Hospital-acquired malnutrition in children in a tertiary care hospital

Date of Award


Document Type


Degree Name

Master of Medicine (MMed)


Paediatrics and Child Health (East Africa)


Introduction: Hospital-acquired Malnutrition occurs as a result of reduction in food intake, increased dietary loss and/or increased calorie requirements as a result of disease-induced high catabolic state. A child's nutritional status often deteriorates after admission to the hospital resulting in longer duration of hospital stay and increased risk of complications, which also increases treatment cost. Hospital-acquired Malnutrition is usually assessed using anthropometric measurements and/or serum pre-albumin.

Objectives: This study sought to investigate the incidence of Hospital-acquired Malnutrition using anthropometric measurements and to determine diagnostic utility of serum pre-albumin in predicting weight change. A secondary objective was to identify factors associated with Hospital-acquired Malnutrition in children admitted at Aga Khan University Hospital, Nairobi.

Methodology: This was a hospital-based, short follow up longitudinal survey carried out on children admitted to Aga Khan University Hospital, Nairobi. One hundred and eighty children were enrolled into the study. Weight and height were taken at admission and discharge. Serum pre-albumin levels were taken at admission and repeated after 48-96 hrs.

Data Analysis: The incidence of Hospital-acquired Malnutrition was estimated from the total number of children showing a decrease in weight-for-height/length or BMI Z-scores from the time of admission to discharge. Sensitivity, specificity, positive and negative predictive values of serum pre-albumin were calculated to determine its diagnostic utility in screening for Hospital-acquired Malnutrition. Levels of serum pre-albumin were summarized using a Pre-albumin Risk Stratification model by Bernstein et al. Logistic Regression Analysis, with calculation of odds ratio, was done for selected variables to look for any association with loss of weight during hospitalization.

Results: The study showed that 103 (60.6%) children lost weight from admission to discharge, giving a mean weight decrease of 0.5kg (SD±3.37), p=0.055. Of the 170 children studied, 109 who were ≤60 months of age, demonstrated a mean decrease in weight-for-height/length Z-score of 0.145 (SD±0.73), p=0.042, and 61 children aged >60 months demonstrated a mean decrease in BMI Z-score of 0.152 (SD±0.39), p=0.004. Nine percent (10/109) of children ≤60 months and 3.0% (2/61) of children >60 months had worsening of nutritional status as determined by WHO classification. The greatest frequency of weight loss was observed among children diagnosed with gastroenteritis (81.2%), gastritis (64.3%) and pneumonia (55.6%). LOS was demonstrated to be the most significant risk factor for weight loss (OR 1.37, p=0.003, C.I 1.11-1.69). Children who had a hospital stay of 5-7 days had about 4.5 fold risk for weight loss (OR 4.67, 95% C.I. 1.34-16.24). A total of 97 (57.1%) patients had a drop in their serum pre-albumin levels within 48-96 hrs following admission. Of the 103 patients, who demonstrated a drop in weight from admission to discharge, 70 (67.8%) had a drop in serum pre-albumin (p-value<0.001). Using a serum pre-albumin cut off point of <0.15g/L, sensitivity and specificity of serum pre-albumin in predicting weight loss were 82.5% and 26.9% respectively. Negative and positive predictive values were 50.0% and 63.4% respectively. Positive and negative likelihood ratios were low at 1.12 and 0.65. The best cut off point for serum pre-albumin in predicting weight loss in the diagnosis of Hospital-acquired Malnutrition was determined to be 0.11g/L. Using this value, sensitivity dropped to 64.1% (95% C.I. 56.9-71.3) with an increase in specificity to 49.3% ( 95% C.I. 41.7-56.8). Majority of the patients (72.3%) at admission were already at risk of malnutrition as determined by the Pre-albumin Risk Stratification model by Bernstein et al. Only 1.8% of patients had an increase of ≥2 mg/dL (0.02g/L) per day, the cut-off point suggesting adequate nutritional support. Although serum pre-albumin was not a significant predictor of hospital stay, patients were four and half times more likely to have a longer hospital stay.

Conclusion: The incidence of Hospital-acquired Malnutrition among children at Aga Khan University Hospital is high irrespective of the nutritional assessment method used. Long duration of hospitalization significantly increases risk of acquiring malnutrition. Children with gastroenteritis, gastritis and pneumonia are especially prone to suffering deterioration of nutrition status during hospitalization. While serum pre-albumin level estimation is moderately sensitive in detection of acute weight loss during hospitalization, it would however need to be used alongside anthropometric measures in confirmation of Hospital-acquired Malnutrition. It is useful in identifying patients at risk of malnutrition at admission and during hospitalization.

This document is available in the relevant AKU library