within three months were considered as relapse'. The ability of individual parameters to predict a relapse' was analysed by construction of ROC curves and contingency tables.
Results: Nine sites with PD increase > 0.5 mm were identified. Clinical parameters (CAL, PD, BOP, GCF volume) were unable to predict a suchlike relapse' (area under the curve AUC<0.6, p>0.05). However, both calprotectin concentration (AUC=0.793, p=0.01) and total amount/sample (AUC=0.776, p=0.02) significantly predicted a subsequent increase of PD with a corresponding positive predictive value of 44% and a negative predictive value of 94%.
Conclusions: In contrast to clinical parameters, calprotectin in GCF did significantly predict a subsequent increase of PD at the monitored site.