By David Wright and Joshua Parker.
The Pharmacogenetics to Avoid Loss Of Hearing (PALOH) trial has created some interesting and important discussions. Questions regarding what should be considered “routine care”, whether parental choice should alter routine care and the fundamental question of whether consent for genetic testing should be considered differently to non-genetic testing.
However the trial and the following discourse did not specifically address the potential harm of changing antibiotics and why that potential harm creates a moral dilemma that is incredibly difficult and incredibly personal. This is the challenge we raise to the roundtable authors.
We know that some neonates have a single nucleotide polymorphism that predisposes to deafness if given gentamicin. Typically, a baby with suspected neonatal sepsis would be given gentamicin, and a second first line antibiotic, within one hour. PALOH seeks to address whether rapid point of care testing can be used to identify these babies at risk before they are given gentamicin. The idea is that timely identification of these neonates can prevent gentamicin-induced deafness. Some have argued that this testing should simply be considered part of standard care for neonatal sepsis and that there is no reason to gain specific parental consent.
The issues start with the assumption that gentamicin is still the “best” antibiotic to treat neonatal sepsis generally as suggested by the NICE guidance. “Best” treatments will often be evaluated as a combination of a) the treatment most effective in achieving its therapeutic aim and b) the treatment with the most acceptable side effect profile. Indeed that is what NICE did when assessing treatments for neonatal sepsis and recommended gentamicin as first line. The therapeutic aims evaluated included cure rates for neonatal infection, mortality, neonatal adverse events and long-term outcomes. The side effect profile included risk of hearing loss.
The PALOH trial highlights that for a specific group of patients, those with m.1555A>G genetic variant, have an increased risk of hearing loss caused by gentamicin use. By suggesting that an individualised antibiotic regime would be used instead of gentamicin, the premise is that gentamicin is no longer the “best” treatment, and that the increased rate of deafness is of a magnitude that the previously considered balance of efficacy and side effects has tipped away from gentamicin. If the efficacy of gentamicin is measured in cure rates for neonatal infection, mortality, neonatal adverse events and long-term outcomes then the decrease in rates of deafness must be balanced by an increase in the failure of therapy to achieve those aims.
This brings us to a second assumption made when suggesting that changing antibiotics is right; that avoiding deafness is important enough that, despite other potential negative outcomes from alternative antibiotics, it remains a worthwhile aim. There is an open question as to whether deafness is a disadvantage, a disability or a subculture in its own right. As an unresolved question, it makes even more complex answering whether avoiding deafness at all costs is, all things considered, good.
If the first assumption can be challenged, that gentamicin is no longer the “best” treatment for neonatal sepsis, then a move away from gentamicin should be encouraged. This would resolve the issue with large numbers of babies being exposed to gentamicin and the accompanying risk of hearing impairment.
Presuming gentamicin should remain first line then the issue becomes who decides between the decreased risk of deafness or increased antibiotic efficacy? This is the dilemma that has been highlighted, but not created, by the testing for m.1555A>G. If it was clear that either option were substantially better than the other, then parental preference is likely to be less influential as the best interests of the child will determine choice. This would appear to remain the position for babies who do not have the m.1555A>G variant and who would receive gentamicin. But it is not clear, so parental preference may be important, as would be the option not to voice a preference. For parents who do not wish to make the choice between deafness and antibiotic efficacy the knowledge of genetic variation may create a weight of expectation that they would avoid the risk of deafness, regardless of their underlying feelings about the value of deafness as an outcome. The very phrase “Pharmacogenetics to Avoid Loss of Hearing” suggests that a positive genetic test will unambiguously lead to a change in antibiotic regime. If a positive genetic test did not lead to antibiotic change then what would be the point of the genetic test? This means that the step which requires parental involvement, indeed parental consent, is the genetic testing itself.
Paper title: Terrible choices in the septic child
Authors: Dr David Wright.1 and Dr Joshua Parker.2
1) Paediatrics; East Cheshire NHS Trust
2) Wythenshawe Hospital; Manchester University Foundation Trust
Competing interests: None