Prevalence and outcome of anemia among children hospitalized for pneumonia and their risk of mortality in a developing country


Study site

The study was performed in Dhaka Hospital of International Center for Diarrhoeal Disease Research, Bangladesh (icddr,b), Dhaka, Bangladesh. The description of Dhaka Hospital has been provided somewhere else10.

Study population

Children aged 0–59 months with pneumonia/severe pneumonia, admitted to Dhaka Hospital of ICDDR,b between August’2013 and December’2017, whose hemoglobin (Hb) level was measured, constituted the study population.


The data were extracted from electronic database of Dhaka Hospital of icddr,b. We stratified the study children into an anemic and non-anemic group. Anemia was defined if blood Hb values ​​were ≤ 10.7 g/dL, ≤ 9.4 g/dL, ≤ 9.5 g/dL, and ≤ 11 g/dl among children aged ≤ 1.0, > 1.0 to 2.0, > 2.0 to < 6.0, and ≥ 6.0 to 59 months of age, respectively11,12. The attending physician decided the selection for Hb investigation and was done in our icddr,b laboratory. An initial baseline comparison was made between the children who had anemia and those who did not have anemia. The final comparison was made between the study children who had a fatal outcome and those who survived during hospitalization.

Patient management

Study children received management according to standard treatment following hospital guideline that has been described elsewhere13.


We have used pre-tested case report forms for initial acquisition of study relevant data from the electronic database before incorporating them in to a personal computer. We have evaluated baseline characteristics (Table 1) and outcome (Table 2) of children hospitalized for pneumonia/severe pneumonia with anemia compared to those without anemia. We further evaluated the characteristics (including anemia) of children who survived or expired after being hospitalized with pneumonia/severe pneumonia (Table 3).

Table 1 Baseline characteristics of under-five children with anemia hospitalized for pneumonia/severe pneumonia.
Table 2 Outcome of under-five children with anemia hospitalized for pneumonia/severe pneumonia.
Table 3 Factors associated with deaths for under-five children hospitalized for pneumonia/severe pneumonia who had their hemoglobin done.

Operational definition

Pneumonia/severe pneumonia, severe acute malnutrition, severe underweight, invasive diarrhea, dehydration, hypoxemia, bacteremia, and anemia were defined according to WHO guideline5. We have also used evidence-based definition of severe sepsis14. Respiratory failure was defined if the proportion of arterial oxygen saturation (measured by using a pulse oximeter) and fraction of inspired oxygen (SpO2/FiO2) was < 31515th. Abnormal mental status was defined if a child had restlessness, or lethargy, or coma, or unconsciousness. These variables were shown in Tables 1, 2 and 3.

Statistical analysis

We analyzed the data using SPSS for Windows (version 20.0) and STATA(version 15.0). For initial baseline analysis (Table 1), anemia was the dependent variable whereas sex, lack of immunization, mental status, severe acute malnutrition, invasive diarrhea, dehydration, hypoxemia, bacteremia, and severe sepsis were the main independent variables. For the outcome table (Table 2) still anemia was the dependent variable whereas respiratory failure, duration of hospital stay and death were the independent variables. The dependent variable of our next analysis (Table 3) was death and main independent variables remained the same in addition to anemia. Normality test was done to understand the distribution of data. We have performed a non-parametric test showing median (IQR), if data having continuous variables were not normally distributed. For normally distributed continuous variable we were intending to perform student ‘t’ test showing mean ± SD but we did not have any variable that require this test. We also performed frequencies and percentages for categorical variables. Differences in proportion for categorical variables were compared by the bivariate analysis. Those who had significant associations between the groups (study children who died compared to those who survived) by the bivariate analysis were put into a multivariable logistic regression model. In the multivariable logistic regression model of death (during hospital course) was the dependent variable and significantly associated factors (on admission) with death that included anemia were the independent variables. We have also performed Chi-square for linear trend to understand the relation of anemia with increasing age. A p value of less than 0.05 was considered statistically significant. Strength of association was determined by calculating risk ratio (RR) (however, odds ratio [OR] for Table 3) and their 95% confidence intervals (CIs). Additionally, the power of the analysis (anemia vs. non-anemia for outcome of deaths) calculated by using Stata (version 15.0) was 96% with an effect size of 29%.

Institutional Review Board statement

The name of the ethics committee of icddr,b is “Ethical Review Committee”. All the study protocol was approved by the “Ethical Review Committee” of icddr,b following all methods in accordance with the relevant guidelines and regulations. The approval date of the project was 12 January 2020.

Informed consent

As it was a retrospective chart review where data were anonymous and no care-giver interview was required, thus, no informed consent was taken. It is important to note that “Ethical Review Committee” of icddr,b waived the need for informed consent of the study.


Please enter your comment!
Please enter your name here