Prevalence of Malnutrition in Rural Karnataka, South India: A Comparison of Anthropometric Indicators

Article excerpt


Malnutrition is widely recognized as a major health problem in developing countries. Growing children in particular are most vulnerable to its consequences. The frequency of malnutrition cannot be easily estimated from the prevalence of commonly-recognized clinical syndromes, such as marasmus and kwashiorkor because these constitute syndromes only, the proverbial tip of the iceberg. Cases with mild-to-moderate malnutrition are likely to remain unrecognized because clinical criteria for their diagnosis are imprecise and are difficult to interpret accurately.

It is widely accepted that, for practical purposes, anthropometry is the most useful tool for assessing the nutritional status of children.

There are many anthropometric indicators in use, such as mid-upper arm circumference (MUAC), MUAC-for-height, weight-for-age, height-for-age, weight-for-height, and body mass index of Quetlet. Most of these indicators need to be used along with specific reference tables, e.g. National Center for Health Statistics (NCHS) tables, for interpreting data. This might not be possible in over-crowded outpatient departments of common tertiary care hospitals. Therefore, to estimate the expected weight or height of a child rapidly, especially in emergency situations, many field workers and clinicians use formulae first introduced by Weech, using age as a variable. Paediatricians now widely use these formulae in clinical practice (1).

Each of the above indicators has advantages and disadvantages. Some have a high sensitivity, while others have a high specificity. An ideal anthropometric indicator should have a high sensitivity to detect malnutrition accurately. At the same time, its specificity should be good so that the government resources and facilities meant for malnourished population may reach only those in need of them.

The study was carried out to estimate the prevalence of wasting and stunting among children aged 12-60 months and to compare the commonly-used anthropometric indicators in terms of their sensitivity and specificity.


This study was carried out in 11 villages in the Singhasandra Primary Health Centre area, at the outskirts of the city of Bangalore in rural Karnataka. A non-government organization (NGO) runs the health centre.

Cases were taken from among those children who visited the hospital along with their mothers during the antenatal clinic--a specialist facility provided for pregnant mothers at the health centre. Cases were also taken from various aanganwadis and creches in the area. The children essentially belonged to one of the 11 villages served by the NGO. Children aged 12-60 months were included in the study. Age of child was recorded using birth/delivery records or aanganwadi/ school/creche records. Children whose age could not be accurately known were excluded from the study. No other inclusion or exclusion criteria were applied. Age of child was estimated to the most recently-attained month (2). In total, 295 children were examined, but only 256 were included in the study since the age of 39 children could not be known accurately.

MUAC was measured to the nearest millimetre at the exact midpoint of the left arm using a narrow, flexible, and non-stretchable tape made of plastic (3,4). Weight of children was taken using a stand-on scale, the accuracy of which was established on a daily basis. Height of child was measured to the nearest millimetre using a right-angled head-plate non-stretchable tape fixed to the wall. Recumbent length was taken for children under 85 cm and standing height for children over 85 cm. All measurements were taken thrice and averaged for the final reading.

Each observer (one of the authors) was assigned to take one particular measurement for the whole study to reduce inter-observer bias. Being medical graduates, they did not undergo any special training. …