KEY WORDS:
anemia
iron deficiency
pregnancy
Nepal
ABSTRACT Anemia and iron deficiency during pregnancy are prevalent in developing countries, but their causes
are not always known. We assessed the prevalence and severity of anemia and iron deficiency and their
association with helminths, malaria and vitamin A deficiency in a community-based sample of 336 pregnant women
in the plains of Nepal. Hemoglobin, erythrocyte protoporphyrin (EP) and serum ferritin were assessed in venous
blood samples. Overall, 72.6% of women were anemic (hemoglobin 110 g/L), 19.9% had moderate to severe
anemia (hemoglobin 90 g/L) and 80.6% had iron deficiency (EP 70 mol/mol heme or serum ferritin 10
g/L). Eighty-eight percent of cases of anemia were associated with iron deficiency. More than half of the women
(54.2%) had a low serum retinol concentration (1.05 mol/L), 74.2% were infected with hookworms and 19.8%
had Plasmodium vivax malaria parasitemia. Hemoglobin, EP and serum ferritin concentrations were significantly
worse and the prevalence of anemia, elevated EP and low serum ferritin was increased with increasing intensity of
hookworm infection. Hookworm infection intensity was the strongest predictor of iron status, especially of depleted
iron stores. Low serum retinol was most strongly associated with mild anemia, whereas P. vivax malaria and
hookworm infection intensity were stronger predictors of moderate to severe anemia. These findings reinforce the
need for programs to consider reducing the prevalence of hookworm, malaria infection and vitamin A deficiency
where indicated, in addition to providing iron supplements to effectively control anemia. J. Nutr. 130: 25272536,
2000.
1
Funded through cooperative agreement HRN-A-00-97-00015-00 between
The Johns Hopkins University and the Office of Health and Nutrition, U.S. Agency
for International Development. This study was conducted as a collaboration
between the Center for Human Nutrition and the Sight and Life Institute in the
Department of International Health at The Johns Hopkins School of Hygiene and
Public Health and the National Society for Comprehensive Eye Care (Nepal Netra
Jyoti Sangh), Kathmandu, Nepal.
2
To whom correspondence should be addressed at Department of Nutrition,
Harvard School of Public Health, 665 Huntington Avenue, Boston, MA 02115.
DREYFUSS ET AL.
2528
3
Abbreviations used: MUAC, mid-upper arm circumference; AOR, adjusted
odds ratio; CI, confidence interval; EP, erythrocyte protoporphyrin; LMP, last
menstrual period.
TABLE 1
Characteristics of study sample of Nepalese
pregnant women1
Frequency
%
Age, y (n 334)
20
2029
30
Socioeconomic status (n 316)
Literacy
Radio owned
Activities during the past 7 d
Smoked cigarettes (n 330)
Drank alcohol (n 329)
Parity (n 342)
0
12
34
5
Gestational age, wk
12
1324
25
Height (145.0 cm)
Mid-upper arm circumference (21.0 cm)
Serum retinol, mol/L (n 334)
0.70
0.701.04
1.051.39
1.40
Iron supplement use in the past month (n 331)
Hookworm infection (n 190)
Negative (0 epg2)
Light (11999 epg)
Moderate (20003999 epg)
Heavy (4000 epg)
Ascaris lumbricoides infection (n 190)
Negative (0 epg)
Light (19999 epg)
Moderate (10,00049,999 epg)
Heavy (50,000 epg)
Trichuris trichiura infection (n 190)
Plasmodium vivax malaria parasitemia (n 288)
1 n 336 except where indicated.
2 epg, eggs/g of feces.
19.5
64.0
16.5
12.0
29.1
18.5
1.5
21.3
42.5
23.9
12.3
15.5
63.7
20.8
17.0
20.2
20.7
33.5
32.3
13.5
0.3
25.8
61.6
6.3
6.3
41.1
53.7
4.7
0.5
5.3
19.8
2529
RESULTS
Characteristics of study sample. Women ranged in age
from 15 to 40 y, with 64% between 20 and 29 y (Table 1). Few
women were literate, and only 29% came from a household in
which a radio was owned. Approximately one fifth of the
women smoked cigarettes, but alcohol consumption was rare.
DREYFUSS ET AL.
2530
TABLE 2
Indicators of anemia and iron deficiency in Nepalese women by trimester of pregnancy
Erythrocyte
protoporphyrin,1
mol/mol heme
Hemoglobin, g/L
Trimester of
pregnancy (n)
110
90
80
Means
SD
%
First (52)
Second (214)
Third (70)
Total (336)
9.6
16.8
37.14
19.9
Means2
(1 SD, 1 SD)
%
3.8
6.1
14.33
7.4
104 13
101 15
94 185
100 16
65.4
60.6
82.96
66.0
Means,2
(1 SD, 1 SD)
%
83 (48, 142)
82 (52, 129)
109 (65, 186)7
87 (53, 143)
n 335.
Geometric means.
2 for trend by trimester of pregnancy, P 0.05.
2 for trend by trimester of pregnancy, P 0.005.
Test for linear trend, P 0.0005.
Significantly different from first and second trimesters combined, 2 test, P 0.005.
Significantly different from first and second trimesters combined, Students t test, P 0.0001.
10
42.3
55.6
79.74
58.5
1
2
3
4
5
6
7
63.5
72.4
80.03
72.6
70
2531
TABLE 3
Indicators of anemia and iron deficiency in Nepalese pregnant women by risk factor1
Erythrocyte
protoporphyrin,
mol/mol heme
Hemoglobin, g/L
Risk factor
110
90
80
Means
SD
Means,2
(1 SD, 1 SD)
10
Means,2
(1 SD, 1 SD)
55.1
68.2
75.9
75.04
8.2
12.5
24.1
33.35
2.0
2.3
3.4
25.05
106 11
104 12
99 14
90 226
55.1
56.3
72.4
75.04
75 (47, 119)
75 (51, 111)
86 (54, 139)
96 (55, 169)7
40.8
60.2
65.5
79.25
66.7
67.0
15.4
16.1
6.4
4.5
102 15
102 14
59.0
62.2
78 (50, 124)
80 (51, 125)
62.8
55.4
66.7
70.0
16.1
10.0
5.6
0
102 15
104 10
60.9
60.0
80 (51, 125)
72 (42, 123)
58.9
50.0
68.4
77.29
15.6
28.110
4.8
12.310
102 14
97 18
63.5
66.7
84 (52, 134)
90 (54, 150)
55.8
68.4
63.4
80.79
12.4
26.09
4.6
9.4
104 13
96 1612
57.2
72.99
77 (50, 119)
95 (57, 159)12
52.3
64.110
sity of infection (Table 3). For example, hemoglobin concentration decreased from 106 g/L among uninfected women to 90
g/L among women with moderate to severe infection (P
0.0005). Hookworm infection intensity was strongly associated with anemia at various levels of severity. At 2000
eggs/g of feces, the prevalence of moderate to severe anemia
was four times that of uninfected women, and severe anemia
was 12 times more prevalent. The prevalence of elevated EP
and of low serum ferritin concentrations also increased with
increasing hookworm infection intensity. Neither A. lumbricoides nor T. trichuris infection was related to any iron status
indicator.
P. vivax malaria parasitemia was associated with anemia,
especially more severe anemia (Table 3). There also was
evidence that malaria was associated with iron deficiency.
Serum ferritin concentration was lower (P 0.05) among
women with P. vivax malaria parasitemia, and the proportion
of women with low serum ferritin tended to be higher.
Low serum retinol concentration was strongly associated
with all three iron status indicators in this cohort (Table 3).
Women with low serum retinol were more likely to be anemic,
to have iron-deficient erythropoiesis and to be iron depleted.
Risk factors for anemia and iron deficiency: multivariate
analyses. AOR for anemia and iron deficiency, calculated
from multivariate logistic regression models, are presented in
Table 4. As in the bivariate analyses, hookworm infection was
Hookworm infection
0 epg3 (n 49)
1999 epg (n 88)
10001999 epg (n 29)
2000 epg (n 24)
Ascaris lumbricoides infection
Negative (n 78)
Positive (n 112)
Trichuris trichiura infection
Negative (n 180)
Positive (n 10)
Plasmodium vivax malaria
parasitemia
Negative (n 231)
Positive (n 57)
Serum retinol
1.05 mol/L (n 153)
1.05 mol/L (n 181)
70
the strongest predictor of poor iron status for all three indicators. The strongest risk factor for anemia varied by the severity
of the anemia being modeled. For example, the risk of anemia
associated with hookworm infection increased as hemoglobin
cutoffs for more severe anemia were used. In contrast, low
serum retinol concentration was most strongly associated with
mild anemia. The relative odds of anemia by any cutoff were
approximately doubled with P. vivax malaria parasitemia, but
the 95% CI included 1. In stratified analyses, malaria was
strongly associated with moderate to severe anemia among the
small proportion of women not infected with hookworms
(25.0 versus 2.9%, P 0.08 by Fishers exact test). However,
the small numbers in some of the subgroups did not allow for
adequate analysis of this interaction.
Hookworm infection was most strongly associated with
depleted iron stores. The adjusted odds of serum ferritin of
10 g/L were approximately three to nine times greater for
increasingly heavy hookworm infection compared with odds
for women without infection. In contrast, P. vivax malaria
parasitemia and low serum retinol were not associated with
increased odds of low serum ferritin after multivariate adjustment. The AOR of iron-deficient erythropoiesis for categories
of hookworm egg counts of 1000 eggs/g of feces was 2
compared with women without hookworm infection, but this
association was not statistically significant because the CI for
the AOR included 1.
DREYFUSS ET AL.
2532
TABLE 4
Adjusted odds ratios (AOR) for anemia, iron-deficient erythropoiesis and depleted iron stores associated with hookworms,
Plasmodium vivax malaria and low serum retinol1
Anemia (low hemoglobin), g/L
Risk factor
110 g/L
100 g/L
90 g/L
Iron-deficient
erythropoiesis2
Depleted iron
stores3
1.0
1.82 (0.83, 3.96)
2.86 (0.97, 8.47)
2.97 (0.93, 9.52)
1.0
1.81 (0.76, 4.30)
2.25 (0.76, 6.71)
7.47 (2.26, 24.68)
1.0
1.97 (0.50, 7.70)
4.71 (1.04, 21.32)
8.24 (1.73, 39.21)
1.0
0.99 (0.46, 2.10)
1.97 (0.70, 5.56)
2.35 (0.76, 7.29)
1.0
2.75 (1.25, 6.08)
3.33 (1.17, 9.45)
8.81 (2.54, 30.49)
1.0
2.24 (0.91, 5.52)
1.0
2.01 (0.90, 4.50)
1.0
2.28 (0.86, 6.09)
1.0
1.67 (0.74, 3.76)
1.0
1.33 (0.59, 3.03)
1.0
2.52 (1.28, 4.98)
1.0
1.96 (1.01, 3.83)
1.0
2.07 (0.84, 5.06)
1.0
1.69 (0.89, 3.19)
1.0
1.06 (0.55, 2.04)
1 Analysis restricted to women who provided a stool specimen (n 190). AOR and their SE estimated from multivariate logistic regression models
that included all three risk factors and trimester of pregnancy.
2 Iron-deficient erythropoiesis defined as erythrocyte protoporphyrin 70 mol/mol heme.
3 Depleted iron stores defined as serum ferritin 10 g/L.
4 CI, confidence interval.
5 epg, eggs/g of feces.
2533
TABLE 5
Attributable fractions of anemia among Nepalese pregnant women
Prevalence of
anemia in
nonexposed group
(p1)
Adjusted
odds ratio1
Adjusted
prevalence
ratio2
Proportion of
cases exposed to
risk factor (pd)
%
Iron deficiency4
All anemia5
Moderate to severe anemia6
Hookworm infection
All anemia
Moderate to severe anemia
Plasmodium vivax malaria parasitemia
All anemia
Moderate to severe anemia
Low serum retinol7
All anemia
Moderate to severe anemia
Attributable
fraction3
44.6
3.1
...
...
1.788
7.808
88.1
97.0
39
85
55.1
8.2
2.15
3.15
1.32
2.68
78.7
86.7
19
54
63.49
13.49
2.24
2.28
1.25
1.95
23.7
32.1
5
16
58.09
11.09
2.52
2.07
1.34
1.85
54.3
63.3
14
29
Downloaded from jn.nutrition.org by guest on November 27, 2012
1 Adjusted odds ratio (AOR) estimated from multivariate logistic regression models that included hookworm infection, P. vivax malaria parasitemia,
vitamin A deficiency (serum retinol 1.05 mol/L) and trimester of pregnancy.
2 Adjusted prevalence ratio (APR) calculated from AOR using the following conversion formula: APR AOR/[1 p1(AOR 1)], where p1
prevalence of anemia in the nonexposed group (Osborn and Cattaruzza 1995).
3 Defined as [pd(RR 1)/RR] where pd is proportion of cases of anemia exposed to a risk factor, and RR is relative risk or APR (Kleinbaum et al.
1982, Rockhill et al. 1998).
4 Erythrocyte protoporphyrin 70 mol/mol heme or serum ferritin 10 g/L.
5 Hemoglobin 110 g/L.
6 Hemoglobin 90 g/L.
7 Serum retinol 1.05 mol/L.
8 Defined as prevalence of anemia among those with iron deficiency divided by prevalence among those without iron deficiency and not adjusted
for other risk factors.
9 Prevalence in the nonexposed group from the subsample of respondents with complete risk factor data, n 190.
worst in the third trimester, suggesting that the high prevalence of anemia was caused by underlying iron deficiency. The
extremely low serum ferritin concentrations observed in this
study are evidence that rural Nepalese women enter pregnancy
with depleted iron stores.
Risk factors for anemia and iron deficiency
Hookworm infection. We previously reported that hookworm infection is endemic among women in the rural plains of
Nepal (Navitsky et al. 1998), and the present results suggest
that it is responsible for 54% of cases of moderate to severe
anemia during pregnancy. Hookworm infection was associated
with all three indicators of iron status in a density-dependent
manner. Mature hookworms cause intestinal bleeding, leading
to fecal blood loss proportional to the intestinal worm burden
(Roche and Layrisse 1966). The intensity of hookworm infection that causes iron deficiency anemia varies according to the
species and the iron status of the population. Hookworms in
this study sample were exclusively A. duodenale (Navitsky et
al. 1998), the hookworm species that causes the greatest blood
loss (Pawlowski et al. 1991), and iron deficiency was severe.
Thus, hookworm infection exacerbated iron deficiency and
anemia in this setting.
Hookworm infection has been established as a strong predictor of iron deficiency and anemia in other populations
(Hopkins et al. 1997, Layrisse and Roche 1964, Roche and
Layrisse 1966, Stoltzfus et al. 1997), but few studies have
examined these relationships in pregnant women. Hookworm
infection was associated with severe but not moderate anemia
among women receiving antenatal care at a hospital in Kathmandu, Nepal (Bondevik et al. 2000). A Kenyan study of
2534
DREYFUSS ET AL.
a synergistic negative effect of vitamin A deficiency and malaria parasitemia on erythropoiesis. However, the mechanism
of this interaction is unclear. Vitamin A deficiency may inhibit erythropoiesis directly (West and Roodenburg 1992) or
through inhibition of iron mobilization or transport (Bloem et
al. 1989, Wolde-Gebriel et al. 1993). Malaria attacks hemolyze
both infected and uninfected red cells, and the immune response of the body to infection results in further red cell
destruction and progressive anemia due to phagocytosis (Markell et al. 1992).
Implications for control of anemia and iron deficiency
The very high prevalence of iron deficiency and anemia
during pregnancy in this study points to an urgent need for
anemia treatment and control in the plains of Nepal. Although improvement in the amount and quality of dietary iron
intake among pregnant women is an important long-term goal,
it is well established that this alone cannot meet the iron
needs of pregnancy, particularly among populations who consume diets of low iron bioavailability. Thus, the need remains
to supplement the iron intake of women during pregnancy
(Stoltzfus and Dreyfuss 1998, World Health Organization,
UNICEF and UNU 1998) and possibly throughout the reproductive years to build up the iron stores necessary to meet the
high iron demands of pregnancy and ubiquitous infections.
Anthelminthic therapy is inexpensive and is safe during
pregnancy after the first trimester (World Health Organization
1996). The World Health Organization recommends anthelminthic therapy for women to control hookworm infection in
areas in which the prevalence of infection is 20 30% and
anemia is prevalent (World Health Organization 1996). The
more severe anemia attributable to malaria may also be reduced by antimalarial chemoprophylaxis during pregnancy, as
has been demonstrated among primigravidae in several studies
of P. falciparum malaria (Bouvier et al. 1997, Fleming et al.
1986, Gilles et al. 1969, Greenwood et al. 1989, Mutabingwa
et al. 1993, Nosten et al. 1994). However, chemoprophylaxis
did not affect hemoglobin concentrations in trials conducted
in areas of lower malaria prevalence (Hamilton et al. 1972,
Jackson and Latham 1982), and its impact is not known when
only P. vivax malaria is present.
Finally, improving the vitamin A status of vitamin A
deficient pregnant women in addition to iron supplementation
may reduce the risk of mild anemia. Vitamin A supplementation of women before or during pregnancy, or both, may be an
effective intervention with multiple benefits for the health and
nutritional status of women (Christian et al. 1998a, West et al.
1999). Our study was conducted as part of a randomized trial
of vitamin A and -carotene supplementation of women of
childbearing age, which will allow us to examine the treatment effect of supplementation on anemia and iron deficiency.
Efforts should be made to prevent pregnancy anemia and its
damaging consequences using an appropriate mix of interventions that address the multiple causes of anemia and iron
deficiency in the population.
nulliparous women (Brabin et al. 1990, Fleming 1989, Matteelli et al. 1994, McGregor 1984, Shulman et al. 1996),
suggesting that the epidemiology of malaria and anemia in
pregnant women may differ by species of malaria parasite. For
example, P. vivax parasites only infect reticulocytes (2% of
red blood cells) rather than invading red blood cells of all ages,
and their erythrocytic asexual phase in the human host remains in peripheral circulation rather than entering capillaries
of internal organs. These differences result in heavier parasitemia with P. falciparum associated with serious complications and even death (Gilles and Warrell 1993, Markell et al.
1992). However, unlike P. falciparum malaria, relapses of P.
vivax malaria can occur for years after an initial attack, leading
to further red cell destruction and worsening anemia (Gilles
and Warrell 1993, Markell et al. 1992). South Indian men
infected with P. vivax malaria showed a progressive decrease in
hemoglobin concentration, packed cell volume and red blood
cell level with increasing number of malaria attacks, and this
relationship held across all levels of parasitemia (Selvam and
Baskaran 1996).
Vitamin A deficiency. The prevalence of vitamin A deficiency in this sample was very high. Twenty-one percent of
women had a serum retinol concentration of 0.70 mol/L,
and 54% had a serum retinol concentration of 1.05 mol/L,
virtually the same as previously reported in other concurrent
studies in the same population (Christian et al. 1998b).
Women with a low serum retinol concentration were more
than twice as likely to be anemic compared with those with a
higher serum retinol concentration, suggesting that vitamin A
deficiency decreases hemoglobin synthesis. Numerous studies
that examined the effect of vitamin A deficiency on iron status
have found an association between serum retinol and hemoglobin concentrations in pregnant women (Bondevik et al.
2000, Suharno et al. 1992), adolescent girls (Ahmed et al.
1996) and children (Meja et al. 1977, Mohanram et al. 1977,
Wolde-Gebriel et al. 1993). Suharno et al. (1992) and Ahmed
et al. (1996) found 4- to 10-g/L increases in hemoglobin
associated with a 1-mol/L increase in serum retinol concentration in multivariate linear regression models adjusted for
confounders. When we adjusted for other risk factors, we
found a relationship of similar magnitude between serum retinol and hemoglobin (see Results).
Randomized trials to examine the effect of vitamin A
supplementation on iron status have reinforced the findings
from observational studies. A single, massive dose of vitamin
A significantly increased hemoglobin among vitamin A deficient Thai children (Bloem et al. 1990) and among Indonesian children who were both vitamin A deficient and anemic
(Semba et al. 1992). In a randomized, placebo-controlled trial
among anemic Guatemalan children, daily vitamin A supplements for 2 mo increased hemoglobin concentration by 9 g/L
compared with an increase of 3 g/L in the placebo group
(Meja and Chew 1988). The consumption of vitamin Afortified monosodium glutamate increased the hemoglobin of
Indonesian children by 10 g/L after 5 mo compared with no
change among control children (Muhilal et al. 1988). A
randomized, placebo-controlled trial of vitamin A and iron
supplementation among pregnant Indonesian women found
that 2 mo of daily vitamin A supplementation significantly
increased the hemoglobin concentration and reduced the
prevalence of anemia by 23% (Suharno et al. 1992). Among
women who received both vitamin A and iron, the positive
effect on hemoglobin was even greater, and anemia was almost
completely eliminated.
We observed a significant interaction of serum retinol and
P. vivax malaria on hemoglobin and EP. These findings suggest
ACKNOWLEDGMENTS
We thank the members of the Nepal Nutrition Intervention
Project-Sarlahi (in addition to the authors) for their dedication and
assistance in the implementation of this study, especially D. N.
Mandal, T. R. Shakya, N. N. Acharya, G. Subedi, U. Shankar, A.
Bhetwal, D. B. Khadka, B. J. Thapa and all the members of the
NNIPS clinic staff. We also thank Paul Connor for his assistance with
data management and Kerry Schulze, Laura Caulfield and Anu Shankar for their input and discussion of the study findings.
LITERATURE CITED
2535
2536
DREYFUSS ET AL.
West, K. P., Jr., Pokhrel, R. P., Katz, J., LeClerq, S. C., Khatry, S. K., Shrestha,
S. R., Pradhan, E. K., Tielsch, J. M., Pandey, M. R. & Sommer, A. (1991)
Efficacy of vitamin A in reducing preschool child mortality in Nepal. Lancet
338: 6771.
World Health Organization (1991) Basic Laboratory Methods in Medical Parasitology. World Health Organization, Geneva.
World Health Organization (1992) The Prevalence of Anaemia in Women: A
Tabulation of Available Information. World Health Organization, Geneva.
World Health Organization (1994) Bench Aids in the Diagnosis of Intestinal
Parasites. World Health Organization, Geneva.
World Health Organization (1996) Report of the WHO Informal Consultation on
Hookworm Infection and Anaemia in Girls and Women. World Health Organization, Geneva, WHO/CTD/SIP/ 96 1.
World Health Organization, UNICEF & UNU (1998) Iron Deficiency: Indicators
for Assessment and Strategies for Prevention. World Health Organization,
Geneva.
Wolde-Gebriel, Z., West, C. E., Haile Gebru, Tadesse, A. S., Fisseha, T., Gabre,
P., Ayana, G. & Hautvast, J. G. A. J. (1993) Interrelationship between
vitamin A, iodine and iron status in school children in Shoa Region, Central
Ethiopia. Br. J. Nutr. 70: 593 607.
Zerfas, A. J. (1975) The insertion tape: a new circumference tape for use in
nutritional assessment. Am. J. Clin. Nutr. 28: 782787.
Zhou, L. M., Yang, W. W., Hua, J. Z., Deng, C. Q., Tao, X. & Stoltzfus, R. J. (1998)
Relation of hemoglobin measured at different times in pregnancy to preterm
birth and low birth weight in Shanghai, China. Am. J. Epidemiol. 148: 998
1006.