Rapid Point-of-Care Genotyping to Avoid Aminoglycoside-Induced Ototoxicity in Neonatal Intensive Care

This pragmatic prospective implementation trial uses a rapid genotyping platform for the m.1555A>G variant to assess whether this technology could be implemented to avoid aminoglycoside-induced ototoxicity without disrupting normal clinical practice on neonatal intensive care units.

This supplemental material has been provided by the authors to give readers additional information about their work. eAppendix.

Background
The Genedrive® MT-RNR1 ID Kit was developed for use as In vitro diagnostic (IVD) molecular assay for the detection of the single nucleotide polymorphism (SNP) m.1555A>G affecting the mitochondrial gene  in human buccal cells. The Genedrive® platform is a portable, rapid thermocycling, point of care test (POCT) instrument which we programmed to perform loop-mediated isothermal amplification (LAMP) nucleic acid amplification followed by end-point fluorescent hybridisation probe-based melt analysis detection to enable allelic discrimination.
The CE-IVD assay was designed to amplify the m.1555A>G region from a buccal swab sample suspended in a buffer solution, following reconstitution of lyophilised assay reagents contained in a sample cartridge (2-30 o C storage), and LAMP before genotyping wild type (WT) and mutant (VAR) alleles by automated discrimination of the respective allele hybridisation probe melt temperature (Tm) position (eFigure 1). Sample analysis is automated without any user interpretation, providing the user with a simple "detected" or "not detected" actionable result in 26 mins of initiating sample analysis (~30 mins from sample collection to actionable result). The IFU and instructional video is available at https://www.genedrive.com/assays/rnr1-product.php).
Each individual Genedrive MT-RNR1 ID Kit contains all the reagents and accessories required for the detection of (SNP) m.1555A>G affecting the mitochondrial gene MT-RNR1 in human buccal cells. The Genedrive MT-RNR1 ID Kit is designed for use in conjunction with the Genedrive instrument and an android smartphone device running the Genedrive Connect application (in an integrated device version for commercial roll out). The Genedrive MT-RNR1 ID Kit contains a buccal swab and sample collection tube containing lysis buffer for the collection of buccal cells.
The Genedrive assay cartridge contains all the reagents needed to perform amplification and detection, in a lyophilised format. The lyophilised reagents are reconstituted with the lysis buffer using the Minivette. The cartridge is then closed and inserted into the Genedrive instrument, where the required metadata is inputted via the Genedrive Connect Application, prior to initiating the test. The test consists of a target DNA amplification step and an end-point melt curve analysis. The data is interpreted by the instrument software which displays the result interpretation clearly on the screen.
Genedrive MT-RNR1 ID Kit should be stored between 2°C to 30°C. The expiry date for each Genedrive MT-RNR1 ID Kit is identified on the kit label. Expired kits must not be used. Once the Genedrive assay cartridge has been removed from the mylar pouch and the strip cap has been removed it is stable up to 40°C for 3 hours.

Assay Inclusivity
Using a panel of 92 genomic DNA samples (Public Health England, Ethnic Diversity panel EDP-1, Catalogue No: 07020701) representing a number of ethnic backgrounds including: African, Oriental, Australian Aborigine, Thai, Italian, French, South American Indian, Japanese and Ashkenazi Jewish, the MT-RNR1 assay was shown to have 100% inclusivity across all ethnicities.

Analytical sensitivity
The analytical sensitivity, or Limit of detection (LoD), is defined as the lowest concentration which >95% of the tested samples generate a positive result and was determined to be 16 cells.
Whilst the clinical significance of reported heteroplasmy is unclear, the Genedrive MT-RNR1 ID Kit can detect the variant allele (m.1555G) at a level of 10% in a background of non-variant (m.1555A), with 100% detection rate.

Interfering substances
The effect of potentially interfering substances on the Genedrive MT-RNR1 ID Kit was evaluated using 18 potentially interfering substances spiked into adult and neonatal/infant buccal samples of non-variant (1555A) at a level predicted to be above the concentration likely to be found in a buccal sample. The substances tested are shown in eTable 1.

Exclusivity (Cross reactivity)
The effect of potential cross reactive commensal organisms was evaluated on the Genedrive MT-RNR1 ID Kit  • Group 1 CONTROLS -Individuals who are already known to not carry the m. 1555A>G genetic variant, as previously tested on a blood sample through the normal clinical laboratory process. This group comprised seventy-four individual participants and 138 individual specimens.
• Group 2 CASES -Individuals who are already known to carry the m.1555A>G genetic variant as this will have previously been tested on a blood sample through the normal clinical laboratory process. These individuals collected buccal cells themselves remotely following instructions supplied. This group comprised thirty-two individual participants and sixty-two individual specimens.
• Group 3 CONTROLS -Children on the neonatal intensive care unit (NICU). These samples will be used to ensure there are no factors specific to neonatal swab sampling that would impair the assay. This group comprised fifty-five individual participants and 110 individual specimens.
In this case-control, the Genedrive MT-RNR1 ID Kit was validated for use in adult and neonatal populations to detect the MT.1555A>G SNP associated with aminoglycoside induced hearing loss. The sensitivity of the test is determined to be 100 % (95 % CI 93.9 to 100.0) and the specificity is determined to be 100 % (98.5 to 100.0), using sanger sequencing as the reference method. The samples which had failed genotyping ("Test Fail") during the PALOH trial were retrospectively reviewed.

Cohort following Excluded samples
There was no association between failed genotyping and participant age, length of time between swab and testing, system, system user, assay batch or the time of day the test was performed. Subsequently, the melt peak data for those samples which had failed genotyping were retrospectively reviewed. A consistent feature of those samples which failed genotyping was a reduced signal intensity below acceptance threshold, meaning the peaks were not being identified by the system software and genotyping would fail. This appeared to be a root-cause of the increased "Test Fail" rate. Rather than lowering the threshold for a peak to be considered detected, which could reduce assay sensitivity, the assay buffer formulation was modified to improve analytical sensitivity, and subsequently lowered the assay LoD from 16 cells to 0.4 cells.
This version 2.0 (V2.0) chemistry was then tested on 35 patient samples which had all previously failed to generate a genotype (i.e. Test Fail) during the PALOH implementation trial. All samples irrevocably anonymised prior to repeated testing. The peak heights and outcome from each genotyping attempt was monitored. All samples were tested in a controlled laboratory setting and in the clinical setting, where the PALOH trial was performed.
Repeated testing of the 35 previously failed samples showed a significant improvement in the performance of the system, both in the laboratory and clinical settings. The average peak height was significantly increased using V2.0 of the MT-RNR1 assay (eFigure 2) which translated into a decreased Test Fail rate. In the laboratory setting, all of the previously failed samples were successfully genotyped (failure rate 0.0%) and in the clinical setting only 2 of 35 samples failed genotyping (failure rate 5.7%)

False Positive Event Investigations
Repeat testing of samples leading to false positive events ruled out sample contamination. False positive results were associated with irregular melting profile traces, with spikes and dips introduced into the fluorescent melt profile. Fluorescence spikes resulted in artefactual "peaks" passing subsequent quality metrics and fluorescent dips or dips resulted in possibility of masking true melt peaks, and therefore also potentially contributing to "test fail" results observed. These were determined to arise from incomplete assay cartridge insertion, resulted in vibrational derived artefacts and light ingress to the optical reader system during assay melt phase. The cartridge and cartridge inlet design were improved, with physical clip and auditory feedback to the user to ensure that these were inserted correctly. Following verification and confirmation of new design improvements, these were validated in the laboratory and intended user setting as above. Instructions for use (IFU) and a user video, demonstrating ideal use of the system including cartridge insertion, are available at https://www.genedrive.com/assays/rnr1-product.php.