Article Text

Download PDFPDF

BTS guidelines for home oxygen in children
  1. I M Balfour-Lynn,
  2. D J Field,
  3. P Gringras,
  4. B Hicks,
  5. E Jardine,
  6. R C Jones,
  7. A G Magee,
  8. R A Primhak,
  9. M P Samuels,
  10. N J Shaw,
  11. S Stevens,
  12. C Sullivan,
  13. J A Taylor,
  14. C Wallis,
  15. on behalf of the Paediatric Section of the Home Oxygen Guideline Development Group of the BTS Standards of Care Committee
  1. Dr I M Balfour-Lynn, Department of Paediatric Respiratory Medicine, Royal Brompton & Harefield NHS Trust, Sydney Street, London SW3 6NP, UK; i.balfourlynn{at}ic.ac.uk

Statistics from Altmetric.com

“… as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns – the ones we don’t know we don’t know.” D Rumsfeld, 2002

1. INTRODUCTION

1.1 Aims and target audience

The aims of these guidelines are to present the evidence base for the practice of administering supplemental oxygen to children outside hospital and to make recommendations for best practice. For many aspects high-quality evidence is lacking, and suggestions are made based on clinical experience. It is hoped the guideline will highlight areas where research is needed to further inform clinicians. The target audience is clinicians who prescribe home oxygen for children, principally those in hospital practice. It is also intended for other professionals involved with the whole process, which may include community paediatricians, paediatric neurodisability specialists, nurse specialists, school nurses, occupational therapists and physiotherapists; this is reflected by the multidisciplinary nature of the guideline committee (section 13).

1.2 Methodology for generation of the guidelines

The initial literature search was carried out by the Centre for Reviews and Dissemination at the University of York. Further searches were then carried out by members of the working group who concentrated on their own topics. Details of the search strategy are given in Appendix 1 available online.

Each section of the guideline was researched and drafted by a subgroup of the Paediatric Section of the British Thoracic Society (BTS) Home Oxygen Guideline Development Group (itself a subcommittee of the BTS Standards of Care Committee). Publications were rated according to the SIGN 50 criteria for the calibre of the methodology of the research to give levels of evidence (see box 1). Once all parts were merged into one document, the whole group then met to discuss the first draft before redrafting took place. This draft was based, where possible, on the published evidence, but this was then combined with clinical expertise as required. The resulting draft is therefore a blend of published evidence and clinical experience.

This was sent to a group of specialist reviewers listed in the Acknowledgements. The manuscript was then amended in the light of their comments and the document was reviewed by the BTS Standards of Care Committee and the Quality of Practice Committee of the Royal College of Paediatrics and Child Health. After a further redrafting and final approval from the BTS Standards of Care Committee, the guidelines were submitted for publication.

Box 1 Revised SIGN grading systems for grades of recommendation and levels of evidence (Annex B of SIGN 50 available at www.sign.ac.uk)

Levels of evidence

1++ High quality meta-analyses, systematic reviews of randomised controlled trials (RCTs) or RCTs with a very low risk of bias

1+ Well-conducted meta-analyses, systematic reviews or RCTs with a low risk of bias

1− Meta-analyses, systematic reviews or RCTs with a high risk of bias

2++ High quality systematic reviews of case-control or cohort studies, or high quality case-control or cohort studies with a very low risk of confounding or bias and a high probability that the relationship is causal

2+ Well-conducted case-control or cohort studies with a low risk of confounding or bias and a moderate probability that the relationship is causal

2− Case-control or cohort studies with a high risk of confounding or bias and a significant risk that the relationship is not causal

3 Non-analytical studies (eg, case reports, case series)

4 Expert opinion

Grades of recommendations

A At least one meta-analysis, systematic review or RCT rated as 1++ and directly applicable to the target population; or a body of evidence consisting principally of studies rated as 1+, directly applicable to the target population and demonstrating overall consistency of results

B A body of evidence including studies rated as 2++, directly applicable to the target population and demonstrating overall consistency of results; or extrapolated evidence from studies rated as 1++ or 1+

C A body of evidence including studies rated as 2+, directly applicable to the target population and demonstrating overall consistency of results; or extrapolated evidence from studies rated as 2++

D Evidence level 3 or 4; or extrapolated evidence from studies rated as 2+

Good practice points

✓ Recommended best practice based on the clinical experience of the Guideline Development Group.

Background facts are shown in the text in italics. Recommendations are shown in bold and placed above the text accompanied by the grade for that recommendation.

1.3 Conflict of interest

All the members of the Guideline Committee submitted a written record of possible conflicts of interest to the Standards of Care Committee of the BTS. IMB-L, BH, RAP and NJS are involved with the Children's Home Oxygen Record Database (CHORD) which has received funding from British Lung Foundation and Carburos Metallicos, the research arm of Air Products (based in Spain). These are available for inspection on request from the Chairman of this Committee.

1.4 Acknowledgements

Funding for the literature search and travel to the guideline meeting was kindly provided by the British Thoracic Society. The authors thank Lisa Stirk at the Centre for Reviews and Dissemination at the University of York for the literature search; Professor Fenella Kirkham, Paediatric Neurologist at the Institute of Child Health, London for advice on sickle cell disease; Dr Gerry Coghlan, Consultant Cardiologist at the Royal Free Hospital, London for his comments on pulmonary hypertension; and Dr Renee McCulloch, Consultant in Palliative Care at Great Ormond Street Hospital, London for comments on palliative care. The following acted as specialist reviewers: Clinical Associate Professor Dominic Fitzgerald, Paediatric Respiratory Medicine, Children’s Hospital, Westmead, Sydney; Professor Sheila G Haworth, Professor of Developmental Cardiology, Institute of Child Health, London; Dr Jeremy Hull, Consultant in Paediatric Respiratory Medicine, Oxford Children’s Hospital; Professor Neena Modi, Professor of Neonatal Medicine, Imperial College London and Honorary Consultant, Chelsea and Westminster NHS Foundation Trust; Dr Win Tin, Consultant Paediatrician and Neonatologist, James Cook University Hospital, Middlesbrough; Dr Jane Williams, Nottingham University College Hospital NHS Trust on behalf of the British Academy of Childhood Disability. The authors also thank the Quality of Practice Committee of the Royal College of Paediatrics and Child Health for reviewing the guidelines.

SUMMARY OF BACKGROUND FACTS

Normal oxygen saturations (Section 2.4)

  1. Oximeters from different manufacturers may give different oxygen saturation readings depending on whether fractional or functional oxygen saturation is being measured.

  2. The median baseline saturation in healthy term infants during the first year of life is 97–98%.

  3. In only 5% of healthy infants is the arterial oxygen saturation measured by pulse oximetry (Spo2) <90% for >4% of the time.

  4. The median baseline Spo2 in healthy children ⩾1 year old is 98% with a 5th centile of 96–97%.

  5. A healthy child aged 5–11 years spends no more than 5% of the time below a Spo2 of 94% while asleep.

Consequences of chronic low oxygen saturation (Section 2.5)

  1. Hypoxaemia causes pulmonary hypertension but the precise severity and duration of hypoxaemia needed to do this are not known. The factors affecting individual susceptibility are also unknown.

  2. Spo2 levels >94–95% appear to reduce pulmonary hypertension, while levels <88–90% may cause pulmonary hypertension. This does not apply to children with congenital cardiac defects and idiopathic pulmonary arterial hypertension.

  3. Hypoxia may have adverse effects on cognition and behaviour at Spo2 levels of ⩽85%, but the effects of milder hypoxia are less clear.

  4. In infants with chronic neonatal lung disease (CNLD), Spo2 <90% is associated with an increased risk of apparent life-threatening events while Spo2 ⩾93% is not.

  5. In infants with CNLD, Spo2 <92% may be associated with suboptimal growth.

  6. In infants with CNLD, Spo2 ⩽90% impairs sleep quality but Spo2 >93% does not.

SUMMARY OF RECOMMENDATIONS

Consequences of excess oxygen therapy (Section 2.6)

  1. Excess arterial and intra-alveolar oxygen concentrations are toxic in preterm infants and must be avoided by appropriate monitoring and adhering to the target Spo2 level; there are no data in older children. [D]

Indications for long-term oxygen therapy (LTOT) (Section 3)

Chronic neonatal lung disease (Section 3.1) (fig 1)

  1. Supplementary oxygen should be given to infants with chronic neonatal lung disease:

    1. to reduce or prevent pulmonary hypertension, reduce intermittent desaturations, reduce airway resistance and promote growth; [C]

    2. as it is likely to be beneficial for neurodevelopment in infants with CNLD; [D]

    3. as it may reduce the associated risk of sudden unexplained death in infancy; [D]

    4. as oxygen at home is preferable to a prolonged hospital stay for both quality of life and psychological impact for the infant, parents and family; [D]

    5. as it saves days in hospital due to earlier discharge despite a significant readmission rate. [C]

Other neonatal lung conditions (Section 3.2)

  1. Home LTOT should be offered to infants with other oxygen-dependent neonatal lung conditions who are otherwise ready for hospital discharge. [✓]

Congenital heart disease (Section 3.3)

  1. Home oxygen should not be used for cyanotic congenital heart disease unless accompanied by other respiratory problems. [✓]

  2. In acyanotic heart disease there is no role for LTOT. [✓]

Pulmonary hypertension (Section 3.4)

  1. In idiopathic pulmonary hypertension, supplementary oxygen is recommended for sleep-associated desaturations and for emergency use. [D]

  2. In pulmonary hypertension associated with congenital cardiac defects, some children may gain symptomatic benefit and a small open study has suggested it may improve survival. However there is a lack of good evidence that LTOT is of benefit and it is not recommended. [D]

  3. LTOT is recommended for pulmonary hypertension secondary to pulmonary disease. [D]

Intrapulmonary shunting (Section 3.5)

  1. The benefits of LTOT in non-cardiac intrapulmonary shunting are unknown with no relevant publications; however, it should be considered if it leads to symptomatic improvement. [✓]

Recurrent cyanotic-apnoeic episodes (Section 3.6)

  1. LTOT should be considered for infants and children who have recurrent cyanotic-apnoeic episodes severe enough to require cardiopulmonary resuscitation, assuming any anaemia has been corrected. [D]

Interstitial lung disease (Section 3.7)

  1. LTOT should be offered to hypoxic children with interstitial lung disease who are otherwise ready for hospital discharge. [✓]

Obliterative bronchiolitis (Section 3.8)

  1. LTOT should be offered to hypoxic children with obliterative bronchiolitis who are otherwise ready for hospital discharge. [✓]

Cystic fibrosis (Section 3.9)

  1. LTOT should be considered for hypoxic children with cystic fibrosis as a means to improve school attendance [B], and for those who obtain symptomatic relief. [D]

  2. In cystic fibrosis, monitoring of CO2 levels should be carried out when oxygen therapy is initiated. [C]

Obstructive sleep apnoea (Section 3.10)

  1. In obstructive sleep apnoea, continuous positive airway pressure (CPAP) or occasionally non-invasive ventilation (NIV) is the therapy of choice if the upper airway obstruction cannot be relieved surgically. If this is not possible, LTOT should be used to improve the Spo2, but CO2 levels need to be monitored at initiation of treatment. [C]

Chronic hypoventilation (Section 3.11)

  1. LTOT should be given in addition to ventilatory support if there is a hypoxaemic component of hypoventilation (assuming the child is optimally ventilated). On occasions when ventilatory support is not possible, supplemental oxygen may be the only alternative. [✓]

Sickle cell disease (Section 3.12)

  1. LTOT should be considered for children with sickle cell disease and persistent nocturnal hypoxia to reduce the risk of stroke and painful crises. [C]

Palliative care (Section 3.13)

  1. LTOT should be considered for hypoxaemic children undergoing palliative care who obtain symptomatic relief from supplemental oxygen. [✓]

Special situations (Section 4)

Intermittent LTOT (Section 4.1)

  1. In children with neurodisability, oxygen may be given in the presence of hypoxia secondary to an acute lower respiratory tract infection. Children will usually be hospitalised but, where families opt for home treatment, facilities for home oxygen may be required if the infections are recurrent. [✓]

  2. The use of home oxygen in children with severe neurodisability and low Spo2 should be driven by quality of life issues rather than oxygen saturation targets. [✓]

Intermittent emergency oxygen therapy (Section 4.2)

  1. Although most children with asthma should receive bronchodilators via a spacer device, for those using a home nebuliser, unless there is a significant co-morbidity or the child has life-threatening acute exacerbations, it should be run off room air. [✓]

  2. Intermittent acute oxygen therapy at home should be considered for the few children with recurrent episodes of severe life-threatening asthma, as a temporary therapy prior to ambulance transfer to hospital. [✓]

  3. Intermittent acute oxygen therapy at home is not routinely recommended for seizures as there is no evidence that it reduces their duration, reduces harm from prolonged seizures or improves quality of life for the child or family. [✓]

Miscellaneous situations (Section 4.3)

  1. Infants with bronchiolitis requiring oxygen (Spo2 ⩽92%) should be admitted to hospital and can be considered for discharge when their Spo2 is >94% and they no longer require oxygen (for at least 8–12 h). [✓]

Assessment of need for LTOT and target oxygen saturations (Section 5)

  1. Suitability for home oxygen therapy should be assessed by a specialist with appropriate experience. [✓]

  2. Pulse oximetry should be used for assessing children rather than arterial blood sampling. [C]

  3. Children should be assessed for at least 6–12 h and during all levels of activity, including sleep and feeding. [D]

  4. Lower limit target Spo2 should be met for at least 95% of the stable recording period. [✓]

  5. There is no need to regularly assess CO2 levels in infants with CNLD who are at home [✓], but it may be useful in some neonates with other conditions [✓] and older children [C], especially when initiating home oxygen therapy.

  6. In CNLD, oxygen therapy should be given to maintain an Spo2 of ⩾93%. [C]

  7. There are no data to guide target levels for Spo2 in children with other respiratory conditions, but the recommendation is to maintain Spo2 at ⩾93%, although ⩾94% may be appropriate for sickle cell disease and ⩾90% for cystic fibrosis. [✓]

  8. In infants with CNLD, prior to discharge an ECG or echocardiogram is useful to assess the right heart in order to exclude significant pulmonary hypertension. [✓]

Ordering and provision of oxygen (equipment) (Section 6)

  1. The decision that a child requires home oxygen and its ordering should be undertaken by paediatric specialists rather than primary care. [✓]

  2. Oxygen concentrators should be provided for LTOT, unless it is likely that the child will only require low flow oxygen for a short while. [✓]

  3. While low weight cylinders are easier to handle, they empty more quickly. Parent choice should be considered. [✓]

  4. Portable equipment should be available for all children as part of the provision of home oxygen unless oxygen is only required at night. [✓]

  5. Continuously delivered liquid oxygen cannot be used at flows of <0.25 l/min, although breath-activated systems can allow lower flows. It has limited applications for children, so is generally not recommended. [✓]

  6. Low flow meters are preferable, so very low flow meters are not recommended. [✓]

  7. Oxygen conservers are not indicated for young children but can be considered for older children capable of triggering the device. [✓]

  8. Humidification should be considered for high oxygen flows when given by face mask, especially for cystic fibrosis; a cold water bubble-through humidifier may be adequate for this purpose. [✓]

  9. When oxygen is given via a tracheostomy, heated humidification is generally recommended; a heat-moisture exchanger with an oxygen attachment may be an adequate alternative. [✓]

  10. Nasal cannulae are preferable for infants and young children for flows of ⩽2 l/min. Patient choice should be considered for older children. [✓]

  11. There is no evidence on whether the routine use of a saturation monitor at home is of benefit or harm, and it cannot be recommended. Nevertheless, some clinicians and parents may find it helpful in certain circumstances. [✓]

Discharge planning (Section 7)

  1. A comprehensive written parent-held discharge plan with multidisciplinary follow-up is recommended to ensure a safe and smooth transition into the community and to avoid repeated or unnecessary hospitalisations. [✓]

  2. Children can be discharged from the neonatal unit when their oxygen requirement is stable with a mean Spo2 of ⩾93% and without frequent episodes of desaturation. This usually corresponds with an oxygen flow ⩽0.5 l/min. [D]

  3. The Spo2 should not fall below 90% for more than 5% of the artefact-free recording period. [✓]

  4. There should be no other clinical conditions precluding discharge and the child must be medically stable. [✓]

  5. Careful preparation with a structured educational programme should be implemented. [D]

Follow-up after discharge (Section 8.1)

  1. The community children’s nurse or nurse specialist should visit the child within 24 h of discharge. [D]

  2. Infants with CNLD should have their Spo2 monitored within a week of discharge, with subsequent recordings as clinically indicated (but not usually less often than 3–4 weekly); monitoring should include various activity states. [D]

  3. Older children with other conditions who are clinically stable are likely to need home Spo2 recordings performed less often than infants with CNLD. [D]

Withdrawal of supplemental oxygen (Section 8.2)

  1. Once the oxygen requirement is down to 0.1 l/min, consideration should be given to withdrawing supplemental oxygen. [✓]

  2. The same target saturations used to decide initiation of supplementation should be used for withdrawal purposes (ie, ⩾93%). [✓]

  3. Children can be weaned from continuous low flow oxygen to night-time and naps only, or remain in continuous oxygen throughout the 24 h until the child has no requirement at all. It is not possible to recommend which strategy is superior. [✓]

  4. Oxygen equipment should be left in the home for at least 3 months after the child has stopped using it. If this is in a winter period, it is usually left until the end of winter. [✓]

  5. In CNLD, failure to reduce oxygen supplementation after 1 year should lead to a specialist review to rule out concomitant conditions. [✓]

Oxygen outside the home (Section 9)

  1. An appropriately trained individual should be present while the child is using the oxygen, but this does not necessarily have to be a school nurse or health professional. [✓]

  2. Children will need higher oxygen flows during air flights or at high altitude, which should be determined by a fitness-to-fly test. [B]

  3. If a child has stopped supplemental oxygen within the last 6 months, they will need a fitness-to-fly test. [✓]

Potential disadvantages (Section 10)

  1. Parental/carer smoking must be strongly discouraged. [✓]

  2. Parents/carers (and older children) must be made aware of the potential hazards of home oxygen. [✓]

  3. It is critical that parents and carers receive sufficient emotional support from their family, friends and the healthcare services. [✓]

Figure 1

Long-term oxygen therapy (LTOT) pathway for an infant with chronic neonatal lung disease. SpO2, oxygen saturation measured by pulse oximetry.

2. BACKGROUND

2.1 Definitions

Although domiciliary refers to the home, in the context of oxygen therapy it refers to delivery of supplemental oxygen outside the hospital as it may also be used outside the home, especially by children. Modes of delivery fall into several categories (fig 2).

Figure 2

Modes of home oxygen. Long-term oxygen therapy (LTOT) is the provision of oxygen for continuous use at home for patients with chronic hypoxaemia. It may be required 24 hours per day (continuous LTOT) or during periods of sleep only (sleep-related LTOT). The latter may be given at night alone (nocturnal LTOT) or, in young children, for daytime naps as well. Portable oxygen therapy is the provision of oxygen to deliver LTOT outside the home; when carried by the patient it is known as ambulatory oxygen therapy. Intermittent oxygen therapy is a less common situation whereby the child receives oxygen in an episodic manner but, because of the recurrent nature of the underlying condition, oxygen needs to be permanently available in the child’s home. Intermittent oxygen use may last days or weeks (intermittent LTOT), or it could be used during an acute emergency situation only (intermittent emergency oxygen therapy).

Long-term oxygen therapy (LTOT) is defined as the provision of oxygen for continuous use at home for patients with chronic hypoxaemia (due to any cause) in order to maintain target oxygen saturations. It may be required 24 h per day (continuous LTOT) or during periods of sleep only (sleep-related LTOT). The latter may be given at night alone (nocturnal LTOT) or, in young children, for daytime naps as well.

Portable oxygen therapy is LTOT outside the home (or in the garden). It refers to the provision of oxygen that can be wheeled on a trolley or pram, worn in a backpack or carried. When carried by the patient it is termed ambulatory oxygen. All children on LTOT require facilities for portable oxygen unless they only use it at night. There are no situations where a child receives portable oxygen that is not part of an LTOT regimen. Children are rarely housebound, and it is important to enable them (and their parents) to go outside the home in order to lead as normal a family life as possible.

Intermittent oxygen therapy describes a less common situation whereby the child receives oxygen in an episodic manner but, because of the recurrent nature of the underlying condition, oxygen needs to be permanently available in the child’s home.

  • An example would be a child with neurodisability who requires oxygen for aspiration pneumonia being treated at home (section 4.1), and who usually receives it for 1–2 weeks every few months. This is known as intermittent LTOT.

  • In another situation an acutely hypoxaemic child may receive oxygen for a short while for an emergency situation at home, for example,. life-threatening asthma waiting for an ambulance (section 4.2). This is known as intermittent emergency oxygen therapy.

Hypoxaemia refers to low oxygen tension or partial pressure in the blood. Hypoxia is less specific and refers to lack of oxygen in a particular compartment (eg, alveolar or tissue hypoxia). It is usually as a result of hypoxaemia (hypoxaemic hypoxia), decreased tissue blood flow (stagnant hypoxia), anaemia (anaemic hypoxia) or an inability of the tissues to utilise oxygen (histotoxic hypoxia).

2.2 What differences are there between adult and paediatric practice?

  • Diagnosis. The range of conditions seen in children is quite distinct from adults. There is a tendency for children’s diseases to improve with time, whereas with adults they tend to deteriorate. Exceptions in children include cystic fibrosis and neuromuscular disease.

  • Ordering oxygen. In children almost all oxygen therapy is prescribed by hospital specialists (consultant paediatricians) rather than in primary care.

  • Assessment. In children, almost all oxygen assessments are done by pulse oximetry and not arterial blood sampling.

  • Growth and neurodevelopment. These are important considerations in children.

  • Equipment. Specific equipment is required to allow for low oxygen flows. Almost all children receiving LTOT also require portable oxygen therapy (they are rarely housebound). Many older children have LTOT for <15 h per day.

  • Care and safety considerations. All children require supervision from a parent/carer.

  • Preschool/school. Provision of oxygen may be necessary at nursery or school.

2.3 What is current UK practice in prescribing home oxygen?

Data are available from the BTS Home Oxygen Database which receives anonymised data for England and Wales from the four oxygen suppliers; and also from the Children’s Home Oxygen Record Database (CHORD) which receives copies of the Home Oxygen Order Form once the parents have signed consent. In June 2007 there were 3136 children under 17 years of age in England and Wales receiving home oxygen, which represents 4% of all patients (adult and children) receiving it.1 From CHORD incidence data (available December 2008) on 828 children, the commonest underlying diagnoses are chronic neonatal lung disease (60%), neurodisability (7%), paediatric cardiac disease (5%), neuromuscular disease (3%) and interstitial lung disease (2%).

2.4 What is the normal oxygen saturation in a healthy infant aged <1 year and a healthy child aged ⩾1 year?

2.4.1 Methodological issues

  • Oximeters from different manufacturers may give different oxygen saturation readings depending on whether fractional or functional oxygen saturation is being measured.

Normal oxygen saturation (Sao2) is an oversimplification of a complex measurement. Assessment of Spo2 (oxygen saturation measured by pulse oximetry) can be made in different behavioural states and using different machines. An oximeter which measures fractional oxygen saturation (eg, Ohmeda 3470) may read 1.6% lower than one measuring functional saturation (eg, Nellcor N3000).2 Fractional saturation refers to the ratio of oxyhaemoglobin to all haemoglobin measured including dyshaemoglobins (eg, carboxyhaemoglobin or methaemoglobin), whereas functional saturation refers to the ratio of oxyhaemoglobin to all haemoglobin that is capable of carrying oxygen. The normal values from published data can be confusing as they may be derived from data which averages the summary data generated from a number of subjects where each recording generates its own mean, median and centiles. Thus, reference values might refer to the median of mean oxygen levels or even the 5th centile for 5th centile oxygen levels. Generation of the lower limit of normal is complicated by the fact that data are often given as mean and standard deviation (SD), but Spo2 is not normally distributed which means that the SD cannot be used reliably to generate lower 95% confidence intervals for a population. Many studies generating reference values are particularly focused on desaturation episodes, and baseline levels may be selected from periods of normal breathing.

Concerning children on home oxygen, the main purpose is to maintain a stable baseline level of oxygenation, and it has been recommended that assessments should be made of summary data on a recording. Thus, the most relevant studies are those which include all data from different behavioural states and give summary data to allow the generation of lower limits of normal values. Artefact rejection is addressed by some studies, usually by visual inspection of the waveform or by disparity between oximeter heart rate and ECG. In more recent studies, automated artefact rejection by commercial monitors is more usual.

Studies have been included for this section if they were of healthy children and where summary data were available for mean, median or distributions of values for Spo2. Studies which were restricted to desaturation indices or other measures of episodic desaturation were not included without the above measures. In all, 20 studies were relevant for normative values, 14 in infants and 6 in older children. Seven of the 14 studies in infants are from the same group of investigators.

2.4.2 Healthy infants aged <1 year

  • The median baseline saturation in healthy term infants during the first year of life is 97–98%.

  • In only 5% of healthy infants is the Spo2 <90% for >4% of the time.

There are two longitudinal studies of infants in the first 6–12 months.3 4 Masters et al4 found a slight increase in baseline levels with age but, in contrast, Hunt et al3 found no change in baseline with age but a decrease in variability. The five infant studies from the Stoke group59 used a modified oximeter, operating in a beat-to-beat mode (not available for clinical use), and restricted analyses of oxygen baselines to periods of regular breathing (table 1). This makes their data difficult to apply to studies in clinical practice using averaging oximeters, and where all data are included irrespective of breathing pattern. In turn, this makes comparison of clinical studies difficult when they use different oximeters and different averaging, thus including pauses, sighs and periodic breathing to different degrees. The median oxygen baseline in these studies ranged from 97.6% to 99.8% with a 5th centile ranging from 91.9% to 95.5%. Three of the studies were on infants born prematurely and without significant lung disease when ready for discharge at term, and the results were similar to those in healthy term infants.

Table 1 Normal Spo2 (%) levels in healthy children measured by the same group of investigators using Nellcor N200 (or N100 with N200 software) pulse oximetry in beat-to-beat mode

One paper11 looked at normal values in healthy preterm infants prior to term, reporting median values of 97%, measuring fractional saturation. The remaining four studies1215 were of healthy term infants, using either short-term or overnight monitoring with different monitors. Median Spo2 levels ranged from 97% to 98.2%.

The mean percentage of time spent <90% has been reported in three studies of term infants,3 4 14 and ranges from 0% to 2%; in preterm babies it was approximately 2.5%.11 However, the 95th centile of the time spent <90% was 0.1–4% in term babies14 and approximately 11.5% in preterm babies.11

2.4.3 Healthy children aged ⩾1 year

  • The median baseline Spo2 in healthy children ⩾1 year old is 98% with a 5th centile of 96–97%.

  • A healthy child aged 5–11 years spends no more than 5% of the time below a Spo2 of 94% while asleep.

Studies in this age group were all confined to sleep recordings or recordings done while in bed. One of the six studies used a beat-to-beat oximeter and analysed only regular breathing periods; this found median values of 99.5% with a 5th centile of 96.6%.10 Another study, measuring fractional saturation, gave only median figures for the 5th centile of Spo2, which was 96% with a 1st centile of 95%.16 Three more studies gave only mean data, ranging from 97% to 97.8%.1719 The most definitive study in this area is that of Urschitz et al,20 which is a carefully validated study setting out to generate normal values for primary school children. The median value in this study is 98%, with a 5th centile of 97% and range 94–100%. However, this study also gives the 5th and 2.5th centiles for the value below which the subjects spent 5% and 10% of the recording (SAT5 and SAT10). These are particularly useful for assessment of normal oxygenation and were 95% and 94% for SAT5 and 96% and 94% for SAT10.

2.5 What are the consequences of chronic low oxygen saturation in children?

2.5.1 Pulmonary arterial hypertension

  • Hypoxaemia causes pulmonary hypertension but the precise severity and duration of hypoxaemia needed to do this are not known. The factors affecting individual susceptibility are also unknown.

  • Spo2 levels >94–95% appear to reduce pulmonary hypertension, while levels <88–90% may cause pulmonary hypertension. This does not apply to children with congenital cardiac defects and idiopathic pulmonary arterial hypertension.

  • Hypoxia may have adverse effects on cognition and behaviour at Spo2 levels of ⩽85%, but the effects of milder hypoxia are less clear.

Chronic hypoxaemia is a well-established cause of pulmonary hypertension. In animals, the critical level of alveolar Po2 at which the hypoxic pulmonary vasoconstriction reflex is triggered is 100 mm Hg (13.3 kPa).21 In human adults the Sao2 at which pulmonary hypertension occurs is believed to be 88–90%,22 although the duration of hypoxaemia required is not known. The development of pulmonary hypertension in children who have intermittent nocturnal hypoxaemia due to obstructive sleep apnoea suggests that hypoxia does not have to be continuous.23 In the past, pulmonary hypertension was often observed in children with CNLD,24 and in one series it was fatal in 5/17 subjects.25 Supplemental oxygen acutely ameliorates pulmonary hypertension in CNLD24 2628 and, in a prospective case series, right ventricular hypertrophy resolved in infants on a home oxygen programme with saturations maintained above 94–95%.29

2.5.2 Neurodevelopment

  • Hypoxia may have adverse effects on cognition and behaviour at Spo2 levels of ⩽85%, but the effects of less severe hypoxia are less clear.

A systematic review of the cognitive effects of chronic hypoxia in children was conducted in 2004.30 The evidence came almost exclusively from studies of sleep-related breathing disorders and from cyanotic heart disease in children. The conclusion was that there was strong evidence for adverse cognitive and behavioural effects of hypoxia. A subsequent re-analysis of a community-based study of nocturnal oximetry in 995 primary school children found that mildly abnormal nadirs of Spo2 (91–93%) were associated with worse academic performance in mathematics, although this effect was not significant when habitual snoring was excluded.31 There are problems with the extrapolation of these conclusions to children with lung disease. Sleep-related breathing disorders may cause neuropsychological effects from sleep fragmentation or deprivation as well as from hypoxia, and the mean saturation of the children with cyanotic heart disease in the studies cited was 85%, which may be lower than that to which pulmonary patients are exposed. The Benefits of Oxygen Saturation Targeting (BOOST) study, which compared target Spo2 levels of 91–94% vs 95–98%, did not find any differences in developmental status at 1 year, although this does not exclude more subtle later effects on cognition.32 See also section 3.1.4.

2.5.3 Apnoeas/apparent life-threatening events/sudden unexplained death in infancy

  • In infants with CNLD, Spo2 <90% is associated with an increased risk of an apparent life-threatening event while Spo2 ⩾93% is not.

In both term33 and preterm infants,34 reduced fraction of inspired oxygen (Fio2) may lead to an increase in periodic breathing, hypoventilation and central apnoeas, thus hypoxaemia may predispose to apparent life-threatening events (ALTEs). When a group of infants with chronic lung disease was kept at a higher Spo2 (94–96%), they experienced fewer desaturations to <80% compared with when their baseline Spo2 was maintained at 87–91%.35 Baseline hypoxaemia (<95%) was found in 25% of 91 preterm infants who had suffered ALTE/cyanotic episodes, and abnormal hypoxaemic episodes were found in 40%.36 In a small study of 10 infants with bronchopulmonary dysplasia (BPD) who had recently stopped supplemental oxygen (within 7 days), their mean Spo2 was significantly lower and they had significantly more central apnoeas compared with 10 healthy preterm babies.37 When the Spo2 was improved with supplemental oxygen, both central apnoea and periodic breathing densities declined.37 A cohort study of 78 infants with CNLD on home oxygen (target Spo2 93–97%) and 78 matched preterm controls found no difference in the incidence of ALTEs in the two groups (8.9% and 10.5%), and no sudden infant deaths in either group.38 These figures were felt to compare favourably with historical controls from other authors.

2.5.4 Growth

  • In infants with CNLD, Spo2 <92% may be associated with suboptimal growth.

Two observational case series have found normal growth along centiles in babies with CNLD where saturations were maintained at or above 92%39 or 93%.40 In both studies, weight gain faltered if supplemental oxygen was discontinued prematurely. In one of the studies, faltering growth was seen at a Spo2 of 88–91%.39 The BOOST study did not find any advantage in growth at a corrected age of 12 months in those whose Spo levels had been maintained at 95–98% compared with 91–94%.32 See also section 3.1.3.

2.5.5 Sleep

  • In infants with CNLD, Spo2 ⩽90% impairs sleep quality but Spo2 >93% does not.

Infants with CNLD who have Spo2 levels around 90% had impaired sleep quality which improved when supplemented with 0.25 l/min via nasal cannulae;41 this effect was not seen in infants with CNLD in whom the Spo2 was increased from >93% to >97%.42

2.6 What are the consequences of excess oxygen therapy in children?

  • Excess arterial and intra-alveolar oxygen concentrations are toxic in preterm infants and must be avoided by appropriate monitoring and adhering to the target Spo2 level; there are no data in older children. [D]

Oxygen toxicity can broadly be divided into two components: the effects of high arterial blood oxygen concentrations (Pao2) and the effects of high intra-alveolar oxygen concentrations (Pao2). There are no trials that address the question of whether a high Pao2 in term children on home oxygen is harmful. Most literature on high arterial blood concentration focuses on infants who are still premature; for example, the effects on the developing retina are well established.43 High alveolar oxygen levels in premature infants can inhibit lung healing and contribute to ongoing lung injury, possibly through the formation of reactive oxygen intermediates and peroxidation of membrane lipids.44 Oxidative stress from a high oxygen concentration may be a contributing factor to the development of BPD,45 and it is suggested that an Fio2 of 0.8–1.0 for 24 h is associated with the occurrence of BPD.46 The BOOST study (see section 5.2 for fuller critique) showed a non-significant excess of deaths from pulmonary causes in the premature babies kept at a higher Spo2.32 The STOP-ROP study (see section 5.2 for fuller critique) found an increased rate of adverse pulmonary sequelae (pneumonia and exacerbations of CNLD)—although not deaths—in the high saturation group; this group also had more infants still requiring supplemental oxygen at 3 months.43 In summary, six studies of extremely low birthweight infants have shown that retinopathy of prematurity and chronic lung disease are significantly reduced if the Spo2 is kept <93–95% compared with higher saturations (when under 36 weeks gestation).47 It is plausible that some of the adverse effects attributed to hypoxaemia may in fact be due to fluctuating rather than low oxygen delivery, and avoidance of fluctuations in Spo2 also seems to be important.47

There is some evidence that high oxygen levels can also be toxic to the term and mature lung. There has been a case report of a newborn infant with a massive left to right shunt (secondary to a cerebral arteriovenous malformation) who required continuous oxygen therapy in high concentrations.48 Despite a high alveolar Pao2, the infant maintained low to normal arterial Pao2 concentrations. Light and ultrastructural studies of the lungs demonstrated typical changes of acute pulmonary oxygen toxicity. These observations may confirm earlier experimental animal studies which demonstrated that the Pao2 concentration and not the Pao2 is the major factor contributing to oxygen toxicity within the lungs.48 The pathogenesis of oxygen toxicity remains unknown but may involve leucocyte-mediated injury and leukotriene B4 production.49

3. INDICATIONS FOR LTOT

For a variety of conditions, we have assessed the evidence (accepting it is often lacking) as to whether supplemental oxygen is beneficial to patients and home oxygen is preferable to hospital-based oxygen. Benefit is considered in terms of survival, symptoms (breathlessness, respiratory distress, exercise tolerance), growth and neurodevelopment, school attendance and hospitalisation rates, quality of life and psychological impact. Obviously these parameters are not applicable to all patient groups.

3.1 Chronic neonatal lung disease

  • Supplementary oxygen should be given to infants with chronic neonatal lung disease:

    • to reduce or prevent pulmonary hypertension, reduce intermittent desaturations, reduce airway resistance and promote growth; [C]

    • as it is likely to be beneficial for neurodevelopment in infants with CNLD; [D]

    • as it may reduce the associated risk of sudden unexplained death in infancy; [D]

    • as oxygen at home is preferable to a prolonged hospital stay for both quality of life and psychological impact for the infant, parents and family; [D]

    • as it saves days in hospital due to earlier discharge despite a significant readmission rate. [C]

For the purposes of these guidelines, the diagnosis of CNLD is defined as an infant requiring supplemental oxygen at a corrected age of 36 weeks gestation who is at least 28 days old.

The pathophysiological effects of chronic hypoxia support the use of supplementary oxygen in infants with CNLD. Pulmonary hypertension is a relatively common complication of CNLD that can cause diminished right ventricular performance and, eventually, cor pulmonale.50 Infants with CNLD who have pulmonary hypertension generally have reactive pulmonary vascular beds responsive to supplemental oxygen.26 51 Right ventricular hypertrophy and pulmonary hypertension can resolve with home oxygen therapy.29 Supplementary oxygen in CNLD may also reduce the frequency of intermittent desaturations.37 52 Oxygen given to mildly hypoxic infants may cause a decrease in total pulmonary resistance.53 The pathogenesis of left ventricular hypertrophy has been attributed to the metabolic effects of chronic hypoxaemia in addition to hypercarbia and acidosis. If the hypertrophy is severe enough, it may cause an increase in left atrial pressure, thereby potentially contributing to pulmonary oedema and the severity of CNLD.54

3.1.1 Survival

Home oxygen might in part be responsible for improved survival in CNLD through its role in the treatment or prevention of pulmonary hypertension. In addition, because patients with CNLD may have an abnormal response to hypoxia that can lead to prolonged apnoea and bradycardia, maintaining the Spo2 at appropriate levels may decrease the higher incidence of sudden unexplained death in infancy in this patient group.55 For infants who have had an ALTE while already on home oxygen, an insufficient amount of supplementary oxygen may have been given.56 Previously, preterm infants with CNLD not on home oxygen at discharge but who subsequently suffered an ALTE have been shown to have episodic or baseline hypoxaemia which improves with home oxygen.36 In this study, 33 premature babies who had suffered ALTEs were given supplemental oxygen (0.1–1.0 l/min via nasal cannulae) for up to 17 months (median 3.9 months). There were no further ALTEs in 24 and a reduction in severity in 7 (in 2 of whom other causes were found).36 The provision of home oxygen when appropriate should eliminate episodes of unrecognised and untreated hypoxaemia so that preterm infants with CNLD are no longer at increased risk from sudden infant death compared with other preterm infants. It is thought that infants with CNLD who die suddenly may have had clinically unrecognised periods of hypoxaemia.38 57

3.1.2 Symptoms

Supplemental oxygen can reduce the demands on an already stressed respiratory system by decreasing the respiratory rate and the work of breathing needed to provide improved oxygenation, thus reducing symptoms.42

3.1.3 Growth

It is suggested that supplemental home oxygen improves growth in infants with CNLD when the saturations are kept above 92%.40 58 59 Also eliminating sleep-associated hypoxia improves growth in infants with CNLD.39 60 See also section 2.5.4.

3.1.4 Neurodevelopment

As so many factors act as confounders for neurodevelopment in this group of children, the relative contribution of CNLD (with or without oxygen supplementation) remains uncertain. Gestational age, birth weight, hypoxic-ischaemic events and intracranial haemorrhage are all independently associated with worse neurodevelopmental outcomes.61 62 Many studies fail to ensure adequate length of follow-up. This is a crucial methodological problem given the increased recognition that preterm children are at risk of a range of more subtle neurodevelopmental problems which have considerable impact on daily living and school performance. These will not be picked up reliably by routine follow-up or even specific testing until around school entry age at 5 years.46

Nevertheless, supplementary oxygen is likely to be beneficial for neurodevelopment.63 Although early assessment at 1 and 2 years of age show lower developmental scores in infants with CNLD discharged home on oxygen, by 4 years and above their development did not differ significantly from controls.62 This was despite severity of illness, duration of oxygen therapy and feeding problems often being greater among those sent home on oxygen.62 However, in this paper, 7.5% of the BPD group were excluded from the study due to “disability” compared with 4.4% of the non-BPD group; children in the non-BPD group also had a higher mean birth weight and gestational age than the two groups of children with BPD. See also section 2.5.2.

3.1.5 Hospitalisation rates

During the first year all infants with CNLD are at increased risk for readmission to hospital, but some studies have shown no further increase in those on home oxygen. Such infants tend to have more frequent and longer hospital stays with non-respiratory problems (eg, failure to thrive) than extremely low birthweight infants without CNLD.58 64 Other studies show that infants with CNLD who require home oxygen have more frequent and longer hospital admissions over the first 2 years after discharge and more clinic attendances for the first 4 years than those with CNLD sent home without oxygen.65 66 This may reflect community support and parental education, and any readmissions to hospital should be analysed for problems related to teaching or the discharge plan.67 When comparing centres with a high rather than a restricted use of home oxygen therapy, early discharge and high use was not associated with increased morbidity,68 and an earlier initial discharge saved significantly more hospital days despite the frequent need for readmission.29

3.1.6 Quality of life and psychological impact

Although there have been no randomised trials, it is suggested that caring for infants on supplementary oxygen at home is preferable to a prolonged hospital stay.69 It reduces the risk of nosocomial infection (although friends and rel