CLINICAL PREDICTORS OF HYPOXAEMIA IN UNDER FIVE CHILDREN WITH PNEUMONIA AT THE UNIVERSITY OF ILORIN TEACHING HOSPITAL

Abstract

Hypoxaemia constitutes a possible complication of severe respiratory illness which is often
under-reported in developing countries. In view of this, the current study was carried out to
determine the prevalence and clinical predictors of hypoxaemia in hospitalized under-five
children with pneumonia in Ilorin. An association between the risk factors of pneumonia and
hypoxaemia, as well as the relationship between various levels of hypoxaemia and the
management outcome of pneumonia were also explored.
This is a descriptive cross-sectional study of 200 children aged between two months and five
years with pneumonia recruited consecutively as subjects. Socio-demographic,
anthropometric, clinical and laboratory data were obtained, while the admission diagnosis
was based on clinical features. The pulse oximetry measurement was recorded after a stable
reading for at least one minute while the child was breathing room air. Hypoxaemia was
defined as an arterial oxygen saturation of less than 90% as recorded by pulse oximetry
Blood samples were collected for determining the packed cell volume, total and differential
white blood cell (WBC) count and blood cultures. Also, chest radiographs were obtained in
all subjects. Data was analyzed using the IBM®SPSS 20.0 (2011) software package.
There were 119 males and 81 females (M:F=1.5:1). Severe pneumonia was present in 188
(94.0%) children while 12(6.0%) children had moderate pneumonia. Bronchopneumonia was
identified in 168(84.0%) of the children while lobar pneumonia was diagnosed in 32(16.0%)
children. The prevalence of hypoxaemia in the children with pneumonia was 41.5%.
Using a linear regression analysis, the clinical features that were significantly associated with
hypoxaemia were restlessness, lower chest wall indrawing, bronchial breath sounds and
tender hepatomegaly (p <0.05). Restlessness had a sensitivity of 22.9%, specificity of 91.5%,
positive predictive value (PPV) of 65.5% and a negative predictive value (NPV) of 62.6%,
while chest wall indrawing had a sensitivity of 86.7%, specificity of 53.3%, PPV of 56.7%
and NPV of 84.9% for detecting hypoxaemia. Bronchial breath sound had a poor sensitivity
(16.9%), a high specificity (95.7%), PPV of 73.7% and a NPV of 61.9%, whereas tender
hepatomegaly had a sensitivity of 48.2%, specificity of 82.9%, PPV of 66.7% and a NPV of
69.3%.
There was a negative correlation between the social class of the child and the presence of
hypoxaemia (r= -0.191, p=0.007). Also, each of maternal age (years), maternal literacy level,
birth order of the child and lack of immunization had a negative correlation with the presence
of hypoxaemia (r= -0.151, -0.162, -0.154, -0.148; p=0.032, 0.022, 0.030 and 0.036
respectively).
Seventeen of the children with pneumonia died, giving a corresponding case fatality of 8.5%.
The mean(SD) SpO2 level of 78.3(10.9) percent recorded among the fatal cases was
significantly lower compared to the corresponding value of 91.5(7.8) percent recorded in the
survivors (p=0.001). The mean(SD) duration of hospitalization in hypoxaemic children of
6.9(6.4) days was significantly longer compared to the corresponding value of 4.9(2.7) days
recorded in those without hypoxaemia (p=0.002). Also, the mean duration of hospitalization
increased significantly as the SpO2 levels reduced (p=0.002).
It is concluded that there is a high local burden of pneumonia-associated hypoxaemia and that
hypoxaemic-related pneumonia is frequently associated with a fatal outcome. It is
recommended that pulse oximeters be made available in facilities where pneumonia is
managed. There is also a need to emphasize the potential value of clinical parameters like
restlessness, lower chest wall indrawing, bronchial breath sounds and tender hepatomegaly
for detecting pneumonia-related hypoxemia in poorly equipped health facilities.

CLINICAL PREDICTORS OF HYPOXAEMIA IN UNDER FIVE CHILDREN WITH PNEUMONIA AT THE UNIVERSITY OF ILORIN TEACHING HOSPITAL