Menu Search Donate
clinical guideline banner


This document is only valid for the day on which it is accessed. Please read our disclaimer.

This guideline is intended to assist clinicians with the performance, reporting and interpretation of paediatric overnight pulse oximetry studies.


Since its introduction in the 1970s, pulse oximetry has become a common and valuable clinical tool.  The use of recording continuous oximetry overnight is a more recent application of this technology.

Sleep is a vulnerable time for respiration, involving changes in muscle tone/control, responsiveness to apnoea and respiratory chemoreceptors, and reduction in functional residual capacity, minute ventilation, and potentially, respiratory control stability.  These changes may lead to either sustained or paroxysmal reductions in ventilation/gas exchange and/or repeated arousals resulting in sleep disruption.

The gold standard for the assessment of sleep quality and sleep disordered breathing is polysomnography.  However, limited channel studies such as pulse oximetry may still provide useful data.  As with all tests, it is critical to understand the indications for pulse oximetry, what the instrument measures and reports, its limitations and how to interpret results.


The indication influences both the performance and interpretation of the pulse oximetry.  Examples include as part of the assessment of:

  • Respiratory function (gas exchange) in individuals with known or suspected diseases such as chronic neonatal lung disease, chronic suppurative lung disease, neuromuscular conditions, severe scoliosis, pulmonary fibrosis, airway diseases, and abnormalities in respiratory control.
  • Oxygen therapy.
  • Individuals with known or suspected obstructive sleep apnoea syndrome (OSAS).


Oxygen delivery to tissues is influenced by:

  • Cardiac output
  • Tissue perfusion
  • Haemoglobin (Hb) concentration
  • Haemoglobin oxygen capacity (BO2)
  • Oxygen affinity, and
  • Oxygen saturation (SpO2). 

Oxygen curveBO2 is usually approximately 1.39 mL/g (adult Hb).  SpO2 ('functional saturation') refers to the amount of oxygen bound to Hb in the blood expressed as a percentage of its maximal capacity. 'Fractional saturation' (HbO2),often reported by blood gas analysers and estimated by some pulse oximeters, refers to the percentage of total Hb bound to oxygen and never reaches 100%. The relationship between SpO2 and HbO2 varies and they should not beconsidered interchangeable. Oxygen affinity refers to the relationship between oxygen saturation and oxygenpartial pressure (oxygen dissociation curve, see Figure 1).  The oxygen affinity should be such that haemoglobin reaches almost full saturation in the lungs, yet readily releases oxygen at the relatively lower partial oxygen pressure in the tissue capillaries

Oximetry Technology

Pulse oximetry refers to technology which estimates the arterial oxygen saturation based firstly on the detection of pulsatile blood flow and secondly on differing absorption spectra of oxyhaemoglobin and deoxyhaemoglobin.  Most commercially available pulse oximeters have two light-emitting diodes (LEDs) which emit light at 660nm (red) and 940nm (infrared).  A photoreceptor is positioned opposite to the LEDs with patient tissue, usually a finger, toe or ear lobe, held between.  The oximeter's signal processor compares the ratio of absorption of the two spectra against a set of stored reference values obtained previously from volunteers.  While the signal is sampled frequently (egg 25 time per second) the displayed output is usually averaged over a set period (2-16 seconds) that is referred to as 'averaging time.'  Longer averaging times (8-16s) may reduce signal artifact (egg from motion) but also reduce the ability to detect the rapid change in saturation often seen with central or obstructive events (apnoea/hypopnoea).  In addition, the report may be influenced by the study 'resolution' - the frequency that the result is saved into the oximeter's memory (commonly once every 2-10 seconds). 

Sources of oximetry inaccuracy include

  • Motion artefact.  This impairs the oximeter's ability to discern pulsatile blood flow.  Manufacturers utilize different signal processing techniques (including the use of long averaging times) to reduce this error.  In addition, oximeters may indicate signal strength and/or show a plethysmographic waveform to assist the clinician in identifying artifact.
  • Severe hypotension, low cardiac output, vasoconstriction and hypothermia reduce pulse volume and consequent signal strength impairing oximeter performance.
  • The presence of carboxyhaemoglobin, methaemoglobin, fetal haemoglobin, sickle cell haemoglobin, nail polish, and xenon and fluorescent lamps potentially influence results. Hyperbilirubinaemia and anaemia do not.
  • Oximeter technology relies on reference data obtained from volunteers.  Individual variation and the limits of what values can be safely generated in volunteers results in intrinsic inaccuracy, especially when oxygen saturations fall below 70%.  Probes designed for low saturations are available.
  • Variations in the frequency output of individual LEDs (individual sensors).

Most pulse oximeters have quoted accuracy of ± 2-3% at saturations above 90%.  Oximeters don't require calibration however should be well maintained and used according to manufacturer recommendations.


While long averaging times (8-16s) may be adequate for monitoring or for oxygen titration, an averaging time of 2-4s is recommended for most studies and is mandatory for the investigation of obstructive sleep apnoea syndrome. Correspondingly, a high recording resolution is required (every 2 seconds).   Ideally these parameters will be indicated on the report.  Resistance to motion artifact and a reliable system for displaying signal quality is especially important for studies in infants and young children.

Observations (see observation sheet below)

It is very useful for staff / parents to make a record of events / observations during the oximetry study recording.  This helps correlate patient circumstances with oximetry results.  Observations should include events (awake, asleep, feeding, crying, alarm soundings, etc) and respiratory observations (snoring, increased work of breathing, etc). 

Study duration

In general studies should record  > 6 hours of sleep.  More is preferable.

Report (see worked examples)

As a minimum, the report should include patient identification, date, time, study duration, the name of the responsible / requesting clinician, and relevant study conditions (such as oxygen flow).  A graphical representation (time vs. saturation) of the overnight recording is the single most useful output.  It is recommended that the output be > 3 cm/hr and > 1mm/%.  Signal quality should be displayed on the printout. Summary statistics such as mean SpO2, mean pulse rate and percentage time below 90% may be useful.  Printouts of individual events (desaturation or poor signal), bar graphs and descriptions of time during other percentage ranges are not usually necessary.

If a clinician interpretation is being added, then it is recommended to include:

  • Indication
  • Relevant overnight observations
  • A brief description of the study's quality and findings, and
  • The clinician's conclusions.

Suggested approach to reporting oximeter studies - 'CAGE'


  • Clinical context - What is the question?  What are the pre-test probabilities?
  • Study context - What happened during the study (awake / asleep, feeding, crying, etc)?  Is the child on oxygen or other respiratory support? 
  • Technical context - What model of oximeter (any peculiarities to be aware of)? What averaging time?  What resolution?  What report format?


  • Identify and exclude obvious artefact.  Is the study of adequate quality overall?- if not do not proceed further (repeat the study).


  • Estimate average saturations from the graphical representation.  While summary statistics provided by the oximeter such as mean and nadir may be useful, they are often skewed by artefact. 
  • For some indications, estimating the time spent below 90% may also be useful.


  • Consider averaging time again - is it possible to report on events?
  • Consider the graphical representation.  Note frequency and severity of desaturation.  Do they occur in clusters?  Is there recovery between? 

Interpretation - specific contexts 

Chronic lung disease - neonatal or other (see worked examples)

Please refer to local guidelines, the TSANZ position statement ( for Infants with Chronic Neonatal Lung Disease: Recommendations for the Use of Home Oxygen Therapy and British Thoracic Society Guidelines ( for discussions on oxygenation targets, frequency of oximetry studies and weaning strategies.

Following the above assessment (CAGE) the following aspects may be considered:

  • Is the baseline saturation in the target range?  - refer to guidelines.
  • Is the percentage time below 90% excessive? - refer to guidelines.
  • Is the baseline stable?  Frequent desaturation may represent obstructive or central events (apnoeas) or hypoventilation.  They may be associated with undesirable cardiovascular instability or sleep disturbance.  - refer to guidelines.

Obstructive sleep apnoea syndrome (OSAS) (see worked examples)

Oximetry studies cannot exclude OSAS and therefore are not screening tests.  Oximetry may, however, support the diagnosis of OSAS, provide a crude estimate of severity, assist in triaging, and contribute to peril-operative risk assessment.  An oximeter with suitable properties (averaging time 2-4 seconds) and graphical output (see above) are vital.  Please see the National Paediatric Sleep Medicine Network guideline ( Assessment of Sleep Disordered Breathing in Childhood) for further discussion regarding the investigation and management of OSAS.

The McGill Oximetry Score provides a validated approach to oximetry studies for OSAS. The graphical report is evaluated and studies designated as 'Positive' or 'Inconclusive' as described in Table 1 below.

Table 1:  Oximetry study classification according to McGill criteria.

Description Definition
A 'desaturation' > 4% fall in saturation
A 'cluster' > 5 desaturation within a 30 minute period
A 'positive' study > 3 clusters with > 3 desaturation to <90%
An 'Inconclusive' study Not a positive study (i.e. <3 clusters or <3 desaturation below 90%)

In children over one year of age, suspected of having OSAS, a positive overnight oximetry has a positive predictive value of 97% and a negative predictive value of 47% (i.e. almost half of those with a normal/inconclusive oximetry actually had OSA).  Nixon et al evaluated the score further and made management recommendations based on it (see Table 2).

Table 2.  Oximetry study severity and recommendations adapted from Nixon et al and National Paediatric Sleep Medicine Network Guidelines

McGill Number of desaturations Recommendations
Interpretation Score <90% <85%  
Inconclusive 1 < 3  0 Additional evaluation of breathing during sleep is required to rule out OSAS. Proceed to intervention (e.g. adenotonsillectomy) if sufficient clinical suspicion. Day case surgery may be appropriate.
Positive 2 > 3  < 3 Recommend intervention (e.g. Adenotonsillectomy) on the waiting list. Day case surgery may be appropriate.
Positive, higher risk 3 > 3  > 3 Recommend urgent intervention  (e.g. Adenotonsillectomy). Post operative overnight stay on continuous monitoring is recommended (see below). 

Risk of post-operative complication

Adenotonsillectomy may be performed safely on a day case basis.  However, young children, those with significant co-morbidities, and those with moderate to severe OSAS are at greater risk of post-operative complication including critical airway compromise. 

Children with OSAS who meet the following criteria should be admitted overnight with continuous oximetry monitoring in a facility with the capacity to manage critical paediatric airways:

  • Age <3 years
  • Significant co-morbidity or OSAS complication
  • McGill score of >3.

Observation Sheet for Oximetry Study

Click on the image below to open a printable pdf

 Obs sheet oximetry

Worked Examples

Click on the image below to open a printable pdf

Worked examples

Information for Families

Kidshealth factsheets on:

Obstructive Sleep Apnoea



Tonsillectomy and Adenoidectomy


This guideline is the product of consultation with Drs I Asher, J Brown, E Edwards, D Elder, J Larkin, D McNamara, G Nixon, P Pattemore, B Taylor, and J Vyas.


McMorrow RC, Mythen MG. Pulse oximetry. Current opinion in critical care 2006;12(3):269-71

Toffaletti J, Zijlstra WG. Misconceptions in reporting oxygen saturation. Anesth Analg 2007;105(6 Suppl):S5-9.

American Academy of Sleep Medicine. The AASM Manual for the Scoring of Sleep and Associated Events:  Rules, terminology and technical specifications. 2007.

Fitzgerald DA, Massie RJ, Nixon G, et al. Infants with chronic neonatal lung disease: recommendations for the use of home oxygen therapy. Position Statement from the Thoracic Society of Australia and New Zealand, MJA, 2008; 189(10):578-582.

Balfour-Lynn IM, Field DJ, Gringras P, et al.  BTS guidelines for home oxygen in children, Thorax, 2009:64(suppl II)

National Paediatric Sleep Medicine Network: New Zealand Guidelines for the assessment of sleep disordered breathing in children (2015).

Brouillette RT, Morielli A, Leimanis A, et al. Nocturnal pulse oximetry as an abbreviated testing modality for pediatric obstructive sleep apnea. Pediatrics, 2000:405-12.

Nixon G, Kermack A, Davis G, et al. Planning Adenotonsillectomy in Children With Obstructive Sleep Apnea: The Role of Overnight Oximetry. Pediatrics, 2004:e19-e25.

Did you find this information helpful?

Document Control

  • Date last published: 01 October 2009
  • Document type: Clinical Guideline
  • Services responsible: Paediatric Respiratory
  • Owner: Jacob Twiss
  • Editor: Greg Williams
  • Review frequency: 2 years

More From Starship