Community and International Nutrition

World Health Organization Hemoglobin Cut-Off Points for the Detection of

Anemia Are Valid for an Indonesian Population1
Helda Khusun, Ray Yip,*2 Werner Schultink3 and Drupadi H. S. Dillon
SEAMEO-TROPMED Regional Center for Community Nutrition, University of Indonesia, Salemba, Jakarta
10430, Indonesia; *UNICEF Indonesia, Jakarta 12920, Indonesia; and Deutsche Gesellschaft fur Technische
Zusammenarbeit, Eschborn, Germany

ABSTRACT The study was designed to determine whether population-specific hemoglobin cut-off values for
detection of iron deficiency are needed for Indonesia by comparing the hemoglobin distribution of healthy
young Indonesians with that of an American population. This was a cross-sectional study in 203 males and 170
females recruited through a convenience sampling procedure. Hemoglobin, iron biochemistry tests and key
infection indicators that can influence iron metabolism were analyzed. The hemoglobin distributions, based on
individuals without evidence of clear iron deficiency and infectious process, were compared with the National
Health and Nutrition Survey (NHANES) II population of the United States. Twenty percent of the Indonesian

Downloaded from jn.nutrition.org by guest on July 28, 2016

females had iron deficiency, but no male subjects were iron deficient. The mean hemoglobin of Indonesian
males was similar to the American reference population at 152 g/L with comparable hemoglobin distribution.
The mean hemoglobin of the Indonesian females was 2 g/L lower than that of the American reference
population, which may be the result of incomplete exclusion of subjects with milder form of iron deficiency.
When the WHO cutoff (Hb , 120 g/L) was applied to female subjects, the sensitivity of 34.2% and specificity
of 89.4% were more comparable to the test performance for white American women, in contrast to those of
the lower cut-off. On the basis of the finding of hemoglobin distribution of men and the test performance of
anemia (Hb , 120 g/L) for detecting iron deficiency for women, it is concluded that there is no need to develop
different cut-off points for anemia as a tool for iron-deficiency screening in this population. J. Nutr. 129:
1669 1674, 1999.

KEY WORDS: hemoglobin anemia iron deficiency humans screening

Iron deficiency is the most prevalent nutritional problem
worldwide; an estimated 2.15 billion individuals are anemic
because of iron deficiency (FAO/WHO 1992). Most affected
are children and women in the developing world. Considering
the magnitude of the problem and the multitude effects of iron
deficiency, assessment of the iron status of the population is
important for every country.
The most commonly used screening methods for the pres-
ence of iron deficiency in a population are the measurements
of hemoglobin or hematocrit concentration for the presence of
anemia (WHO 1994). These measurements are relatively sim-
ple and cheap, can be carried out under field conditions, and
values below a certain cut-off point indicate or define that
anemia is likely to exist. The cut-off value defining anemia has
been determined by convention as the value at 22 SD from the
mean or the 2.5th percentile of the normal distribution of a
healthy iron-replete population. Because iron deficiency is
often the most common cause of anemia, the presence of
anemia is also used as a screening tool for iron deficiency.
1
2
Supported by a grant from UNICEF Indonesia.
Current address: UNICEF China, Beijing 100600, China.
3
To whom correspondence should be addressed at UNICEF Headquarters,
New York, NY.
Although other iron-related tests are required for the confir-
mation of iron deficiency, it is reasonable to assume that a
population with a high anemia prevalence is likely to also have
a high prevalence of iron deficiency (Freire 1989, Yip 1994).
In view of the close relationship between anemia and iron
deficiency for either individual-based screening or for defining
the burden of iron deficiency on a population basis, it is very
important to ensure the validity of the hemoglobin cut-off
point for the detection of iron deficiency. It is well known that
there are a number of physiologic characteristics such as age
(Garn et al. 1981a, Yip et al. 1984), sex (Garn et al. 1981a)
and stage of pregnancy (WHO 1994) influence hemoglobin
concentration; thus, an appropriate anemia cut-offthat takes
into account the normal variations is indicated. There are
some environmental factors that also influence hemoglobin
distribution such as changes in altitude (Miale 1982) and
smoking habits (Nordenberg et al. 1990, Stonesifer 1978).
Vitamin A deficiency (Bloem 1995) and inflammation (Farid
et al. 1969) also influence the hemoglobin concentration. In
addition, several investigators (Garn et al. 1981b, Jackson et
al. 1983, Johnson-Spear and Yip 1994, Perry et al. 1993,
Williams 1981 and Yip 1996) found that hemoglobin distri-
bution varies among races or ethnic backgrounds. The general
world-wide application of the common cut-off for anemia may

0022-3166/99 $3.00 1999 American Society for Nutritional Sciences.

Manuscript received 26 October 1998. Initial review completed 1 December 1998. Revision accepted 27 April 1999.

1669
1670 KHUSUN ET AL.
be questioned. An analysis of data from the National Health
and Nutrition Survey (NHANES)4 II by Johnson-Spear and
Yip (1994) showed that individuals of African extraction in
the U.S. have hemoglobin concentrations that are on average
8 g/L lower than those of European extraction, with the
difference not due to iron nutriture. To have a similar screen-
ing performance for iron deficiency in terms of sensitivity and
specificity, the hemoglobin cut-off point for those of predom-
inantly African extraction is 10 g/L lower than for those of
European extraction. A survey report in Vietnam showed that
the healthy Vietnamese population had mean hemoglobin
values 10 g/L lower than the mean Hb of the Caucasian
population, which resulted in a 10 g/L reduction of cut-off
values (Yip 1996).
The correct interpretation of hemoglobin values requires
the application of appropriate cut-offs and knowledge of the
influencing factors. Application of a single inappropriate cut-
off will result in misclassification and exaggeration or under-
estimation of the iron-deficiency problem in a community.
More information is therefore required on the validity of the
use of hemoglobin cut-off values as a screening for iron defi-
ciency because the frequently used WHO cut-off may not be
universal.
Iron deficiency is common in Indonesia, and it is important
to estimate the problem adequately. It was the aim of the study
to examine whether hemoglobin distribution of healthy young
Indonesians was similar to that of an American population and
whether population-specific hemoglobin cut-off values for de-
tection of iron deficiency were required. This study might
serve as model or basis for further studies of this issue.
SUBJECTS AND METHODS
The subjects were selected among male and female students of the
University of Indonesia, Jakarta, using a nonprobability sampling
procedure (convenience sampling). The potential subjects were re-
cruited by distributing a written announcement about the research.
Subjects were volunteers. Potential subjects were first questioned with
the use of a precoded questionnaire. A total of 210 male and 200
female students were interviewed. Information was collected about
sociodemographic background (ethnicity, level of education of par-
ents and possession of luxury items), physiologic condition (age, sex,
menstruation, pregnancy or lactation), health status and life style
(presence of disease, use of drugs and supplements, smoking habit, use
of contraceptives). After establishing that the potential subjects did
not suffer from any obvious illnesses as indicated by the question-
naire, anthropometric measurements were made and blood was col-
lected. The data collection took place for ;3 wk. A complete data set
became available for 203 male and 170 female students. Weight was
measured to the nearest 0.1 kg using an electronic weighing scale
(SECA 770), and height was measured to the nearest 0.1 cm using a
microtoise.
The ethical committee of the Faculty of Medicine, University of
Indonesia approved the conduct of this study.
Blood samples were drawn by venipuncture into two different
vacutainers between 0800 and 1300 h. Blood (;10 mL) was drawn
into a vacutainer tube with EDTA for determination of hemoglobin
(Hb), hematocrit (Ht), mean corpuscular volume (MCV), mean
corpuscular hemoglobin (MCH), mean corpuscular hemoglobin con-
centration (MCHC), red blood cell count (RBC), white blood cell
count (WBC), erythrocyte sedimentation rate (ESR) and zinc pro-
toporphyrin (ZP). The tubes with EDTA-treated blood were stored in
4
Abbreviations used: ESR, erythrocyte sedimentation rate; Hb, hemoglobin;
Ht, hematocrit; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular
hemoglobin concentration; MCV, mean corpuscular volume; NHANES, National
Health and Nutrition Survey; RBC, red blood cell count; SF, serum ferritin; SI,
serum iron; TIBC, total iron-binding capacity; WBC, white blood cell count; ZP,
zinc protoporphyrin.
a cool box and analyzed within 4 h of collection. Blood (;4 mL) was
drawn into a plain vacutainer tube for determination of serum iron
(SI), total iron-binding capacity (TIBC) and serum ferritin (SF).
Blood was allowed to clot at room temperature (25C) and was
centrifuged at 3000 3 g for 15 min. Each serum sample was divided
into two tubes and stored at 220C for mo 1 and then at 280C for
mo 2. Serum ferritin determination was done within 1 mo of blood
collection, and SI and TIBC were measured within 12 mo.
Hb, Ht, WBC, RBC, MCV, MCH and MCHC were determined
using a Coulter counter (Coultert AC-T10 Hematology Analyzer;
Coulter Electronic, Miami, FL). ESR was analyzed by the Wester-
green method (Widmann 1983). Serum ferritin was determined with
the use of a microparticle enzyme immunoassay procedure with a
commercial kit (IMX Ferritin Assay, Abbott, Abbott Park, IL).
Serum iron and TIBC were determined by a colorimetric procedure
(Gibson 1990) using a commercial kit (Hoffman-la Roche, Basel,
Switzerland). All of the above assays were done once. Zinc protopor-
phyrin was measured fluorometrically in duplicate in red blood cells
(Hematofluorometer model 206D, AVIV Biomedical, Lakewood,
NJ), which were obtained by centrifuging the EDTA-treated blood
samples (Hastka et al. 1992). The Coulter counter and SI/TIBC
results were analyzed at the Clinical Pathology Department, Cipto
Mangunkusumo Hospital, Faculty of Medicine, University of Indo-
nesia; the other measurements were done at the SEAMEO-
Downloaded from jn.nutrition.org by guest on July 28, 2016
TROPMED Center.
Choice of cutoff points for abnormal values of iron status indi-
cators and ESR. Three tests were used to assess the iron status of the
subjects. The respective criteria for each test,indicating low iron
status, were as follows: serum ferritin , 12 mg/L (Dallman et al.
1996), transferrin saturation , 16% (Dallman et al. 1996) and zinc
protoporphyrin . 40 mmol/mol heme (Hastka et al. 1992). A subject
was considered to be iron deficient when at least two of the three test
values were beyond the cut-off value, indicating deficiency (Dallman
et al. 1996). For hemoglobin, the cut-off criterion indicating anemia
was the WHO cut-off of 120 g/L for females and 130 g/L for males
(WHO 1994). Hematocrit was considered to be abnormal at values
, 0.36 for females and , 0.41 for males (Gibson 1993). RBC for
females was considered normal in the range of 4200 5800/mm3 and
for males, 3600 5600/mm3 (Gibson 1993). The cut-off values for the
red blood cell indices were as follows: MCV , 80 fL, MCH , 27 pg
and MCHC , 320 g/L (Gibson 1993). For serum iron (SI) and total
iron-binding capacity (TIBC) the cut-off points were 60 mg/dL (10.74
mmol/L) and 410 mg/dL (73.39 mmol/L), respectively (Cook and
Finch 1979).
ESR and WBC were used as indicators of the presence of a
possible infection because the NHANES II survey, in which percen-
tile values were used for comparison, also used ESR and WBC as
indicators of inflammation (Expert Scientific Working Group 1985).
ESR was considered to be abnormal at .15 mm/h for males and .20
mm/h for females (Widmann 1983), whereas WBC values ,3400/
mm3 or .11500/mm3 were judged to be abnormal (Expert Scientific
Working Group 1985). To conserve sample size, the hemoglobin
concentration of smokers was adjusted downward according to the
number of cigarettes smoked per day (Centers for Disease Control
1989).
Statistical analysis. ANOVA and the Kruskall-Wallis test were
used to detect differences in the characteristics of men and women
(Snedecor and Cochran 1980). Comparison of percentiles values and
analysis of means and confidence intervals were used to compare the
distribution of the present data set with that from the NHANES II
and III surveys (Dallman et al. 1996, Gibson 1993). For the hema-
tological and biochemical indices, normality was tested by the one-
sample Kolmogorov Smirnov test. WBC, ESR, RBC, MCV, MCH,
MCHC, serum ferritin and zinc protoporphyrin were not normally
distributed; thus medians were used as the measure of central ten-
dency. Because Hb, Ht, SI, TIBC and transferrin saturation were
normally distributed, means were used as the measure of central
tendency.
The performance (sensitivity and specificity) of different cut-off
criteria for anemia as a screening tool for iron deficiency was esti-
mated in female subjects. Sensitivity was defined as the proportion of
cases of iron deficiency correctly identified by Hb as anemic and
 DETECTION OF ANEMIA IN INDONESIAN POPULATION 1671

TABLE 1
Characteristics of subjects1

Characteristic n Men (194) Women (140) P-value

Age, y 21.6 6 1.9 22.0 6 1.8 NS2

Weight, kg 58.8 6 9.0 50.4 6 7.0 ,0.001
Height, cm 167.0 6 5.8 155.9 6 5.3 ,0.001
BMI, kg/m2 21.0 6 2.8 20.7 6 2.5 NS2
Parents ethnicity, % 0.004
Both from Java 64.9 51.1
Both from Sumatra 19.3 27.1
One from Java, other from Sumatra 7.9 9.1
Other 7.9 12.7
Education background of father, % NS
Elementary school 7.7 3.6
High school 41.7 39.3
University degree 40.3 55.0
Possession in household, %
Television 90.2 97.9 0.006
Phone 73.7 87.9 0.001
Video 36.1 60.9 ,0.001
Car 39.7 68.6 ,0.001
Smoking, % 19.6 2.1 ,0.001
Number of cigarettes/d, %

Downloaded from jn.nutrition.org by guest on July 28, 2016

19 80.0 3
1019 14.3
2039 5.7
Taking micronutrient supplement, % 29.4 35.7 NS
Frequency of taking supplement, %: NS
Weekly or more 37.5 48.0
Less than weekly 62.5 52.0
Menstruating at blood drawing, % 25.0
Usual period of menstruation, %
13 d 3.7
47 d 77.8
.8 d 10.7

1 Values are means 6 SD or %.

2 NS 5 Not significantly different, P . 0.05.
3 Female smokers refused to give information about the number of cigarettes they smoked per day.

specificity as the proportion of cases of iron adequacy correctly
identified by Hb as nonanemic.
RESULTS
Of the group of 373 subjects for whom data were collected,
6 men and 25 women had abnormal ESR values, and 4 men
and 4 women had abnormal WBC values. To avoid the con-
founding influence of a possible infection on the iron status
indicators and their relationship, these 39 subjects were ex-
cluded from the analysis. Selected characteristics of the re-
maining 334 subjects are presented in Table 1. The subjects
ranged in age from 18 to 27 y with a mean age of 21.6 y for
men and 22.0 y for women (Table 1). Fifteen percent of the
men and 18.6% of the women had a body mass index , 18.5
kg/m2. Ethnicity of the subjects was defined by the origin of
their parents, the majority of whom came from the island of
Java. A small percentage of the sample (7.9% males and 12.7%
females) had parents originating from parts of Indonesia other
than Java and Sumatra. This study therefore refers mainly to
the western part of Indonesia. There was no significant differ-
ence in mean Hb concentration among the different ethnic
groups. Considering the educational level of the father and
household possession of selected commodities, all subjects
belonged to the middle or high socioeconomic class (Table 1).
Hemoglobin (P , 0.001), hematocrit (P , 0.001), serum
iron (P , 0.001), serum ferritin (P , 0.001), transferrin
saturation (P , 0.001) and zinc protoporphyrin (P , 0.001)
were different in men and women. Men had a better iron
status than women because no male subject was considered
iron deficient, whereas 20.0% of the women were deficient and
14.3% had hemoglobin concentrations , 120 g/L (Table 2).
The prevalence of high zinc protoporphyrin values (.40
mmol/mol heme) was higher than the prevalence of low seum
ferritin or low transferrin saturation. When 50 mmol/mol
heme was used as a cut-off point, the prevalence of high values
among women became 28.8% in stead of 51.8%. Hemoglobin
concentrations and values of iron status indicators did not
always correlate (Table 3). Among anemic women, 40% were
iron deficient, whereas 16.7% of the nonanemic women were
iron deficient (using 40 mmol/mol heme zinc protoporphyrin
as a cut-off point).
Hemoglobin concentration of nondeficient subjects was
152 6 11 g/L for men and 131 6 9 g/L for women (Fig. 1). Of
the noniron-deficient subjects, 3.1% of men and 9.9% of
women had a hemoglobin concentration ,130 and 120 g/L,
respectively. The hemoglobin distribution of the Indonesian
subjects was compared with the distribution of the nonblack
population from the United States (Table 4). The mean
hemoglobin concentration of American men and women was
within the 95% confidence intervals for the mean of the
respective concentration of the Indonesian subjects. The
mean value for the American women was just below the upper
border of the confidence interval for the Indonesian mean.
The percentiles of the hemoglobin distribution were also com-
1672 KHUSUN ET AL.

TABLE 2
Hematological and biochemical characteristics of male and female students of the University of Indonesia1

Men (n 5 194) Women (n 5 140)

Central tendency Subjects at Central tendency Subjects at

Indicator and deviation abnormal value,2 % and deviation abnormal value,2 %

WBC,3 31023/mm3 7.1 (4.9;9.9) 0 7.2 (5.4;9.5) 0

ESR, mm/h 5.0 (2.0;10.3)* 0 10.0 (4.1;20.0) 0
Hemoglobin, g/L 152 6 11* 3.1 131 6 10 14.3
Hematocrit 0.45 6 0.03* 4.6 0.38 6 0.03 25.7
RBC, 31023/mm3 5.2 (4.7;6.1)* 21.1 4.4 (3.9;5.3) 24.3
Mean corpuscular volume, fL 85.7 (77.2;92.4) 8.2 85.6 (74.1;91.3) 15.7
Mean corpuscular hemoglobin, pg 29.9 (26.5;32.4)* 5.7 29.6 (25.0;32.0) 13.6
Mean corpuscular hemoglobin concentration, g/L 349 (337;359)* 1.0 345 (322;355) 3.6
Serum ferritin, mg/L 97.2 (31.0;228.1)* 0 36.7 (6.3;131.6) 12.1
Serum iron, mmol/L 17.8 6 3.34 3.1 14.0 6 4.3 25.7
TIBC, mmol/L 69.6 6 6.4 23.7 69.7 6 7.9 27.9
Transferrin saturation, % 25.8 6 5.1* 3.6 20.4 6 6.5 25.0
Zinc protoporphyrin, mmol/mol heme 32.0 (21.0;54.3)* 12.9 41.0 (27.0;83.0) 51.8
Iron deficiency, % 0 20.0

1 Values are means 6 SD or medians.

Downloaded from jn.nutrition.org by guest on July 28, 2016

2 Cut-offs were explained in Methods.
3 Abbreviations: WBC, white blood cell count; ESR, erythrocyte sedimentation rate; RBC, red blood cell count; TIBC, total iron-binding capacity.
* Significantly different from females, P , 0.05.

pared. The percentile values for men were largely similar in
the Indonesians and Americans. The percentile values of the
Indonesian women were 3 8 g/L lower than the values of
American women. Using the mean hemoglobin concentration
22 SD as a definition for anemia in the Indonesian subjects,
the cut-off point below which a person would be considered
anemic was 113 g/L for women and 130 g/L for men. The
cut-off point for men was similar to that defined by WHO,
whereas the cut-off point for the women was lower than the
WHO cut-off point of 120 g/L.
The sensitivity and specificity of different hemoglobin cut-
off points for anemia to detect iron deficiency were assessed in
the Indonesian women (Table 5). The calculated cut-off point
for the Indonesian women of 113 g/L had lower sensitivity but
a slightly higher specificity for detecting iron deficiency than
the WHO cut-off point for women of 120 g/L. The sensitivity
and specificty when 120 g/L was used to detect iron deficiency
were lower in the Indonesian subjects compared with the
American subjects. From the relatively low sensitivity
(,50%), it becomes clear that anemia is not related only to
iron deficiency as defined in this study. Among the anemic
women, only 40.0% were iron deficient, whereas in the non-
anemic subjects, 15.4% were iron deficient. Six men had
anemia without iron deficiency.
DISCUSSION
Anemia prevalence in this study population was much
lower than the estimated prevalence for the whole Indonesian
population. Among women of child bearing age, the preva-
lence of anemia in Indonesia was ;30 40% (Helen Keller
International 1997). On the basis of the WHO cut-offs points,
in this study, the prevalence of anemia was 15.9% for women,
whereas only 3.9% of the men were anemic. For men, the

TABLE 3
Prevalence of abnormality of several iron status indicators in male and female students of the University of Indonesia stratified by
hemoglobin concentration

Prevalence

Low serum High zinc Low transferrin Iron

Hemoglobin Level g/L ferritin1 protoporphyrin2 saturation3 deficiency4

Women
Hb , 120 (n 5 20) 25.0 (n 5 5) 80.0 (n 5 16) 45.0 (n 5 9) 40.0 (n 5 8)
Hb . 120 (n 5 120) 10.0 (n 5 12) 47.1 (n 5 56) 21.7 (n 5 26) 16.7 (n 5 20)
Men
Hb , 130 (n 5 6) 0 0 25.0 (n 5 2) 0
Hb . 130 (n 5 188) 0 13.3 (n 5 25) 2.7 (n 5 5) 0

1 Serum ferritin , 12 mg/L.

2 Zinc protoporphyrin . 40 mmol/mol heme.
3 Transferrin saturation , 16.
4 More than one iron biochemistry tests (serum ferritin, zinc protoporphyrin and transferrin saturation) was abnormal.
 DETECTION OF ANEMIA IN INDONESIAN POPULATION 1673

and transferrin saturation). It is very likely that there were

FIGURE 1 Hemoglobin distribution curves for healthy and iron-
sufficient male (n 5 194) and female (n 5 112) students of the University
of Indonesia, aged 18 27 y. Iron deficiency was defined as the occur-
rence of two or more abnormal values for serum ferritin, zinc protopor-
phyrin and transferrin saturation.
some women who had milder forms of iron deficiency but
whose values did not meet the study definition. This finding of
the positive predictive value of anemia for detecting iron
deficiency (;40%) is similar to the previously reported value
for American women (Johnson-Spear and Yip 1994). The
relatively low positive predictive value of anemia for detecting
iron deficiency suggests that anemia is not a perfect screening
tool for iron deficiency, especially when the anemia is mild.
The remaining 60% includes subjects with mild iron defi-
ciency or other conditions that did not meet the study criteria
such as mild hereditary anemia, normal variations and mild
infections not excluded on the basis of the ESR criteria, or
vitamin A and folate deficiency. Furthermore, in a perfectly
healthy population, 2.55% of the people would be anemic by
definition.
The mean hemoglobin concentration of American men
was within the 95% confidence interval of the mean hemo-
globin for Indonesian men, indicating similarities in mean
values. The comparison of percentile values also suggested that
there is no difference in mean hemoglobin concentrations of
healthy Indonesian and American men. Among women, the

Downloaded from jn.nutrition.org by guest on July 28, 2016

observed prevalence was near the expected level of 2.55% for
an iron-replete population, based on the definition of WHO
anemia criteria. This was confirmed by the other iron bio-
chemistry tests. The lower prevalence of anemia for the
women studied compared with the overall prevalence in In-
donesia is not surprising considering the fact that subjects were
university students and most of them came from a relatively
high socioeconomic background. It was also to be expected
that women would have a higher deficiency prevalence than
men for all iron-related tests. This can probably be attributed
to the 50% higher iron requirement of women than men due
to monthly menstrual blood losses (Cheong et al. 1991, Hall-
berg et al. 1995) and a lower energy and iron intake from food.
This study also confirms a common finding that, in many parts
of the world, even when substantial anemia and iron defi-
ciency are prevalent among women, men do not suffer from
iron deficiency because of their lower iron requirement. This
male and female differential in iron deficiency indicates that
dietary iron intake is the main factor responsible for the
anemia observed among women (Yip, 1994).
Among the anemic women, 40.0% were iron deficient on
the basis of a strict criterion of having abnormal results for two
or more of the three tests (serum ferritin, zinc protoporphyryn
mean hemoglobin concentration of the Americans was exactly
at the higher border of the 95% confidence interval of the
Indonesian population. The percentile values of the Indone-
sian population were lower than those of the American pop-
ulation, suggesting that there is difference between the two
groups.
Because there was substantial iron deficiency among the
women studied, it is not certain whether the recommended
criterion for iron deficiency fully excluded most of those with
some degree of iron deficiency. Therefore, the postexclusion
hemoglobin distribution may not be a truly iron-replete sam-
ple. For this reason, it would be more accurate to use the male
subsample of the study to contrast with the iron-replete sample
from the United States. In doing so, we found the two distri-
butions nearly identical. This finding strongly suggests that it
would be appropriate to use the common anemia criterion
recommended for those of European extraction for Indone-
sians also.
For the purpose of identifying the proportion of individuals
at risk of iron deficiency for possible intervention, a higher
cut-off value with greater sensitivity is generally desirable
(Himes et al. 1997). Using different hemoglobin cut-off points
for the assessment of iron deficiency showed that, compared
with the WHO cut-off points (120 g/L), the population-

TABLE 4
Comparison of mean and percentiles of hemoglobin concentration of the Indonesian and American populations

Percentile

5 10 25 50 75 90 95 Mean 6 SD 95% CI Mean 2 2 SD

g/L

Men
Indonesian (n 5 194) 136 141 146 152 158 165 168 152 6 11 151154 130
American1 137 140 146 153 159 165 168 152 6 9 133
Women
Indonesian (n 5 112) 116 118 126 131 136 142 145 131 6 9 129133 113
American1 120 123 129 135 142 148 151 133 6 9 115

1 For percentiles from NHANES II population, Source: Pilch and Senti (1984) as copied by Gibson (1993); for mean 6 SD, 95% CI and Mean 2 2
SD from Dallman et al. (1996), derived from NHANES III data.
1674 KHUSUN ET AL.
TABLE 5
Comparison of sensitivity and specificity of cut-off criteria for
anemia of women for the screening of iron deficiency1
Sensitivity2 Specificity3
% deficiency in infants and children. In: Iron Nutrition in Health and Disease
Indonesian (n 5 170)
Hb , 113 g/L 13.2 96.2
Hb , 116 g/L 26.3 96.2
Hb , 120 g/L 34.2 89.4
White American4
Hb , 120 g/L 35.7 94.6
1 Iron deficiency is defined as having two or more abnormal values
for iron-related tests (serum ferritin, zinc protoporphyrin and transferrin
saturation).
2 Percentage of subjects with hemoglobin (Hb) below the cut-off
point who had iron deficiency.
3 Percentage of subjects with Hb above the cut-off points who did
not have iron deficiency.
4 From Johnson-Spear and Yip (1994).
specific anemia criterion of 113 g/L for Indonesian women, the
mean 22 SD, had a very low sensitivity for detecting iron
deficiency. Only when the cut-off approached that of the
WHO criterion did the test performance become similar to
that for American women, which is based on the NHANES II
survey (Johnson-Spear and Yip, 1994). Because the prevalence
of high zinc protoporphyrin was not similar to the prevalence
values of the other iron status indicators, we also considered
using 50 mmol/mol heme as a cut-off point instead of 40
mmol/mol heme. This higher cut-off point resulted in a lower
estimated percentage of iron-deficient women. However, when
this elevated cut-off was also used for zinc protoporphyrin, the
sensitivity using 120 g/L as a borderline value for hemoglobin
concentration to detect iron deficiency was higher (40.6%)
than when using 116 g/L (31.3%) or 113 g/L (15.6%). For the
Indonesian population studied, the application of the WHO
anemia criterion will yield results for defining the extent of
iron-deficiency problems comparable to those in populations
of mainly European extraction.
This finding is similar to that of Charoenlarp and Polpothi
(1987) in Thailand. They investigated the distribution of
hemoglobin concentration in healthy Thai children and found
that, after excluding those having abnormal hemoglobin types,
the hemoglobin distribution was the same as that of the U.S.
population.
This result suggests that there is no need to define separate
cut-off criteria for anemia in the Indonesian population stud-
ied, most of whom originated from the western part of Indo-
nesia.
LITERATURE CITED
Bloem, M. W. (1995) Interdependence of vitamin A and iron: an important
association for programmes of anemia control. Proc. Nutr. Soc. 54: 501508.
Centers for Disease Control (1989) CDC criteria for anemia in children and
childbearing age women. Morb. Mortal. Wkly. Rep. 38 : 400 403.
Charoenlarp, P. & Polpothi, T. (1987) The distribution of haemoglobin concen-
tration in healthy Thai children. Southeast Asian J. Trop. Med. Public Health
18: 567568.
Cheong, R. L., Kuizon, M. D. & Tajaon, R. T. (1991) Menstrual blood loss and
iron nutrition in Filipino women. Southeast Asian J. Trop. Med. Public Health.
22: 595 604.
Cook, J. D. & Finch, C. A. (1979) Assessing iron status of a population. Am. J.
Clin. Nutr. 32: 21152119.
Dallman, P. R., Looker, A. C. Johnson, C. L. & Carrol, M. (1996) Influence of
age on laboratory criteria for the diagnosis of iron deficiency anemia and iron
(Hallberg, L. & Asp, N.-G., eds.), pp. 6574. John Libbey & Company, London,
UK.
Expert Scientific Working Group (1985) Summary of a report on assessment of
the iron nutritional status of the United States population. Am. J. Clin. Nutr.
42: 1318 1330.
FAO/WHO (1992) Preventing micronutrient deficiencies. ICN: Fact Sheet
Number One. Supporting Paper of the International Conference on Nutrition,
December 1992, Rome, Italy.
Farid, Z., Patwardhan, V. N. & Darby, W. J. (1969) Parasitism and anemia.
Am. J. Clin. Nutr. 5: 498 503.
Freire, W. B. (1989) Hemoglobin as a predictor of response to iron therapy and
its use in screening and prevalence estimates. Am. J. Clin. Nutr. 50: 1442
1449.
Garn, S. M., Ryan, A. S., Abraham, S. & Owen, G. (1981) Suggested sex and
age appropriate values for low and deficient hemoglobin levels. Am. J.
Clin. Nutr. 34: 1648 1651.
Garn, S. M., Ryan, A. S., Owen, G. M. & Abraham, S. (1981) Income-matched
black-white hemoglobin differences after correction for low transferrin satu-
ration. Am. J. Clin. Nutr. 34: 16451647.
Downloaded from jn.nutrition.org by guest on July 28, 2016
Gibson, R. (1990) Principles of Nutritional Assessment. Oxford University
Press, New York, NY.
Gibson, R. (1993) Nutritional Assessment: A Laboratory Manual. Oxford Uni-
versity Press, New York, NY.
Hallberg, L., Hulthen, L., Bengston, C., Lapidus, L. & Lindstedt, G. (1995) Iron
balance in menstruating women. Eur. J. Clin. Nutr. 49: 200 207.
Hastka, J., Lassere, J. J., Schwarzbeck, A., Strauch, M. & Hehlmann, R. (1992)
Washing erythrocytes to remove interferents in measurements of zinc proto-
porphyrin by front-face hematofluorometry. Clin. Chem. 38: 2184 2189.
Helen Keller International (1997) Iron Deficiency Anemia in Indonesia. Report
on Policy Workshop, April 12, 1997. Helen Keller International, Jakarta,
Indonesia.
Himes, J. H., Walker, S. P., Williams, S., Bennet, F. & Grantham-McGregor, S. M.
(1997) A method to estimate prevalence of iron deficiency and iron defi-
ciency anemia in adolescent Jamaican girls. Am. J. Clin. Nutr. 65: 831 836.
Jackson, R. T., Sauberlich, H. E., Skala, J. H., Kretsch, M. J. & Nelson, R. A.
(1983) Comparison of hemoglobin values in black and white male US
military personnel. J. Nutr. 113: 165171.
Johnson-Spear, M. A. & Yip, R. (1994) Hemoglobin difference between black
and white women with comparable iron status: justification for race-specific
anemia criteria. Am. J. Clin. Nutr. 60: 117121.
Miale, J. B. (1982) Laboratory Medicine Hematology. The C. V. Mosby Com-
pany, St. Louis, MO.
Nordenberg, D., Yip, Y. & Binkin, N. (1990) The effect of cigarette smoking on
hemoglobin levels and anemia screening. J. Am. Med. Assoc. 264: 1556
1559.
Perry, G. S., Byers, T., Yip, R., & Margens, S. (1993) Iron nutrition does not
account for the hemoglobin differences between black and whites. J. Nutr.
123: 597599.
Snedecor, G. W. & Cochran, W. G. (1980) Statistical Methods. The Iowa State
University Press, Ames, IA.
Stonesifer, L. D. (1978) How carbon monoxide reduces plasma volume.
N. Engl. J. Med. 299: 311312.
Widmann, F. K. (1983) Clinical Interpretation of Laboratory Tests. F. A. Davis
Company, Philadelphia, PA.
Williams, D. M. (1981) Racial differences of hemoglobin concentration: mea-
surements of iron, copper, and zinc. Am. J. Clin. Nutr. 34: 1694 1700.
World Health Organization (1994) Indicators and Strategies for Iron Deficiency
and Anemia Programmes. Report of the WHO/UNICEF/UNU Consultation.
Geneva, Switzerland, 6 10 December, 1993.
Yip, R. (1994) Iron deficiency: contemporary scientific issues and international
programmatic approaches. J. Nutr. 124: 1479S1490S.
Yip, R. (1996) A Recommended Plan of Action for the Control of Iron Defi-
ciency for Vietnam. Final Report of the 1995 Vietnam National Nutrition
Anemia and Intestinal Helminth Survey, October 1, 1996.
Yip, R., Johnson, C. & Dallman, P. R. (1984) Age-related changes in laboratory
values used in the diagnosis of anemia and iron deficiency. Am. J. Clin. Nutr.
39: 427 436.