SIR: Nonsteroidal anti-inflammatory drugs (NSAIDs) are well known to cause significant side effects, including delirium, particularly in older adults. The new cyclo-oxygenase-2 (COX-2) selective agents are generally considered to cause fewer gastrointestinal side effects and perhaps less renal dysfunction. We report delirium induced by both celecoxib and rofecoxib.
An 81-year-old woman was prescribed celecoxib 100 mg/day. Over the next 2 weeks she developed confusion, disorientation, and auditory and visual hallucinations. She was seen in the emergency room where delirium was diagnosed, metabolic causes were ruled out, and the celecoxib was stopped. The symptoms resolved over several days. Six months later she was prescribed rofecoxib 12.5 mg but took only a few doses. After another 2 months she was encouraged to take rofecoxib regularly and began doing so. One month later, she was taken to the emergency room agitated, with visual hallucinations, and disoriented to person, place, and time. She was unable to complete a Mini-Mental State Examination and did not recognize her children. She was placed in restraints in the emergency room.
She suffered from several chronic conditions, including atrial fibrillation (for which she was prescribed sotalol, cilazapril, and furosemide) and hypothyroidism (levothyroxine), and she also took conjugated estrogen tablets (Premarin) and nizatidine.
Physical examination suggested no cause of the delirium other than rofecoxib. Chest X-ray, electrocardiogram, and basic laboratory data were normal. A CT scan of the head showed mild atrophy and periventricular leukoaraiosis. The only abnormal laboratory result was a TSH level of 25 μU/ml, which was confirmed by repeat testing. Her pharmacy reported that her thyroid supplement was dispensed at appropriate intervals for the number of tablets.
The rofecoxib was stopped, and over the next 2 days her delirium resolved. It is difficult to know if the hypothyroidism made her more vulnerable to delirium. Although her levothyroxine dose was increased on the second hospital day, this would not have cleared her delirium.
On follow-up 3 months later she was cognitively intact, with an MMSE of 27/30 and normal clock drawing.
This report demonstrates delirium induced by two COX-2 inhibitors. Delirium caused by nonselective NSAIDs has been documented.1 In some cases, the putative mechanism was a link to the indolic moiety in certain NSAIDs.1 Consistent with the serotonin hypothesis of delirium, the similarity of this moiety to serotonin was proposed as the reason why certain NSAIDs can cause delirium. Neither celecoxib nor rofecoxib has such moieties, leaving the mechanism in this case open to speculation. The role of COX-2 in the brain remains unknown, as does the effect of selective COX-2 inhibitors.2 Interestingly, some studies have documented that nonselective NSAIDs worsen cognitive function in elderly people.3,4
This case illustrates the importance of several prescribing principles in geriatrics, including "Any drug can cause any side effect" and "Any new symptom should be considered to be due to a drug until proven otherwise."