Article Text

Download PDFPDF

Nitrous oxide analgesia during retinopathy screening: a randomised controlled trial
  1. Romain Mandel1,
  2. Nabeel Ali1,
  3. John Chen2,
  4. Ivan John Galic2,
  5. Linda Levesque3
  1. 1Department of Pediatrics, Division of Neonatology, McGill University Health Centre, Montreal, Quebec, Canada
  2. 2Department of Ophthalmology, McGill University Health Centre, Montreal, Quebec, Canada
  3. 3Department of Respiratory Therapy, Jewish General Hospital, Montreal, Quebec, Canada
  1. Correspondence to Dr Nabeel Ali, Department of Pediatrics, Division of Neonatology, McGill University Health Centre, Royal Victoria Hospital, Room c7.80, 687 Pine Avenue West, Montreal, QC H3A 1A1, Canada; nabeel.ali{at}


Objectives To determine if the addition of an inhaled equimolar mixture of nitrous oxide (N2O) and oxygen (EMONO) would produce superior pain relief to standard pharmacological and non-pharmacological measures during eye examination screening for retinopathy of prematurity (ROP) in premature infants.

Study design A randomised, double-blind controlled trial was conducted.

Setting Royal Victoria Hospital, a tertiary neonatal intensive care unit in Montreal, Canada.

Patients Stable spontaneously breathing premature infants with birth weights less than 1500 g or gestation of 30 weeks and less.

Intervention During the eye examination, all infants were swaddled, received oral sucrose and topical anaesthetics. Control group infants received a mixture of 50% oxygen and 50% nitrogen (n=18) administered by nasal cannula, while the intervention group received EMONO (50% oxygen and 50% N2O).

Main outcome measures Pain was assessed by the premature infant pain profile (PIPP).

Results The mean PIPP score at speculum insertion in the control group (8.4, 95% CI 7.6 to 9.3) was comparable with the EMONO group (8.5, 95% CI 7.3 to 9.8) with a p value of 0.94. There were no significant differences in heart rate or saturation between the two groups. EMONO inhalation was tolerated without any measured side effects.

Conclusion EMONO does not produce any additional pain relief over currently used measures during ROP screening eye examinations. Systematically combining pharmacological and non-pharmacological treatment modalities appears to be the best option until newer treatments are proven effective.

Clinical trials registration number NCT00623220

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Early detection of retinopathy of prematurity (ROP) in preterm infants can allow for interventions that may improve visual outcomes. The American Academies of Pediatrics and Ophthalmology and the American Association for Pediatric Ophthalmology and Strabismus recommend retinal examinations performed by a trained ophthalmologist for the screening of ROP.1 In accordance with these recommendations, preterm newborns with birth weights <1500 g or gestational age of ≤30 weeks are screened. Retinal examinations can be repeated as needed.1

Multiple studies confirm that ROP examinations are painful and have deleterious physiological effects that can last up to 24 h after the examination.2,,8 Furthermore, it seems that the act of lid speculum insertion is the most painful part of the examination.2 4 6 8,,10 The pain associated with retinal examination is universally recognised and warrants treatment.11

What is already known on this subject

  • Eye examination for screening of retinopathy of prematurity (ROP) is known to be painful to premature infants.

  • Current pharmacological and non-pharmacological pain control measures are only partially effective.

  • Nitrous oxide (N2O) is commonly is used in paediatrics for minor procedures but has not been studied for ROP screening.

What this study adds

  • Inhaled nitrous can be administered during eye examination and is well tolerated.

  • Inhaled N2O does not provide any additional pain relief over current available methods.

Non-pharmacological and pharmacological interventions have been studied as methods for reducing the pain associated with ROP examinations. Non-pharmacological interventions including replacement of binocular indirect ophthalmoscopy with the Retcam 120, nesting the newborn and using the Newborn Individualized Developmental Care Assessment Program have shown contradictory results.12,,16 Topical anaesthetics have shown mixed results.8 10 17 Three of five studies looking at the efficacy of oral sucrose have shown significant reduction in pain.2 6 18,,20 A systematic review by Sun et al21 of these studies showed a statistically significant beneficial effect with sucrose. Although an average reduction of the premature infant pain profile (PIPP) score of 1.38 was noted, the scores remained above 10, which is relatively high.22 The clinical relevance of such a small improvement is debatable.

Inhaled nitrous oxide (N2O) is a γ aminobutyric acid agonist with analgesic and anaesthetic properties, with minimal effects on respiratory drive.23 Animal studies have not shown any significant neuronal apoptosis or cellular hypoxia when N2O is used alone.24 An equimolar mixture of nitrous oxide and oxygen (EMONO) is composed of 50% oxygen and 50% N2O. EMONO has been widely used for paediatric procedural sedation.25 In premature infants, it has been studied as premedication for intubation, and as an analgesic in conjunction with spinal anaesthesia during inguinal hernia repair.26 27 Side effects including agitation, vomiting and apnoea were described in 7 of 26 premature infants in the study by Milesi et al.26 EMONO is easily accessible and does not require intravenous access.

The main objective of our study was to determine if inhaled EMONO would reduce pain associated with retinal examination in premature infants when compared with current standard treatments (swaddling, sucrose and topical anaesthesia).



We conducted a randomised controlled trial at the Royal Victoria Hospital, McGill University Health Center, Montreal, Canada. Local institutional research ethics board approval was obtained. Written informed consent was obtained from parents of participants. All inborn or outborn infants admitted to the neonatal intensive care unit (NICU) with birth weights <1500 g or gestational ages ≤30 weeks who were still alive at 72 h of life were screened for eligibility. Patients were excluded if they had craniofacial malformations, significant cardiac lesions, pneumothorax, congenital pulmonary malformations or neuromuscular disease. Extremely premature infants receiving palliation were excluded.

Study inclusion was delayed if the infant had received opiates or benzodiazepines within the week preceding the eye examination. A minimum period of 7 days free from sedatives was required before inclusion. Infants were not studied if they were receiving nasal or intratracheal ventilation. In such cases, the study was deferred until the next eye examination because our EMONO delivery system was not compatible with mechanical ventilators.


Patients were studied in the NICU. Both control and intervention groups received care in accordance with the standards for pain control in our unit at the time of study. All infants were swaddled and restrained by a nurse to avoid accidental dislodgement of the speculum.28 All infants received a 24% solution of sucrose directly in the mouth without a pacifier. The first dose of 0.1 ml (24 mg) was given immediately before the local anaesthetic. Repeat doses were given at the discretion of the infant's nurse, up to a maximum of 0.3 ml (75 mg). No minimal interval between sucrose doses was required. All infants received one drop of proparacaine HCl 0.5% in each eye 1 min before examination.

Intervention group infants received inhaled EMONO available in premixed tanks (Praxair Canada, Mississauga, Ontario, Canada). Gas was delivered by nasal cannula (Airlife Infant Cushion Nasal Cannula; Cardinal Health, McGraw Park, Illinois, USA) at a flow rate of 4 l/min. The infant's mouth was gently held closed during inhalation. The control group received a mix of 50% oxygen and 50% nitrogen via the same device at 4 l/min. Since there was no established neonatal dosing in the literature, we chose a flow rate similar to that used by Milesi et al.26 A dose of 4 l/min puts the gas delivery well above the infant's spontaneous minute ventilation and provides saturation of N2O delivery assuming there was minimal leakage through the mouth or around the cannula. A pilot study of six patients prior to the main trial was conducted to determine the optimum method of gas delivery. After evaluating the nasal cannula, the clear plastic resuscitation mask and the plastic non-rebreather mask, we chose the cannula as the preferred delivery method since it was the only device that allowed us to satisfactorily evaluate the infant's face and permitted the ophthalmologist sufficient freedom to adequately conduct the examination. A scavenging hood was placed around the head, connected to suction at −120 mm Hg. This procedure was required by our occupational health and safety department to keep ambient N2O levels in the NICU below provincial allowances.

Infants' pupils were dilated 30–60 min before eye examination with topical phenylephrine 2.5% and cyclopentolate 1.5%, one drop in each eye. Infants were placed on cardiorespiratory monitoring and continuous pulse oximetry. Eye examinations were conducted by one of two retinal specialists with more than 5 years of experience in examining preterm infants. Gas inhalation was initiated 5 min before eye examination, and the examination was started once the infant was calm. The eyelid was kept open with a metal lid speculum, and the retina was visualised by binocular indirect ophthalmoscopy. A sterile cotton-tipped applicator was used for scleral depression. We evaluated only the pain response during the examination of the first eye, because both eyes were examined in rapid succession, and the pain score might not return to baseline levels between both eyes. We decided to stop inhalation after the examination of the first eye as a precaution to limit unnecessary exposure to N2O or oxygen.

Primary outcome

The effect of EMONO on pain during eye examination was measured using the PIPP score. The PIPP score is a multidimensional composite pain score developed and validated in clinical settings used for evaluating acute procedural pain in preterm neonates.22 It measures seven different elements including physiological parameters, facial expression, behaviour and gestational age. Each is evaluated on a scale of 0–3, yielding a combined score ranging from 0 to 21. Two trained nurses independently scored the infant's pain response at the time of lid speculum insertion in the first eye and 30 min after the eye examination. The results were averaged. Due to the number of personnel surrounding the infant as well as the scavenging system, it was not possible for us to videotape the infant's facial expression in a reliable manner.

Sample size and randomisation

Prior to study initiation, patient charts from our unit were reviewed over a 6-month period, and routine PIPP scores determined by nurses during eye examination were collected, yielding an average of 13.4±3.4 in 30 patients. Based on these results, we calculated that 20 patients were required in each group to show a 25% reduction in PIPP using a two-tailed α level of 0.05 and a power of 0.8. Although this is a somewhat large effect size, a PIPP score reduction below 10 is considered significant and clinically relevant.22

Randomisation sequence was predetermined using a computerised random number generator and allocation was unknown to investigators. Allocation was known only to a respiratory therapist in charge of gas administration but not involved in pain scoring. Apparatus for delivering control or treatment gas was hidden from study personnel by a shroud.

Secondary outcomes

PIPP score was measured the morning of the eye examination before mydriatic drop instillation, after eye drop instillation and 24 h after the examination. Heart rate, respiratory rate and cuff blood pressure were monitored using neonatal monitors: Draeger Infinity Gamma XL (Draeger Medical Canada, Richmond Hill, Ontario, Canada) or HP Agilent V29C/HP 895 (Hewlett Packard Company, Palo Alto, California, USA). Data were manually recorded from the monitor the morning of the eye examination before eye drops, after eye drop instillation, during eye examination, 30 min after eye examination and 24 h after eye examination. Blood pressure was not measured during eye examination due to excessive movement artifact.

Continuous pulse oximetry was performed for the 24 h preceding and following the eye examination with a Masimo Radical 7 pulse oximeter (Masimo, Irvine, California, USA) using a 2-s averaging time. Data were downloaded using Profox oximetry software 2006 (Escondido, California, USA).

Apnoea was defined as the absence of chest movement on the impedance monitor for 15 s or more.29 Low saturation was reported as total time spent at a saturation below 88%. Our unit policy is to maintain oxygen saturation levels for infants on supplemental oxygen between 88% and 92% until 34 weeks corrected gestational age, and above 94% for older infants.

The Clinical Risk Index for Babies II score at admission was calculated to compare severity of illness in both groups.30 Intraventricular haemorrhage severity was defined using Papile's classification.31

Statistical methods

Continuous outcome variables were compared between the two groups using the two sample t test. Categorical variables were compared using the χ2 test. Differences were considered significant if p<0.05. Inter-rater agreement for PIPP scores was evaluated using the intraclass correlation coefficient.


Patients were recruited from February 2007 to April 2010. One hundred and forty patients were screened and 40 were randomised (figure 1). Two patients were excluded due to hydrocephalus and one had pulmonary hypoplasia. Twenty of the patients died, 10 in the delivery room or within the first 72 h after admission due to extreme prematurity. The other 10 died during hospitalisation but before their first eye examination. Twenty-one patients were transferred to another centre before their first examination, either for subspecialised care or to be relocated closer to home. One infant could not be randomised because she remained on mechanical ventilation until her last eye examination was performed. Two eligible infants could not be randomised due to lack of study personnel on the day of the examination. One infant randomised to the control group has incomplete data since she required emergent surgery for necrotising enterocolitis. Data from this subject were not analysed. Baseline characteristics for both groups are shown in table 1.

Figure 2 shows the mean PIPP scores and 95% CIs. We did not find a significant difference in PIPP scores between control (8.4, 95% CI 7.6 to 9.3) and treatment groups (8.5, 95% CI 7.3 to 9.8) at the time of speculum insertion (p=0.94) or at any other time. Intraclass correlation between the two nurses was 0.79.

Figure 2

Premature infant pain profile (PIPP) scores (mean (95% CI)). *EMONO, equimolar mixture of nitrous oxide and oxygen.

Table 1

Baseline characteristics of the patients

Neither heart rate (figure 3) nor saturation (figure 4) differed significantly when comparing the two groups at any time point. Mean blood pressure and respiratory rate did not differ significantly between either group at any time (results not shown). The number of apnoea episodes in the 24 h following the eye examination was comparable (Control 1.55±1.95, EMONO 1.27±1.69, p=0.70). Desaturation episodes expressed as the percentage of time spent with saturation below 88% in the 24 h following the eye examination was comparable (control 9.9±10.4, EMONO 12.6±8.7, p=0.36). Gas inhalation time was comparable (control 12.2±4.7 min, EMONO 12.9±7.3 min).

Figure 3

Heart rate (mean (95% CI)). *EMONO, equimolar mixture of nitrous oxide and oxygen.

Figure 4

Saturation (mean (95% CI)). *EMONO, equimolar mixture of nitrous oxide and oxygen.

In both groups, the number of examinations performed by ophthalmologist #1 versus ophthalmologist #2 was similar (10 vs 8 control group, 10 vs 12 EMONO group, p=0.52). The total amount of sucrose (in mg) per infant used for pain control was similar (control 33.3±16.7, EMONO 29.4±20.8, p=0.68).


We did not find that EMONO delivered by nasal cannulae produces any additional pain relief during ROP screening examination in infants already receiving standard care (swaddling, topical anaesthetics and sucrose). It should be noted that our PIPP scores during eye examination in both allocation groups were lower than those in most studies by two to three points.6 8 13 18 Infants in both groups of our study systematically received a combination of swaddling, sucrose and topical anaesthetics. Combining multiple modalities focused on pain control may have produced a lower PIPP score than expected. Studies that used all three of these measures simultaneously showed PIPP scores similar to ours, between 8 and 9.2 19 Furthermore, since our study used trained retinal specialists, examinations may have been more efficient resulting in less eyeball manipulation. We did not evaluate the effects of EMONO when compared to no treatment (pharmacological or non-pharmacological) since this was not considered ethical.

The PIPP scores in our study were lower than the scores measured during the pretrial phase (13.4±3.4). This reduction may be due to standardisation of the eye examination during the trial and reduction of variability. All examinations were performed by the same two ophthalmologists and use of sucrose, swaddling and topical anaesthetics was systematic.

Our failure to show any additional beneficial effect with N2O may have several explanations. First, EMONO is a minimally anaesthetic and analgesic agent.32 It may be insufficient to achieve adequate pain control. Stronger analgesia requires delivery of N2O in concentrations greater than 50% using specialised gas outlets not available in most NICUs. Second, nasal cannula may not be as effective for gas delivery as an airtight facemask. However, our ophthalmologists found it difficult to perform the eye examination with a facemask. Third, it is possible that EMONO has some beneficial effect but the PIPP score may not be the most appropriate tool to measure this. It may not be able to adequately discriminate between pain, discomfort from being restrained, photophobia or stress. Salivary cortisol as a stress indicator may be a useful adjunct measurement in future studies.13

We recognise that during inhalation, infants in our study were subjected to relative hyperoxia due to the admixture of 50% oxygen. However, the duration of this hyperoxia is relatively short (about 12 min) and all of the infants in the study were past 32 weeks corrected age. No other major side effects were observed.

Our study has some limitations. Our population may be healthier than other neonates ≤30 weeks, since we excluded infants who died before the first examination. Information bias was minimal since our study was prospective and loss to follow-up was not an issue. All infants received their allocated treatment. Confounding factors are unlikely, given that we used simple randomisation, and both groups had similar characteristics.

Our aim was to find a non-invasive method of controlling the pain, stress and discomfort elicited by ophthalmological examination in premature infants. No single intervention seems to adequately achieve this. Considering the available evidence, systematically combining currently available pharmacological and non-pharmacological treatment modalities appears to be the best option.


The authors would like to thank Ms Kim Lampron for her help in data collection. The authors would also like to thank the nurses and respiratory therapists of the Royal Victoria Hospital NICU for taking care of the study patients.



  • Funding This research was funded by a grant from the Montreal Children's Hospital Research Institute.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Montreal Children's Hospital Research Ethics Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.