• hypoxaemia
• respiratory illness

Duke et al are to be commended for their interesting report aimed to determine normal oxygen saturation values in healthy infants and children and to assess the performance of clinical signs for predicting hypoxaemia in sick neonates and children with and without acute lower respiratory infections (ALRI).¹

Acute lower respiratory infections (ALRI) account for a substantial burden of disease in children and adults, pneumonia being the leading cause of deaths in children under five, particularly in developing countries. Tachypnoea and chest retraction have been shown to be the most useful clinical signs for determining the presence of pneumonia and thus they are widely used in the diagnosis and management of this condition in children.² The World Health Organization pneumonia case detection and management programme,³ which relies on these simple signs, seems to be justified by the existing body of evidence.

Varying degrees of hypoxaemia may be present in children with pneumonia. However, surprisingly few studies have been performed to assess normal values of haemoglobin oxygen saturation (SpO2) through the use of transcutaneous pulse oximetry, at both sea level and high altitude. Singhi’s response to Duke et al rightly emphasises that altitude of studies reported must be taken into account in the interpretation of their results.⁴ There are some reports on SpO2 values at mid- and high altitude settings in healthy and sick children.^5–8 We previously reported normal values of SpO2 in 1264 healthy children and adolescents living at 4100 m.⁹

The main conclusions of these studies performed at different altitudes are: firstly, values considered abnormal at sea level are very frequently found at high altitude in healthy children; secondly, normal values vary for different altitudes; thirdly, recommended SpO₂ cut offs for giving supplementary oxygen to sick children at sea level are clearly not applicable to high altitude settings, as according to these recommendations oxygen should be administered for values below 92%.² There is a need to perform more studies for determining which cut off values for supplementary oxygen are related to better outcomes in sick children living at high altitude. Moreover, our study at 4100 m revealed that SpO₂ values may be different according to different ethnic groups and history of exposure to high altitude. Higher SpO₂ values in Quechua children suggest a better degree of adaptation to high altitude in native populations with a longer time to exposure to high altitude. This latter finding has obvious practical implications, as high altitude native children, with higher baseline oxygen saturation levels than newcomers or resident non-native children, may need oxygen at higher cut off SpO₂ values when they are sick.

Singhi is justifiably concerned on the cost of giving oxygen to children who may not need it. Oxygen may be unacceptably expensive for health services in developing countries, particularly at primary level, where most sick children seek health care. However, hypoxaemia may be a serious, life threatening problem in sick children, particularly at high altitude, and thus we need to extend the study of Duke et al for different altitudes, in healthy and sick infants and children, to determine normal values of SpO₂ and to identify highly predictive clinical signs of hypoxaemia. The potential aggravating role of co-existing prevalent childhood diseases other than ALRI, namely diarrhoea, malnutrition, malaria, and HIV/AIDS, is also an area that warrants more attention. These data will allow providing both good quality and cost effective health care to sick children with and without ALRI.

Millions of children and adults live at high altitude. Developing a medicine based on scientific evidence that can be applicable to this setting is a major public health challenge for all of us working in those parts of the world.