![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
*
Department of International Health, The Johns Hopkins School of Hygiene & Public Health, Baltimore, MD 21205,
Ivo de Carneri Foundation, Milan, Italy,
Schistosomiasis and Intestinal Parasites Unit, Division of Control of Tropical Diseases, World Health Organization, Geneva 27, Switzerland,
§
Nepal Netra Jyoti Sangh, Kathmandu, Nepal, and
¶
Ministry of Health, Zanzibar, United Republic of Tanzania
3To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
KEY WORDS: • humans • Zanzibar • Nepal • anemia • hemoglobin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Anemia is
usually defined as hemoglobin concentration below normative cut-off
values that are based on age, sex, physiologic state and altitude
(WHO, UNICEF and UNU 1998
). Anemia has many causes,
including nutrient deficiencies, chronic blood loss, disease states
that cause hemolysis or suppress erythropoiesis, toxic exposures and
hemoglobinopathies. Although iron deficiency is believed to cause the
largest part of anemia globally (WHO, UNICEF and UNU
1998
), current international recommendations for prevention and
treatment of anemia include the use of anthelminthic and antimalarial
drugs and folic acid, in addition to supplemental iron
(Stoltzfus and Dreyfuss 1998
).
Severe anemia is variously defined as hemoglobin concentration below
cut-off values of 50–80 g/L, but the most widely accepted
definition is <70 g/L (WHO, UNICEF and UNU 1998
). In
population groups where the majority of people are anemic, the
prevalence of severe anemia is highly variable, for example, as low as
<1% in pregnant Chinese women (Stoltzfus et al. 1997
)
or as high as 39% in pregnant Indian women (Gujral et al. 1989
). Severe anemia is particularly common among pregnant
women and young children in south Asia and sub-Saharan Africa.
The health risks of severe anemia are profound. Severely anemic Kenyan
children admitted to hospital were 2.25 times more likely to die than
were children without severe anemia. And among children with severe
anemia, children who were not transfused were three times more likely
to die than children transfused and with similar diseases and age
(Lackritz et al. 1992
). Among British adult surgical
patients who refused transfusion, severely anemic patients were 26
times more likely to die than were those without anemia (Carson et al. 1996
). Severe anemia in pregnancy was strongly
associated with perinatal death and low birth weight in Malaysian women
(Llewellyn-Jones 1965
) and increased the risk of
maternal death by a factor of around 5 in both Malaysia
(Llewellyn-Jones 1965
) and Nigeria (Harrison and Rossiter 1985
). Thus, severe anemia combined with stress of
illness (especially respiratory illness), surgery, or childbirth
frequently leads to death. In the absence of these concurrent stresses,
physical work capacity is dramatically reduced in severely anemic
individuals (Gardner et al. 1977
, Viteri and Torun 1974
).
Because of the excessive morbidity burden and mortality risk that
accompanies severe anemia, international recommendations advocate for
the detection and treatment of severe anemia in primary care settings
where the prevalence in populations groups such as pregnant women
exceeds 2% (Stoltzfus and Dreyfuss 1998
), and as part
of management of the sick child (WHO and UNICEF 1995
).
However, in many primary care settings, hemoglobin or hematocrit cannot
be determined on a routine basis, even in high-risk groups such as
sick children or pregnant women. Clinical pallor in anatomical sites
where capillary beds are visible through the skin or mucosa is one
potential method for detecting severe anemia in public health practice.
Because of its very low cost and feasibility, the World Health
Organization has included evaluation of palmar pallor to detect severe
anemia in its algorithm for management of the sick child (WHO
and UNICEF 1995
). However there has been a need to more
carefully evaluate its utility in the primary care settings where it is
most needed, and by nonphysician health workers (Kalter et al. 1997
, Zucker et al. 1997
).
We measured hemoglobin and assessed clinical pallor by similar protocols in two populations where severe anemia is common, the plains of southern Nepal and the islands of Zanzibar. In both settings, assessments were carried out by local nonphysician health workers. The five studies we report include young children, older children, and pregnant and nonpregnant women. We address the following questions: When implemented by primary health care workers in field settings, is there a strong association between clinical pallor and hemoglobin concentration? What is the sensitivity and specificity of clinical pallor to detect severe anemia in individuals? And finally, is this performance good enough to be useful in programmatic settings?
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Data were collected from five different population samples. In Nepal,
pregnant and postpartum women were assessed in the context of a trial
of nutritional interventions to women of reproductive age in the plains
district of Sarlahi (West et al. 1999
). Malnutrition is
common in this area, particularly protein-energy malnutrition and
vitamin A and iron deficiencies (Dreyfuss 1998
).
Ancylostoma duodenale (hookworm) and Plasmodium
vivax malaria infections are endemic in the region. All women in
30 village development communities (subdistricts) were invited to
participate in the intervention study. A clinical substudy was
conducted in three of those subdistricts. In the substudy area, women
who became pregnant in the course of the trial were invited to come to
a clinic for a health and nutrition assessment. The present analysis
includes women who identified themselves as pregnant between August
1994 and July 1996. Of 1,521 pregnant women eligible for the clinical
examination during this time period, 1,062 (69.5%) came to the clinic
and 1,006 (94.7%) were confirmed to be pregnant. Some women (28)
visited the clinic for two different pregnancies. Only data from their
first pregnancies are included in these analyses. Of the remaining
women, 945 had complete data for hemoglobin concentration and clinical
pallor. Most women (64%) were examined in their second trimester of
pregnancy, but the sample also included women in their first trimester
(17%) and third trimester (19%). These women form the Nepal—pregnant
women sample in this paper. All women who had a live birth were invited
to visit the clinic at 3 mo postpartum. In the above time period, 735
women completed the postpartum examination, and 720 had complete data
on hemoglobin concentration and clinical pallor. These women form the
Nepal—postpartum women sample in this paper.
In Zanzibar, studies were carried out on Pemba Island, the smaller of
the two islands that comprise modern-day Zanzibar. These islands
lie just below the equator, near the east coast of Africa, and are part
of the United Republic of Tanzania. Plasmodium falciparum
malaria, both common species of hookworms, and Schistosoma
haematobium are highly endemic (Stoltzfus et al. 1997
). The school children in the present analysis were
evaluated in 1995, as a part of an ongoing evaluation of the
government's school-based deworming program. Although data were
collected in several years, 1995 is the only year in which clinical
pallor was assessed as part of the evaluation. Primary schools (12) on
Pemba were randomly selected, and 3,605 children in grades 1–4 were
included in the baseline survey conducted in 1994. In 1995, 3,316 of
those children were reevaluated, 3,302 (99.6%) of whom had complete
data for hemoglobin concentration and clinical pallor.
A sample of preschool children was selected based on a census of Kengeja municipality on Pemba Island. All censused children aged 6–59 mo were invited to participate in a 12-mo trial of anemia prevention strategies. A few children aged 4–5 mo or 60–71 mo came to the clinic and were also examined and included in this analysis. A baseline survey was carried out on 614 children in September 1996, of whom 613 had complete data for hemoglobin and clinical pallor. These children form the Zanzibar—preschool 96 sample. Of these children, 538 participated in a follow-up clinic in September 1997. The Zanzibar—preschool 97 sample is the 537 children who had complete data for hemoglobin and clinical pallor.
Data collection.
The protocol for measuring hemoglobin and clinical pallor was the same in all studies. Blood was collected by antecubital venipuncture into vacutainers. One drop of blood was immediately used to fill a cuvette for determination of hemoglobin using the HemoCue method (HemoCue AB, Angelhom, Sweden). The accuracy of the HemoCue machines was checked daily using control cuvettes provided with the machines. All reported values are from machines that met the quality control standards recommended by HemoCue.
Clinical pallor was assessed by local public health practitioners in both settings. In Nepal, the examiner was an ophthalmic assistant who provides primary eye care to the study area. He was experienced in conducting clinical examinations, especially involving the eye. In Zanzibar, the examiners were two staff of the helminth control team of the Ministry of Health. Their normal responsibilities included the implementation and evaluation of helminth control activities. The same two people were responsible for assessing clinical signs in both the Zanzibar preschool and schoolchildren studies.
Training protocols were similar to what might be carried out if screening for severe anemia was incorporated into primary care activities in these two settings. Screening for severe anemia was not an ongoing activity in either study site prior to these studies. Training was carried out for 1–2 d, during which time observations of clinical pallor were compared to hemoglobin concentrations in selected individuals at high risk for severe anemia. In Nepal, this was an antenatal clinic, in which around 40 individuals were examined during the training period. In Zanzibar, training was conducted in a local hospital. Around 15 individuals were examined, of whom around half were severely anemic. Pallor was assessed in three sites: the inferior conjunctiva, the palm (especially the fleshy part at the base of the thumb), and the nail beds. In the case of the palm and nail bed, the examiner manually pressed and released the site, to observe the appearance of reddish color as the pressure was released. During the training period, pallor was assessed by the local staff person(s) and three study investigators. Only one investigator, RJS, was present during the training sessions in both Nepal and Zanzibar. These four individuals first assessed pallor independently from one another and then shared their observations with the patient still present so that they could reexamine as needed until they came to a consensus. This consensus assessment of pallor was then compared with the hemoglobin determination so that during the training period the clinical judgment of pallor could be adjusted in light of the gold standard. The health workers in Zanzibar were trained in 1995 and retrained once in 1996. The health workers in Nepal were trained in 1994 and were not retrained during the 2-y period of data collection.
During the period of data collection, clinical pallor was assessed independently and prior to the hemoglobin determination in a room removed from the phlebotomist. Data were recorded in an identical way in all studies. For each anatomical site, the observation was coded as "normal," "a little pale," "very pale," or "unreadable." The latter designation was used mainly for the conjunctiva if the eye was inflamed, or for the nail beds if the nails were painted.
All studies were reviewed and approved by the Committee on Human Research of The Johns Hopkins School of Hygiene and Public Health. The Nepal studies were also approved by the Nepal Health Research Council, and the Zanzibar studies received ethical approval from the Ministry of Health of Zanzibar and the World Health Organization.
Data analysis.
Data on clinical pallor were recorded to combine the categories "a little pale" and "very pale," because the number of observations coded "very pale" was very small in all studies. Thus, any degree of paleness is considered pallor in the following tables. Unreadable observations were recoded as missing. A new variable, pallor at any site, was coded as present if either conjunctiva, palm or nail beds was pale, and normal otherwise.
Anemia was defined as hemoglobin <110 g/L in Nepalese pregnant women,
and <120 g/L in Nepalese postpartum women (WHO, UNICEF & UNU
1998
). No adjustment was made for altitude as the study site is
near sea level. Anemia was defined as <110 g/L in school children and
<100 g/L in preschool children in Zanzibar. These cutoffs for Zanzibar
are below those recommended by the World Health Organization, because a
cutoff around 10 g/L lower was recommended for black populations
(Himes et al. 1997
, Johnson-Spear and Yip 1994
). Severe anemia was considered to be hemoglobin <70 g/L
in all groups studied.
The hemoglobin concentration in individuals with and without clinical
pallor was compared by Student's t-test. Sensitivity and
specificity of pallor were calculated using hemoglobin below a
specified cutoff as "true anemia." Thus, sensitivity was the
proportion of truly anemic individuals who were found to have clinical
pallor. Specificity was the proportion of truly nonanemic individuals
who were found not to have clinical pallor (Lilienfeld and Lilienfeld 1980
). Confidence limits for these proportions were
calculated with continuity correction (Snedecor and Cochran 1980
).
Positive predictive value was calculated as the proportion of persons with clinical pallor who were truly anemic. Positive predictive value depends on both the sensitivity and specificity of the test (i.e., pallor) and the prevalence of disease (i.e., anemia) within the population, and directly relates to the cost efficiency of the screening tool. To illustrate this, we present the inverse of positive predictive value, which may be regarded as the multiplier per unit cost spent to treat one true case of anemia due to the inefficiency of the test. For example, if the test has a positive predictive value of 0.30, then the cost multiplier due to the inefficiency of the test would be 3.33. This means that if the health care cost to treat one case of anemia is U.S. $1.00, U.S. $3.33 would be spent for each truly anemic person treated, because many nonanemic people will also receive treatment. Data were analyzed using Systat (SPSS Inc., Chicago, IL).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Hemoglobin concentrations were highest in the Zanzibari schoolchildren,
in whom the prevalence of anemia was 32.3% (Table 1
).Hemoglobin concentrations were lower in the Nepalese women, and
Zanzibar—preschoolers 97 sample. Anemia was very prevalent and severe
in the Zanzibar—preschoolers 96 sample, in whom the mean ± SD hemoglobin concentration was only 87 ± (16) g/L.
|
. The prevalence of pallor at any site was highest in the
Zanzibar—preschoolers 97 sample, and lowest in the Zanzibar—schoolers
sample.
In all five groups, pallor of the conjunctiva, palm, nail beds or at
any site was associated with a significantly lower hemoglobin
concentration (Table 2
).In each study, the mean hemoglobin concentration of individuals without
pallor at any site was near the sample mean hemoglobin in Table 1
. The
mean hemoglobin concentrations of individuals with pallor at any site
were 10–24 g/L lower than those without pallor.
|
The sensitivity of pallor to detect low-hemoglobin concentration in
individuals was low at higher cutoffs and increased greatly at lower
hemoglobin cutoffs (Table 3
).Specificity was 
82% at all hemoglobin cutoffs and in all studies.
The sensitivity of pallor to detect severe anemia (hemoglobin <70 g/L)
was highest in Nepalese postpartum women (81%, 95% confidence limits:
62–100%), intermediate in Nepalese pregnant women (63%, 46–79%),
Zanzibari schoolchildren (65%, 44–87%) and Zanzibar—preschoolers 97
(61%, 37–81%), and lowest in Zanzibari—preschoolers 96 (29%,
19–39%).
|
).An exception was that nail beds performed poorly in Zanzibari
preschoolers. The conjunctiva had the highest sensitivity in Nepalese
pregnant and postpartum women and in one study of Zanzibari
preschoolers.
|
The performance of a screening system based on clinical pallor to
detect severe anemia in the five population groups is summarized in
Table 5.
The positive predictive value of clinical pallor was highest in the
Zanzibar—preschoolers 96 sample, despite its low sensitivity, because
of the higher prevalence of severe anemia in this group. Thus, the
number of cases of severe anemia detected and treated, and also the
economic efficiency of the screening system, was highest in this
sample.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In all five study samples and in all three anatomical sites examined,
clinical pallor was strongly associated with hemoglobin concentration.
This has been a consistent finding in all of the published studies of
the clinical assessment of anemia that we were able to identify
(Gjorup et al. 1986
, Jacobs et al. 1979
,
Kalter et al. 1997
, Luby et al. 1995
,
Nardone et al. 1990
, Strobach et al. 1988
, Thaver and Baig 1994
, Wurapa et al. 1986
, Zucker et al. 1997
), except for one
(Gujral et al. 1989
). In the latter study, only tongue
and lips were assessed for pallor, and the clinical finding was not
associated with hemoglobin <80 g/L in pregnant Indian women.
Our findings confirm that pallor is useful to detect severe anemia, but
is insensitive to detect mild anemia. At descending hemoglobin cutoffs,
sensitivity of clinical pallor increased greatly while specificity
decreased only slightly. The sensitivity of clinical pallor to detect
mild anemia (defined by age and population-specific criteria in
Table 2
) was 
23% in all groups. Hemoglobin <70 g/L is the most
commonly used definition for severe anemia (WHO, UNICEF and UNU
1998
, Stoltzfus and Dreyfuss 1998
). At this
cutoff, the sensitivity was 61% in all groups except the
Zanzibar—preschoolers 96 sample. The specificity at this cutoff
remained relatively high, 84% in all groups.
The performance of clinical pallor to detect severe anemia in the Zanzibar—preschoolers 96 sample was significantly lower than in the other four samples. This is not due to a difference in the skills of the examiners, because the examiners were the same for the Zanzibari schoolers (assessed in 1995) and for the preschoolers reassessed in 1997. We hypothesize that the performance is affected by the range of hemoglobin concentrations in the sample examined. The average hemoglobin concentration in the preschoolers in 1996 was exceptionally low, only 87 g/L. Only 5.5% of children (34 children) had hemoglobin concentrations >110 g/L, a level that might be considered to provide a normal reddishness of skin and mucosa. This is a lower proportion of normals than in any other published study that we reviewed. Assuming that degree of pallor is a continuous phenomenon directly related to hemoglobin concentration, the examiners of these children were comparing between degrees of abnormality rather than between normals and abnormals. It is possible that examiners need to be regularly reminded what normal mucosa look like, to be able to distinguish pallor.
In support of this hypothesis, the relationship of pallor sensitivity
to hemoglobin level in the preschoolers in 1996 was very similar to
that seen in the schoolers but shifted down 20 g/L in hemoglobin cutoff
(Table 3)
. Pallor had similar sensitivity and specificity to detect a
hemoglobin level <50 g/L in the preschoolers and <70 g/L in the
schoolers—cut-offs values 37 and 47 g/L below the respective
population means.
Given that a true relationship exists between pallor and hemoglobin, is
pallor a practically useful assessment? Our experience and that of
others (Kalter et al. 1997
, Luby et al. 1995
, Nardone et al. 1990
, Shah et al. 1984
, Wurapa et al. 1986
, Zucker et al. 1997
) demonstrate that the method can be learned with a
training protocol in 1–2 d. Kalter et al. (1997)
found
that pallor performed reasonably well to detect severe anemia in young
children when implemented by Ugandan medical officers who received only
written instructions, with no supervised training protocol.
Luby et al. (1995)
, who examined 1,104 clinically ill
young children in Malawi, were the only other investigators who relied
solely on nonphysician health care workers to make the clinical
assessments. It is an important finding, therefore, that both
Luby et al. (1995)
and we conclude that the assessment
is feasible and useful in typical clinical facilities in Africa and
Asia. Zucker et al. (1997)
evaluated assessments by a
study physician and also by health workers, although not in the same
children, and also concluded that treatment algorithms based on pallor
were feasible to implement in African settings. Kalter et al. (1997)
found that including additional respiratory signs such
as grunting (i.e., pallor or grunting) improved the detection of severe
anemia, but this has not been tested in the hands of nonphysicians.
Based on the numbers summarized in Table 5
, a system to screen and
treat severe anemia based on pallor appears to be worthwhile, although
imperfect. It is especially useful in pregnant women and preschool
children in whom severe anemia is most prevalent. According to
international recommendations, these groups are target groups for
universal supplementation with iron, and anthelminthic or antimalarial
drugs appropriate to the epidemiology of parasitic infections in the
population (WHO 1996
, WHO, UNICEF and UNU
1998
, Stoltzfus and Dreyfuss 1998
). Where these
recommendations are being implemented, screening in these groups would
detect individuals in need of therapeutic treatments, more intensive
counseling, follow-up, and possible referral (Stoltzfus and Dreyfuss 1998
). Detection of a clinical sign of illness might
also be a useful motivational tool for women, who are sometimes
reluctant to take iron tablets or give them to their children if they
perceive themselves or their children to be well.
Where universal iron supplementation is not being implemented, as is the case in both Zanzibar and Nepal, screening for and treating severe anemia represents a modest but important first step in reducing the morbidity and mortality from anemia. Where supplies are very limited, it will enable the most effective targeting of therapies (i.e., iron and folic acid supplements, anthelminthic and antimalarial drugs).
Although assessment of clinical pallor is a useful screening strategy
for severe anemia, some of the numbers in Table 5
are discouraging. Row
2 is the good news: a fair number of severely anemic people would
receive treatment in this screening system. Since assessment of
clinical pallor costs almost nothing (some training and supervision,
plus several minutes examination time), row 2 represents substantial
benefit at almost no cost. The next two rows show the weaknesses of the
system. Many people who are not severely anemic will receive treatment
for it (row 3). In the particular case of severe anemia, this is not
very problematic. First, the treatments (oral iron and folic acid,
possibly combined with antimalarial or anthelminthic medications) are
safe. Second, they are relatively inexpensive. The drug costs for
treatment of severe anemia (A. Montresor, World Health Organization,
personal communication) according to INACG/WHO/UNICEF recommendations
are U.S. $0.07 for a Zanzibari child <2 yr old (oral iron and folic
acid plus chloroquine), U.S. $0.16 for a Zanzibari school child (oral
iron and folic acid plus anthelminthic) and U.S. $0.29 for a Nepalese
woman (oral iron and folic acid plus anthelminthic). The price of
treating nonseverely anemic people is reflected in row 6 of Table 5
. In
Nepal for example, the true drug cost to treat one severely anemic
pregnant woman is 3.8 times the price of the one treatment (3.8 x $0.29 = $1.10), because treatments are dispersed among some people
who are not severely anemic. Third, people diagnosed with clinical
pallor (and therefore treated) but who do not meet the criterion for
severe anemia are likely to be at least moderately anemic, and thus
will benefit from therapy. These treatments are not targeted with
maximum efficiency, but they are not wasted.
The major weakness of the system is the number of people severely
anemic and not treated (row 4 of Table 5
), which reflects the
relatively low sensitivity of the test. It is disappointing to have 104
severely anemic children per thousand go through the system untreated.
Fourteen per thousand severely anemic pregnant Nepalese women would
also go undetected, and face the prospect of home birth with severe
anemia. There is need for a more accurate method for detecting severe
anemia in settings without laboratories or electricity at very low
cost. Development of a standardized, ready-to-use product based on the
copper-sulfate specific gravity method (Politzer et al. 1988
), and further development of a filter paper visual
colorimetric method (Stott and Lewis 1995
) should be
pursued.
The relative performance of different anatomical sites was not
consistent among the five studies reported here, nor among the various
published studies that we reviewed. Given that assessing multiple sites
is feasible and not very time-consuming and that sensitivity is the
limiting factor in performance of clinical pallor to detect anemia, we
recommend that multiple sites be assessed in clinical practice. All the
recent studies (Kalter et al. 1997
, Luby et al. 1995
, Zucker et al. 1997
) arrived at this
conclusion. This should motivate change in the World Health
Organization's algorithm for the management of the sick child
(WHO & UNICEF, 1995
), which recommends the use of palmor
pallor alone.
In conclusion, where hemoglobin or hematocrit cannot be directly determined, assessment of clinical pallor is feasible and is clinically useful for detecting severe anemia. From these results, where severe anemia is relatively rare (<10%), the sensitivity of clinical pallor to detect hemoglobin <70 g/L is 60–80%, with specificity of 92–94%. In populations where hemoglobin concentrations are uniformly low, the sensitivity may be lower, as we observed in the 1996 Zanzibar preschoolers. This might be remedied by having the examiners regularly see some normal individuals. Nonetheless, the positive predictive value of pallor was highest in this group, because of the high prevalence of severe anemia. To maximize sensitivity, we recommend that several anatomical sites (at least the palm and the conjunctiva) be assessed, with pallor at any site used to define a positive test.
|
FOOTNOTES |
|---|
2 This manuscript received institutional approval
from the World Health Organization.
Manuscript received February 18, 1999. Initial review completed May 18, 1999. Revision accepted June 2, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. Carson J. L., Duff A., Poses R. M., Berlin J. A., Spence R. K., Trout R., Noveck H., Strom B. L. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet 1996;348:1055-1060[Medline]
2. Dreyfuss, M. L. (1998) Anemia and iron deficiency during pregnancy: etiologies and effects of birth outcomes in Nepal. (Doctoral Dissertation) The Johns Hopkins University, Baltimore, MD.
3.
Gardner G. W., Edgerton V. R., Senewiratne B., Barnard R. J., Ohira Y. Physical work capacity and metabolic stress in subjects with iron deficiency anemia. Am. J. Clin. Nutr. 1977;30:910-917
4.
Gjorup T., Bugge P. M., Hendriksen C., Jensen A. M. A critical evaluation of the clinical diagnosis of anemia. Am. J. Epidemiol. 1986;124:657-665
5. Gujral S., Abbi R., Anderson M. A., Christian P., Gopaldas T. Agreement between haemoglobin estimation and anaemia recognition card in assessment of anaemia in pregnant women. Eur. J. Clin. Nutr. 1989;43:473-475[Medline]
6. Harrison K. A., Rossiter C. E. Maternal mortality. Brit. J. Obstet. Gynaecol. Supp 1985;5:100-115
7.
Himes J. H., Walker S. P., Williams S., Bennett F., Grantham-McGregor S. M. A method to estimate prevalence of iron deficiency and iron deficiency anemia in adolescent Jamaican girls. Am. J. Clin. Nutr. 1997;65:831-836
8. Jacobs H. D., Farndell P. R., Grobbelaar P. S., Smith D. J., Bromfield M. E. Observer bias and error in the integumentary clinical diagnosis of chronic anaemia. South Afr. Med. J. 1979;55:1031-1034[Medline]
9.
Johnson-Spear M. A., Yip R. Hemoglobin difference between black and white women with comparable iron status: justification for race-specific anemia criteria. Am. J. Clin. Nutr. 1994;60:117-121
10. Kalter H. D., Burnham G., Kolstad P. R., Hossain M., Schillinger J. A., Khan N. Z., Saha S., de Wit V., Kenya-Mugisha N., Schwartz B., Black R. E. Evaluation of clinical signs to diagnose anaemia in Uganda and Bangladesh, in areas with and without malaria. Bull. WHO 1997;75(Supp 1):103-111
11. Lackritz E. M., Campbell C. C., Ruebush II T. K., Hightower A. W., Wakube W., Steketee R. W., Were J.B.O. Effect of blood transfusion on survival among children in a Kenyan hospital. Lancet 1992;340:524-528[Medline]
12. Levin H. M., Pollitt E., Galloway R., McGuire J. Micronutrient deficiency disorders. Jamison D. T. Mosley W. H. eds. Disease Control Priorities in Developing Countries 1993 Oxford University Press (World Bank) New York, NY.
13. Lilienfeld A. M., Lilienfeld D. E. Foundations of Epidemiology 2nd ed. 1980 Oxford University Press New York, NY.
14. Llewellyn-Jones D. Severe anaemia in pregnancy. Aust New Zealand J. Obstet. and Gynaecol. 1965;5:191-197
15. Luby S. P., Kazembe P. N., Redd S. C., Ziba C., Nwanyanwu O. C., Hightower A. W., Franco C., Chitsulo L., Wirima J. J., Olivar M. A. Using clinical signs to diagnose anaemia in African children. Bull. World Health Org. 1995;73:477-482[Medline]
16. Nardone D. A., Roth K. M., Mazur D. J., McAfee J. H. Usefulness of physical examination in detecting the presence or absence of anemia. Arch. Intern. Med. 1990;150:201-204[Abstract]
17. Politzer W. M., Myburgh W. M., van der Merwe J. F. Haemoglobin estimation—reliability of the copper sulphate specific gravity v. the cyanmethemoglobin colorimetric method. South Afr. Med. J. 1988;73:111-112[Medline]
18. Shah U., Pratinidhi A. K., Bhatlawande P. V., Alekar S. V. Using community health workers to screen for anaemia. World Health Forum 1984;5:35-36
19. Snedecor G. W., Cochran W. G. Statistical Methods 7th ed. 1980 Iowa State University Press Ames, IA.
20.
Stoltzfus R. J., Chwaya H. M., Tielsch J. M., Schulze K. J., Albonico M., Savioli L. Epidemiology of iron deficiency anemia in Zanzibari school children. Am. J. Clin. Nutr. 1997;65:153-159
21. Stoltzfus R. J. Rethinking anaemia surveillance. Lancet 1997;349:1764-1766[Medline]
22. Stoltzfus, R. J. & Dreyfuss, M. L. (1998) Guidelines for the Use of Iron Supplements to Prevent and Treat Iron Deficiency Anemia. International Nutritional Anemia Consultative Group (INACG), Washington, D.C.
23. Stott G. J., Lewis S. M. A simple and reliable method for estimating haemoglobin. Bull. World Health Org. 1995;73:369-373[Medline]
24. Strobach R. S., Anderson S. K., Doll D. C., Ringenberg Q. S. The value of the physical examination in the diagnosis of anemia. Arch. Intern. Med. 1988;148:831-832[Abstract]
25. Thaver I. H., Baig L. Anaemia in children: Part. I. Can simple observations by primary care provider help in diagnosis?. J Pak Med Assoc 1994;44:282-284[Medline]
26. Viteri F. E., Torun B. Anaemia and physical work capacity. Clin Haematol 1974;3:609-626
27.
West K. P., Jr, Katz J., Khatry S. K., LeClerq S. D., Pradhan E. K., Shrestha S. R., Connor P. B., Dali S. M., Pokhrel R. P., Sommer A. Double blind, cluster randomised trial of low dose supplementation with vitamin A or beta carotene on mortality related to pregnancy in Nepal. The NNIPS-2 Study Group. Brit Med J 1999;318:570-575
28. World Health Organization (1996) Report of the WHO Informal Consultation on Hookworm Infection and Anaemia in Girls and Women. WHO/CTD/SIP/96.1. WHO, Geneva, Switzerland.
29. World Health Organization & UNICEF (1995) Management of Childhood Illness. WHO, Geneva, Switzerland.
30. World Health Organization, UNICEF & United Nations University (1998) Iron Deficiency: Indicators for Assessment and Strategies for Prevention. WHO, Geneva, Switzerland.
31. Wurapa F. K., Bulsara M. K., Boatin B. A. Evaluation of conjunctival pallor in the diagnosis of anaemia. J. Trop. Med. Hyg. 1986;89:33-36[Medline]
32. Zucker J. R., Perkins B. A., Jafari H., Otieno J., Obonyo C., Campbell C. C. Clinical signs for the recognition of children with moderate or severe anaemia in western Kenya. Bull. WHO 1997;78 (Supp 1):97-102
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
