Use of chloramphenicol as topical eye medication: time to cry halt?

Bone marrow aplasia also occurs with ocular use

Chloramphenicol accounts for over half the general medical services prescriptions for ocular antibiotics in the Republic of Ireland, while in the United Kingdom 55% of patients presenting to general practitioners with “red eyes” are treated with chloramphenicol eye ointment.1 The British National Formulary currently recommends chloramphenicol as the drug of choice for superficial eye infections. It has the advantage that it has a broad spectrum of activity and rarely causes local irritation or hypersensitivity, which may be a problem with other antibiotics. Yet on our wards we no longer prescribe topical ocular chloramphenicol. Why?

Since 1950, when Rich et al highlighted the relation between oral chloramphenicol and bone marrow aplasia,2 doctors have been well aware of this side effect. The first death resulting from bone marrow aplasia induced by chloramphenicol eye drops was described by Rosenthal and Blackman in 1955.3 Numerous subsequent articles directly implicated chloramphenicol eye drops in causing bone marrow aplasia,4 5 6 7 and after the publication of Fraunfelder et al's paper6 in 1982 the sale of chloramphenicol eye drops in the United States fell by 80% in two years.8 By 1993 the national register of drug induced ocular side effects (in Oregon) had received reports of 23 patients with blood dyscrasias that might have been related to topical ocular administration of chloramphenicol.8

This caution over the association is important since there are several problems in directly implicating ocular chloramphenicol as a cause of blood dyscrasias. Firstly, clinicians do not usually ask about eye drops or ointment when taking a drug history; secondly, many patients do not realise that they have used ocular chloramphenicol; thirdly, blood dyscrasias may develop weeks to months after treatment has been stopped9; and, finally, some of the reactions are idiosyncratic. All these difficulties underline the importance of an extensive drug history.

Oral chloramphenicol causes two types of bone marrow toxicity.6 The first is a dose related reversible depression generally affecting erythroid cells, which is thought to be mediated by mitochondrial injury; the second is an idiosyncratic reaction, which affects all three cell lines and is generally fatal. In the second type a nitroreduction derivative of chloramphenicol produced by a predisposed patient is thought to induce damage to DNA in stem cells, leading to blood dyscrasias. The ultimate toxic derivative of chloramphenicol may be produced in the marrow in situ, thus making the marrow both the site of metabolic conversion and the target of injury. Genotoxicity in stem cells may make them non-proliferative or, alternatively, may initiate marrow damage mediated by the immune system. The occurrence of aplastic anaemia with the small doses characteristic of topical chloramphenicol is thus consistent with the hypothesis of an individual metabolic predisposition.8 Ocular chloramphenicol may therefore cause this idiosyncratic reaction only in genetically predisposed people, and a personal or family history of blood dyscrasias may be very important.

At present there are no accurate figures to estimate the precise risk of marrow aplasia after the ocular administration of chloramphenicol. We do, however, know that the overall risk of developing aplastic anaemia after oral administration of chloramphenicol is 1 in 30000 to 1 in 50000,6 which is 13 times greater than the risk of idiopathic aplastic anaemia in the population as a whole. Since topical administration achieves systemic effects by absorption through the conjunctival membrane10 or through drainage down the lacrimal duct with eventual absorption from the gastrointestinal tract, we can postulate that the risk may be similar to that after oral administration of the antibiotic.

Although the numbers of documented cases of aplastic anaemia associated with topical chloramphenicol are few, the tragic consequences in previously healthy patients cannot be ignored.11 From our review of published reports we find it difficult to justify subjecting patients to this potential risk—except when the ocular infection is resistant to all other available antibiotics. The prescribing habits of doctors on the other side of the Atlantic support this approach. In the United States the Physician Desk Reference emphasises with repeated boxed warnings the importance of not using ocular chloramphenicol unless there is no alternative. Framycetin and fusidic acid are but two of many safer and equally effective preparations in the treatment of most superficial eye infections. All doctors should follow the example of their American colleagues in restricting their use of ocular chloramphenicol.