The year is 2030. Another busy shift at St. Emlyn’s and you’re asked about your plan for the patient in cubicle 4. You scan your list. The patient attended a few hours ago and you’re planning to discharge home with some pain relief. You tell them you should be able to get that sorted soon. You just need to run the genetics; it should only take a few minutes.
For those of you reading that thinking we’ve lost the plot, bear with us. Today, findings from the Pharmacogenetics to Avoid Loss of Hearing (PALOH) Trial have been published in JAMA Pediatrics.1 This study, the first of its kind anywhere in the world, looked to implement a rapid genetic point of care test (POCT) in the acute setting to guide prescribing practice. The ability to rapidly detect clinically relevant genetic variation by the bedside has the potential to impact how we all practice medicine. Here, we outline the major findings from the PALOH Trial and explain why the concept of “Acute Genetics” isn’t as much of an oxymoron as it first might sound.
The Road to PALOH
Sequencing the long string of letters which make up our genetic code to look for clinically relevant genetic changes has traditionally taken an extremely long time. In the early 1990s, genetic technology had scaled to a point which allowed academics to launch the Human Genome Project.2 This initiative set about to map the DNA sequence of the entire Human genome. The process took 13 years and cost approximately $3 Billion. Less than 20 years on from that point, the same process can now be performed in under 24 hours for less than $1000. We have now sequenced many millions of genomes around the world, giving us a far greater understanding of how genetic changes can influence human health and disease.3
A large proportion of clinically relevant genetic variation is related to rare disease or cancer predisposition syndromes. However, it is increasingly recognised that gene changes can also impact our response to certain medicines, a concept known as pharmacogenetics.4 Furthermore, many gene changes in combination across the genome can contribute to our risk of developing common diseases such as coronary artery disease, asthma, or atrial fibrillation – this is known as our polygenic risk.5
With an increasing awareness of how our genetics can influence our response to medicine, or predict the risk of certain common disorders, the relevance of delivering these data to specialities beyond clinical genetics becomes greater. Many of us carry common genetic changes which influence our response to routinely prescribed medicines, including opioids, antiplatelet medicines, antidepressants, proton pump inhibitors, and antibiotics.6–11 Two pharmacogenetic genes which are particularly important are CYP2C19 and CYP2D6.
Imagine for a moment that the modality or rapidity of genetic testing wasn’t an issue. As a thought experiment, suppose that genomic data was available in an accessible and useful format for all patients either within their health records or via a rapid test. If that were the case, is there a reason you wouldn’t want to know about whether an individual will respond to opioids (CYP2D6) in ED?8 Is there a reason you wouldn’t want genomic data which predicted the most appropriate antiplatelet (CYP2C19) therapy following an ischemic event?6 Or what if you had access to genetic data which could help determine your choice of antibiotic therapy, avoiding severe aminoglycoside induced hearing loss (MT-RNR1)?7 The barrier to implementing this data in the ED arguably isn’t the utility of the data, but our ability to make it available in a clinically interpretable format, in a clinically relevant timeframe.
A perfect example of this is the relationship between aminoglycoside induced ototoxicity (AIO) and a specific change in a mitochondrial gene called RNR1, which was first recognised in 1993.12 Since then, there has been robust and consistent evidence to show that individual’s carrying this gene change are at very high risk of deafness if they receive the antibiotic. This gene change is present in 1 in 500 individuals. Guidance from the Clinical Pharmacogenetics Implementation Consortium (CPIC) and from the UK Medicines and Healthcare products Regulatory Agency (MHRA) states that individuals with the m.1555A>G variant should not receive aminoglycoside antibiotics unless the high risk of permanent hearing loss is outweighed by the severity of infection and lack of safe or effective alternative therapies.7,13 The reason testing hasn’t been deployed in clinical practice isn’t because it’s not clinically important but because, until now, there has been no way to test for the m.1555A>G gene change in the acute setting, where aminoglycosides are frequently prescribed.
The Pharmacogenetics to Avoid the Loss of Hearing (PALOH) trial looked to address this issue. For the study we chose to focus on neonates admitted to Neonatal Intensive Care Unit (NICU), a cohort where aminoglycoside antibiotics are commonly prescribed.14 We developed a rapid genotyping platform for the m.1555A>G variant and assessed whether this technology could be implemented to avoid AIO, without disrupting normal clinical practice on NICUs, via a prospective implementation trial.
The PALOH Study
It’s important to make clear that we are authors on the PALOH manuscript. We were both involved in the conception, design, management, analysis and write up of the project. This is an NIHR funded trial and we worked closely with an industry partner, genedrive, to develop the technology. A complete list of authors, funding and conflicts of interest can be found with the manuscript via JAMA Pediatrics.
What did the study aim to do?
Given the relatively low prevalence of the m.1555A>G variant in the UK, approximately 0.2%, an unfeasibly large sample size would be required to estimate the sensitivity of the rapid test using a traditional prospective diagnostic test accuracy study. Because of this, we evaluated the test in two stages. Firstly, the assay underwent pre-clinical validation via a case-control study followed by an investigator-initiated, pragmatic prospective implementation trial to evaluate the real-world impact of implementing the POCT in the neonatal setting.
Developing a Rapid Point of Care Test
The Genedrive® platform is a rapid thermocycling instrument which programmed to detect the m.1555A>G variant (Figure 1). The input to the system is DNA extracted from a buccal (cheek) swab. The DNA is extracted using a buffer solution which can all be done in less than a minute at the bedside. This short video shows how the system works. The system is able to discriminate between the wild type (m.1555A) and mutant (m.1555G) forms of the RNR1 gene by first amplifying the region containing the gene change using a process known as loop-mediated isothermal amplification (LAMP), before undertaking a second step known as melt analysis. Pre-clinical validation demonstrated an assay sensitivity of 100% (95 % CI 93.9 to 100.0), a specificity of 100 % (95% CI 98.5 to 100.0), and time to generate a genetic result of 26 minutes. The gold standard test, which it was validated against, has a turnaround time of 2-3 weeks.
Figure 1. A) The genedrive platform and B) MT-RNR1 kit contents.
Where was the Technology Implemented?
The trial was undertaken at two large NICUs in the North West of England. Both centres follow NICE guidelines for the treatment of neonatal infection. This means that a large proportion of the babies being admitted will have been being treated with a combination of gentamicin and benzylpenicillin. All neonates admitted to NICU across the two participating sites were eligible for enrolment. Neonates requiring antibiotics immediately, as determined by the admitting clinician, with established intravenous access, were excluded.
How was the Technology Implemented?
We wanted to integrate the system into routine clinical practice, causing as little disruption to normal care as possible, i.e. antibiotics should be administered within one hour (the “golden hour”) of admission to the NICU. The m.1555A>G buccal swab was performed on admission, undertaken by the admitting nurse at the same time as skin swabs which are standardly performed at both sites. Once the MT-RNR1 result was available, it was used to guide antibiotic prescribing, avoiding aminoglycoside antibiotics, and prescribing an alternative (cephalosporin based) regimen, consistent with existing national guidelines, if the m.1555A>G variant was detected.
What were the Outcome Measures?
This study looked to assess how the platform was used in the acute setting. The primary outcome was the number of neonates successfully tested for the m.1555A>G variant as a proportion of all babies who received aminoglycoside antibiotics. We also went onto assess the total number of babies testing positive for the m.1555A>G variant and whether that data could be used to perform tailored prescribing. Critically, we assessed clinical timings related to antibiotic administration to ensure that time-to-antibiotics were equivalent to previous practice. We monitored prescribing practice for a month-long period before implementation as a baseline.
What did the study show?
During the 11-month study window, 751 neonates were recruited from across two centres. The majority, 713, were recruited from a single site as recruitment was paused at the second center and did not re-commence due to the SARS-CoV-2 outbreak. This is important to remember when considering the generalizability of the findings, as discussed below. 80.6% of all neonates prescribed aminoglycoside antibiotics were successfully tested for the m.1555A>G variant. Three neonates with the m.1555A>G variant were identified, and none of these patients went onto receive aminoglycoside antibiotics, avoiding AIO. Critically, before and after implementation of the MT-RNR1 assay there was no significant difference in the mean times to antibiotic therapy.
Hence, the platform was able to detect the m.1555A>G variant in a clinically relevant timeframe. Data from the trial allowed iterative adaptations to the technology, improving efficiency and accuracy. The performance of this updated system will be monitored during any future deployment. Based on our experience, we recommended that implementation assessments such as this should form part of any in vitro diagnostic manufacturer’s development pipeline.
What were the limitations of the trial?
The main limitation is related to the generalizability of the findings. Both NICUs in this study were situated in large academic Teaching Hospitals, providing Level 4 care for the most critically unwell neonates. Both sites engage in regular research activity, which facilitated delivery of the trial, but the same expertise is unlikely to be available at every center. Furthermore, antibiotic prescribing practices show considerable variation both within and between countries. As such, the utility and cost effectiveness of m.1555A>G testing will be context dependent. Local value assessments and implementation programmes should be undertaken as part of a multidisciplinary strategy with stakeholders from NICU, Microbiology, Pharmacy and Clinical Genetics.
The Relevance to Emergency Medicine
Findings from PALOH have relevance to ED and other Acute Care specialities beyond the prescription of aminoglycoside antibiotics. The demonstration that clinicians can use genetic technology in the acute setting to tailor management, without disrupting normal clinical practice is a significant finding. There are a large number of commonly prescribed medicines with so called “gene-drug interactions” where knowledge of genetic variation could influence prescription. The Clinical Pharmacogenetics Implementation Consortium (who can be thought of like the NICE for pharmacogenetics) have published guidelines for the pharmacogenetic guided prescribing of 26 medicines, a figure which is likely to increase in years to come.
Medicines with pharmacogenetic dosing guidelines include opioids, antiplatelets, antibiotics, anti-virals, proton pump inhibitors, anti-depressants, NSAIDs and anti-epileptics.6,8,9,15,16 Not all of these drug groups are regularly prescribed in ED, but many will be. There is a clear argument to be made that having this data in the ED could improve patient outcomes.
Imagine a patient presenting to ED following a fall. You establish they have a soft tissue injury and can mobilise independently. After thorough counselling, you decide to discharge the patient with paracetamol and a limited supply of codeine for breakthrough pain. The aim here is provide sufficient analgesia for the patient to recover and rehabilitate in the community, preventing the need for reattendance with ongoing pain. However, codeine isn’t an active drug. It needs to be metabolised to morphine before it can deliver its analgesic effect. The CYP2D6 enzyme is the most important component of this metabolism. Approximately 1 in 10 people have substantially reduced or absent CYP2D6 activity, caused by a genetic change in the CYP2D6 gene.8 These patients get substantially poorer analgesia from codeine compared to other members of the population.
Imagine if you knew that patient who presented with a fall had reduced CYP2D6 activity. Would that impact your management decision? You might decide to prescribe an alternative analgesic, not metabolised by CYP2D6. This would ensure the patient has better pain relief, promoting rehabilitation and may prevent reattendance. This clinical scenario is just one of many examples where having genetic data to hand could influence treatment decisions and potentially improve patient outcomes.
Pharmacogenetic data has clinical utility in the acute setting and PALOH shows that, packaged correctly, clinicians can use this data without it disrupting normal practice. The challenge is therefore getting relevant genetic information into the hands of clinicians in a useable, timely and cost-effective format. This will most likely involve a multi-faceted strategy, including the development and deployment of more rapid POCTs. The COVID-19 pandemic has seen companies invest heavily in their rapid testing portfolios. It is highly likely that at least some of this innovation will be re-directed towards the detection of clinically relevant human genetic variation.17 As such, the next few years may see the emergence of novel POCT technologies with even faster turnaround times.
Aside from POCTs, it is highly likely that genetic data will increasingly become integrated into patients’ electronic health records (EHRs). This would eliminate the need for a test to be done at the point of prescription. Rather, pharmacogenetic data would be available pre-emptively and smart decision tools built within EHRs could dramatically reduce the requirement for clinical interpretation. If you tried to prescribe codeine to that patient after their fall, the system would alert you that it was unlikely to be efficacious. Again, this may seem farfetched, but this is already happening at many centres around the world and there is a commitment to make this a reality in the NHS.18–21
There is strong evidence which shows that genotype guided prescribing can improve patient outcomes and avoid harm. The issue we have is getting the data to the clinicians in a clinically relevant format and a clinically relevant timeframe. It’s not about utility, it’s about time. The success of the PALOH trial demonstrates that we’re on the cusp of overcoming this issue. Over the next decade is highly likely we will see genetic data made increasingly available to support clinical decision across healthcare, including the ED.
1. McDermott JH, Mahaveer A, James RA, et al. Rapid Point-of-Care Genotyping to Avoid Aminoglycoside-Induced Ototoxicity in Neonatal Intensive Care. JAMA Pediatrics [Internet] 2022 [cited 2022 Apr 11];Available from: https://doi.org/10.1001/jamapediatrics.2022.0187
2. Hood L, Rowen L. The Human Genome Project: big science transforms biology and medicine. Genome Med 2013;5(9):79.
3. Koch L. Exploring human genomic diversity with gnomAD. Nat Rev Genet 2020;21(8):448–448.
4. Daly AK. Pharmacogenetics: a general review on progress to date. Br Med Bull 2017;124(1):65–79.
5. Khera AV, Chaffin M, Aragam KG, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet 2018;50(9):1219–24.
6. Lee CR, Luzum JA, Sangkuhl K, et al. Clinical Pharmacogenetics Implementation Consortium Guideline for CYP2C19 Genotype and Clopidogrel Therapy: 2022 Update. Clin Pharmacol Ther 2022;
7. McDermott JH, Wolf J, Hoshitsuki K, et al. Clinical Pharmacogenetics Implementation Consortium Guideline for the Use of Aminoglycosides Based on MT-RNR1 Genotype. Clinical Pharmacology & Therapeutics [Internet] [cited 2021 Dec 21];n/a(n/a). Available from: https://onlinelibrary.wiley.com/doi/abs/10.1002/cpt.2309
8. Crews KR, Gaedigk A, Dunnenberger HM, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for codeine therapy in the context of cytochrome P450 2D6 (CYP2D6) genotype. Clin Pharmacol Ther 2012;91(2):321–6.
9. Lima JJ, Thomas CD, Barbarino J, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for CYP2C19 and Proton Pump Inhibitor Dosing. Clin Pharmacol Ther 2020;
10. Hicks J, Bishop J, Sangkuhl K, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for CYP2D6 and CYP2C19 Genotypes and Dosing of Selective Serotonin Reuptake Inhibitors. Clin Pharmacol Ther 2015;98(2):127–34.
11. Hicks JK, Sangkuhl K, Swen JJ, et al. Clinical pharmacogenetics implementation consortium guideline (CPIC) for CYP2D6 and CYP2C19 genotypes and dosing of tricyclic antidepressants: 2016 update. Clin Pharmacol Ther 2017;102(1):37–44.
12. Prezant TR, Agapian JV, Bohlman MC, et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic–induced and non–syndromic deafness. Nature Genetics 1993;4(3):289–94.
13. Medicines and Healthcare products Regulatory Agency. Drug Safety Update. 2021;14(6).
14. McDermott JH, Mahood R, Stoddard D, et al. Pharmacogenetics to Avoid Loss of Hearing (PALOH) trial: a protocol for a prospective observational implementation trial. BMJ Open 2021;11(6):e044457.
15. Theken KN, Lee CR, Gong L, et al. Clinical Pharmacogenetics Implementation Consortium Guideline (CPIC) for CYP2C9 and Nonsteroidal Anti-Inflammatory Drugs. Clin Pharmacol Ther 2020;108(2):191–200.
16. Karnes JH, Rettie AE, Somogyi AA, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for CYP2C9 and HLA-B Genotypes and Phenytoin Dosing: 2020 Update. Clin Pharmacol Ther 2021;109(2):302–9.
17. McDermott JH, Burn J, Donnai D, Newman WG. The rise of point-of-care genetics: how the SARS-CoV-2 pandemic will accelerate adoption of genetic testing in the acute setting. Eur J Hum Genet 2021;29(6):891–3.
18. Pulley JM, Denny JC, Peterson JF, et al. Operational implementation of prospective genotyping for personalized medicine: the design of the Vanderbilt PREDICT project. Clin Pharmacol Ther 2012;92(1):87–95.
19. Hicks JK, Dunnenberger HM, Gumpper KF, Haidar CE, Hoffman JM. Integrating pharmacogenomics into electronic health records with clinical decision support. Am J Health Syst Pharm 2016;73(23):1967–76.
20. Rasmussen-Torvik LJ, Stallings SC, Gordon AS, et al. Design and Anticipated Outcomes of the eMERGE-PGx Project: A Multi-Center Pilot for Pre-Emptive Pharmacogenomics in Electronic Health Record Systems. Clinical pharmacology and therapeutics 2014;96(4):482.
21. Hoffman JM, Haidar CE, Wilkinson MR, et al. PG4KDS: A Model for the Clinical Implementation of Pre-emptive Pharmacogenetics. Am J Med Genet C Semin Med Genet 2014;0(1):45–55.