This study has shown that a deep neural network can detect referable diabetic retinopathy with sensitivities and specificities close to or better than 80% from both an international and a domestic (New Zealand) database. We believe that deep neural networks can be integrated into community screening once they can successfully detect both diabetic retinopathy and diabetic macular oedema.

AimsTo determine ifVery low dose mydriatic eye microdrop regimen sufficiently dilates the pupil (above 4.1 mm) compared with the currently used low dose mydriatic eye microdrop regimen.Cardiovascular, gastrointestinal and respiratory adverse effects occur following eye drop instillation.MethodsSeventeen premature infants were recruited into this prospective, randomised controlled pilot trial in January 2017 to November 2018. Data were collected from the single-centre Neonatal Intensive Care Unit, Dunedin Hospital, New Zealand. The inclusion criteria were birth weight less than 1500 g or gestational age less than 31 weeks, or any premature infant requiring red reflex testing. Infants were randomised to receive either phenylephrine 1% or 0.5% and cyclopentolate 0.2% or 0.1%, 1 microdrop in both eyes. Efficacy outcome measures were pupil size at retinopathy of prematurity eye examination (ROPEE) and ophthalmologist rating of ease of screen.ResultsAll participants had sufficient pupillary dilation for a successful ROPEE. Ophthalmologists rated the ROPEE as easy for 90% of all examinations. Pupil dilation measurements at the time of examination, mean±SD, 4.8±0.2 (95% CI 4.5 to 5.2) mm for treatment A and 5±0.2 (95%CI 4.6 to 5.4) mm for treatment B (p=0.61). There were no statistically significant differences between the groups for safety data.ConclusionsVery low dose microdrop administration of phenylephrine and cyclopentolate appears to be effective at sufficiently dilating the neonatal pupil for ROPEEs. Low dose and very low dose microdrop mydriatic regimens may also reduce the risk of unwanted adverse effects associated with these medicines.Trial registration numberAustralian New Zealand Clinical Trials Registry (reference ACTRN12616001266459p).

No abstract

Objective This study was aimed to determine mydriatic regimen(s) used in neonatal units in Aotearoa New Zealand (NZ) and Australia and to estimate the frequency of adverse drug events following mydriatic administration in preterm neonates. Study Design A cross-sectional survey was sent to neonatal nursing staff listed in the Australian and New Zealand Neonatal Network contact list. Participants were asked to state what mydriatic regimen they use, and to estimate the frequency of adverse drug events when eye drops were administered for retinopathy of prematurity eye examinations (ROPEE). Results Thirteen different mydriatic regimens were identified; phenylephrine 2.5% and cyclopentolate 0.5% (1 standard drop of each) was the most commonly used regimen. Two of the regimens exceeded adult doses and five regimens included a mydriatic that is equivalent to an adult dose. Following mydriatic instillation, the three most common adverse effects were apnea, tachycardia, and periorbital pallor. Conclusion Low-concentration single-microdrop regimens are currently in use and resulting in successful ROPEE, yet doses exceeding adult doses are in use throughout Aotearoa NZ and Australian units. We know from this dataset that neonates are experiencing unwanted and potentially preventable, adverse effects associated with mydriatics, and every effort should be made to minimize this risk. Key Points

ObjectiveTo determine if very low dose (VLD, 0.5% phenylephrine, 0.1% cyclopentolate) mydriatic microdrop (approximately 7 μL) administration (up to three doses) is non-inferior to low dose (LD, 1% phenylephrine, 0.2% cyclopentolate) mydriatic microdrop administration for ophthalmologist-determined successful retinopathy of prematurity eye examination (ROPEE).DesignMulticentre, prospective, randomised controlled, non-inferiority clinical trial.SettingFour neonatal intensive care units in Aotearoa, New Zealand from October 2019 to September 2021.PatientsInfants with a birth weight less than 1250 g or gestational age less than 30+6 weeks and who required a ROPEE.InterventionsThe intervention: microdrop (approximately 7 μL) of VLD (0.5% phenylephrine and 0.1% cyclopentolate) to both eyes, or the comparison: microdrop of LD (1% phenylephrine and 0.2% cyclopentolate) to both eyes. Up to three doses could be administered.Main outcome measuresThe primary outcome measure was an ophthalmologist-determined successful ROPEE.ResultsOne hundred and fifty preterm infants (LD mean GA=27.4±1.8 weeks, mean birth weight=1011±290 g, VLD mean GA=27.5±1.9 weeks, mean birth weight=1049±281 g,) were randomised. Non-inferiority for successful ROPEE was demonstrated for the VLD group compared with the LD group (VLD successful ROPEE=100%, LD successful ROPEE=100%, 95% CI no continuity correction −0.05 to 0.05) and for Māori (95% CI no continuity correction −0.02 to 0.19).ConclusionVLD microdrops enable safe and effective screening for ROPEE in both Māori and non-Māori preterm infants.Trial registration numberACTRN12619000795190.