Real-time grading of bladder urothelial carcinoma (UC) is clinically important, but the current standard for grading (histopathology) cannot provide this information. Based on optical coherence tomography (OCT)-measured optical attenuation (µt), the grade of bladder UC could potentially be assessed in real time. We evaluate ex vivo whether µt differs between different grades of UC and benign bladder tissue. Human bladder tissue specimens are examined ex vivo by 850-nm OCT using dynamic focusing. Three observers independently determine the µt from the OCT images, and three pathologists independently review the corresponding histology slides. For both methods, a consensus diagnosis is made. We include 76 OCT scans from 54 bladder samples obtained in 20 procedures on 18 patients. The median (interquartile range) µt of benign tissue is 5.75 mm−1 (4.77 to 6.14) versus 5.52 mm−1 (3.47 to 5.90), 4.85 mm−1 (4.25 to 6.50), and 5.62 mm−1 (5.01 to 6.29) for grade 1, 2, and 3 UC, respectively (p = 0.732). Interobserver agreement of histopathology is "substantial" [Kappa 0.62, 95% confidence interval (IC) 0.54 to 0.70] compared to "almost perfect" [interclass correlation coefficient (ICC) 0.87, 95% CI 0.80 to 0.92] for OCT. Quantitative OCT analysis (by µt) does not detect morphological UC changes. This may be due to factors typical for an ex-vivo experimental setting.