We observed significantly higher deaths in children with diarrhea and pneumonia and metabolic acidosis compared to those without metabolic acidosis. The children having diarrhea and pneumonia with metabolic acidosis more often had younger age and frequently presented with severe wasting compared to those without metabolic acidosis. Young infant compared to their elder counterpart, usually do not have adequate immunity, more vulnerable to severe infection, and likelihood of fatal outcome 
. Moreover, malnutrition in children often causes depressed cell mediated and humoral immune responses, frequently associated with impairment of IgA production, inefficient chemotaxis, reduced mature T cells, and compromised phagocytic activity 
. As a result, they become highly susceptible to severe infection, often leading to sepsis 
and subsequent severe metabolic acidosis and death. Furthermore, significantly higher proportion of diarrheal children with pneumonia and metabolic acidosis compared to those without metabolic acidosis had clinical dehydration which is often associated with high case-fatality 
. This observation might have an additional impact on the higher deaths in diarrheal children with pneumonia and metabolic acidosis. Thus, these explain our observation of frequent deaths in children with diarrhea and pneumonia with metabolic acidosis. Our infrequent observation of fever among the diarrheal children with pneumonia and metabolic acidosis might be due to the fact that most of the children presented with severe malnutrition which results in poor febrile response.
Our observation of more frequent low systolic blood pressure despite correction of dehydration among the diarrheal children with pneumonia and metabolic acidosis compared to those without metabolic acidosis is also understandable. Low systolic blood pressure even after correction of dehydration is one of the markers of poor peripheral perfusion, which is a feature of severe sepsis in under-five children 
. Severe sepsis is often associated with vasodilatation and capillary leakage as a result of amplified cytokines or other inflammatory stimuli 
. This leads to disordered microcirculation and increased lactate production as a by-product of anaerobic cellular respiration leading to metabolic acidosis 
Our observation of frequent clinical dehydration among the diarrheal children with pneumonia and metabolic acidosis compared to those without metabolic acidosis is very classical. Children with dehydrating diarrhea often have fecal loss of bicarbonate and often develop metabolic acidosis 
. This finding is consistent with our earlier observations 
Thus the observation of clinical dehydration and low systolic blood pressure even after correction of dehydration as independent predictors for metabolic acidosis in children under five with pneumonia and diarrhea after adjusting for potential confounders, such as age of the patient, fever on admission, and severe wasting, indicates that metabolic acidosis is a serious feature of underlying pathologies. This observation is very important and has high clinical relevance in diarrheal children with pneumonia especially in resource limited setting where there is a lack of opportunity in performing total serum carbon-dioxide. Early identification of these clinical parameters of metabolic acidosis by the health workers might help in making decision for prompt rehydration and subsequent early referral to tertiary hospitals for further and appropriate management in order to reduce morbidity and deaths in such children.
The observation of lack of influence of metabolic acidosis on cough, respiratory rate, lower chest wall indrawing, nasal flaring, head nodding, grunting respiration and cyanosis in diarrheal children with pneumonia is very important. High respiratory rate and lower chest wall indrawing are the basis of the WHO algorithm for diagnosing pneumonia in under-five children 
. It is expected that more children with metabolic acidosis and pneumonia would be tachypnoeic. In this study population, however, a significantly higher proportion of children with diarrhea, pneumonia and metabolic acidosis presented with severe acute malnutrition compared to those without metabolic acidosis, and this may have confounded the results. The poor inflammatory response in severely malnourished children impedes an adequate rise of respiratory rate even when there is metabolic acidosis 
. Moreover, severely malnourished children have deficient total body electrolytes (except sodium), especially potassium and magnesium 
, and this may be partly responsible for the reduced work of the respiratory accessory muscles, even in the presence of metabolic acidosis, and resulting in the lack of fast breathing and lower chest wall indrawing in such population. The observation of lack of influence of metabolic acidosis on nasal flaring and head nodding might be due to the same reason. However, the comparable distribution of cough, grunting respiration, cyanosis and hypoxemia in both the groups might be due to the fact that both the groups had pneumonia and these features are the consequence of pneumonia. The failure to find the association of sex, poor socio-economic condition, vomiting, non-breast-fed, and abnormal mental status among the groups might be due to small sample size.
Although we have underscored the importance of metabolic acidosis measured by total serum carbon-dioxide in diarrheal children with pneumonia, the actual cause of acidosis in our study cases may be mixed rather than metabolic only. All of our study children had both pneumonia and diarrhea. Acidosis in pneumonia is usually respiratory in origin 
and on the other hand, diarrhea leads to metabolic acidosis 
. Thus, the acidosis in our study population is likely to be mixed acidosis.
In conclusion, metabolic acidosis in under-five children having diarrhea and radiological pneumonia is associated with increased deaths compared to those without metabolic acidosis. Diarrheal children with radiological pneumonia presenting with clinical dehydration, and low systolic blood pressure even after correction of dehydration are more likely to have metabolic acidosis. Metabolic acidosis, however, has no impact on the diagnostic clinical features of radiological pneumonia in such children. It underscores the importance of early initiation of appropriate antibiotics in under-five diarrheal children with clinical features of pneumonia irrespective of the presence or absence of metabolic acidosis and simultaneously it is important to rule out severe sepsis in under-five diarrheal children with pneumonia who have metabolic acidosis, in an effort to reduce morbidity and deaths.