Manifestations of VAD are most common in those that experience rapid growth, hence vitamin A deficiency, especially clinical manifestation, is most often seen in preschool aged children and pregnant women. Preschool children being among the most sensitive to vitamin A deficiency are often used as a surveillance group to indicate the presence of VAD in the population. Some of the clinical signs of deficiency (unlike those for anemia) are specific to VAD; therefore, they are used as diagnostic indicators. These indicators include the clinical signs of the eye such as nightblindness and physical changes to the eye, such as Bitot's spots and xeropthalmia. Although these signs are widely recognized and specific, they are relatively rare in the population, developing only in the late stages of VAD when body stores are largely depleted. Some other clinical signs resulting from an earlier stage of deficiency are not specific to only VAD, including increased illness and infection due to a compromised immune system. The less specific clinical signs will be considered in the data analysis, but not usually as the determining outcome variable. When trying to assess an earlier stage of VAD, usually sub-clinical indicators (e.g. serum retinol) are preferable, but often not collected due to increased financial and technical burdens.

CLINICAL INDICATORS:

Nightblindness (XN)
Bitot’s spots (X1B)
Conjunctival xerosis (X1B)
Active corneal lesions
Corneal xerosis (X2)
Corneal ulcers (X3A)
Keratomalacia (X3B)
Corneal scars (XS)

Because sub-clinical indicators are often not available, WHO recommends that a combination of at least two clinical indicators for determining VAD status of the population (e.g. XN and X1B are often reported together). If only one clinical indicator is available, then support this with four or more of the ecological and illness related indicators, where two are nutrition related (to review the indicators, return to VAD section 1: Causes, Manifestations, and Indicators). This allows for verification of findings through the use of several indicators, a routine part of any good analysis.

SUB-CLINICAL INDICATORS:

Serum retinol
Breast milk retinol
Relative dose response test (RDR)
Modified relative dose response test (MRDR)
Conjunctival impression cytology (CIC)

 

When sub-clinical data is available on VAD, that most commonly collected indicator is serum retinol taken from blood. Serum retinol is still not seen routinely in large or national level surveys because it is relatively difficult to collect, expensive, and analysis facilities are not often found in district or regional hospitals or facilities. Some other sub-clinical indicators are used, although much less frequently and in very specific and often smaller studies. RDR and MRDR are now becoming preferable to serum retinol because they give a better indication of liver stores of vitamin A, but the tests are even more complex and therefore rarely seen. In CIC, which might gain popularity in the near future but is still not often seen, conjunctival cells are examined under the microscope with a stain to identify level of VAD. When serum retinol is collected, it provides specific identifying information for the early stages of vitamin A deficiency. Serum retinol, like hemoglobin is a normally distributed continuous variable, which can be visualized using a histogram. Those out of range values can easily be seen using the graphing techniques (the N, mean, std. deviation can be listed at the side).

For tabulation purposes (comparing by strata), it can be useful to produce categorical variables using serum retinol cut-offs.

 Cut-off for categorizing Serum retinol (Children 6-71 months)

Indicator	Individual level cut-off
Serum Retinol- 6 mo- 71 mo- mild 6mo- 71mo-moderate 6mo- 71mo-severe	1.05 umol/l (30 ug/dl) 0.70 umol/l (20 ug/dl) 0.35 umol/l (10 ug/dl)

Process Indictors and Other Independent Variables Related to VAD

Whether or not the survey has included clinical or sub-clinical indicators, it is always important to look in the data for risk factors of VAD. Risk factors start to reveal causes of the deficiency and help in the formulation of intervention strategies. The assessment of risk is vital (especially in the case that sub-clinical measures are not available) because children with clinically visible VAD represent only a small portion of those at risk increased morbidity and mortality, making death the ultimate concern (beyond vision impairment or blindness). As mentioned above, when relying on one clinical indicator it is always important to support it with at least four supporting ecological indicators (e.g. prevalence of breastfeeding, general malnutrition, low birthweight, household food availability, etc.). It is important to note that this method only indicates the population level risk for VAD though, not the individual level risk.

One method of assessing risk is through the use of a food frequency questionnaire (e.g. the Helen Keller International Vitamin A Food Frequency Questionnaire), which allows for assessment of risk at the community and public health level. The technique measures consumption of Vitamin A foods including common global sources as well as regionally specific vitamin A sources. These types of questionnaires are semi-qualitative and semi-quantitative so that it gives a sense of meeting the needs in general at the population level, but it does not give a diagnosis for the individual. Link to HKI www.hki.com

Many variations of food frequency and dietary recall surveys are used around the world to assess consumption of vitamin A foods for children and households. Because of the variety in style as well as the biases involved in collecting recall information, analyzing this type of data takes special care and introduces many challenges. There are no clear recommendations for adequate quantity or frequencies of consumption because creating a standard is not possible with such a non-specific measure—depending on the person and the foods consumed. Some general guidelines have been made regarding analysis of food consumption information, albeit they are only loose suggestions. Despite the dubious introduction a lot can be learned from food consumption data, especially through looking at trends in market surveys to indicate food availability.

Many other variables will contribute to better understanding the causes and manifestations of VAD. To begin looking at some VAD data, start here:

First, using Cambd.sav make a list of the indicators of VAD in the data set (including all process and outcome variables), using the code book that was printed from the dataset. The indicators of most interest include those that involve supplementation and dietary intake. Also the variables collected on child illness should be combed for errors.

 Vitamin A Indicators