Clinical predictors of hypoxaemia in radiology confirmed pneumonia in 2-60 months old children
Forrukh Ahammad1, M Ruhul Amin2
Background: Pneumonia is one of the commonest causes of death in children in
developing countries. One of the major fatal complications of pneumonia is hypoxaemia.
However detection of hypoxaemia by use of pulse oxymetry is not feasible in most
situations in developing countries. Therefore it is important to accurately identify
hypoxaemic children by use of clinical signs alone.
Objectives: The study was conducted to determine the clinical predictors of
hypoxaemia in children with radiology confirmed pneumonia.
Methods: This is a cross-sectional study conducted at Dhaka Shishu Hospital. Total
164 cases of childhood pneumonia (2-60months) confirmed by chest radiograph were
included in this study. Clinical symptoms and signs were evaluated for their ability
to predict hypoxaemia. Oxygen saturation was determined by pulse oximeter.
Hypoxaemia was defined as oxygen saturation less than 90%. Clinical predictors
were evaluated by (χ2 test) and also by determining sensitivity, specificity, positive
predictive value (PPV) and negative predictive value (NPV). Few combination models
were also evaluated for their ability to predict hypoxaemia.
Results: Of 164 cases, 114 (69.5%) children were hypoxaemic. Fast breathing, chest
indrawing and nasal flaring were highly associated with hypoxaemia. Rhonchi,
crepitation and lethergy had significant association with hypoxaemia. The sensitivity
and specificity of fast breathing was 73.6%, 60.0% and that of chest indrawing was
95.6% and 24.0%. Nasal flaring had 35.9% sensitivity and 84.0% specificity. Various
combination models were also tried to find out their ability to predict hypoxaemia. A
simple model using the presence of fast breathing (WHO criteria) or nasal flaring had
a sensitivity of 82.4% and specificity of 56.0% for detecting hypoxaemia. Other models
were respiratory rate cutoff of >60/min in children 2-12 months, >50/min in 12-60
months age group or nasal flaring had a sensitivity of 70.1% and specificity of 66.0%;
fast breathing or lethergy or nasal flaring (sensitivity 85.0% and specificity 40.0%)
and finally fast breathing or chest indrawing or lethergy (sensitivity 95.6% and
specificity 18.0%) for detecting hypoxaemia. No individual clinical symptom/sign or
a combination had both desirable sensitivity and specificity to identify hypoxaemia.
Conclusion: The study concluded that fast breathing, chest indrawing and nasal
flaring were significantly associated with hypoxaemia. Combination of clinical signs
improved the sensitivity and specificity to predict hypoxaemia.
Key words: Pneumonia, Predictors, Hypoxaemia