Background: This study aimed to evaluate the gene expression of cyclooxygenases (COXs) in an oral model of preemptive analgesia. Material and Methods: Gingival tissue was collected during extraction of lower third molars from a randomized, triple-blind, split-mouth and placebo-controlled study. The eligible patients were randomly sorted to receive a single dose either of ibuprofen 400mg, or etoricoxib 120 mg or a placebo, one hour prior to surgery. The temporal course of RNAm was evaluated for COX-1 and -2 by means of a quantitative polymerase chain reaction in real time (RT-qPCR) at time zero and 30 minutes after the surgical procedure began, and it was correlated with clinical parameters (pain and maximum mouth opening). Results: There was a significant increase in COX-1 expression between T0 and T30 in ibuprofen (p=0.004) and eto-ricoxib (p=0.010) groups. As regards COX-2, there were increases from T0 to T30 in all groups (placebo, p=0.012; ibuprofen, p<0.001; etoricoxib, p<0.001). All groups showed a significant decrease in COX-2:COX-1 ratio from T0 to T30 (placebo, p=0.013; ibuprofen, p<0.001; etoricoxib, p=0.047). Experimental groups showed a significant correlation between COX-1 and COX-2 levels and clinical pain parameters. Conclusions: The present preemptive analgesia study concludes that COX-2 RNAm induction was directly linked to third molar-related tissue inflammation and that the relation between COX-1 and COX-2 levels were inversely proportional to the preemptively administered nonsteroidal anti-inflammatory drugs COX-2 selectivity.