Central nervous system concentrations of cyclooxygenase-2 inhibitors in humans

Anesthesiology. 2005 Feb;102(2):409-15. doi: 10.1097/00000542-200502000-00026.


Background: Cyclooxygenase-2 (COX-2)-selective inhibitors (coxibs) are under investigation for the potential therapy, attenuation, or prevention of neuroinflammatory and neurodegenerative disorders. Coxibs are also a significant advance in pain therapy and are traditionally considered to achieve analgesia via peripheral effects. However, in animals, central nervous system (CNS) COX-2 activity and prostanoid concentrations are increased by peripheral inflammation, central sensitization has been proposed to account for long-term pain-related phenomena, and coxibs achieve significant cerebrospinal fluid (CSF) concentrations and may cause analgesia via CNS action. Nevertheless, it remains unknown whether or which coxibs reach the CNS in humans. This investigation determined whether coxibs can reach the CNS in humans, based on CSF concentrations.

Methods: Ten healthy human volunteers simultaneously received a single oral dose of celecoxib (200 mg), rofecoxib (50 mg), and valdecoxib (40 mg). Blood and CSF were serially sampled for 10 h, and plasma total and unbound and CSF coxib concentrations were quantified by mass spectrometry.

Results: Total plasma concentrations and time to maximum plasma concentration were similar among the three coxibs. In contrast, unbound (free) plasma concentrations differed significantly. Maximum unbound plasma concentrations were 1.4 +/- 0.5, 42 +/- 17, and 6.0 +/- 2.9 ng/ml, respectively, for celecoxib, rofecoxib, and valdecoxib. COX-2 inhibitors rapidly penetrated the CNS. Maximum CSF concentrations were 2 +/- 2, 57 +/- 25, and 10 +/- 4 ng/ml, respectively, for celecoxib, rofecoxib, and valdecoxib. CSF concentrations exceeding the median inhibitory concentration for COX-2 were achieved by rofecoxib and valdecoxib but not celecoxib.

Conclusions: These results show that coxibs do reach the CNS in humans, with rapid penetration, and in concentrations apparently sufficient to inhibit COX-2 activity. There were significant differences among coxibs in CSF penetration. Unbound (free) plasma coxib concentration was the major determinant of CSF concentration. This supports the hypothesis that coxibs may act, in part, in the human CNS, provide important new information on the mechanism and treatment of pain and may guide coxib selection for therapeutic trials when CNS penetration is desirable.

Publication types

  • Clinical Trial
  • Comparative Study
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Administration, Oral
  • Adult
  • Analysis of Variance
  • Blood-Brain Barrier / drug effects
  • Blood-Brain Barrier / metabolism
  • Central Nervous System / drug effects*
  • Central Nervous System / enzymology*
  • Cyclooxygenase 2
  • Cyclooxygenase 2 Inhibitors
  • Cyclooxygenase Inhibitors / administration & dosage*
  • Cyclooxygenase Inhibitors / blood
  • Cyclooxygenase Inhibitors / cerebrospinal fluid*
  • Female
  • Humans
  • Male
  • Membrane Proteins
  • Prostaglandin-Endoperoxide Synthases / metabolism*


  • Cyclooxygenase 2 Inhibitors
  • Cyclooxygenase Inhibitors
  • Membrane Proteins
  • Cyclooxygenase 2
  • PTGS2 protein, human
  • Prostaglandin-Endoperoxide Synthases