Ototoxicity caused by aminoglycosidesBMJ 2007; 335 doi: https://doi.org/10.1136/bmj.39301.680266.AE (Published 18 October 2007) Cite this as: BMJ 2007;335:784
- Maria Bitner-Glindzicz, reader in clinical and molecular genetics,
- Shamima Rahman, senior lecturer in mitochondrial medicine
- Clinical and Molecular Genetics Unit, University College London Institute of Child Health, London WC1N 1EH
Aminoglycoside antibiotics are widely used for the treatment of Gram negative sepsis. It is well known that they can cause dose related renal toxicity and ototoxicity, which occur in almost everyone who receives a sufficiently toxic dose.1 It is less well known that some people have an inherited predisposition that renders them highly sensitive to the ototoxic effects of these antibiotics: aminoglycosides taken at levels that are well within the therapeutic range can result in rapid, profound, and irreversible hearing loss. Even a single dose in a predisposed individual can result in permanent hearing loss.2
In countries that use aminoglycosides widely, a quarter of people with hearing loss induced by aminoglycosides have maternal relatives who also have deafness related to drug induced ototoxicity.3 In the familial cases of hearing loss, individuals received antibiotics for a much shorter period than those without a family history of ototoxicity, suggesting the presence of an inherited predisposing mutation. The most common predisposing mutation is now known as m.1555A>G, a mitochondrial DNA mutation. This will be inherited by every child of a mother who has the mutation as a consequence of mitochondrial DNA being exclusively maternally inherited. This mutation accounts for at least 33-59% of aminoglycoside ototoxicity, according to studies from China, where use of aminoglycosides in the community is widespread owing to their low cost.4 The mutations responsible for the remainder are being studied.
Aminoglycosides exert their antibacterial effects by binding to bacterial ribosomes, leading to errors in bacterial protein synthesis. Human mitochondrial ribosomes bear a structural resemblance to bacterial ribosomes. Mutation at position 1555 of human mitochondrial DNA makes the human mitochondrial ribosome even more similar to the bacterial one, which facilitates aminoglycoside binding. Once bound, aminoglycosides have a long half life in the hair cells of the inner ear (several months), which increases the risk of ototoxicity. How common is the m.1555A>G mutation? To date, no large prevalence studies have been performed and data can only be extrapolated from small studies. In the US state of Texas, screening of blood spots from 1161 newborns found one positive case, and in New Zealand there was one positive case among 206 random blood samples screened (0.48%; 95% confidence interval 0.01 to 2.75)5 6 This prevalence is much higher than previously suspected from calculations of its contribution to childhood deafness.
In the United Kingdom about 1 in 1000 children are born deaf; half of these cases have a genetic cause, with about 80-85% caused by recessive genes, 10% by dominant genes, and 2-5% caused by the m.1555A>G mutation.7 This indicates a prevalence of the m.1555A>G mutation of 1 in 40 000. The discrepancy between this and the prevalence in New Zealand and Texas implies that either the prevalence in the UK is very much lower or penetrance of the mutation is very low, meaning that more people have the mutation but are not deaf. As aminoglycosides in the UK are used only in hospitals, penetrance is likely to be low in the absence of exposure to aminoglycosides. A genuine population frequency of between 1 in 206 and 1 in 1161 would have substantial implications for clinical practice in terms of the numbers of people at risk of ototoxicity.
Even in the absence of exposure to aminoglycosides, some families carrying this mutation may also develop deafness, albeit at a later age and with a lower penetrance. The variable penetrance of the m.1555A>G mutation may be attributable partly to the presence of a modifying nuclear genetic mutation.8 In some populations, the m.1555A>G mutation seems to be a common cause of deafness. In Spain, 27% (19/70) of families with at least two deaf individuals were positive for this mutation.9 Everyone with the mutation who was exposed to aminoglycosides became deaf. The probability of becoming deaf by the age of 30 years if an individual had received such antibiotics was 96.5% compared with 39.9% if they had never been treated. Thus aminoglycosides are a major environmental modifier of the m.1555A>G mutation. Because penetrance of the mutation is very low in some families (0-18%), exposure to aminoglycosides may cause drug induced deafness that may be erroneously categorised as sporadic.10 11
Is it cost effective to screen for this mutation before aminoglycosides are given? Cost effectiveness is determined by the cost of a screening test and the prevalence of the mutation versus the cost of not screening. The current cost of testing for this mutation in the UK is about £35 (€52; $71) per test, based on a small number being performed (generally in those who have already lost their hearing after aminoglycoside administration). However demand for more tests would reduce the unit costs, and single nucleotide genotyping in the commercial sector costs pennies per genotype. Conversely, the cost to the health service of providing a cochlear implant for a child who becomes deaf before acquiring language and of maintaining the implant for 15 years is estimated to be about £47 000 per child, rising to £61 000 over a child's lifetime.12 Educational costs for a profoundly deaf child with a cochlear implant are estimated at about £18 000 a year.13 However, the cost of not providing a cochlear implant to a profoundly deaf child is even greater in terms of educational costs and eventual earning power. In the US, the total lifetime cost to society for a child with prelingual onset of profound deafness has been estimated to exceed $1m.14
Hearing loss induced by aminoglycosides in individuals with the m.1555A>G mutation is, in theory, preventable. The mutation is well known among doctors who see patients who already have hearing loss. However, the general medical community is not aware of this susceptibility and that mutation testing is available through regional genetics centres. We recommend that the true prevalence of the mutation in the UK be ascertained to determine the cost effectiveness of screening everyone prescribed aminoglycoside antibiotics. In the meantime, patients who are likely to receive multiple courses of aminoglycosides—for example, patients with leukaemia and newborns admitted to special care baby units—should be screened.
Genetic testing needs to be turned around rapidly, and consideration should be given to using an alternative antibiotic until the result of genetic testing is known.
Competing interests: The authors have received funding from SPARKS (Sport Aiding Medical Research for Kids) to investigate the prevalence of m.1555A>G in the UK population.
Provenance and peer review: Not commissioned; externally peer reviewed.