The first death resulting from bone marrow aplasia induced by chloramphenicol eye-drops was described in 1955 (17). Chloramphenicol causes two types of bone marrow damage (18).
• A frequent, early, dose-related, reversible suppression of the formation of erythrocytes, thrombocytes, and granulocytes (early toxicity).
• A rare, late type of bone marrow aplasia, a hypersus-ceptibility reaction, which is generally irreversible, and has a high mortality rate (aplastic anemia) (19-21).
Chloramphenicol inhibits mRNA translation by the 70S ribosomes of prokaryotes, but does not affect 80S eukaryo-tic ribosomes. Most mitochondrial proteins are encoded by nuclear DNA and are imported into the organelles from the cytosol where they are synthesized. Mitochondria retain the capacity to translate, on their own ribosomes, a few proteins encoded by the mitochondrial genome. True to its prokar-yotic heritage, mitochondrial ribosomes are similar to those of bacteria, meaning that chloramphenicol inhibits protein synthesis by these ribosomes. Chloramphenicol-induced anemia is believed to result from this inhibition (22). Chloramphenicol can also cause apoptosis in purified human bone marrow CD34+ cells (23).
The early, dose-related type of chloramphenicol toxicity is usually seen after the second week of treatment, and is characterized by inhibited proliferation of erythroid cells and reduced incorporation of iron into heme. The clinical correlates in the peripheral blood are anemia, reticulo-cytopenia, normoblastosis, and a shift to early erythrocyte forms. The plasma iron concentration is increased. Early erythroid forms and granulocyte precursors show cytoplasmic vacuolation. After withdrawal, complete recovery is the rule. Leukopenia and thrombocytopenia are less frequent.
Although there is no evidence that these abnormalities progress to frank bone marrow aplasia, continuation of chloramphenicol after the appearance of early toxicity is thought to be hazardous. Pre-existing liver damage (for example due to infectious hepatitis or alcoholism) and impaired kidney function can lead to reduced elimination of chloramphenicol and its metabolites, thereby aggravating marrow toxicity. As a rule, this is not the irreversible type.
Although bone marrow aplasia has not been related with certainty to either the daily or the total dose of chloram-phenicol or to the sex or age of the patients, it has occurred almost exclusively in individuals who were taking prolonged therapy, particularly if they were exposed to the drug on more than one occasion (24). The condition is rare, occurring about once in every 18 000-50 000 subjects in various countries. These variations may in part depend on ethnic factors (25,26). For example, there have been very few cases reported in blacks (27). Bone marrow aplasia due to chloramphenicol has usually resulted in aplastic anemia with pancytopenia; other forms, such as red cell hypoplasia, selective leukopenia, or thrombocytopenia, are less common.
When bone marrow aplasia was complete, the fatality rate approached 100%. As a rule, it has been found that the longer the interval between the last dose of chloram-phenicol and the appearance of the first sign of a blood dyscrasia, the more severe the resulting aplasia. Nearly all patients in whom the interval was longer than 2 months died as a result of this complication. However, fatal aplas-tic anemia can also occur shortly after normal doses of chloramphenicol (28).
The pathogenesis of bone marrow aplasia after chlor-amphenicol is still uncertain. Compared with normal cells, bone marrow aspirates from patients with bone marrow aplasia are relatively resistant to the toxic effects of chloramphenicol in vitro. This has been explained by the hypothesis that during treatment with chloramphenicol, chloramphenicol-sensitive cells were eliminated, leaving behind only a chloramphenicol-insensitive population of blood cell precursors with poor proliferative capacity (29). Chloramphenicol can induce apoptosis in purified human bone marrow CD34+ cells; however, there was no protection from a variety of antioxidants on chloramphenicol-induced suppression of burst-forming unit erythroid and colony-forming unit granulocyte/monocyte in vitro (30). In contrast, a caspase inhibitor ameliorated the apoptotic-inducing effects of chloramphenicol.
Since thiamphenicol, which causes very few cases of aplastic anemia, differs from chloramphenicol by substitution of the para-nitro group by a methylsulfonyl group, interest has been focused on the para-nitro group and metabolites of that part of the molecule, nitrosochlor-amphenicol and chloramphenicol hydroxylamine. In human bone marrow, nitrosochloramphenicol inhibited DNA synthesis at 10% of the concentration of chloramphenicol required for the same effect, and proliferation of myeloid progenitors was irreversibly inhibited. The covalent binding of nitrosochloramphenicol to marrow cells was 15 times greater than that of chloramphenicol
(31). This has lent support to the hypothesis that abnormal metabolism may contribute to the susceptibility to bone marrow aplasia. The production of reduced derivatives by intestinal microbes may contribute to toxicity, but oral administration of chloramphenicol is not essential for the development of aplastic anemia
(32). There is evidence that genetic predisposition may play a role (33,34). The wide geographical variations in the incidence of aplastic anemia may also reflect environmental factors.
For many years it had been said that there were no cases of aplastic anemia after parenteral administration of chloramphenicol; however, a few cases of aplastic anemia have been reported (35). There have also been reports of bone marrow hypoplasia after the use of chloramphenicol eye-drops (36,37).
There is controversy about the risk of aplastic anemia with topical chloramphenicol. In a prospective case-control surveillance of aplastic anemia in a population of patients who had taken chloramphenicol for a total of 67.2 million person-years, 145 patients with aplastic anemia and 1226 controls were analysed. Three patients and five controls had been exposed to topical chloram-phenicol, but two had also been exposed to other known causes of aplastic anemia. Based on these findings, an association between ocular chloramphenicol and aplastic anemia could not be excluded, but the risk was less than one per million treatment courses (38). In another study, a review of the literature identified seven cases of idiosyncratic hemopoietic reactions associated with topical chloramphenicol. However, the authors failed to find an association between the epidemiology of acquired aplas-tic anemia and topical chloramphenicol. Furthermore, after topical therapy they failed to detect serum accumulation of chloramphenicol by high performance liquid chromatography. They concluded that these findings support the view that topical chloramphenicol was not a risk factor for dose-related bone marrow toxicity and that calls for abolition of treatment with topical chloramphenicol based on current data are not supported (39).
In a study using general practitioner-based computerized data, 442 543 patients were identified who received 674 148 prescriptions for chloramphenicol eye-drops. Among these patients, there were three with severe hematological toxicity and one with mild transient leukopenia. The causal link between topical chloramphenicol and hematological toxicity was not further evaluated in detail (40).
In a small fraction of patients who survive the chronic type of bone marrow damage, myeloblastic leukemia develops (41,42). In most instances this complication has appeared within a few months of the diagnosis of aplasia and was considered to be a sequel of chloramphenicol treatment. Sometimes the delay was shorter. The majority were either children or adults aged 50-70 years.
The occurrence of acute leukemia has been studied in relation to preceding use of drugs (before the 12 months preceding the diagnosis) in a case-control study of 202 patients aged over 15 years with a diagnosis of acute leukemia and age- and sex-matched controls (43). Among users of chloramphenicol or thiamphenicol the odds ratio for any use was 1.1 (0.6-2.2) whereas the odds ratio for high doses was 1.8 (0.6-5.3). Other systemic antibiotics showed no substantial relation with the occurrence of leukemia.
Was this article helpful?