Evolution of Anthropometry in Malnutrition

Various methods are used to detect malnutrition viz. those based on clinical examination, those based on measurement of dietary intake (including breastfeeding, complementary feeding), biochemical methods and anthropometry.

Clinical assessment can easily detect severe forms of malnutrition with obvious signs like visible wasting, oedema, skin changes, hair changes, growth retardation, apathy, mental changes, hepatomegaly etc. Clinical examination involves a lot of subjectivity. Therefore, its specificity and sensitivity is high only in severe forms of malnutrition and diagnosis depends on the astuteness of the clinician.

Detection of malnutrition by biochemical tests involves simple tests like serum albumin or complex ones like serum & urinary creatinine, urea & their ratios. These investigations are time consuming, invasive & yield results with low sensitivity or specificity. They cannot be tried in field settings.

Dietary recall by the parent or caregiver along with weighing the child before and after feeds has also been used in addition to clinical examination to get an idea about malnutrition. Obviously, its success depends on educational status and intelligence of the mother and the time she devotes for the baby. Dietary component in the procedure is also to be combined with history of diarrhea, malabsorption in a given case to determine the degree of malnutrition. All these tests have their own merits & demerits.

Anthropometry has always been an important and the best tool in the diagnosis, management & prognosis of malnutrition in clinic and in community settings. The science of body measurements in children & adults has evolved over the last century. The present paper deals with the development of anthropometry over years and those anthropometric measurements which should form the basis of diagnosis, prognosis & recovery in malnutrition.

A) Age dependant anthropometric measurements -

1) Weight (Wt)

2) Height (Ht)

3) Occipitofrontal circumference (Head circumference)

4) Chest circumference

5) Wt for age, Ht for age

B) Age independent (or partially dependent) anthropometric measurements -

1) Mid upper arm circumference (MUAC)

2) Body mass index (BMI)

3) Skin fold thickness – triceps, subscapular,biceps,suprailiac etc.

4) Indices – Wt. for height, Wt for length

5) Various ratios

Taking weight and height of a child or weight and length of an infant is relatively easy and inexpensive in a clinic setting.It does not need complicated instruments or extensive training of personnel. When combined with age, it gives good indication of the nutritional status of the child. Nearly two decades ago, growth monitoring was done mainly on the basis of wt. for age of the child. Weight gain or weight faltering or even weight loss for age gave an idea about the growth of the baby for that age. This was measured in terms of the percentage of the median of Harvard growth standards. Thus the measure would be in terms of median wt for that age. So, grade-I malnutrition was weight of 71-80% of median, grade-II 61-70%, grade-III 51-60% & grade-IV was below 50% of the median. However, with the development of World Health Organization (WHO) growth charts in 2006 and expressing weight in terms of percentiles, it was seen as a better index to reflect nutritional status in terms of SD or Z-score. Thus, expression of weight for age gave an idea whether the child was normal, moderate underweight (MUW) as <-2SD or severe underweight (SUW) as <-3SD. This classification deals with the chronic nature of weight faltering, it is a composite index and is age dependent. It can broadly categorize children at risk.

However, a better index is a measure of wt. for height of the child, using percentile WHO growth charts¹.By using this method, children are classified as ‘ normal, or with moderate acute malnutrition (MAM) <-2SD or with severe acute malnutrition (SAM) or wasting <-3SD’. This is a measure of acute weight loss and identifies children at a high risk of dying (9 to 20 times increased risk as against wt. for age where in the risk is around 4 times more). Weight for ht. is an age independent index, is used to detect acuteness of the illness and often used now to categorise malnutrition (WHO).

A clinical analogy for difference between wt. for age & wt. for height measure would be the one of lipid profile and ECG with cardiac enzymes in an old person. Lipid profile may identify individuals at a chronic risk of coronary artery disease but ECG and cardiac enzymes are pointers towards acuteness of the coronary artery disease and myocardial infarction. Therefore, in clinical settings now, detection of undernutrition should be done by weight for height.

The third index used is height for age which detects if a child is stunted for age. This is again an age dependent index like the first index and gives an idea about chronicity of nutritional deprivation. All these indices (underweight, wasting & stunting) are used in clinical practice to get an idea about nutritional status of the child, but wasting is the best index to be used to identify children who are at high risk of death and who need to be treated immediately. (Table 1)

 MUAC-(Mid upper arm circumference).-It involves measuring the circumference of the mid upper arm region of the child. It is done by using a measuring tape encircling the mid upper arm region of the child at the level which is midpoint between acromion and olecranon.MUAC of less than 11.5 cm in children between 6 months to 5 years indicates severe malnutrition, between 11.5 to 12.5 cms indicates moderate malnutrition and that above 12.5 cm indicates a normal child. It is a rough measure , is easy to perform and is generally used in emergency settings in the field. It has limited sensitivity.

Various ratios related to height, weight and midarm circumference – Many scientists have devised ratios that may make the results more sensitive and specific.

These are enumerated below:

a) QUAC stick (Quaker Arnold MAC) (Quaker arm circumference measuring stick).

Mid arm circumference for height. A child taller for his arm circumference for height is malnourished (Normal – 75 to 85%).

b) Kanawati and McLaren’s index- MUAC in cm 0.32-0.33 is normal. <0.25 is SAM Head circumference in cm.

c) Kanawati Index – Ratio of MUAC and head circumference. It varies between 0.32 – 0.33. A ratio of less than 0.25 indicates SAM.

d) Rao’s ratio = Wt in gms / height in cm²X100

* Normal is 0.0015.

* Value between 0.0013-0.0015 is moderate malnutrition and below 0.0013 is SAM.

e) Rao and Singh’s index = Wt in kg/ height² in cmX 100

* 0.14 normal.

* 0.12 – 0.14 malnutrition.

f) Dugdale’s index = Wt in kg/ height^1.6 in cm

* 0.88 – 0.97 normal

* Below 0.79 is malnutrition.

g) Jelliffe’s ratio – Head circumference/ chest circumference

* Ratio above 1 in a child more than 1 year is malnutrition.

h) Quetelet index = Wt in kg / height in cm²

i) Ponderal index = Wt in gms/ Length in cm³⁼ X 1000

* Index less than 2 means asymmetric intrauterine growth retardation(IUGR)

* 2 – 2.5 borderline malnutrition & more than 2.5 is appropriate for

* for gestational age (symmetric IUGR)

* Ponderal index is also called as Rohrer index or corpulence index

j) BMI –This is now the most commonly used index..It is used to detect obesity. Obesity has been defined by using BMI (body mass index – wt in kg/height in meter²) in adults.. n recent years, the need to measure childhood obesity is increasing rapidly, considering the emerging epidemic of obesity & syndrome X (Hyperlipidemia, diabetes, hypertensioin, coronary artery disease), especially in low and middle income countries (LMICS). BMI between 18.5 – 24.9 is normal, 25-25.9 is overwt, above 30 is obesity, 15-18.5 is underwt, 13-14.9 is moderate underwt & below 13 is severe underwt.

Various growth charts using Z-scores are being used for children. Those with weight more than +2 Z-score or +2 SD on growth charts are considered to be obese. There has been varied opinion on the use of WHO growth charts for obesity. Use of these charts with cutoffs for under-5 years is likely to misclassify children as obese². Children above 5 years are classified as obese with >1 SD BMI by some scientists. BMI cutoffs are still not considered as ideal for detecting obesity³. So, one is in a dilemma as to which one measurement is to be used to measure wasting and obesity both i.e. the wt for ht or BMI charts.

Laillou Arnand⁵ analyzed data from over 11000 children from Cambodia and found that with current MUAC of 11.5 cms as screening for SAM, over 90% children with a wt for ht. Z-score <-3 would have been missed. Reversely WHZ <-3SD missed 80% children with MUAC of less than 11.5 cms. Both measures thus identify different sets of children.

Fiorentino M, Sophonneary P, Laillou A⁶ in a study on 14173 Cambodian children concluded that boys had higher MUAC cutoffs than girls except in 8-10.9 years’ range. In children below 2 years MUAC cutoff was lower for stunted children compared to non-stunted.

Many studies have corroborated this finding, it has been observed by many workers that specificity and sensitivity of MUAC & Wt for ht is variable, and correlation is not absolute.

Many scientists have tried to use cutoffs to define thinness, overwt and obesity. de Onis & Lobstein⁷ (WHO) in 2010 stated that for children below 5 yrs, thinness is defined as those with BMI below -2SD, overwt as those with BMI above 2SD & obesity as those with BMI above 3 SD. Thus, in one BMI chart measurement, a child could be categorized as thin, overwt or obese. The same workers found that for children above 5 years overwt is BMI >1SD & obesity is BMI >2SD. IO TF (Cole at 2000, 2007)⁸,⁹ defined thinness as percentile equivalent of BMI <18.5, overwt as percentile equivalent of BMI >25 & obesity as >30.

Relationship between weight and height is complex & enigmatic. Ponderal index of Rohrer wt/ht³ is appropriate for infants, but it is not clear as to what formula can one apply to toddlers, young children or adolescents. Many studies have concluded that after 5 years of age wt/height² can be used. It was compared with triceps skinfold thickness and was found useful. Some difference was reported by race, sex and age. It was also seen that height influences BMI in children, with taller children having larger BMI to some extent. In some series, stunting was seen to be associated with overwt/obesity as measured by BMI. The results from different studies, in different ethnic populations have given differing results.

The issue therefore of whether to use BMI only for detection of undernutrion, overwt and obesity is partially resolved when we use WHO charts with SD or Z-score.

The main advantage of MUAC compared to other anthropometric measurements is that it is easy to perform, can be used by Anganwadi worker at the community level and requires just one non-stretchable tape costing few Rs. MUAC of less than 11.5 cms identified children at high risk of death in a series of longitudinal population studies that were undertaken in the 80’s and early 90’s¹⁰. They found MUAC to be a better diagnostic tool as it had the highest ROC (receiver operating characteristic)¹⁰. Aguayo VM¹¹ found that in 3 States of India, MUAC <115 mm appeared to be appropriate criterion to identify children with SAM who are at a greater risk of complications and death particularly 6-23 months’ old children.

Unlike wt, MUAC is not affected by dehydration due to diarrhoea¹² MUAC specifically selects young children who have higher mortality. It also selects females more. MUAC has a relationship with fat mass and to some extent muscle mass. Survival is linked to fat stores during starvation and to muscle mass during infections¹²,¹³. Young children have comparatively low muscle mass that puts them at greater risk of death when they suffer from malnutrition. This may explain why wt for ht is not as effective as MUAC because it classifies in the same category children of different ages with same wt deficit not taking into account that young children with low muscle mass are at higher risk.

MUAC lacks sensitivity but is very specific. If more sensitivity is desired, higher MUAC cutoff may be used.

Wt. for ht. on the other hand detects a different set of people especially older children, boys and those with longer legs. Some workers feel that to increase sensitivity and specificity, one may combine MUAC and wt for htwhile some workers feel it more appropriate to increase MUAC cutoff rather than combination of MUAC and wt for ht. WHO recommends that either MUAC or wt for ht Z-score are to be used to assess prevalence of SAM.

A study by Talapalliwar & Garg showed the sensitivity and specificity of MUAC <11.5 cm was 13.6% & 99.3% respectively. The best cutoff for screening of SAM was obtained at MUAC <12.8 cm where sensitivity and specificity was 50% & 90.8% respectively¹⁴.

One study by Benitez Brito in adults showed that MUAC value ≤22.5 cms presented a sensitivity of 67.7% and specificity of 94.5% when compared to BMI <18.5¹⁵. Abdel Rahman¹⁶ SM used MUAC Z-score in U.S. children 2 months to 18 years to generate data at all ages.

There are a number of studies that have shown that MUAC below 11.5 cms & Wt for ht <-3SD or Z-score identify different sets of children at different risk of deaths. The specificity of MUAC is higher than wt for ht to predict subsequent death⁵. It has been shown that MUAC is strongly related to fat mass in children but is related poorly to fat free mass or overall wt. As against that, wt for ht cannot discriminate between fat & lean body mass and therefore, reflects fat mass & lean body mass. Fat mass is linked to immune function through leptin produced in adipocytes. Leptin favours Th1 immune response. So, low MUAC will predispose a child to be at high risk of infection that needs typical Th1 response eg. viral infection⁵. Lean body mass is linked to immune response through different roles of amino acids in immune system like acute phase response, also glutamate, sulphur containing amino acids which are antioxidants through glutathione, arginine and are important for NO production. So, low wt for ht predisposes a child to get infections requiring humoral immune response (eg malaria)⁵.

This also means that in some seasons when viral infections are common, MUAC measurement may play a role while in seasons when malaria is common wt for ht could be more relevant. Thus, different infection pressure can result in differential survival.

Skinfold thicknesses are measurements that provide an idea about subcutaneous fat deposits under the skin and total body fat. The procedure followed is that the skin is pinched at 8 skin fold sites and precision thickness calipers (Harpenden or Lange Holtaino) are applied. The areas are Triceps, Subscapular, Biceps, Suprailiac, Midaxillary,Quadriceps,Abdominal and Pectoral. The child or adult should be in standing position with shoulders relaxed while taking arm skinfold thicknesses. Right arm is preferred.

The formulae for fat calculation are separate for males and females and give body fat percentage. Body fat percentage above 25 % in males and above 30% in females is considered as obesity. Average is 14% in men and 26.9% in females. Other methods of calculating body fat are bioelectrical impedence analysis, X-ray analysis and the gold standard is water displacement method or hydrostatic weighing.

Four step formula

1. Step 1-Total body mass index(TBM) x % body fat = Fat mass wt (FW)

2. Step 2.TBM – FW=LBM lean body mass

3. Step 3 LBM / (1-GBF%)=TW

4. Step 4 TBM-TW=WL wt. loss TMI

For practical purposes, triceps skinfold thickness at the back of arm, biceps in front of arm,subscapular at 45⁰ just below scapula and suprailiac in front of iliac crest on ant. abdominal wall are the 4 important measurements taken by calipers. All readings in mm are added.

Now , an example is cited.

A sum of 20 mm skinfold thicknesses for a 30 to 49 years old indicates 18.4%fat in females and 12.1% in males.

Similarly,

· 30 mm indicates 23.3% fat in females and 16.9% in males.

· 40mm indicates 26.8% fat in females and 20.3% in males

· 80mm indicates 35.6% fat in females and 28.8% in males.

For 30 to 49 years of age, 11 -17% is a good range for men and 15 - 23 % is a good range for females.

These can be derived from MUAC and skinfold thicknesses

This is an age independent anthropometric nutritional assessment devised by Manuel Cidras¹⁷. It is thought to be better than BMI as exponential index is better than power i.e. squared.

Use of BMI as an anthropometric index of nutritional assessment from 2 to 20 years has always been difficult. Though BMI is considered by and large as the standard screening for nutritional status specially obesity, it is associated with several problems.

Taller children tend to have higher BMI and therefore higher probability of being diagnosed as obese. It may be a mathematical artifact than a biological one.

BMI varies in children on a U shaped line with nadir at 6 years of age.

BMI depends on leg length and in certain races it may result in erroneous values.

It depends on fat & non-fat wt. So athletes with increased muscle mass will have more wt. and will be wrongly labelled as obese.

Allometric growth of body i.e. growth of a part, at a different rate from that of the body as a whole cannot be best expressed as power function i.e. height². If one organ grows exponentially and the other in a linear fashion, the allometric relation should be exponential and not power. (Therefore, BMEI is thought to be better by some workers).

There is lack of consensus on four normal limits of BMI for age 5^th to 85^th centile as per CDC, 18.5 – 25 BMI percentile lines as per IO TF & ±2SD as per WHO

Manuel Cidras undertook statistical analysis and proved that BMEI wt/exp (2* height) is a better index and the curves drawn with this index coincide better with 5^th & 85^th centile. A BMEI of 2 with limits of 1.5 and 2.5 is useful for screening nutritional status during growth and the wt for ht chart is an ideal substitute for BMI for age chart.

Cidras has demonstrated that the wt, ht relationship from 2 to 20 years is better expressed by an exponential function than by a power function on which BMI is based. Wt/exp (2H) is more accurate than wt/ht², wt/ht³ or wt/htρ, in obtaining nutritional index¹⁷.

Quetelet had considered growth coefficient to be ideally wt²/ht⁵ or wt/ht^2.5. The variation limit of 1.5 to 2.5 could be for females and males with 2 as the standard BMEI.

A BMEI of 2 with limits of 1.5 & 2.5 can then be used as nutritional index without requiring age chart.

Stunted children show inconsistent results. Some studies showed that stunting in early childhood was not related to BMI or adiposity. Other studies showed that stunted children were more likely to develop central adiposity³.

Population differences related to ht & maturation can affect BMI.

There is a relationship between BMI & ht. Therefore, short children of China & tall Dutch population will need a scaling approach for adjustment of height & risk of adiposity. Here, wt for ht may be a better approach.

Stunted population may be a predictor of greater BMI. Asians may have a greater risk of chronic disease outcome at a lower BMI. Indian infants have a greater fat mass at any given wt compared to British infants¹⁸. Thus, though BMI continues to be a valuable measure for adiposity & though WHO defined thinness using BMI cutoffs in 2010, all these measurements are not yet ideal. The limitations of MUAC, wt for ht, BMI are discussed in the above paragraphs. In stunted children one may use waist circumference & skinfold thickness.

Percentage of wt. for height : This approach is partially age independent.

Rao & Singh¹⁹,²⁰ & Dugale²¹ have studied it in details.

Rao & Singh showed that ( wt/height^{2 ) x100} was remarkably constant between 1-5 yrs and was same for males & females. The mean value was 0.15 (equaling to a BMI of 15?) and those with signs of malnutrition showed it to be 0.12 to 0.14. Dugdale described that wt/ht^1.6 provided the best index of anthropometric normality and was age independent. It remains to be studied if wt/ht^1.6 could replace BMI calculation of wt/ht². It may hold promise.

Answer to this question needs a lot more evidence to be generated. A low BMI with less body fat, thin lean, asthenic built may predispose to development of infections (Th2 cell dependent, rightly mycobacterium tuberculosis). Should we therefore change the lower limit of normal value of 18.5 BMI to around 20? Similarly, because Asian bodies have more fat than muscle mass¹⁵,¹⁸ should we change the value of upper limit of normality from 24.9 to around 23? Further evidence will answer these questions. Should we use wt/ht^1.6 to replace the conventional BMI and generate one chart that will detect SAM, MAM, normal, overwt, obese children & adults of all ages? The answer will be obtained through large international studies. Indian Academy of Pediatrics²² has already revised growth charts for Ht, Wt & BMI for 5 to 18 year old children. One may think of applying Dugdale Index²¹ afresh and workout its specificity & sensitivity with MUAC in detection of SAM and its relation to BMI cutoffs in detecting obesity so as to arrive at a common measure for all. TMI- Triponderal mass index could be tried out in the same way as Dugdale. All these evidences so generated will probably arrive at one standard measure that could detect under wt and obesity in all age groups.