|
|
Indicators
for Assessing Iodine Deficiency
Disorders and their control through Salt Iodisation: WHO/UNCF/ICCIDD
1 Introduction
1.1 Background
1.
This document is based on a technical consultation 1
that WHO, in collaboration with ICCIDD and UNICEF, convened in November 1992
(see Annex 1 for list of participants). The three organizations have worked
closely for a decade to combat IDD, which is one of the oldest and most
insidious of human health scourges.
2.
Three factors combine to make reassessment of IDD
indicators a timely exercise. First, increasing
scientific knowledge about IDD and accumulating experience of related
control programmes have made it
necessary to review and, in some cases, revise earlier judgements.
Secondly, experience from other areas, e.g. quality control procedures used in
manufacturing and implementation of immunization programmes, provides new tools
that can be applied with advantage t o programmes for combating micronutrient malnutrition.
Thirdly, the adoption by all governments of the ambitious goal of eliminating
IDD as a significant public health problem by t he end of t he century signals
the urgency, even as it provides the impetus, for concerted action.2
3.
Although indicators for assessing IDD might initially
appear to be an uninteresting, even tedious, topic, their review has met with
considerable enthusiasm. Building on existing momentum for eliminating IDD as a
significant public health problem,
participants in the 1992 consultation were unanimous in stressing the
importance of having the necessary tools for measuring whether, and when, the
decade goal is likely to be reached, They concluded that this can be achieved
only by clearly defining appropriate indicators and criteria for assessing both
IDD and related control programmes.
4.
The consultation was marked by willingness to take a
fresh look at the whole subject of IDD indicators, and even WHO's long-standing
goitre classification came under critical review! Participants favoured
involvement by all who are I a position
to identify IDD and to assess, simply but effectively, both its severity
and the adequacy of measures to combat it through the recommended intervention,
salt iodization.3 Thus, while
participants agreed that the classic five-grading of goitres remained valid ,
they concluded that a simpler three-grade classification system would be a more
practical field
----------------------------------------------------------------------------------------------------------
1 Joint WHO/UNICEF/ICCIDD Consultation on
the Indicators for Assessing Iodine Deficiency Disorders and their Control
through Salt Iodization, 3-5 November 1992, Geneva. The report of the
consultation, document WHO/NUT/93.1, was circulated as a review version of the
present text.
2This goal was formally endorsed by the
World Summit for Children (New York, 1990), the World Health Assembly ( Geneva,
1991), the Policy Conference on Ending Hidden Hunger ( Montreal, 1991), and the
International Conference on Nutrition (Rome,1992).
3Considerable progress is now being made
towards meeting the goal of universal salt iodization in the majority of
countries where IDD is a significant public health problem.
----------------------------------------------------------------------------------------------------------
Screening
device to be used, for example, by schoolteachers who have been trained for
this task.
5.
The document provides the most up-to -date internationally
agreed outline of practical procedures for verifying the adequacy of salt
iodization on a national and local scale. The vexing issue of numbers of
subject and biological specimens required for IDD surveys in creatively
handled, in keeping with the various purposes for which surveys are conducted,
Thus, sampling methods and sizes are expected to vary according to survey aims.
For the first time, clear and epidemiological adequate guidance is given on
this subject, while additional information may be obtained in the works cited.
6.
Because of
constant evolution in both IDD assessment methodology and the integration of
indicators, the document will certainly not be the last word on the subject.
Experience in applying the methods and procedures outlined will no doubt permit
their further elaboration and refinement. Nevertheless, the document is an
important milestone because of the practical yardstick it provides for
measuring progress and evaluation the effectiveness of related prevention and
control programmes, While the focus for scientific purposes is on clinical and
biochemical parameters, the ultimate goal is elimination of the developmental
effects of IDD, which are considerably less easily measured.
7.
The 1992 consultation was one of a series held for
each of the micronutrients for which WHO and UNICEF have adopted specific
goals, namely iodine, vitamin A, and iron. While much of the discussion in the
first four section of the present document is also valid for iron and vitamin
A, the last two sections are generally specific to iodine only. In view of the
increasing interest in micronutrient malnutrition, and the potential benefits
from an integrated approach to its monitoring and assessment, it may be
feasible eventually to develop guidelines for simultaneous action in respect of
all three micronutrient deficiencies. The resulting degree of imprecision would
be acceptable if a common sampling frame were to reduce costs significantly and
improve logistical capacity to measure progress towards achieving micronutrient
goals. There is no point in continuing to amass data when a carefully planned
survey covering a minimum number of people produces information that is
sufficiently reliable for this purpose. As participants in the 1992
consultation were reminded, it is more important to be roughly right, than
precisely wrong.
1.2 Magnitude
of the problem
8.
Because they are only evidence of a condition,
indicators cannot be meaningfully discussed until the condition itself has been
clearly identified. Iodine deficiency not only causes goitre; it may also
result in irreversible brain damage in the fetus and infant, and retarded
psychomotor development in the child. Iodine deficiency is the most common
cause of preventable mental retardation; it also affects reproductive functions
and impedes children's learning ability. The cumulative consequences in
iodine-deficient populations spell diminished performance for the entire
economy of affected nations. The impact of iodine deficiency on intellectual
development and the resulting brake on socioeconomic development has played a
significant role in mobilizing scientists, public health administrators and
political leaders the world over to deal effectively with IDD.
9.
Knowledge of
the global magnitude of IDD, and thus its real significance for health and
socioeconomic development, has improved conssiderably during the last decade.
IDD is known to be a significant public health problem in 118 conutries. 1
At least 1572 million people worldwide are estimated to be at risk of IDD, i.e.
who live in areas where iodine deficiency is prevalent (total goitre rates
above 5%), and at least 655 million of these are considered to be affected by
goitre. The regional distribution of goitre prevalence , and estimated rates of
frank. 2
Table 1. Total number of
people and percent of regional population living in areas at risk of IDD, or
affected by IDD and cretinism (1990)
|
WHO
region |
Population |
At
risk of IDD |
Affected
by goitre |
Affected
by cretinism |
||||||
|
Millions |
%
of region |
Millions |
%of
region |
Millons |
%
of region |
|||||
|
Eastern Med Southe-East Western Pacific |
550 727 406 847 1355 1553 |
181 168 173 141 486 423 |
32.8 23.8 42.6 16.7 35.9 27.2 |
86 63 93 97 176 141 |
15.6 8.7 22.9 11.4 13.0 9.0 |
1.1 0.6 0.9 0.9 3.2 4.5 |
0.2 0.9 2.3 1.1 1.3 2.9 |
|||
|
TOTAL |
5438 |
1572 |
28.9 |
655 |
12.0 |
11.2 |
2.0 |
|||
1.3 Purposes
of surveillance
10. IDD
surveillance can be used for a number of distinct purposes, including for
assessing the magnitude and distribution the magnitude and of IDD prevalence,
identifying high-risk populations, evaluating control programmes, and monitoring
progress towards achieving long-range goals. The surveillance design employed,
indicators used, and the approach to data interpretation will vary according to
the specific purpose intended.
11. Assessing IDD prevalence. One of the
fundamental purposes of IDD surveillance is to determine the magnitude and
distribution of IDD within a population. This assessment can provide a baseline
for long- term monitoring, serve as an advocacy tool to
1
ICIDD/UNICEF/WHO. Global prevalence of iodine deficiency disorders. MDIS
Working Paper#1. Micronutrient Deficiency Information System, Geneva, World
Health Organization, 1993.
2Bailey KV,
Clugston GA. Iodine deficiency disorders. In: Murray CJL, Lopez AD,eds. The
global burden of disease and risk factors in 1990. WHO/World bank. Geneva,
World Health Organization (in preparation).
Highlight the extent of IDD problems,
and stimulate action including the appropriate allocation of resources for
eliminating IDD.
12. Identifying
severely affected areas for intervention. Identifying severely affected
communities is also crucial to IDD programme development. This type of
surveillance activity is primarily concerned with identifying priority areas
for intervention, thereby ensuring a more efficient use of resources.
13. Monitoring
and evaluating IDD control programmes. Another fundamental purpose of
surveillance is to evaluated implementation and impact of control programmes.
Indicators can be measured that assess the extent of a programme activity as
well as its impact on specific outcomes.
14. Measuring progress towards achieving long-term
micronutrient goals. Many countries are working towards the goal of eliminating
IDD as a significant public health problem as part of their effort to achieve
child health and development goals for the year 2000. Surveillance activities
can provide a quantitative basis for assessing progress towards meeting those
goals.
2. Selecting target groups
15. A variety of
target groups, including neonates, infants, preschool-age and school -age
children, and certain groups or groups depends on a number of considerations,
including their vulnerability, representativeness, accessibility, and potential
usefulness for surveillance of multiple health problems.
1.4 Selection
criteria
16. Vulnerability.
In order to serve as a sensitive indicator, a target population must be
vulnerable to the deficiency. Three aspects of vulnerability are:
·
extent of exposure to the deficiency;
·
severity of health consequences due to the deficiency;
·
degree of clinical or biochemical responsiveness to
the deficiency and related interventions.
17. Representativeness.
The issue of representativeness, or " generalizability" to the wider
context of a study, is sometimes referred to as" external validity".
·
Is the target group used for surveillance
representative of all persons in the same age/sex group in the community? For
example, if children are examined in school, are they representative of all school-age
children in the community? It may be that children who are in school come from
more advantaged or better-educated families, and consequently their risk of IDD
may be lower
·
Is the status of the target group representative of
the status of the community as a whole? It may be that the apparent prevalence
of IDD in the target group leads to an overestimation or underestimation of
community prevalence.
18. Accessibility.
Another criterion for selecting a target population is its accessibility.
Easily accessible populations, for example children in school, women in MCH
clinics, and neonates in a hospital, may be useful for surveillance purposes. Using these
relatively accessible groups will facilitate surveillance and reduce logistic
costs, even if the most accessible groups may not always prove to be either
fully representative or the most vulnerable.
19. Usefulness
for surveillance of multiple micronutrient malnutrition and other health
problems. It may prove advantageous if the target group selected for IDD
surveillance can also be used to assess other nutritional and health problems.
For example, school-based IDD surveillance might also serve for surveillance of
iron status or helminth infections. On the other hand, school-age children may
be less useful for assessing vitamin A deficiency and child growth using
anthropometry.
1.5 Target
groups
20. Applying the
above criteria to potential target groups makes clear some of their advantages
and disadvantages for IDD surveillance. A framework for considering these characteristics
in relation to various potential target groups is presented in Table2.
21. Neonates.
The use of neonate screening to identify congenital defects is well established
in many developed countries and it is being introduced in some less developed
countries as well. Regular collection of blood-spot specimens, where this is
done, is an important source of information for IDD surveillance given their
use in assessing TSH status. Elevated TSH levels, especially during infancy,
suggest a deficiency of iodine.
22. Surveillance
may also be done practically in iodine-deficient areas by collecting blood-spot
specimens from cord blood in a sample of neonates. Blood spots can be collected
by trained birth attendants in homes, or
at health posts or hospitals. Sample representativeness in this type of
surveillance depends on the health system's degree of coverage of neonates.
This is a potentially highly effective surveillance design that makes efficient
use of existing primary health care
infrastructure.
23. Infants and
preschool -age children.
Infants and
preschool-age children are also highly vulnerable to IDD and may be useful for
surveillance of other health problems as well, for example vitamin A deficiency
and anemia. However , this age group may not be readily accessible except in
MCH clinics, where the question of representativeness may arise. Some IDD
indicators, e.g. goitre by palpation, may be relatively difficult to assess in
this age group. There may be opportunities to coordinate IDD surveillance and
iodine supplementation with the immunization programme.
24. School -age
children. School -age children are a
useful target group for IDD surveillance because of their combined high
vulnerability, easy access, and usefulness for a variety of surveillance
activities. Affected children develop an enlarge thyroid in response to iodine
deficiency and can be readily examined in large numbers in school settings. At
the same time, other health concerns in this age group, including helminth
infections, anemia, and behavioral factors affecting health, can be assessed
and educational interventions
implemented . A major concern arising in school-based surveys is
that children no attending school are not
represented, which possibly leads to biased prevalence estimates. However, on
special occasions it may be possible to use school premises for assembling this
group together with preschool -age children.
25. Pregnant
women in MCH clinics. The iodine status of pregnant women is particularly
crucial because of the susceptibility of the developing fetus to iodine
deficiency. Pregnant women are accessible in primary health care settings, and
a wide variety of other health conditions can be assessed concurrently.
Representativeness may be a problem depending upon the level of access to and
utilization of health care services by women who are at highest risk.
26. Adults in households. Screening adult women
and men through household surveys provides an opportunity to establish a sample
of a population vulnerable to IDD However, after age 30, goitre rates are no
longer reliable indicators of current iodine intake. Accessibility may be
limited because of the expense and logistical
constraints associate with
performing household surveys. Representativeness depends on the extent
to which men and women work outside the home; in particular, it may be
difficult to find men at home during the day.
Table 2. Framework for
considering target groups for IDD surveillance
|
|
Vulnerability |
Representativeness* |
Accessibility |
Usefulness for other surveillance** |
|
Neonates |
High |
Intermediate |
Intermediate |
Intermediate |
|
Preschool-age
children in MCH clinics |
High |
Intermediate/low |
Intermediate |
High |
|
Preschool-age
children in households |
High |
High |
Intermediate |
High |
|
Children
in schools |
High |
Intermediate |
High |
High |
|
Pregnant
women in MCH clinics |
High |
Intermediate |
High |
High |
|
Adult
women in households |
Intermediate |
Intermediate |
Intermediate |
Intermediate |
|
Adult
men in households |
Intermediate |
Low |
Low |
Low |
*Level of representative
ness depends on access or coverage (see
text)
** Usefulness of group for
surveillance of other nutrition and
health problems
3.Surveillance methods
27. As noted in section
1, IDD surveillance serves several purposes, and data collection methods will
vary accordingly. Suggested data-collection schemes are described below. When
designing a surveillance system, it is important to decide whether the primary
purpose of the survey is to derive a prevalence estimate or to identify
high-prevalence areas. Each purpose usually requires a different survey method,
and studies that attempt to fulfil both tend to be inefficient and fail to
answer adequately either question. In general, the first step is to undertake a
prevalence survey. If results show an IDD problem and it is decided that
identifying high-prevalence areas is important, a lot quality assurance
sampling (LQAS) survey would then be useful.
3.1 Assessing IDD prevalence
28. There are
two main requirements in assessing IDD prevalence: collection of the minimal
amount of data that are representative of the target population, and provision
of a stable prevalence estimate within a desired level of precision. The survey
method used depends upon many factors, including the target group (e.g.
neonates or school-age children), the survey site(e.g. household or schools),
and the size of the geographic area of interest. In large countries, it may be
desirable to undertake prevalence surveys by state of province, which can be
further, refined in selected districts. The number of geographic units to be
studies should be kept to the minimum considered necessary to guide
interventions.
29. A common method used for immunization and
anthropometric household -based surveys is the
" probability proportionate to size" (PPS) cluster method.1
While this method requires up -to date population census data, where IDD
surveys concerning mainly stable populations in rural areas are concerned,
outdated census data could be used without undue violation of PPS sampling
requirements. IN general, all villages and cities are listed and a systematic
sampling scheme is used based on the cumulative population. This sampling
scheme assures that larger villages and cities are more likely to be selected
than smaller ones. The number of individuals to be sampled within each cluster
depends upon the prevalence of the condition, the level of precision desired
(based on the type and width of the confidence interval), the design effect ( a
measure of the variability of the prevalence between clusters), whether
inference at the cluster level is desirable, and the biological precision of the surveillance method.
----------------------------------------
1
Cluster sampling is a sampling process in which sampling units are made up of
clusters or groups of study units, e.g. districts, villages and schools.
30. Sample size calculations for
prevalence may be based on relative or absolute precision; relative precision
is recommended here.1 The recommended sample size for initial
school-based goitre surveys is 1200 children ( 30culsters x 40 children per
cluster). This sample size is based on
an assumed goitre prevalence of 50%, 95% confidence intervals, a design effect
of 3, and a relative precision of 10%; or a prevalence of 20% with a relative
precision of 20%. As IDD programme interventions proceed and goitre prevalences
decline, the critical level of goitre prevalence becomes 5%; larger sample size
would then be needed to estimate the prevalence rate with the same relative
precision, e.g.5% prevalence with relative precision of 30% would require
3x811= 2400 subjects. As prevalence becomes smaller, estimating prevalence with the same relative precision requires examination of more children. If
this is difficult for logistical reasons, one could reduce the precision even
further , say to 40%, giving a confidence interval of 3 to 7% in the case of an
estimated prevalence of 5%. However, the design effect may also change ad
interventions proceed. Thirty clusters ( schools, households etc) are intended
to ensure a valid prevalence estimate; examining fewer clusters can lead to estimates that differ
substantially from the true prevalence.
31.The recommended sample size for the
collection of biologic specimens like urine or blood spots is 300 (30 cluster x
10 children per cluster). The sample size for biologic specimens is based on
the assumption of a prevalence of abnormal result of 50%, 95% confidence intervals,
a design effect of 2,2 and a relative precision of 16%. This is the
appropriate sample size when the results are used for monitoring an
intervention's effectiveness, or as the basis for diagnosing endemic IDD.
Smaller sample sizes could be adequate when they are used simply to confirm
iodine deficiency in a clearly goitrous population.
32. The rationale behind these sample
sizes is that palpation is relatively easy and inexpensive to perform , and
therefore palpating 40 children per school should pose little problem. Because
of the costs involved in collecting, transporting, and analyzing biological
specimens, sample sizes may need to be larger or smaller than those recommended
above.
33. School surveys could use either PPS
or a two-stage cluster approach. The main difference between a PPS survey and a
two-stage cluster survey is how schools are selected. With PPS, schools are
systematically selected from a cumulative population list; in two-stage cluster
sampling they are selected randomly. The PPS survey is self-weighting, whereas
in two-stage cluster surveys the school enrolment is required to weight the results. Once schools are chosen, usually
a fixed number of pupils are selected to be in the survey, e.g. 40 pupils are
palpated while 10 provide urine specimens.
--------------------------------------
1
For more information and sample size tables, see Annex2.
2Note
that, generally speaking, the fewer children sampled per cluster, the smaller
the design effect.
34. Sampling of neonates may be
performed randomly or non-randomly , depending on the purpose of the survey. If
neonatal TSH levels are used to track changes in iodine status over time, a
random selection procedure should be devised. Particular care is needed in
populations with low coverage of assisted deliveries, e.g. rural areas in
developing countries.
3.2 Identifying
high-prevalence areas
35. In some situations the surveillance
goal may be to identify areas of high prevalence in order to focus intervention
activities. A difficulty associated with identifying high-prevalence areas is
that IDD tends to occur in geographic foci, and a large number of sites may
have to be surveyed in order to find the high-prevalence areas. As described in
Annex 3 and eslsewhere,1 LQAS is an efficient survey method for screening a
large number of sites, for example when school children are chosen as the
surveillance target group for identifying areas with high goitre prevalence. In
order to find the high-prevalence areas , every school within a geographic area
would be surveyed , and a sample of children in each would have their thyroid
palpated. If a large number of children had goitre, the area would be
identified as" high prevalence".
36. To continue the example, suppose
that the estimated goitre prevalence among school children in a region is 10%,
and the ministry of health (or equivalent) is interested in identifying schools
with a severe problem, i.e. a prevalence greater than 30%, in order to focus
intervention efforts. How many children would have to be palpated in each
school, and at which point would a school be classified as having a severe IDD
problem?
37. Using the tables in Annex 3 (and
assuming a significance level of 5%and power 2 of 90% ), 33 children
would need to be randomly selected and palpated in each school. If five or more
children had goitre, the school would be classified as having a severe IDD
problem (i.e. it would be considered a "rejected" lot). If fewer than
five had goitre, the school would be classified as a not having a severe
problem (i.e. it would be considered an "accepted" lot). There are
two possible errors in this connection. A school could be diagnosed as not
having a severe IDD problem although it truly has one, or it could be diagnosed
as having one although in fact it does not. Using LQAS in this situation is
intended to minimize the first error since in may mean not intervening in areas
that require attention. The second error is of lesser conssequence. In the
above example, if fewer than five student s out of 33 had goitre , there is
only a 5% chance that the true prevalence in the school would be greater than
30%.
38. In order to minimize the number of
schools to be screened, the study area could be reduced to include only those
at high risk of having IDD, such as rural mountainous regions. Since the LQAS-
based prevalence estimates for each school would be the result of randomly
selected children, a weighted average of the school prevalence estimates
---------------------------------------
1
World Health Statistics Quarterly, 1991:44(3),115-132 and 133-139.
2"Power
in this context means the probability of correctly rejecting the null
hypothesis when it is false.
Would be a valid estimate of the overall
area estimate of the prevalence among all school children. The weighting
process is accomplished by multiplying the prevalence found in each school (or
cluster) in the sample examined by the number of children in the school or
cluster.
3.3 Monitoring and evaluating IDD control
programmes
39. The main interventions for
controlling IDD are fortification of salt with potassium iodate (or iodide) and
supplementation with iodized oil. The procedures for monitoring salt iodization
programmes are outlined in paragraphs 95-105 while those concerning iodization
of other foods and iodized oil are found in paragraphs 106-107.
3.4 Measuring progress towards achieving
long-term micronutrient goals
40. Periodic prevalence surveys, as
described earlier in this section, are necessary to measure change in
prevalence over time. Measuring progress towards achieving long-term
micronutrient foals requires that surveys be representative of the population
concerned. Because surveys need to be repeated, they should be simple to
perform and analyze and based on the minimum number of individuals required to
provide stable prevalence estimates
within the desired level of precision.
4. Interpreting and
presenting results
41. Many IDD parameters are measured on
a continuous scale, e.g. urinary iodine levels and TSH, but the results do not
have the normal Gaussian distribution. Thus , the use of means and standard deviations is likely to be
inappropriate. It may be possible to transform some non-normally distributed
data,e.g. by logarithmic means, into a normal distribution and then to
calculate means and standard deviation. Transformations may
not work for other parameters, and presentation of a
median or other centiles, and use of
non-parametric statistics, would then be appropriate.
42. It is recommended that the full
distribution of results be presented, in addition to a measure of the central
tendency ( mean of median), and that cut-off point be used to delineate or
identify the upper or lower tail of the
distribution . The distribution of individuals
at the extremes of a distribution can be
characterized by using standard cut-off points and tabulating the
prevalence of values below or above cut-off values. Several cut-off points may
be used to provide an impression of the magnitude of the problem occurring at
different levels of the distribution.
For example, lower cut-off points may be selected to highlight the most extreme
cases of deficiency, while higher cut-off points may be useful in capturing the
proportion of the population that may be moderately affected, or at risk of
inadequate iodine status.
5. Selecting appropriate indicators
43. There are basically two types of
indicators: outcome indicators that provide a measure of IDD status, and process
indicators that measure the conditions or progress in implementing an IDD
control programme. Out come indicators can be categorized according to whether
the assessments are clinical ((thyroid size, cretinism) or biochemical (urinary iodine and thyroid –related
hormones). Once the target population (s) for assessment are defined, the
selection of particular indicators should be based both on the following
criteria and the specific surveillance objectives.
5.1 Criteria for indicator
selection
44. Acceptability. An indicator’s
acceptability to a given target population is a crucial factor. Some
procedures, e.g. assessment of thyroid size by palpation, may be widely
acceptable. Others, such as drawing venous blood for biochemical
determinations, may be quite unacceptable, especially in certain target groups
such as infants and children. Acceptability is also an issue for field staff
who perform the tests. Thus, for example, drawing blood specimens in
populations having a high prevalence of HIV infection involves some level of
risk, or perceived risk, that should be taken into account during indicator
selection.
45. Technical feasibility. Technical
feasibility involves a number of factors including:
·
ease of data or specimen collection;
·
specimen storage and transport requirements;
·
transportability and ruggedness of field
equipment;
·
availability of personnel to obtain specimens.
46. Cost. Costs associated with the use
of certain indicators include:
·
Capital costs for facilities and equipment;
·
Recurring costs for supplies and reagents;
·
Maintenance costs;
·
Training costs';
·
Personnel, administrative and related costs.
It is useful to estimate the cost per
test, but this should still take into account all the above- listed components.
47. Performance. Another criterion for
indicator selection is performance in identifying IDD status. Useful measures
of indicator performance include sensitivity, specificity, and reliability.
48. Interpretation and availability of
reference data.
Interpretation of IDD status depends on
availability of reference data assist in establishing cut-off values and
prevalence levels for use in identifying public health problems. Reference data
are useful in selecting indicators and target groups for surveillance, ass
these will enhance interpretation across different studies.
49. Using a combination of parameters.
It is generally recommended that at least
two indicators be used. No single parameter reflects the entire iodine- deficiency
picture and resulting change in the thyroid. In cases of severs deficiency ,
initial emphasis may be on lowering rates of
goitre and cretinism. As a programme evolves, however, emphasis is
likely to shift to ensuring adequate iodine intakes, e.g. by monitoring salt
iodization levels and population salt intakes and urinary iodine levels and
normal thyroid function, e.g. by measuring TSH levels, especially in neonates,
To the extent that resources permit , it is thus desirable for a biochemical baseline
to be established at the outset, at least in a sub-sample of a the larger
population
5.2 IDD outcome indicators
Clinical indicators
Thyroid Size
50. The size of the thyroid gland
changes inversely in response to alterations in iodine intake, with lag of 6 to
12 months in children and young adults (i.e.< 30 years of age). The
traditional method for determining thyroid size is inspection and palpation.
Ultrasonography provides a still more
precise and objective method. Both are described below . Issues common to palpation and
ultrasound are not repeated in the section on ultrasound
Palpation
Feasibility
51 Palpation of the thyroid is important
in assessing goitre prevalence. Costs are associated with mounting a survey,
which is relatively easy to conduct, and training of personnel, accessibility
of the population, and the sample size. Feasibility and performance vary
according to target group.
52. Neonates. It is neither feasible nor
practical to assess goitre among neonates, whether by palpation or ultrasound.
Performance is poor.
53.
Preschool- age and school - age children. Preferably children 6-12 years of age
should be studied. There is a practical reason for not measuring very young age
groups: the smaller
the
child,
the smaller the thyroid and the more difficult it is to
perform palpation. It is recommended
that if the proportion of children attending school is less than 50%, spot
surveys should be done on two groups of children of the same age, i.e. those
who attend school and those who do not, to ascertain if there is any
significant difference between the two. If so, both groups should be studied
separately, in all clusters, or an
appropriate adjustment
should be made
in the rate
found among school children.
54. Adults. Pregnant and lactating women are of
particular concern. Pregnant women are a prime target group for IDD control
activities because they are especially sensitive to marginal
iodine deficiency.
Frequently they are
relatively accessible given their
participation in antenatal clinics.
55.
A modification of the previous goitre classification system, which defined five grades, is recommended. The
previously used grades 1 A and 1B are thus combined into one, and grades 2 and
3are combined into another (Table 3)
. Table 4
gives the epidemiologic criteria
for establishing IDD severity
based on goitre prevalence in
school-age children. It should be understood that "mild" is a
relative term; it does not imply that
this category of IDD is of little consequence.
56. It
is recommended that a total
goitre rate (TGR,
goitre grades 1 and 2) of 5% or more in primary school children
(age range approximately 6 to 12 years) be used to signal the presence
of a public health problem. This
recommendation is based on the observation that in a normal,
iodine-replete population the prevalence of
goitre should be quite low. The
cut-off of 5%
allows some
margin of inaccuracy of goitre assessment and
for goitre that may occur in iodine-replete populations due to other causes
such as goitrogens and autoimmune thyroid diseases.
The
previously recommended
10% cut-off level
has been revised downwards since it has been shown that goitre
prevalence rates between
5% and 10%
may be associated
with a range
of abnormalities, including inadequate urinary iodine excretion and /or sub-normal levels of TSH among
adults, children and neonates.
57. For example, in
iodine
in
adults and
adolescents, alterations of
thyroid function in pregnant
and lactating women and
neonates required, and
were corrected by, iodine
----------------------------
1 Beckers
C et al Status of
iodine nutrition and
thyroid function in Belgium. In: Delange F, Dunn JT ,Glinoer
D, eds. Iodine deficiency in
Europe: a continuing concern. New York,
Plenum Press, 1993:359-362.
supplementation.1,2,3 Moreover, several recent studies
in Asian cities have shown abnormally
raised TSH levels (e.g. 32%>5mU/1 in Manila, where goitre rates in children
aged 7-9 years were1% and 4% [personal
communication from G. Maberly, who is preparing a detailed report on these studies]. In
short, even goitre rates5-10% or lower are no guarantee of normal thyroid
function.
58. Finally , the inaccuracy of clinical
assessment of grade 1 goitres
is increasingly recognized. With
more frequent use of ultrasonography ( described below), it is known that once
that rate of thyroid
enlargement exceeds 5%,
there is increasing evidence of biochemical
abnormality.
59. The
specificity and sensitivity of
palpation are low in
grades 0
and 1 due to a
high inter-observer variation.
As demonstrated by studies of experienced examiners, misclassification
can be as high as 40%. If possible, low
goitre rates, especially when found following a sustained intervention,
should be confirmed
by ultrasound, which
provides the only objective measure
of thyroid size (see next
section). In any case,
measurement of urine iodine levels ( in an adequate sample) is essential to
decide whether an iodine deficiency problem is of public health importance.
|
Grade O: |
No palpable or visible
goitre. |
|
Grade 1: |
A mass in the neck that is consistent
with an enlarged thyroid that is in the
normal position. It moves upward in the neck as the subject swallows. Nodular
alteration (s) can occur even when the thyroid is not visibly enlarged. |
|
Grade 2: |
A swelling in the neck that is visible
when the neck is in a normal position and is consistent with an enlarged
thyroid when the neck is palpated. |
-------------------------------
1
Glinoer D. Thyroid regulation during
pregnacy. In: Ibid., 181-188.
2
Delange F. et
al. Neonatal thyroid function in
iodine deficiency. In: Ibid.,199-07.
3 Delange
F. et al. Influence of dietary
goitrogens during pregnancy in
humans on thyroid function of the newborn. In:Nutritional factors
involved in the goitrogenic action
of cassava. Delange F. et al., eds. IDRC Monograph 184e,
Ottawa, International Development Research Centre, 1982:40-50.
Table 4. Epidemiological
criteria for assessing the severity of
|
|
Mild IDD |
Moderate IDD |
Severe IDD |
|
Prevalence
of goitre |
5.0-19.9% |
20.0-29.9% |
> 30.0% |
Ultrasonography
Feasibility
60. Ultrasonography is
a safe, non-invasive specialized technique, which should be
performed by trained operators who can perform
up to 200
examinations per day.
The degree of subjectivity in
assessing results, however,
points up the importance of developing standardized interpretation criteria. Thyroid volume can be easily calculated using a
calculator or a microcomputer during
data entry. Portable ultrasound equipment is relatively rugged but requires electricity; it can be
operated from a car battery with the aid of a transformer.
Cost
61. Portable
ultrasound equipment with a 5
MHz transducer currently costs about US$12000. Prices are expected to
decline with the increasing availability of smaller machines.
Performance
62.ultrasonography
provides a more precise measurement of thyroid volume (Table 5) compared with
palpation. This becomes especially significant when the prevalence of visible
goitres is small , and in monitoring iodine control programmes where thyroid
volumes are expected to decrease over time. For practical reasons, school-age
children between 8-10 years should be examined, although
this range can be extended to 6-12 years if necessary. The thyroid of
younger children is more difficult to examine,
especially in children under
six years of age . For this purpose a 7.5 MHz
transducer
is required to obtain adequate resolution.
Table 5. Comparison between
palpation and ultrasound in grading small thyroids
|
|
Goitre grade |
|||
|
0 |
1a |
1b |
2 |
|
|
Number of subject by
palpation |
105 |
88 |
101 |
10 |
|
Subjects with grade
confirmed by ultrasound |
||||
|
Numbers |
63 |
72 |
89 |
10 |
|
Percentage |
60 |
82 |
88 |
100 |
Interpretation
63.
Results of ultrasonography from a study population should be compared to
normative data from populations
with sufficient iodine intake( average intake> 150U*g iodine per
person per day and urinary iodine>
100 µg/1) . In an iodine-replete population,
the expected prevalence of thyroid sizes greater than the mean + 2 SD would
be 2.3 %, and this figure can be compared
with the observed prevalence.
In addition, the median
(50th centile ) thyroid volume may be useful. Normative
thyroid volume size data from ultrasound
on iodine-replete children
in Europe are
currently being reviewed 1,2 and will be
available on request from
WHO and ICCIDD
about mid -1995.
Comparable data for developing countries are not available. There is
some evidence of a correlation between thyroid size in adults and body weight.3
Children in developing countries commonly
weight substantially less and
are shorter than same-age
European children. It is
possible that normal
thyroid size in such
subjects would be smaller
. Adequate reference values for children of small
body size cannot, therefore, be provided at present.
---------------------------------
1 F.
Delange, personal communications.
2 Vitti P
et al. Thyroid volume measurement
by ultrasound in children as a tool for the assessment of
mild iodine deficiency. J. Clin.
Endocrin. Metab., 1994,79:600-603.
3 Hegediis
et al. The
Determination of Thyroid
Volume by Ultrasound and its Relationship to Body Weight, Age, and
Sex in Normal Subject. J. Cline Endocr.
Metab.,1983,56:260-263.
Cretinism
Definition
63. Endemic cretinism has been well described and
defined by a Pan
American Health Organization
study group 1 and Delange2
in terms of three major features:
(a)
Epidemiology: Cretinism is associated with endemic goitre and severe iodine
deficiency.
(b) Clinical manifestations of cretinism are mental
deficiency and either:
(i)
a predominant neurological
syndrome consisting of hearing and speech defects and varying
degrees of characteristic stance and
gait disorders; or
(ii) predominant hypothyroidism and
stunted growth. In some regions, one of the two types may predominate;
in others a mixture of the two
syndromes will occur.
(c) Prevention:
endemic cretinism has been prevented
in areas where adequate
correction of iodine deficiency has been achieved.
Biological features
64. The
typical neurological cretin3 is
extremely mentally retarded and
most are reduced to a vegetative existence; most are deaf-mute and
many have impaired
voluntary motor activity,
usually involving
paresis or paralysis of
pyramidal origin, chiefly in the lower limbs, with hypertonia, clonus,
and plantar coetaneous reflexes in
extension, and occasionally extra-pyramidal
signs; spastic or ataxic gait (in severe cases,
walking or even standing are
impossible); and strabismus
(squint). In milder cases
there are varying degrees of impaired motor coordination,
inability to play games etc. The prevalence of
goitre in these cretins
is as high as in the
non-creations population of the
areas concerned and they are clinically euthyroid.
--------------------------------
1
Querido AF et al. Definitions of endemic goitre and cretinism, classification of goitre size and
severity of endemic, and survey
techniques. In: Dunn JT, Medeiros-Neto, eds. Endemic goitre and
cretinism. PAHO Scientific
Publication No. 292, Washington DC, 1974:266-272.
2 Delange
FM. Anomalies in physical and intellectual development associated with severe
endemic goitre. In: Dunn JT et al. Endemic goitre cretinism,
and iodine deficiency. Pan
American Health Organization, Scientific
Publication No. 502, Washington DC, 1986:49-67.
3 Buttfield
IH , Hetzel BS. Endemic cretinism in eastern New Guinea. Aust. Ann. Med.
1969,18:217-221.
65.
There is less mental retardation among typical
myxoedematous cretins, as found for example in Zaire, who may be capable
of performing simple manual
tasks.Sings of long-standing
hypothyroidism include
dwarfism, myxoedema, dry skin,
sparse hair, retarded sexual
development, and retarded
maturation of body proportions
and naso-orbital configuration. Goitre prevalence is much lower than in the non-cretinous
population. The clinical picture is confirmed by extremely low levels
of serum T4 and T3, low thyroidal
uptake of radio-iodine, and very
high levels of serum TSH. Lesser degrees
of hypothyroidism have fewer
clinical signs and
biochemical abnormalities. Delong provides additional details of
neurological assessment of cretins 1.
66. In addition to the severe clinical
manifestations of frank cretinism , it is clear that iodine deficiency causes a
spectrum of milder clinical and
subclinical deficits- sensory, motor and
intelectual-
that may be difficult to identify in a typical field setting . These deficits
may be grouped
into three main categories:(a) mild multiple features of
cretinism (e.g. hearing
loss,
squint); (b) motor deficits and/or developmental delay; and (c) impaired
intellectual performance quantified, for example, by improved IQ
scores (5-20 points) in offspring
of mothers who received
iodized oil during pregnancy.2 Hetzel3 also
draws attention to impaired
cerebral functions that occur in areas
of chronic iodine deficiency.
67. In an
effort to quantify the prevalence
of cretinism and milder
clinical and subclinical
deficits in iodine
deficient populations, Clugston et.al.,4reviewed available studies from
a
few countries in Asia, and in Ecuador and Zaire and identified a
functional relationship between
goitre rates cretinism prevalence. These studies also indicated that the prevalence of
milder clinical
and subclinical deficits was approximately
3 times the prevalence
of frank cretinism.
A recent meta- analysis of
effects of iodine
deficiency on cognitive
and
psychomotor development
reported an average
loss of 13.5IQ points.
-------------------------
1 Delong
R. Neurological involvement
in iodine deficiency disorders. In: Hetzel et al. ,
eds. The prevention and control of iodine deficiency disorders. Amsterdam,
Elsevier, 1987:49-63.
2 Fierro
- Benitez R et al. Iodized oil in
the prevention of endemic
goitre and associated defects in
the Andean region of Ecuador. I Program design, effects on
goitre prevalence, thyroid
function and iodine
excretion. In: Stanbury JB, ed.
Endemic goitre. Washington, World Health Organization/Pan American
Health Organization, 1969,193:306-321.
3 Hetzel B. The
story of iodine deficiency : an
international challenge in nutrition. Oxford /New Delhi, Oxford
University Press, 1989:89-93
4 Clugston
GA et al. Iodine deficiency disorders in South - East Asia
. In: Hetzel BS, Dunn JT, Stanbury JS, eds.
The prevention and control of iodine deficiency
disorders. Amsterdam, Elsevier,
1987:273-308. The functional relationship(Fig.2 , p.280) between total goitre rate. The multiple 3.0 ,
relates milder clinical and subclinical
deficits to cretinism prevalence,
includes those individuals with
milder mental, motor and other developmental handicaps (deficits or delays)
and those with multiple or single
cretinous features which are not
identifiable as full cretinism in a
typical field study setting. This includes an
estimated proportion of persons, in endemic areas, who score below the
mean of mental or motor ability tests in excess of the proporation of persons
who score below the mean in a
comparable, but non-endemic, area.
Annex 3 Lot quality assurance sampling
Source: S.K. Lwanga and S. Lemeshow (1991), Sample
size determination in health studies, A practical manual, pp 15-16 and pp
63-71, WHO, Geneva.
Accepting a population prevalence as not exceeding a
specified value
This
section outlines how to determine the minimum sample size that should be
selected from given population so that, if a particular characteristics is
found in no more than a specified number of sampled individuals, the prevalence
of the characteristic in the population can be
Required information and notation
(a) anticipated population prevalence P
(b) Population size N
(c) Maximum number of sampled individuals showing
characteristic d*
(d) Confidence level 100(1-x)
%
Tables A3.1 present minimum sample sizes
for confidence levels of 95% and 90% and values of d* of 0-4.
Example 18
In a
school of 2500 children, how many children should be examined so that if no
more than two are found to have malaria parasitaemia it can be concluded, with
95% confidence, that the malaria prevalence in the school is no more than 10%?
Solution
(a)
Anticipated population prevalence 10%
(b)
Population size 2500
(c)
Maximum number of malaria cases in the sample 2
(d)
Confidence level 95%
Table A3.1 shows that for P=0.10 and
N=2500 a sample size of 61 children would be needed.
Decision rule for “rejection a lot”
This
section applies to studies designed to test whether a “lot” ( a sampled
population) meets a specified standard. The null hypotheses is that the
proportion of individuals in the population with a particular characteristic is
equal to a given value, and a one-sided test is set up such that the lot is
accepted as meeting the specified standard only if the null hypothesis can be
rejected. For this purpose a “threshould value” of individuals with the
characteristic (d*) is the characteristic does not exceed the threshold, the null
hypothesis is rejected (and the lot is accepted), whereas if the threshold is
exceeded, the lot is rejected.
Required information and notation
(a) Test
values of the population proportion under the null hypothesis Po
(b) Anticipated
value of the population proportion P
a
(c) Level
of significance 100x%
(d) Power
of the test 100(1-β)
%
Tables A3.2 present minimum sample size for a level of
significance of 5% and power of 90%, 80% and 50% in one-sided tests.
Example 19
In
large city, the local health authority aims at achieving a vaccination coverage
of 90% of all eligible children. In response to concern about outbreaks of
certain childhood disease in particular parts of the city, a team of
investigators from the health authority is planning a survey to identify areas
where vaccination coverage is 50% or less so that appropriate action may be
taken. How many children should be studied, as a minimum, in each area and what
threshold value should be used if the st6udy is to test the hypothesis that the
proportion of children not vaccinated is 50% or more, at the 5% level of
significance? The investigators wish to be 90% sure of recognizing areas where
the target vaccination coverage has been achieved (i.e. where only 10% of children
have not been fully vaccinated).
Solution
(a) Test
value of the population proportion 50%
(b) Anticipated
value of the population proportion 10%
(c) Level
of significance 5%
(d) Power
of the test 90%
Because
the mistake of accepting groups of children as adequately vaccinated, when in
fact the coverage is 50% or less, is the more important, Po = 0.50 and Po =0.10. Table
A3.2 shows that in this case a sample size of 10 and a threshold value of 2
should be used.
Therefore,
a sample of 10 children should be taken from each of the areas under study. If
more than 2 children in a sample are found not to have been adequately
vaccinated, the lot (the sampled population) should be “rejected”, and the
health authority may take steps to improve vaccination coverage in that
particular area. If, however, only 2 (or fewer) children are found to be
inadequately vaccinated, the null hypothesis should be rejected and the group
of children may be accepted as not being of immediate priority for an
intensified priority for an intensified vaccination campaign.
`Feasibility
68. Because cretinism is a
clinical diagnosis of a disorder with a presentation spectrum from mild to
devastatingly severe, it is difficult to identify all of the affected
individuals in a population. In fact,
the more mildly affected cretins may not be diagnosed except by clinical
experts or by using specialized methods, e.g. audiometric or psychometric
tests. A significant amount of time may
be required to perform the necessary physical examination.
Performance
69. Prevalence of cretinism
is not a sensitive indicator of a population’s iodine status. As noted above, complete identification of
cases is difficult and requires expert clinical skills. The more severely affected cretins, though
easily identifiable, probably represent only the ‘tip of the iceberg’ in terms
of IDD case-finding.
Interpretation
70. Clinical examination for signs of cretinism may be most
interesting as a historical record of a community’s exposure to iodine
deficiency, and their prevalence as an initial indicator of severe endemic
IDD. While cretinism results from iodine
deficiency during intrauterine life and early childhood, it is most easily
diagnosed in later childhood and adulthood.
In a qualitative sense, therefore, the presence of cretins in a
community, even if prevalence is very low, is significant because it
demonstrates that individuals were exposed to a marked environmental iodine
deficiency sometime in the recent past.
While this does not necessarily reflect a population’s current iodine
status, it may have considerable advocacy value. As iodine deficiency decreases, cretins will
no longer be born, cretinism will progressively disappear, and the condition
will lose its value for monitoring IDD.
However, subclinical cretinous manifestations will still have to be
taken into account in assessing the disability-burden in areas of severe iodine
deficiency.
Biochemical
indicators
Biological features
71. Since most iodine that is absorbed is excreted in the urine,
urinary iodine level is a good marker of a previous day’s dietary iodine
intake. However, since an individual’s
level of urinary iodine varies daily and even during a given day, data can be
used only for making a population-based estimate. Experience has shown that the iodine
concentration in early morning urine specimens (child or adult) provides an
adequate assessment of a population’s iodine status; 24-hour samples are not
necessary. It has been found preferable
to express the results per litre of urine1 rather than per gram of
creatinine as was done formerly.
Relating urinary iodine to creatinine is cumbersome, expensive,
unreliable and unnecessary. Generally
speaking, where reports have quoted iodine concentration per gram of
creatinine, the same figure may be taken as the concentration per litre of
urine. This 1:1 relationship is not
always valid, however, particularly where a subject’s protein intake, and
consequent creatinine excretion, is very low, e.g. as in parts of Zaira and
Papua New Guinea.
Feasibility.
72. Acceptability is very high and spot urine specimens are easy
to obtain. Urinary iodine assay methods
are not difficult to learn or use,2 but meticulous attention is required
to avoid contamination with iodine at
all stages. Special rooms, glassware and
reagents should be set aside solely for this purpose.
1 100mg/l is equivalent to 0.79 mmol/l.
2 Dunn JT et al. Methods for measuring iodine in urine. ICCIDD/UNICEF/WHO, 1993
73. Small amounts (0.5-1.0 ml) of urine are required.3 Specimens are collected in tubes, which are
tightly sealed with screw tops;4 they do not require
refrigeration or the addition of a preservative. Iodine content remains stable throughout
transport to the laboratory. The tightly
sealed specimens can be refrigerated in the laboratory for several months
before actual determinations are made.
Should evaporation occur, iodine concentration will increase.
74. Many analytical techniques are used. The simplest methods, quite adequate for
epidemiological surveys, require less than one milliliter of urine. Annex 4 provides a simple method, recommended
for use in most circumstances. The
specimen is digested in chloric acid and its iodine content is measured by its
catalytic action in the reduction of ceric ammonium sulfate (yellow) to the
cerous (colourless) form. The result is
expressed as a concentration (mg I/l or mmol I/l urine). A trained technician can process 150
specimens per day. Total instrument
costs are about US$ 3000, and the total cost per specimen has been estimated to
be $ 0.50-1.00, including labour. Other
methods digest the urine more completely, but are more complicated, time
consuming and costly.
75. Since
casual specimens are used, it is desirable to measure about 300 from a given
population group to allow for varying degrees of subject hydration and other
biological variations between individuals, as well as to obtain a reasonably
small confidence interval as already discussed in paragraphs 30-31. A sample of 200 specimens would give a
relative precision of 20%, e.g. 50 + 10% below 100 mg/l.
Smaller sample sizes are adequate to establish at the outset that iodine
deficiency is the cause of the endemic goiter.
Performance
76. The recommended methods are able to detect urinary iodine
levels as low as 5 to 20 m/l with a
coefficient of variation under 10%.
Laboratory techniques require training, but are not difficult to
apply. As in all surveys for estimating
prevalence, population samples must be representative.
Interpretation
Simple
modern methods make it feasible to process large numbers of samples at low cost
and to characterize the distribution according to different cut-off points and
intervals. The cut-off points proposed
for classifying iodine deficiency into different degrees of public health
significance are shown in Table 6.
Frequency distribution curves are necessary for full interpretation,
since urinary iodine values from populations are usually not normally
distributed, and therefore the median value should be used rather than the
mean. The indicator of iodine deficiency
“elimination” is a median value for iodine concentration of 100 mg/l, i.e. 50% of the samples should be above
100 mg/l, and not more than 20% of samples should
be below 50 mg/l. In principle, the same sample size tables
apply as those given for goitre prevalence rates (see Annex 2). The “design effect” applicable for goitre
prevalence surveys (see Annex 2), i.e. 3, is also applicable for urinary iodine
values, at least in an initial surveys, but may decrease over time with
implementation of a salt iodization
programmes. If there are fewer subjects
per cluster (e.g. 10, the design effect may be reduced to 2.
78. As an IDD prevention programme progresses, goitre rates become
progressively less useful, and urinary iodine levels progressively more useful,
as elimination criteria.
3 However, considerably larger specimens are required if urinary
thiocyanate levels are to be measured as well, which is desirable whenever
cassava is a major part of the diet.
Cyanogenic glucosides are formed from linamarin in cassava, and the
thiocyanates impair utilization of iodine by the thyroid. Smoking is also believed to cause varied
urinary thiocyanate levels (see Delange F et al. Influence of dietary goitrogens during
pregnancy in humans on thyroid function of the newborn. Op. cit.).
4 Such
as Sarstedt tubes (manufacturer’s reference number 60542).
|
Median value (mg/l) |
Severity of IDD |
|
<
20 20-49 50-99 > 100 |
Severe
IDD Moderate
IDD Mild
IDD No
deficiency |
Blood constituents
79. Two blood constituents, TSH and Tg, can be used as
surveillance indicators. In a popultion
survey, the collection of a blood spot on filter paper can be used to measure
TSH and Tg. Issues common to TSH and Tg
will not be repeated in the section on Tg.
The thyroid hormones thyroxine (T4) and triiodothyronine (T3) are not
discussed, since their measurement is not recommended for surveillance
purposes.
Biological features
80. Iodine is essential for synthesis of thyroid hormones, which
are necessary for normal brain and neurological development. The kinetics of the thyroid hormone receptor
in the pituitary gland mimic the kinetics of thyroid hormone receptors in the
brain. When iodine levels are low, the
concentration of thyroid hormones in the pituitary gland stimulates the release
of TSH, which is then detectable in blood.
Serum or whole blood TSH levels thus directly reflect the availability
and adequacy of thyroid hormone. TSH
level is the best diagnostic test for determining hypothroidism. An elevated TSH level in neonate or infant
blood is of considerable concern because this indicates an inadequate thyroid
hormone level during this crucial stage of brain development. However, if used for purposes of screening,
i.e. detecting all hypothyroid neonates, coverage must be universal for the
screening to be considered adequate.
There are limitations in interpreting adult TSH levels, as noted in
paragraphs 82 and 87.
Feasibility
81. There is a wel-established methodology for determining TSH
levels from either dried whole blood spots on filter paper or serum. Whole blood from any site is acceptable for
spotting on to certified grade 1 filter paper.
Because only a few drops of whole blood are required, a dinger, heel, or
earlobe is the most common site for puncture.
It is essential that sterile equipment be used – either lancets for
blood spot collection, or needles and syringes for collecting whole blood from
which the serum is separated. Standard
procedures for handling blood products or objects contaminated with blood
should be followed. The risk of
contracting HIV or hepatitis infection from dried blood spots is extremely low.
82. Timing. For neonates, blood specimens can be
collected from cord blood at the time of birth, or by heel prick after 3 days
(72 hours). Collection from the heel
during the first 3 days (< 72 hours) of life is not recommended because
levels can be elevated as a result of the birth process itself. After the first three days, the timing of
specimen collection is no longer crucial.
Blood specimens may be obtained from pregnant women during prenatal care
visits, or from school-age children during school-based surveys. However, further study of TSH distributions
among these older subjects is needed to improve understanding of the
specificity of their relationship to iodine deficiency.
83. Transport. Blood spots are easy to transport. It is important that the spot be dry before
storage or shipment. Filter papers,
usually stored in a plastic bag, can be transported using the normal postal
system and are stable for periods of up to 6 weeks even in a hostile
environment of high temperature and humidity. Customs clearance may be required
for international transport of dried blood spots.
84. TSH assay. TSH in the blood spot can be measured by
commercially available assay kits. The
enzyme-linked immunosorbant assay (ELISA) methodology is recommended because of
lower equipment cost, longer shelf life of reagents (6 months), and high
sensitivity (< 2 mU/l)1.
Direct linkage of the ELISA system to a microcomputer is recommended
because this allows a high throughput and facilitates data management for
quality assurance and public health decision-making. Laboratory staff may need training in
laboratory management and quality assurance.
Laboratories should participate in an external quality control
programme.2
85. Cost.
Current estimates for the cost of laboratory equipment and reagents are
given below. Labour costs for collection
and processing will vary depending on local circumstances.
US$
15 000
US%
5000
US$ 0.50-1.00/test.
The same equipment can be used to perform
ELISA-based assays for surveillance of various other micronutrients and
infectious diseases.
Performance
86. Sensitivity. The term ‘sensitivity’ has two distinct
meanings depending on whether it is used in a laboratory or epidemiologic
context. In a laboratory context, ‘sensitivity’ refers to the improved ability
of the TSH while-blood spot-assay kits to detect levels of TSH over the full
physiologic range including values as low as 1-2 mU/l. Earlier TSH assay systems were sensitive and
reliable only for values over approximately 20 mU/l. The relatively recent commercial availability
of the ‘sensitive’ TSH blood spot assay kit now permits the determination of
mild-to-moderate UDD, which may occur when TSH levels are under 20 mU/l.
1 Miyai K, Ishibashi K, Kawashima M. Two-site immunoenzymometric assay for
thyrotropin in dried blood samples of filter paper. Clin.
Chem. 1981;27 (8): 1421-1423; Tseng YC, Burman KD, Baker JR, Wartofsky
L. A rapid, sensitive enzyme-linked
immunoassay for human thyrotropin. Clin.
Chem. 1985;31 (7):1131-1134; Waite KV, Maberly GF, Eastman CJ. Storage conditions and stability of
thyrotropin and thyroid hormones on filter paper. Clin.
Chem. 1987;33:853-855; Waitre KV,
Maberly GF, Ma G, Eastman CJ.
Immunoradiometric assay with use of magnetizable particles: measurement
of thyrotropin in blood spots to screen for neonatal hypothyroidism. Clin.
Chem. 1986;32 (19):1966-1968.
2 The Programme Against Micronutrient Malnutrition
(PAMM), which is based at the Emory
University School of Public Health, Atlanta, Georgia, USA, supports such a
programme.
87. In an epidemiologic context, ‘sensitivity’ refers to the
ability of the TSH assay to identify correctly IDD cases among populations.
‘Specificity’ refers to a screening test’s ability to identify correctly
those people who do not have a given condition.
The specificity of TSH for IDD screening has not been clearly
quantified. However, the number of false
negatives – individuals with IDD who test negative – is probably very low. Individuals with mild IDD and mild elevations
of TSH can now be detected by using the newer sensitive whole-blood TSH
assays. Elevation of TSH values in
individuals is associated with all causes of primary hypothyroidism. Causes of TSH elevation other than iodine
deficiency include goitrogen ingestion, congenital hypothroidism (CH), and
autoimmune thyroidiitis. CH is reltively
uncommon, affecting approximately 1 in 4000 neonates (0.275%) worldwide. Autoimmune thyroid disease is less rare,
particularly in Western countries.
Goitrogen ingestion is usually regional and easily identified in the
environment, and in any case results in a relative deficiency of iodine because
requirements are greatly increased. IDD
screening programmes are not designed to follow-up individuals but to direct
population-based interventions. TSH
levels are an excellent indicator for hypothyroidism in neonates, but their
specificity in older groups is less certain.
In some populations (e.g. in Algeria1) there are clear examples of raised goitre
rates and inadequate urinary iodine intakes that are associated with TSH levels
with the normal range.
Interpretation
88. Reference data for TSH are available among neonates because
they are routinely collected as part of neonate congenital hypothyroid
screening programmes. TSH values are
currently reported in whole blood units or serum units. It is crucial that all reports and discussion
of TSH distributions specify the units employed. Congenital hypothyroid screening and IDD
surveillance require different TSH cut-off points. A TSH cut-off of 20-25 mU/l whole blood
(approximately 40-50 mU/l serum) is commonly used to screen for congenital
hypothyroidism. IDD may be present with
TSH levels which are only mildly
elevated. While further study of
iodine-replete populations is needed, a cut-off of 5 mU/l whole blood may be
appropriate for epidemiological studies of IDD.
Populations with a substantial proportion of neonates with TSH levels
above the cut-off could indicate a significant IDD problem (see Table 7).
Thyroglobulin
(Tg)
Biological features
89. Insufficient iodine intake induces a proliferation of thyroid
cells, which results in hyperplasia and hypertrophy. This leads to an enhanced turnover of thyroid
cells, which release Tg into the serum.
Tg in serum changes inversely with iodine intake in all age groups.
Feasibility
90. Acceptability is high, and sample collection and transport are
simple, identical to those for TSH (see previous section). The determination technique is similar to
that for TSH using a Tg antibody instead of one for TSH, although methods are
not yet commercially available. Costs
are comparable to those for TSH.
Training will be required in laboratory techniques.
Performance
91. Available
methods can detect levels as low as 2 ng/ml serum with a coefficient of
variation close to 5%. Tg changes
rapidly after an alteration of iodine intake in all age groups, and may be a
more sensitive indicator than TSH, as the following observations suggest. Tg rises in individuals with an insufficient
iodine intake, even under conditions where TSH falls or is suppressed due to
functional autonomy, as frequently happens with long-standing iodine
deficiency. After iodine depletion, Tg
will rise before TSH shifts to higher values and long before goitre
develops. Following iodine
supplemenation, Tg normalizes before thyroid volume has decreased.
1 Benmiloud M et al. Oral iodized oil for correcting iodine
deficiency: optimal dosing and outcome indicator selection. J. Clin.
Endocr, Metab. 1994, 79:20-24.
Interpretation
92. Individuals (children and adults) with sufficient iodine
intake show a median Tg serum level of 10 ng/ml and a normal limit, for
individuals, of 20 ng/ml in most assay techniques. The results obtained from a survey should be
expressed as a median and as the percentage of Tg levels above 20 ng/ml.
93. Table 7 provides a summary of cut-off points and prevalences
that are considered indicative of a significant public health problem for
IDD. As mentioned in paragraph 49, it is
highly desirable to have more than one parameter of iodine and thyroid
status. Generally speaking, at least two
are required. The foregoing pages
provide information on which to base a suitable choice of indicators.
|
|
|
Severity of public health
problem (prevalence) |
||
|
Indicator |
Target population |
Mild |
Moderate |
Severe |
|
Goitre grade > 0 Thyroid volume > 97 th
centile by ultrasoundb Median urinary iodine level (mg/l) TSH > 5 mU/l whole blood Median Tg (ng/ml serum)c |
SACa SAC SAC Neonates C/Ad |
5.0-19.9 % 5.o-19.9 % 50-99 30-19.9 % 10.0-19.9 |
20.0-29.9 % 20.0-29.9 % 20-49 20.0-39.9 % 20.0-39.9 |
>30.0 % >30.0 % <20 >40.0 % >40.0 |
a SAC = School-aged children. c Different
assays may have different normal ranges.
b Normative tyroid volume size values
will be d C/A = Children and adults.
available from WHO and ICCIDD in 1995.
5.3
Process
indicators for IDD control programmes
94. IDD control programmes
should have built-in monitoring and education mechanisms. Included in such protocols will be both
process indicators associated with programmes implementation, and indicators of
impact achieved through its implementation as a result. Depending upon the specific characteristics
of the IDD control programmes, distinct indicators will need to be considered
and different techniques employed to monitor them. Although most countries will use iodized salt1
as the primary control activity,2 other programmes, where
implemented, will need to be monitored, even if they are used as short-term
measures while salt iodization is being established.
1 The term “iodized salt” refers to salt iodinated with any iodine
compound – usually potassium iodide (KI) or iodate (KIO3).
2 Manner MGV and Dunn JT.
Salt iodization for elimination of iodine deficiency. ICCIDD/MI/UNICEF/WHO, 1995 (in press).
DD public health problem should undertake a
situation analysis of salt available for human and animal consumption, from
points of production (or importation) through distribution channels to actual
consumption. Such salt is refered to as
food-grade salt,3 which includes crude salt for direct edible use by
people and live stock; refined salt for edible use; an
95. All countries with a significant I d salt
used industrially in processed foods such as bread, biscuits, salted foods,
canned foods, and instant foods.
96. The situation analysis
should include a list of the major salt producers or importers,
production/import/export statistics, and information on salt quality,
packaging, transport and storage, retail marketing, prices and household
consumption. This information should be
updated periodically, e.g. every two
years. In addition, periodic monitoring
will have to be undertaken at different points along the distribution chain of
iodized salt to ensure iodine-concentration levels are adequate and, if
necessary, that corrective action is taken.
This could include improvements in packaging, transportation or storage
of iodized salt at different points before it reaches the consumer, and
modification of iodine level at the factory.
97. Salt iodization involves
the addition of a small quantity of iodine (30 to 100 mg of iodine per kg of
salt, or parts per million, usually in
the form of potassium iodide or potassium iodate.1 A joint FAO/WHO expert committee noted in
1990 that “potassium iodate has been shown to be a more suitable substance for
fortifying salt than potassium iodide, because of its greater stability,
particularly in warm, damp or tropical climates”.2
Techniques for measuring salt iodine levels
98. There are essentially two techniques for
measuring iodine levels in salt:
(a). Standard titration
method. It is necessary for measuring
this method be conducted in a laboratory.
Iodine is liberated using sulphuric acid. The free iodine is titrated with sodium
thiosulphate, using starch as an indicator. Slighlty different techniques are
employed, depending on whether the iodine was in the form of iodate or iodide. Facilities for this method are normally
available in most countries in a public health or standards laboratory, but
some other laboratory may need to be equipped and helped to develop competence
in its operation. The titration method
is preferred for checking salt batches produced in factories, or on arrival in
the country and in general where accurate testing is required. However, it is too time-consuming and
expensive for purposes of routine national monitoring, for which the second
method may be more suitable.
(b). Rapid-test kits. These comprise bottles of starch solution
(stabilized), one drop of which is placed on the salt. If the salt is alkaline a neutralizing
solution is first applied. The intensity
of the blue colour which develops indicates the salt iodine level, up to 50 or
100 ppm, depending on the kit used, with an accuracy of + 10 ppm. Most of the rapid-test kits presently
available can detect the presence of iodine only. Details of available kits may be obtained
from WHO, UNICEF or ICCIDD.
Monitoring salt iodine levels
3 Food-grade salt is a crystalline product destined for human or
animal consumption. It should contain at
least 97% sodium chloride on a dry-matter basis, excluding additives. All additive should be of food-grade
quality. The production, packaging,
storage and transportation should be such as to avoid any risk of
contamination. Contaminants such as
arsenic, copper, lead, cadmium and mercury should not exceed the levels of 0.5,
2, 0.5, and 0.1 mg/kg, respectively, as specified by the FAO/WHO Codex
Alimentarius Commission. Codex
Alimentarius, Vol. 1, General requirements, section 5.5. Food and Agriculture Organization of the
United Nations, Rome, 1992.
1 Iodine and health: eliminating iodine deficiency disorders safely
through salt iodization. A statement by the
World Health Organization. Document
WHO/NUT/94.4.
2 Joint FAO/WHO Expert Committee on Food Additives. Geneva, World
Health Organization, 1991 (WHO Technical Report Series, No. 806, Annex 5).
99. Iodization may take
place inside the country, at the main production or importation sites, or
outside, with salt being imported already iodized. There are inevitably some
losses of iodine from salt between production and consumption, and these are
often much higher in non-industrialized countries, where packaging, storage and
transportation conditions are sub-optimal.
Countries should establish the expected iodine content of salt at
different points along the distribution system, taking into account climatic
conditions, types of packaging, and habitual daily salt consumption. Guidelines are given in Table 8, calculated
in each case to provide individuals with 150 mg
iodine per day, and allowing for an average loss of 30% during cooking. The iodine concentration in salt needs to be
monitored regularly at two or three levels of the distribution system. The overall responsibility for quality
control within a country should be clearly identified. Often it is vested in the ministry of health
or equivalent authority through its primary health care unit and regional or
provincial and district health departments.
At the point of (internal) production, however, the ministry of industry
is likely to be involved; at the point of importation, the customs services;
and within the marketing system, the ministry of commerce or trade. A bureau of standards may also be responsible
for fixing and controlling the level of iodine in iodized salt. Criteria for
assuming programme adequacy are provided in Table 9.
Table 8. ICCIDD-UNICEF-WHO recommended levels of iodine in
salt
Examples of desirable
average levels at various points on the salt distribution chain, depending on
climate, salt intake, and conditions affecting packaging and distribution.
Parts of iodine per
million parts of salt, i.e. micrograms per gram, milligrams per kilogram or
grams per tonne.
|
Climate and daily salt consumption (g/person) |
Requirement at factory outside the country |
Requirement at factory inside the country |
Requirement at retail salt (shop/market) |
Requirement at household level |
||||||
|
|
Packaging |
|
||||||||
|
|
Bulk (sack) |
Retail pack (<2
kg) |
Bulk (sack) |
Retail pack (<2
kg) |
Bulk (sack) |
Retail pack (<2
kg) |
|
|||
|
War moist 5 g 10 g |
100 50 |
80 40 |
90 45 |
70 35 |
80 40 |
60 30 |
50 25 |
|||
|
Warm dry or cool moist 5 g 10 g |
90 45 |
70 35 |
80 40 |
60 30 |
70 35 |
50 25 |
45 22/5 |
|||
|
Cool dry 5 g 10 g |
80 40 |
60 30 |
70 35 |
50 25 |
60 30 |
45 22.5 |
40 20 |
|||
Source:- Adapted from World Summit for
Children – mid-decade goal: iodine deficiency disorders. Geneva, 1994.
UNICEF-WHO Joint Committee on Health Policy, document JCHPSS/94/2.7 and
document WHO/NUT/93.1.
N.B. 168.6 mg of KIO3
contains 100 mg of iodine.
N.B. These are indicative initial
levels, which should be adjusted in the light of urinary iodine measurement.
Table 9. Criteria for assessing adequacy of salt iodization
programmes
|
Process indicator |
Criterion of adequacy |
|
A. Factory or
importer level 1.
Percent of food-grade salt claimed to be iodized 2.
Percent of food-grade salt effectively iodized 3.
Adequacy of internal monitoring process* 4.
Adequacy of external monitoring process |
100 % > 90 % > 90 % 10-12 monthly checks
per producer/importer, per year |
|
B. Consumer and
district level 1. Percent of monitoring sites with
adequately iodized salt (i)
households (or schools) (ii)
district headquarters (including major markets) 2. Adequacy
of monitoring process** |
Adequate in 90% of
samples 90% or more |
* Corrective action
systematically taken within 3 hours in 90% of cases using LQAS methodology
** Monitoring undertaken in 90% of
districts in each province at both household and district levels.
Monitoring procedures
100. Factory level. The
responsibility for routine monitoring rests with the factory itself (internal
salt monitoring). The recommended
procedure is to carry out hourly monitoring with the rapid-test kit, and at
least once daily by titration.
Observations should be recorded systematically in a register indicating
the dte, time,batch number and iodine content of the salt. However, the responsible government authority
(e.g. ministry of health, industry or commerce, or the standards office or
other designated body) should also undertake periodic checks (external salt
monitoring) at least once monthly, by titration, and compare results with the
factory;s own test. During inspection,
the manufacturers’s records should be verified for adequacy of internal
monitoring and variations in iodine levels.
101. Importation level.
Principles similar to those at factory level apply. Each batch of imported salt should be monitored
using rapid-test kits, and periodic checks should also be made using a chemical
method. Responsibility for monitoring
may be delegated to the customs authorities, ministry of commerce or health, or
the standard board. The action to be
taken if the iodine level is too low depends on circumstances and facilities,
e.g. if there is no alternative iodation plant, which is desirable, the salt
should be sent there for obligatory iodization at the importer’s expense. Unless appropriate penalties are defined and
rigorously imposed, the whole system will break down. Hence ensuring cooperation in monitoring the
system requires the motivation of importers as much as producers. Large quantities of imported salt should be
checked at the source by a monitoring agency – several multinational firms
undertake such monitoring – to ensure that salt sent from external factories
complies with requirements at the source.
Imports should be divided into “lots”, e.g. train wagons or truckloads,
for sampling purposes. LQAS methods
apply to this type of sampling.
102. Wholesale
and retail level. The major
distributions should be sensitized to the subject and provided with rapid-test
kits to check the salt iodine levels before it is released for retail
sale. This is especially important in
larger countries or in situations where transportation results in a long time
lag between production and consumption.
Regular monitoring at three-monthly intervals may be advisable. If there are deficiencies, they should be
notified to the provincial or district health authorities.
103. Community
and district level.1
Public health inspectors or nurses at district level will often be
responsible for this monitoring activity, the objective of which is to verify,
using rapid-test kits, that adequate concentrations of iodine are attained in
salt, especially when it reaches the consumer.
Salt monitoring at the district and community level should be used as an
opportunity to disseminate information regarding the importance of consuming iodized salt. When checks at those levels show inadequate
salt iodine concentrations, further spot checks should be made at successive
levels of the distribution
system to identify at
what point losses are occurring.
104. Three approaches are recommended for
monitoring salt iodization quality at the district level, based on surveys in
markets, schools and households. In the
first two, more specific information may be obtained about distribution of
inadequately iodized salt samples within the district, which may be used to
highlight ‘hot spots’ where problems are likely to be occurring. For schools with an enrolment of 100 to 1000
pupils at least 35 salt samples should be collected and tested to detect if
more than 20% of the population served have access to inadequately iodized salt
(which is the case if 4 or more samples are found to be inadequate). In monitoring based on household sampling, it
will be possible to ascertain whether there is a problem for a district as a
whole, but it would not be possible to gain any information about patterns
within a district. For each district at
least 10 houses in each of 10 remote villages should be randomly selected for
spot testing, with a new selection if possible every 4 months.
105. More detailed guidelines on salt monitoring
will be included in a salt-monitoring kit, which is being developed by PAMM and
others.
106. Other potential IDD
control programmes include iodization of water, fortification of foods other
than salt, and supplementation with iodized oil. If other foods are fortified, then an
evaluation of that food, similar to the one described for iodized salt, should
be devised. In supplementation
programmes, it is important to ensure that the high-risk popultion is
adequately covered. LQAS surveys can be
useful (see paragraphs 35-38 and Annex 3).
107. Iodized oil has been
widely used in a number of national IDD control programmes. As salt iodization becomes more widespread,
iodized oil programmes are likely to be limited to areas that cannot be easily
reached by iodized salt distribution.
The essential points for monitoring purposes are two: ensuring that
normal thyroid function is attended and maintained in the target population; and
that operational coverage is adequate, i.e. at least 95% of the at-risk
population should receive iodized oil annually, or at an alternative frequency
set by programme managers in the light of local conditions. Operational coverage may be determined from
health center or mobile team records, or by rapid cluster-sample surveys to
determine the proportion of target
groups that actually received the supplement.
The latter approach requires that it be possible to identify accurately
whether or not an individual received a supplement. In areas where iodized oil is recommended for
specific target groups, e.g. young children and women of child-bearing age,
LQAS surveys can be useful in determining population coverage. Details of such monitoring are outside the
scope of the present document, which
focuses on universal salt iodization.
1 A district is the smallest administrative unit of local
government in which all the major government departments are represented.
6. Criteria for monitoring progress towards
eliminating IDD as a significant public health problem
108. Table 10 presents
recommended criteria for use as core indicators in monitoring progress towards
the goal of eliminating IDD as a significant public health problem. Criteria include both IDD status indicators
and a control programme process indicator since it is important to ensure
sustained control of iodine deficiency for an entire population rather than
focus on reaching goals based on measuring the IDD status of a single group. Moreover, monitoring salt iodization is
useful first step in tracking progress towards meeting the goal of IDD
elimination. Where thyroid size
indicators are concerned, a prevalence of thyroid enlargement above 5% signals a
public health concern. However, since
there a re several causes of thyroid enlargement, iodine status should be confirmed by assessing urinary iodine
concentration.
|
Indicator |
Goal |
Proportion of households consuming
effectively iodized salt |
>90% |
Proportion below 100 mg/l Proportion below 50 mg/l |
<50% <20% |
In school children 6-12 years of age: Proportion with enlarged thyroid, by
palpation ultrasound |
<5% |
Proportion with levels >5mU/l whole
blood |
<3% |
Annex 1 List of participants in
the consultation1
Dr D. Alanwick
Senior Nutrition
Advisor (Micronutrients)
UNICEF
DR J. Dunn
Box 511,
1 Joint WHO/UNICEF/ICCIDD Consultation on Indicators for Assessing
Iodine Deficiency Disorders and their Control Programmes, 3-5 November 1992,
Geneva.
Prof. F. Delange*
Department of
Pediatrics
Hospital Saint-Pierre
322 rue Haute
1000
Dr J. Gorstein
Community Systems
Foundation
Dr P. Greaves
2 The
London SE3 0AB
Dr R. Gutekunst
Im Felde 10
W-2430 Neustadt
Dr B.S. Hetzel
(Chairman)
C/o Health Department
Foundation
8th Floor,
Adeladide Medical
Centre for Women and Children
5006
Dr G. Maberly
Dr V. Mannar
Etobicoke
Ontario M9B 4X8
Dr G. Ndossi
Tanzania Food and
Nutrition Centre
United
Ms S. Pak
Community Systems
Foundation
* Unable to attend but made
substantial contributions by other means both during and after the
consultation.
Dr C.S. Pandav
Centre for Community
Medicine
All India Institute of
Medical Sciences
Ansari Nagar
New Delhi 110029
Dr C. Thilly
Ecole de Sante
Publique
Case postale 590
Route de Lennik 808
1070 Bruxelles
Dr F. Trowbridge
Director, Division of
Nutritioin
National Centers for
Disease Control and Prevention
DR K.V. Bailey,
Nutrition,
Dr G.A. Clugston,
Nutrition,
Dr N. Cohen, Expanded
Programme on Immunization,
Dr A. Verster,
Regional Advisor,
Nutrition,
Regional Office,
Confidence interval. The
interval is determined with a specific level of confidence (or probability)
that will include the population parameter being estimated. Thus 95% confidence interval is an interval
for which one would be 95% confident that it will include the true prevalence
level.
Precision is a measure of how close an estimate is, or is required to
be, to the true population value.
Table A2.1 gives the sample sizes in terms of relative precision, i.e.
expressed as a proportion of the mean value (e)
expected or obtained in the survey. The
top row marked “P” gives the anticipated prevalence. For example, 0.05 means a prevalence of 5%
and so on. The first column labeled “e” gives the desired relative precision. For example, a relative precision of 0.10 for
an expected prevalence of 0.50 implies a precion of 0.10 x 0.50 (or 5
percentage points) with a minimum sample size of 384 randomly selected
individuals to be 95% confident that the range 45% to 55% includes the true
prevalence.
Depending on the type of circumstances of a survey, the degree of
precision required may very, e.g. in an initial goitre survey, with expected
goitre rates of about 50%, a relative precision of 20% may be appropriate, i.e.
50% + 10%. If a salt iodization
programme has been implemented for some years and goitre rates are though
likely to be around 5%, a relative precision of 30% (which would mean an
eventual “width” of + 1.5%) may be adequate.
If the anticipated prevalence is entirely unknown, sample size should
be estimated assuming that the result will be at whatever the critical level is
for decision-making, i.e. 30% prevalence indicating severe endemicity, or 5%
for existence of a public health problem.
Design effect. The figure in
Table A2.1 refer to a survey with strict random sampling of a give
population. Very often, cluster sampling
procedures are used, and are more practical – this is the procedure followed
for example in many EPI and CDD surveys, as well as in anthropometric and
goitre surveys. For phenomena such as
goitre whose distribution is patchy, cluster sampling may produce misleading results – if one
happens to fall on cluster of high prevalence, for example. To avoid errors of this sort a larger total
number of subject should be examined.
The numbers in the basic Table A2.1 should be multiplied by a factor
called the “design effect” to allow for this possible lack of homogeneity in
the population studied. It is an
indication of the variation due to clustering.
It is estimated by the ratio of the variance when cluster
sampling is used, to
established at 3 for goitre surveys.
In all cases where the parameter is not normally distributed (i.e. Gaussian distributioin), e.g. for
urinary iodine excretion values, it is wiser to give median rather than mean
values, or another form of presentation of the distribution of values, rather
than the mean, as indicated in paragraph 35 of this document.
For other confidence intervals and more details, including guidance on
other types of studies than simple prevalence surveys, including lot quality
assurance sampling, please see:
S.K. Lwanga & S. Lemeshow (1991) Sample size determination in
health studies, a practical manual, WHO Geneva (this book is a practical guide
to the subject, with a minimum of background theory).
S. Lemeshow et al. (1990)
Adequacy of sample size in health studies, Chichester, John Wiley (this book
includes the statistical methodology of sample size determination).
Annex 4 Recommended
method for determining iodine in urine1
Brief description. Urine is digested
with chloric acid under mild conditions and iodine determine manually by its
catalytic role in the reduction of ceric ammonium sulfate in the presence of
arsenious acid. The method described is fast and inexpensive , and the
emphasizes urinary iodine concentrations in the range of 0-150 µg/l (0-1.19
µmol/l) but can be expanded to cover a wider range of values.
Equipment. Heating block, colorimeter
(or simple spectrophotometer), vented fume hood with perchloric acid trap (or
the simple substitute apparatus described in the text), thermometer, test tubes
(13 mm x 100 mm), reagent flasks and bottles, pipettes and a laboratory
balance.
Reagents (analytical grade
only)
1.
KCIO3 (potassium chlorate), dry powder
2.
HCIO4 (perchloric
acid, 70%) comes as 70% liquid solution-do not dilute
3.
As2O3 (arsenic trioxide) , dry powder
4.
NaCl (sodium chloride), dry powder
5.
H2SO4 (sulfuric acid,
concentrated- 100%, 36 N, liquid)
6.
Ce(NH4)4(SO4)4 2H2O (ceric ammonium sulfate), dry
powder
7.
H2O (deionized water- must be free of
iodine and other contaminants)
8.
KIO3 (potassium iodate), dry powder
Solutions
1.
Chloric acid solution: In a 2000 ml Erlenmeyer flask
dissolve with heating 500 g KCIO3 in 910 ml H2O until
solubilized. This may take several hours and does not always dissolve
completely into solution. Then add slowly (about 15 ml/minute) 375 ml in a
vented fume hood. Store in a freezer overnight. The next day filter with filter
paper (Whatman #1 or similar product), preferably on a Buchner funnel. The
volume of filtrate is approximately 850 ml. Store in refrigerator (4 degree
celcius).
1 The
method described is a modified version of that developed by Wawshinek O, Eber
O, Petek P, Wakonig P and guraker A, 1985: Bestimmung der Harnjodaussheidung
mittels einer modifizierten Cer-Arsenitmethode. Berchte der OGKC 8:13-15. This
modified method and additional methods are described in: Dunn JT, Cructchfield
HE, Gutekunst simple methods for measuring iodine in urine. Thyroid , 1993,
3:119-123.
2.
5N H2SO4: Slowly add 139 ml
concentrated (36 N) H2SO4 to about 700 ml
deionized water [careful- this generates heat!] and when cool, adjust with
deiozed water to a final volume of 1 litre.
3.
Aresenious acid solutions:- In a 2000 ml Erlenmeyer
flask., place 20 g As2O3 and 50 g NaCl, then slowly add
400 ml 5N H2SO4. Add water to about 1 litre, heat gently
to dissolve, cool to room temperature, dilute with water to 2 litres, filter,
store in dark bottle away from light at room temperature. The solution is
stable for months.
4.
Ceric ammonium sulfate solution: Dissolve 48 g ceric
ammonium sulfate in 1 litre 3.5 N H2SO4 is made by
slowly adding 97 ml concentrated (36 N) H2SO4 to
about 800 ml deionized water [careful- this generates heat!], and when cool,
adjusting with deionized water to final volume of 1 litre ). Store in a dark
bottle away from light at room temperature. The solution is for months.
5.
Standard iodine solution, 1 µg iodine/ml (7.9 µmol/l):
Dissolve 0.168 mg KIO3 in deionized water to a final volume of 100
ml. (1.68 mg KIO3 contains 1.0 mg iodine; KIO3 is
preferred over KI because it is more stable, but KI has been used by some
laboratories without apparent problems). It may be more convenient to make a
more concentrated solution, e.g. 10 or 100 mg iodine/ml, then dilute to 1 µg
/ml. Store in a dark bottle. The solution is stable for months.
Standards curves for each assay can
either be prepared fresh each time by appropriate dilutions of the 1 µg/ml
solution of KIO3, or individual stock solutions of the desired
iodine concentrations can be made. The following are useful standards: 2, 5,
10, 15 µg/dl.
Procedure. Mix the urine sample to
suspend evenly any sediment, then pipette 250 µl of each urine sample into a 13
x 100 mm test tube. Prepare standards from the 1 µg/ml (7.9 µmol/l) iodine
solution by taking aliquots of 0, 5, 12.5, 25, or 37.5 µl, then add H2O
to a final volume of 250 µl for each tube. This gives iodine standards
corresponding to 0, 2, 5, 10, and 15 µg/dl (0, 0.16, 0.40, 0.79 and 1.19
µmol/l, respectively). Additional standards can be prepared if desired.
Add
750 µl of chloric acid solution to each tube (samples, blank and standards),
mix gently, and heat all tubes for 50-60 minutes under a heating block at
110-115 degree celcius in a fume hood with a perchloric acid trap. ( A fune
hood can be made using an inverted glass funnel suspended just above all the
tubes in the heating block and attached through a water trap to an aspitrator
suction.) The exact time and temperature are not critical as long as all tubes
are heated the same way. Usually there will be very little volumechange during
heating. If the volume has decreased, make to 1.0 ml with deionized water
(pre-marked tubes are useful for this). Some samples may become faintly yellow.
Cool
the tubes to room temperature, then add 3.5 ml arsenious acid solution to each
tube, mix (by inversion or vortex) and let stand for about 15 minutes.
Add
350 µl of ceric ammonium sulfate solution to each tube and quickly mix by
vortex or other means. Use a stopwatch or other precise timer to keep a
constant interval between additions to successive tubes, usually 15-30 seconds (20 seconds is
recommended as a convenient interval).
Exactly
20 minutes after addition of ceric ammonium sulfate to the first tube, read its
absorbency at 405 µM in a colorimeter (or spectrophotometer), and read
successive tubes at the same interval as that used for addition of the ceric
ammonium sulfate (i.e15-30 seconds, or
at 20 seconds as recommended above), so that time between addition of ceric
sulfate and reading is exactly the same for each tube (e.g. 20 minutes).
Calculation of results. Construct a
standard curve on graph paper by plotting iodine concentration of each standard
on the abscissa. This is the urinary iodine concentration in mg/dl.
NOTES