Intended for healthcare professionals

Clinical Review Lesson of the week

Celecoxib, rofecoxib, and acute temporary visual impairment

BMJ 2003; 327 doi: (Published 20 November 2003) Cite this as: BMJ 2003;327:1214
  1. David Morris Coulter (david.coulter{at}, director1,
  2. David Walter John Clark, senior research fellow1,
  3. Ruth Lesley Savage, senior research fellow2
  1. 1Intensive Medicines Monitoring Programme, New Zealand Pharmacovigilance Centre, Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, PO Box 913, Dunedin 9000, New Zealand
  2. 2Centre for Adverse Reactions Monitoring, New Zealand Pharmacovigilance Centre
  1. Correspondence to: D M Coulter
  • Accepted 6 October 2003


We present evidence of acute severe temporary visual disturbance with two of the new selective anti-inflammatory cyclo-oxygenase-2 (COX 2) inhibitors. We report two cases, one of temporary blindness and one that suggests a visual field defect together with five other less specific reports of blurred or abnormal vision (table 1). These reports were received while monitoring celecoxib and rofecoxib in the New Zealand intensive medicines monitoring programme. This national programme monitors selected newly introduced drugs.1 Monitoring of the COX 2 inhibitors celecoxib and rofecoxib started in December 2000.

Table 1

Case reports of visual disturbance with COX 2 inhibitors

View this table:

Case reports

Case 1

After knee replacement surgery, an 81 year old man in good health was taking 100 mg of celecoxib every morning for analgesia. Three weeks after starting celecoxib, he told his doctor that he had central loss of vision in a jellybean-like shape for a few hours after each dose. His only other treatment was weekly eye drops for blepharitis, which contained sulfacetamide, prednisolone, and phenylephrine. His blood pressure was normal. Celecoxib was discontinued, and the problem has not recurred in seven months.

Case 2

A 78 year old man was prescribed rofecoxib for shoulder pain. He took a first dose of 50 mg one evening and next day took 25 mg in the morning and evening with good relief of pain. The next morning he awoke unable to read the newspaper because his vision was blurred. He attempted to drive to his doctor but ran his vehicle into the back of a truck because he could not see; he fortunately avoided injury. His general practitioner found his visual acuity to be “nil useful vision” in the right eye and 6/18 in the left eye. Later that day the man was reviewed by an ophthalmologist and told that he had only a minor cataract and retinal degeneration consistent with his age. His visual acuity was normal. He took no further rofecoxib and five months later has had no recurrence. There was no rechallenge. At the time of his visual disturbance he was taking 90 mg slow release diltiazem twice a day for mild angina, 10 mg atorvastatin a day, and 100 mg aspirin a day. His total concentration of serum cholesterol was < 4 mmol/l and his blood pressure was normal.

Other cases

General practitioners described three further cases as “blurred vision” and one as “vision reduced” (table 1). The patient in case 4 had had similar symptoms many years before while taking indometacin. The onset of these events was between one week and four months after starting the course. Four of the patients recovered rapidly after stopping the drug, and the outcome was not stated for the fifth. The reason for use was given for all patients. Only one case was for an inflammatory condition (psoriatic arthritis). In one other, the ophthalmologist used the not otherwise specified term “arthritis.”

The patient in case 3 had sudden loss of vision in her right eye, which she described as a central black smudge, and was unable to read. This occurred 26 days after starting celecoxib. She had a complicated ocular history with angle closure glaucoma and peripheral iridectomy of the right eye in 1977, herpes zoster associated keratitis with significant scarring in the right eye also in 1997, long standing corneal surface drying, a right cataract extraction with intraocular lens implantation February 1999, minor opacification of the posterior capsule, and long standing posterior vitreous detachment with floater but no macular damage. The patient was reviewed by her ophthalmologist five days after her sudden loss of vision, and he found no change in the long standing conditions or reason for the sudden change in vision. The ophthalmologist's report made no reference to the visual fields. The patient was still taking celecoxib. A week later she stopped celecoxib and, after two days, was able to read as before.


Case characteristics

These cases show a close association between visual impairment and using celecoxib or rofecoxib. Four of the seven patients had an onset time of one week or less. One patient regularly had problems for a few hours after each dose. Recovery occurred within one or two days of stopping. There appeared to be no confounding by changes to drug therapy. With regard to indication and comorbidity, one patient had inflammatory arthritis and may have been at increased risk of vasculitis or arterial thrombosis, and one patient had mild ischaemic heart disease. Their rapid recovery, however, suggests the absence of vascular embolism or thrombosis in these patients. The symptoms of cases 1 and 3 suggest a visual field defect. In only one patient (case 3) was only one eye affected, but the patient had a complex history of problems in that eye, perhaps making it more susceptible. The ages of the patients indicates that the impairment was not confined to elderly people. Two of the patients saw an ophthalmologist soon after the event. No reason for the acute reduction in vision was established in either.

Reports from other sources

We found a previous report of a 79 year old woman who developed orange spots in both visual fields while taking celecoxib.2 The spots disappeared when she stopped the drug. Of relevance to case 4, in which the patient developed symptoms with both indometacin and celecoxib, we found two published accounts of four cases of visual disturbance associated with the conventional non-selective COX inhibitor ibuprofen.3 4

The database of the World Health Organization Collaborating Centre for International Drug Monitoring contained 230 reports of visual disturbance with celecoxib and 244 with rofecoxib (table 2). Similar events occur with non-selective COX inhibitors. The proportion of all reports with these events is lower for celecoxib and rofecoxib, but this may be due to confounding.

Table 2

Reports of visual impairment with selected non-steroidal anti-inflammatory drugs in the database of the World Health Organization Collaborating Centre for International Drug Monitoring, May 2003

View this table:

Possible mechanism

A likely mechanism for the visual impairment is the inhibition of synthesis of prostaglandins and other related compounds that control retinal blood flow. The vascular endothelium of retinal blood vessels produces compounds, such as prostacyclin, which may increase blood flow,5 and others which may decrease it, such as thromboxane A2 and prostaglandin H2. Both COX 1 and COX 2 mediate synthesis of prostacyclin and other vasoactive prostanoids. Thus inhibition of either COX 1 or COX 2 may alter the cyclo-oxygenase pathway and in turn alter regulation of retinal blood flow with potential changes in vision. The characteristics of the cases are consistent with this mechanism.


Use of celecoxib and rofecoxib is associated with visual impairment. Loss of vision may be marked or severe, but is likely temporary. This effect has also been reported with conventional non-steroidal antiinflammatory drugs, which are non-selective COX inhibitors. Although there are relatively large numbers of reports of visual disturbances associated with use of COX 2 inhibitors in the World Health Organization's database, we have identified only one published account in the medical journals. This reaction is under recognised.


We thank the reporting doctors.


  • Contributors DMC evaluated most of the original case reports, identified the signal, accessed and prepared the data, and revised the manuscript. DWJC made follow up inquiries of the reporters, did literature searches, prepared the references, identified the possible mechanism, and wrote the first draft. RLS evaluated some of the case reports, recognised that there were similar reports with conventional non-selective COX 2 inhibitors, designed the presentation of the case reports, and revised the manuscript.

  • Funding No additional funding.

  • Competing interests DMC has obtained unconditional programme grants for the Intensive Medicines Monitoring Programme from a number of pharmaceutical companies, especially Merck Research Laboratories and Novartis

  • Cyclo-oxygenase-2 inhibitors and conventional non-steroidal anti-inflammatory agents may cause severe visual impairment


View Abstract

Log in

Log in through your institution


* For online subscription