Preclinical data indicates that cyclooxygenase-2 (COX-2) inhibition impairs plasma cell growth and potentially synergizes with thalidomide. We performed a trial in previously treated patients with myeloma using thalidomide up to a maximum dose of 800 mg/d with celecoxib (400 mg bid). Outcomes were compared with a prior trial of thalidomide. Sixty-six patients with median age of 67 (range, 43-85) received a median dose of thalidomide and celecoxib of 400 and 800 mg/d, respectively, with median durations of treatment of 27 and 13 weeks, respectively. The most common toxicities associated with premature discontinuation of celecoxib (n = 30 of 53, 57%) were fluid retention and deterioration of renal function. Overall response rate (RR) was 42% and with 20 months median follow-up; the actuarial median progression-free survival and overall survival were 6.8 and 21.4 months, respectively. Unlike our prior study, age >65 years was not predictive of inferior RR due to improvement in RR in older patients with the combination (37% versus 15%, P = 0.08). The RR was superior in patients who received a total dose of celecoxib exceeding 40 g in the first 8 weeks of therapy (62% versus 30%, P = 0.021). Progression-free survival and overall survival were also improved. Other predictors for inferior progression-free survival were age >65 years (P = 0.016) and elevated beta(2)-microglobulin (P = 0.017). This study provides evidence that the addition of high-dose celecoxib adds to the antimyeloma activity of thalidomide but this comes with unacceptable toxicity. Future studies should use newer COX-2 inhibitors with thalidomide, or their respective derivatives.