The serotonin-norepinephrine reuptake inhibitor duloxetine is appropriate for this patient with osteoarthritis. The pharmacologic management of osteoarthritis pain can be difficult because therapy for symptom relief does not halt or reverse the disease process. Furthermore, all potential agents have side effects, some severe, and elderly patients can be at particularly high risk for developing them. The optimal choice of therapy relies on the risks and benefits of each agent in the context of the patient's comorbidities. This patient has already tried an NSAID (ibuprofen) as well as glucocorticoid and hyaluronic injections without symptomatic relief and has recently experienced an episode of gastrointestinal bleeding. Duloxetine is a reasonable choice given the patient's comorbidities and recent history of gastrointestinal bleeding. Compared with placebo, duloxetine significantly reduces pain and improves physical functioning in patients with knee osteoarthritis. In short-term studies, duloxetine was not associated with an increase in the adverse event rate compared with placebo. Unlike NSAIDs, duloxetine does not increase the risk for a recurrence of peptic ulcer disease. There are no medication interactions that contraindicate its use in this patient.

In an elderly patient who has had a recent bleeding peptic ulcer, the use of any NSAID, including the cyclooxygenase-2 inhibitor celecoxib, is inadvisable because the risk of recurrence with repeated exposure is high.

Narcotic use in the elderly, although not associated with ulcer risk or gastrointestinal bleeding, should also be approached cautiously. Short-acting narcotics should be tried first, and long-acting agents such as fentanyl may be used when other agents have failed. However, in this elderly patient at risk for falling due to her recent hospitalization, debilitation, quadriceps weakness, and a history of osteoporosis, fentanyl is not the optimal choice.

Although intra-articular glucocorticoids can be of benefit in the management of individual joints in osteoarthritis, there is no evidence that oral glucocorticoids would be of benefit. Furthermore, the likelihood of adverse effects with extended use is high in this patient.