This was a prospective observational study conducted over a period of one year from Septmber 2013 to August 2014 in a SCB Medical College, Cuttack, Odisha, India. Informed consent was obtained from the parents prior to inclusion of subjects (infants) into the study. Ethical committee approval was obtained from the Institutional Ethical Committee prior to the study.

Inclusion criteria involved patients between two-months to five-year of age having features of severe and very severe pneumonia. The diagnosis of severe and very severe pneumonia was based on WHO criteria [6]. A patient was considered to have severe pneumonia when chest in-drawing was present along with fast breathing. Very severe pneumonia was considered when patient presented with any one of the following signs i.e., cyanosis, severe chest in drawing, feeding difficulty, along with fast breathing. Fast breathing was considered to be present when the respiratory rate was ≥ 50 breaths per minute for infants of 2 to 12-months of age and ≥ 40 breaths per minute for children between 12-months to five-years of age.

Exclusion criteria: Children below two-months of age and more than five-years, with underlying heart disease or pulmonary tuberculosis presenting as pneumonia, were excluded.

After admission of a child; detailed history was taken from the parents regarding the onset, duration and progress of symptoms. Education and socio-economic status, population and overcro-wding in the family, parental smoking etc., were noted as per the case record proforma specially designed for the study. A thorough clinical examination including respiratory as well as other systems were done. Anthropometric measurements like height, weight, chest, head and mid arm circumference were recorded in each case; which were then compared with the available norms of 50^th percentile of National Centre for Health and Statistics (NCHS) standards. The weight-for-age criteria was taken for the diagnosis of different grades of malnutrition, as per the nutrition subcommittee of Indian Academy of Paediatrics in the year 1972. Socioeconomic class was identified using Modified Kuppuswamy Scale.

Haemoglobin, Erythrocyte Sedimentation Rate (ESR), Total Leucocyte Count (TLC), peripheral smear for differential leucocytes count and blood picture were done in each patient. Chest X-ray was done to examine the site, type and associated complications. Blood gas analysis was done in anticipation of assisted ventilation. Pulse oximetry was done to detect the degree of hypoxia. CT scan of thorax was done if obstruction or tumour was suspected. Bronchoscopy was done when there was evidence of an obstructing tumour or foreign body.

Blood culture was collected from a vein under strict aseptic measures using povidone-iodine, rectified spirit for skin sterilization. About 5 ml of blood was collected into 50 ml of glucose broth (the ratio being 1:10) bottle, which was observed for growth of organisms after 24 hours and 72 hours.

Aetiological diagnosis was established on the basis of isolating the organism from the culture and sensitivity reports. Clinical monitoring and recording of data were done daily until discharge or death. All the data were serially tabulated and critically analyzed.

Data analysis was done using the Microsoft excel, SPSS for Windows version 23.0. Values for continuous data were expressed as mean ± SD (if normally distributed) and median (range) (if not normally distributed). Categorical variables were reported as proportions. Continuous variables with normal distribution were compared using Student t test while those not normally distributed were analyzed using Mann-Whitney U test. Categorical data were analyzed using Pearson Chi square test or Fischer Exact test.

Out of 10300 patients admitted to our hospital during the study period 141 patients were included in the study based on the inclusion and exclusion criteria.

[Table/Fig-1] shows the demographic profile of the studied population. The median age was five months (range 2-60 months). Most of the patients were male 86 (61%). More number of cases were admitted during the months of January, February and August. One hundred nineteen (84%) patients were in the age group of 2- 12 month. Only 36 (25%) patients were born with a birth weight ≤ 2500 gm. Most of the patients 127 (90%) included in the study had grade IV malnutrition. About 137 (97%) patients were completely immunized according to their age. Eighty five (60%) patients were exclusively breastfed. Eighty one (58%) patients were from middle socio-economic status. Ninety six (68%) patients did not have any history of exposure to smoke.

[Table/Fig-2] shows severe and very severe pneumonia were present in 84 (60%) and 57 (40%) of patients respectively. Seventeen percent of patient had blood culture positive and Staphylococcus aureus was the most common organism isolated. Twenty four (17%) patients developed some form of complication during hospital stay. Thirty one (22%) patients died while undergoing treatment.

[Table/Fig-3] shows that both the groups (severe pneumonia and very severe pneumonia) were comparable to each other in respect to age, sex, nutritional status, immunization status, Socio Economic Status (SES), feeding practice and exposure to smoke.

[Table/Fig-4] shows mortality was higher in the very severe pneumonia group and the difference was statistically significant (p=0.003). Mortality was higher among the patients who were admitted to hospital beyond 72 hours of onset of illness (p=0.048).

In our study, most cases of severe pneumonia and very severe pneumonia, i.e., 119 (84%) were reported in the age group of 2 months-12 months. In the age group of 12 months to 60 months, 22 (16%) of cases were present. Debnath D et al., [7] in their study found occurrence of severe pneumonia /very severe illness in 38.2% cases in the age group 2 months-12 months and 29.9% in the age group 12 months to 60 months. Banstola A et al., [8] also found occurrence of pneumonia more in the age group below one year as compared to age group one year-five year. Major occurrence of severe pneumonia cases in less than one year age group could be attributed to low immunity, smaller and narrower airways and frequent exposure to infection, poor nutritional status as well as more susceptibility of infants to viral and general infections [9-11] .

Male to female ratio in our study was 1.56:1. Similar observations were noticed by Debnath D et al., [7] in a study in Pune, India and found male to female ratio to be 1.4:1. Study by Banstola A et al., [8] in Nepal found male to female ratio to be 1.5:1. Divyarani D C et al., [12] in a similar study in Karnataka, India, observed 62.6% were males and 37.4% were females (ratio 1.67:1). More number of male cases were admitted for treatment of pneumonia in the age group of under five. This might be due to preferential treatment given to male child in family. Another possibility is that boys are more exposed to the outer environment though under five male and female children are usually equally exposed. The lower incidence in girls might be due to some inherent immunity in terms of extra ‘X’ chromosome present with their genotype as per article of Libert C et al., [13].

In the current study, increased number of cases was seen in January, February and August i.e., in winter and rainy season respectively. Banstola A et al., [8] also observed increased number of cases in winter and rainy season. Rudan I et al., [5] also reported peak occurrence of pneumonia in the rainy season in tropical climates. There was similar trend in our study as per seasonal distribution. Monthly distribution of cases in our study was different and it could be due to difference in geographical location and delay in reaching of monsoon.

In our study, 36 (25%) of cases were Low Birth Weight (LBW) at birth, 105 (75%) cases were having normal weight at the time of birth. Divyarani DC et al., [12] in a similar study in Karnataka, India, reported 84% of cases were LBW at birth in their study. This disparity was because of our study population which is two months to 60 months whereas in the study of Divyarani DC et al., the age group was birth to 60 months [12].

In the present study, 7 (4.9%) of cases were having normal nutritional status whereas 134 (95%) were malnourished and 127 (90%) cases had Grade IV malnutrition. Debnath D et al., [7] in a similar study in Pune, India found no malnutrition in 9.5%, mild malnutrition in 28.9% and moderate /severe to very severe malnutrition in 44.7% of their studied cases.

Various community based study also marked malnutrition as a major contributor of pneumonia. A Delhi based study observed that severe malnutrition was a significant contributor of acute respiratory infection in children under-five years [14] and a study from Kolkata showed that malnutrition was significantly associated with occurrence of Acute Respiratory Infection (ARI) among under-five children [15]. Malnutrition causes defective cell mediated immune response thus predisposes to infection [5,16].

In our study, 137(97.16%) cases were immunized as per National Immunisation Schedule (NIS) completed for age, and only 4(2.84%) were unimmunized. Debnath D et al., [7] in found 23.8% cases were immunized complete for age as per NIS and in 51.2% cases immunization was incomplete as per age. This difference might be due to better immunization (as per NIS) coverage in our study area as well as increased awareness among parents. Immunization has a major role in preventing pneumonia as it is one of the common complication of both pertussis and measles.

In our study, 85 (60.28%) cases were exclusively breast fed, 21 (14.89%) cases were on spoon feeding and 35 (24.82%) cases were on bottle feeding. Divyarani DC et al., [12] in a similar study in Karnataka, India, found 94.6% cases were on exclusive breast feeding. This low figure of exclusive breast feeding in our studied cases could be due to poor awareness of benefits of exclusive breast feeding among mothers.

Majority of cases belonged to upper lower (IV), lower middle (III) class i. e., 51 (36.17%) and 42 (29.78%) respectively in our study and 39 (27.65%) cases were from upper middle (II) class. Kumar AMK et al., [17] in their study from Karnataka observed 76.5% of occurrence of pneumonia in lower social status (class III-V). Thorn LK et al., [18] also reported inverse relationship between cases of pneumonia and SES class.

In the current study, 31 (21.98%) cases were exposed to firewood cooking and 14 (9.92%) were exposed to passive smoking, but in 96 (68.08%) of studied cases there was no exposure to smoking. Literature says parental smoking has direct impact on childhood clinical pneumonia in the community in developing country [5].

In our study, cough, tachypnea, and chest in-drawing were present in 141 (100%) of studied cases, fever in 137 (97.16%), severe respiratory distress in 57 (40.42%) and central cyanosis in 1(0.7%), inability to take food or refusal to food in 57 (40.42%) cases. Yaguo Ide LE et al., [19] reported cough in 75.9%, fever in 70.7%, fast breathing in 53%, respiratory distress in 83.6% of pneumonia cases. Ahmad Al Najjar S et al., [20] reported fever in 87.4%, tachypnea in 73.5%, cough in 98% and retraction in 80% cases of pneumonia which is almost tallying with our reports.

In the present study, blood culture was positive in 24(17.02%) of cases out of which Staphylococcus aureus was 12 (8.51%) cases, Streptococcus pneumoniae in 9 (6.38%) cases and Haemophilus influenzae in 3(2.1%) cases. Mudhusudhan K et al., [21] in their study from Telengana, India, observed blood culture positive in 23.63% cases with predominance of Staphylococcus aureus. Banstola A et al., [8] in their descriptive study found Streptococcus pneumoniae in 15.8% cases, Haemophilus influenzae type b (Hib) in 9.2%, Respiratory Syncytical Virus (RSV) in 12.6%, Influenzae A in 7.5%, Parainfluenzae type 3 in 4.0% and Influenzae B in 3.5%. Viral isolation method could not be adopted because of lack of PCR facility as well as inaffordability of most of the parents.

In the present study, 90 (63.83%) cases reported to the hospital within first 72 hours, while 51 (36.17%) cases reported to hospital after 72 hours of onset of illness. Mortality observed among patients admitted before and after 72 hours of illness was 15 (10.63%) and 16 (11.34%) respectively.

Average hospital stay among survivors was 10.75±6.7 days where as among the non survivors it was 3.94±4.32 days. Kumar AMK et al., [17] in their study from Karnataka observed that duration of hospital stay was more in very severe pneumonia group. The duration of hospital stay was less in very severe pneumonia group in our study, because more number of deaths occurred early due to more serious nature of illness.

Complications in our study were septicaemia 8 (5.6%), CCF 6 (4.2%), empyema 6 (4.2%) and pleural effusion 4 (2.8%). Kumar AMK et al., [17] in their study from Karnataka observed similar trends in complications.

Overall mortality in our study was 31 (21.98%). Mortality seen in severe pneumonia was 9 (6.38%) and in very severe pneumonia was 22 (15.6%). Kumar AMK et al., [17] in their study observed mortality of 3%. The higher mortality rate observed in our study could be due to more number of very severe pneumonia cases in contrast to their study population. More seriousness of the illness as well as delay in reaching in tertiary center could be another attributing factor.

This was a hospital based study and does not represent the true incidence of pneumonia and its different forms and aetiologies of the entire population. Viral pathogens were not routinely identified in this study as the diagnosis and treatment was based on WHO case management protocol and non-availability of PCR facility for viral isolation in our centre.

The incidence of severe and very severe pneumonia was high in infancy. More number of patients died who arrived at hospital after 72 hours of onset of illness and most death were recorded in children who had very severe pneumonia. Hence, early referral is of paramount importance as it will not only halt the progression of severity of disease but death as well. A detailed prospective study is needed to further confirm the findings.