Objective-To determine agreement between arterial partial pressures of carbon dioxide (Paco(2)) and end-tidal partial pressures of carbon dioxide (P-ETCO2) measured with a nasal catheter in spontaneously breathing, critically ill dogs. Design-Validation study. Animals-26 client-owned dogs admitted to an intensive care unit for various conditions. Procedures-Paco(2) was measured with a commercial blood gas analyzer, and P-ETCO2 was measured with a sidestream capnograph attached to a nasal catheter. Measurements were obtained twice (ie, with and without supplemental oxygen). Paired values were compared by means of the Pearson correlation method. Level of agreement was assessed by means of the Bland-Altman method. Results-Mean difference between Paco(2) and P-ETCO2, when dogs did not receive supplemental oxygen (mean +/- SD, 3.95 +/- 4.92 mm Hg) was significantly lower than mean difference when dogs did receive supplemental oxygen (6.87 +/- 6.42 mm Hg). Mean difference in dogs with a condition affecting the respiratory system (8.55 +/- 5.43 mm Hg) was significantly higher than mean difference in dogs without respiratory tract disease (3.28 +/- 3.23 mm Hg). There was a significant linear correlation and good agreement between measured values of Paco(2) and P-ETCO2. Catheter size, ventilatory status, and outcome were not significantly associated with mean difference between Paco(2) and P-ETCO2. Conclusions and Clinical Relevance-Results suggested that nasal capnography is a clinically relevant method of estimating Paco(2) in spontaneously breathing, critically ill dogs, but that values should be interpreted with caution in dogs receiving supplemental oxygen and in dogs with conditions affecting the respiratory system. (J Am Vet Med Assoc 2009;235:1314-1318)