European Journal of Clinical Nutrition (2001) 55, 538±546
ß 2001 Nature Publishing Group All rights reserved 0954±3007/01 $15.00
www.nature.com/ejcn
Original Communication
Postnatal iron status of Hong Kong Chinese women in a
longitudinal study of maternal nutrition
SM Chan1, EAS Nelson1*, SSF Leung1 and CY Li2
1
Department of Paediatrics, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong, People's Republic of
China; and 2Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong
Kong, People's Republic of China
Objective: To report postnatal iron nutritional status of Hong Kong Chinese women during the ®rst 6 months
postpartum.
Design and subjects: A longtitudinal study examining postnatal calcium and iron status of Hong Kong Chinese
breastfeeding and formula-feeding women was conducted during 1998. Postpartum women aged 20 ± 40 y, with no
bone or blood disorders were recruited and interviewed at 0 (baseline), 2, 6 weeks, 3 and 6 months postpartum.
Dietary intake was assessed by a 3 day dietary record and cross checked by a 24 h recall. Complete blood count
and serum ferritin level were measured to assess anaemia and iron status. In this report, subjects were divided into
an anaemic group (haemoglobin level < 10 g=dl) and a non-anaemic group (haemoglobin level 10 g=dl)
according to baseline haemoglobin levels.
Results: At baseline, 13=47 (27.7%) subjects were anaemic. Two of these 13 anaemic subjects were still anaemic
at 3 and 6 months postpartum. Anaemic subjects showed signi®cantly (P < 0.01) greater amounts of blood loss and
a higher rate of primary postpartum haemorrhage than the non-anaemic subjects. Daily food intake and dietary
nutrient intake did not differ signi®cantly between the two groups. During the ®rst 6 weeks postpartum, subjects in
both groups consumed more poultry and egg, and comparable amounts of meat, compared with women in the
Hong Kong general population. Iron and vitamin C intakes for the majority of subjects reached 60% of the US
Recommended Daily Allowances. Regression analysis suggested that the rate of change in haemoglobin level in
the ®rst 6 weeks postpartum was positively correlated with baseline MCV level and serum ferritin level, but
negatively correlated with baseline haemoglobin level.
Conclusions: Blood loss at delivery is an important factor for postpartum anaemia. Postnatal recovery of iron
status of this group of women appeared to be more related to physiological factors than to dietary factors. The role
of diet as well as other physiological changes in postpartum women requires further investigation. Finding ways to
minimise blood loss at delivery could be the most practical strategy to reduce the rate of postpartum anaemia.
Sponsorship: CSM was supported by a research studentship from the Research Grants Council, Hong Kong.
Descriptors: Chinese; postpartum; anaemia; iron; ferritin
European Journal of Clinical Nutrition (2001) 55, 538±546
Introduction
Iron is the most abundant trace element in humans. Total
body iron averages about 3.8 g in men and 2.3 g in women.
*Correspondence: EAS Nelson, Department of Paediatrics, The Chinese
University of Hong Kong, 6=F Clinical Science Building, Prince of Wales
Hospital, Shatin, Hong Kong, People's Republic of China.
E-mail: [email protected]
Guarantor: EAS Nelson.
Contributors: SSFL and SMC initiated and designed the study. Most of the
subject recruitment, data collection, laboratory analysis and data analysis
were performed by SMC. All authors contributed to the original
manuscript and the subsequent revisions of the paper.
Received 22 February 2000; revised 2 January 2001;
accepted 10 January 2001
In addition to serving as a component of haemoglobin (Hb),
myoglobin and many enzymes, iron is stored in the reticulo-endothelial system as ferritin and haemosiderin (Great
Britain. Panel on Dietary Reference Values of the Committee on Medical Aspects of Food Policy, 1991). Iron
de®ciency ultimately results in defective erythropoiesis
leading to a normocytic or microcytic hypochromic anaemia (Anonymous, 1985). However, functional consequences of iron de®ciency may occur in the absence of
anaemia. These include impaired work performance (Cook
et al, 1994; Lozoff et al, 1998), behaviour and intellectual
performance (Moffatt et al, 1994). In addition, impaired
thermoregulatory mechanisms (Beard et al, 1984) and an
increased risk of infection (Dallman, 1987) have been
documented in iron de®ciency status. Iron de®ciency has
Postnatal iron status
SM Chan et al
also been found to be associated with adverse pregnancy
outcomes (Yip & Dallman, 1996).
Pregnant women and women of childbearing age are at a
high risk of iron de®ciency. Permanent iron losses during
pregnancy are up to 840 mg (Viteri, 1997). Measured blood
loss during normal delivery is signi®cantly greater than the
estimated blood loss, indicating that the visual estimation
of blood loss is inaccurate (Duthie et al, 1990). Although
iron loss in breast milk during lactation is relatively small
and menstruation rarely occurs during the ®rst few months
of lactation (Institute of Medicine (US), 1990), the total
iron cost of pregnancy and menstruation is signi®cant.
Iron absorption is a highly regulated process, in which
absorption increases as iron stores become depleted and
vice versa (Cook, 1990; Herbert, 1987). Unlike haem iron,
non-haem iron absorption is easily affected by dietary
components. Several dietary factors have been found to
affect the non-haem iron absorption. The presence of meat
and organic acids promotes absorption, whereas polyphenols, phytate, dietary ®bre complex, calcium, phosphorus
and dietary protein inhibit non-haem iron absorption
(Bothwell et al, 1989).
In traditional Chinese culture, women are believed to be
`very weak' after delivery and a special diet during the ®rst
month after delivery is recommended. Extra meat, chicken
and eggs are amongst the main dietary items included in
this diet. As these food items contain comparably high
amounts of iron, it can be speculated that this dietary
practice may help restore the iron stores of postpartum
women. In addition, a special dish composed of ginger,
pig's trotter, eggs and sweet vinegar is also recommended
(Chan et al, 2000). Other soups and drinks, such as a date
tea prepared from soaking dates in boiled water, and a
chicken wine prepared from cooking chicken with glutinous rice wine are also consumed. Special Chinese tonics
composed of herbs and Chinese medicine are recommended to help restore the health status of women after
delivery.
This paper reports the iron status of a sample of Hong
Kong Chinese women during the ®rst 6 months postpartum,
and assesses the possible effect of these traditional Chinese
dietary practices on iron status.
Subjects and methods
Subjects
Results of this report are part of a longitudinal study
examining both the calcium and iron nutritional status of
Hong Kong Chinese postpartum women. Subjects were
recruited in the postnatal wards of the Obstetrics and
Gynaecology Department, Prince of Wales Hospital
between February and December 1998. Interviews were
undertaken at 0 week (7 days postpartum), 2 1 weeks,
6 1 weeks, 3 0.5 and 6 0.5 months postpartum. The 2
week interview was undertaken at home and other interviews were performed at the hospital.
Basic personal and medical information were ®rst
checked. Women who were Hong Kong Chinese citizens
aged between 20 and 40 y, and who had delivered a fullterm (gestational age 37 weeks) healthy singleton infant
were eligible for inclusion in the study. Exclusion criteria
included a history of smoking, drinking, metabolic bone
disease, osteoarithritis, haematological disorder, oestrogen
or other steroid hormone medications, treatment with
heparin, thyroxine, thiazide or anticonvulsants, or an
antenatal haemoglobin level of < 11 g=dl at the ®rst or
second antenatal checkup. Since one main aim of the study
was to compare the calcium and iron status between
formula-feeding mothers and breastfeeding mothers (lactation for 3 months) in Hong Kong, postpartum women
who ful®lled the above criteria were approached by the
research staff to ask about the intended infant feeding
practice. Mothers who intended to breastfeed were further
asked about the intended duration of breastfeeding and only
those mothers intending to exclusively breastfeed for 3
months were included. De®nition of breastfeeding was
based on Labbok's de®nition (Labbok & Krasovec,
1990). The purpose and nature of the study was explained
to the mothers and informed written consent obtained from
those agreeing to participate. The study protocol was
approved by the Ethics Committee of The Faculty of
Medicine, The Chinese University of Hong Kong. Primary
postpartum haemorrhage was de®ned as blood loss
500 ml at delivery (Hayashi, 1998). The amount of
blood loss is routinely recorded after all deliveries. After
vaginal deliveries, all blood, including blood clots and
placenta, is collected in a standard kidney dish. After
examination of the placenta, the volume of the remaining
blood and clots in the kidney dish is estimated and recorded
by the midwife or doctor. The number of blood-soaked
gauze swabs are also included in the estimation. The
attending anaesthetist will estimate blood loss during caesarean sections by viewing all blood and clots collected in a
standard kidney dish together with blood-soaked gauze
swabs or pads. Anthropometric data, dietary intake and
biochemical indices of iron status were measured and the
date of resumption of the ®rst postpartum mense was noted.
539
Dietary data and nutrient analysis
Dietary data were assessed by trained researchers at all
interviews, using a 3-day dietary record cross-checked by a
24 h recall. Bowls, cups and spoons were displayed to help
recall. A food photo album containing pictures of commonly consumed foods was also shown to subjects (Leung
et al, 1997). Nutrient intakes were estimated using a
computerised food table with food items compiled from
Britain (McCance et al, 1978), China (Institute of Health,
1980), Taiwan (Tung et al, 1961), and US (Church &
Church, 1975; US Department of Health, 1972). Chemical
analyses of some local Chinese foods from the government
European Journal of Clinical Nutrition
Postnatal iron status
SM Chan et al
540
non-pregnant and non-lactating females aged 12 y (Institute of Medicine (US), 1993). Hb and serum ferritin levels
were used as biochemical indices for the diagnosis of iron
de®ciency as shown in Table 3 (Cook & Skikne, 1989).
chemist, and product information from food manufacturers
are also included. In addition, chemical analysis of the
calcium and iron contents of `ginger vinegar soup', a
specially prepared soup recommended for Chinese
women in the early postpartum, was carried out and results
included (Chan et al, 2000).
Statistical analysis
Biochemical measurement of iron status
Data were analysed by SPSS Version 8.0 for Windows
(SPSS Inc. Chicago, IL, USA). Longitudinal data were
plotted extensively as a guide for data analysis and distribution of all variables was checked. Since a proportion of
subjects failed to attend the 6 week interview, not all
subjects were included in the data analysis. Only those
who had complete data sets at both 0 week and 6 weeks
postpartum were included and their data at 3 months and 6
months postpartum were also presented. Subjects were
divided into an anaemic group (Hb < 10 g=dl) and a nonanaemic group (Hb 10 g=dl) for data analysis based on
the Hb level measured at 0 week.
Independent Student t-test and Fischer's exact test were
used to compare demographic data and biochemical indices
between the two groups, whereas the Mann ± Whitney Utest was used to compare dietary intakes. Longitudinal
regression models using stepwise regression analysis were
performed. Plots of the postpartum changes in Hb suggested that the Hb level increased gradually from baseline
to 6 weeks postpartum and remained fairly constant afterwards. Therefore, the rate of change in Hb level from
baseline to 6 weeks postpartum was chosen as the dependent variable. It is also during this 6 week time period that
Ten millilitres of venous blood were drawn at the 0 week, 6
week and 3 and 6 month interviews. Serum and plasma
samples were aliquoted after centrifugation and stored at
ÿ80 C until analysis. EDTA specimens were sent immediately to the Departments of Haematology and Chemical
Pathology of the Prince of Wales Hospital for complete
blood count (CBC) measurement, which includes haemoglobin (Hb), mean cell volume (MCV), mean cell haemoglobin concentration (MCHC) and mean cell haemoglobin
(MCH). Blood samples for CBC measurement at 0 weeks
postpartum were collected at least 1 day after delivery to
minimise the postnatal haemodilution effect. Serum ferritin
at 0 week and 6 weeks postpartum were checked in
duplicate by an immunoassay method (Catalogue no.:
L1.4023 M SEROZYME-M Ferritin Assay kit, IFCI Clonesystems, Spain).
At 0 week postpartum, anaemia was de®ned as a Hb
level below 10 g=dl due to the effect of haemodilution
(Taylor, 1981). This de®nition was also used in a previous
Hong Kong study (Lao et al, 1996). At 6 weeks, 3 months
and 6 months postpartum, anaemia was de®ned as an Hb
level below 12 g=dl, this being the cut-off value used for
Table 1
Demographic data of Hong Kong postpartum Chinese women classi®ed as anaemic and non-anaemic postpartum
Anaemic groupa
n
Age (y)
Infant's gestational age (week)
Day ®rst postpartum mense returned
Numbers of pregnancies
Parity
Placental weight (g)
Haemoglobin level at bookingc (g=dl)
Day baseline postpartum blood taken
Height (cm)
Baseline weight (kg)
Mode of delivery (%)
Normal
Instrumental
Caesarean section
Infant feeding practice (%)
Bottle-feeding
Lactating < 3 months
Lactating 3 months
a
Mean (s.d.)
30.8
40.2
69.6
2.0
1.5
576.2
12.5
4
157.3
62.8
(4.0)
(1.4)
(24.1)
(1.2)
(0.8)
(102.3)
(0.8)
(1.5)
(6.6)
(9.2)
n
Mean (s.d.)
34
34
32
34
34
34
31
34
33
32
30.4
39.6
96.5
1.9
1.5
524.0
12.8
3.7
157.4
59.3
(3.2)
(1.3)
(76.1)
(1.0)
(0.6)
(94.9)
(0.8)
(1.3)
(4.7)
(7.0)
9 (69.2)
3 (23.1)
1 (7.7)
26 (76.5)
7 (20.6)
1 (2.9)
5 (38.5)
6 (46.2)
2 (15.4)
12 (35.3)
13 (38.2)
9 (26.5)
De®nition of anaemia according to baseline Hb level (anaemic group: Hb level< 10 g=dl; non-anaemic group: Hb level 10 g=dl).
Tested by independent Student t-test.
Hb level at ®rst or second trimesters of pregnancy.
b
c
13
13
13
13
13
13
13
13
13
13
Non-anaemic groupa
European Journal of Clinical Nutrition
Pb
0.72
0.14
0.41
0.87
0.86
0.11
0.19
0.52
0.96
0.16
0.71
0.84
Postnatal iron status
SM Chan et al
traditional special dietary practices are followed. Potential
descriptive variables and dietary variables were entered
into the model for regression analysis. The statistically
signi®cant level was set at P < 0.05.
541
Table 2 Changes in haemoglobin (Hb) and mean cell volume (MCV)
levels in anaemic group (Hb < 10 g=dl at baseline) and non-anaemic
group (Hb 10 g=dl at baseline)
Anaemic group
n
Results
Descriptive data of subjects
Sixty-three subjects were recruited in the study. Fortyseven (74.5%) subjects (13 anaemic, 34 non-anaemic)
had complete data at both 0 week and 6 weeks postpartum.
Eleven anaemic subjects and 29 non-anaemic subjects
completed the 6 month interview. Two subjects were
pregnant again before the 6 month interview and were
excluded. No signi®cant differences in demographic data
were detected between those excluded and those who
remained in the study.
Demographic data for the anaemic group and the nonanaemic group are summarised in Table 1. Maternal age,
gestational age, time of return of ®rst postpartum mense,
number of pregnancies, parity, placental weight, antenatal
haemoglobin level and anthropometric data were similar
between the two groups. The anaemic group showed a
signi®cantly (P < 0.001) greater amount of estimated blood
loss at delivery, 400 (s.d. 170) ml vs 196 (s.d. 73) ml, and a
higher rate of primary postpartum haemorrhage, 6 (46%) vs
0 (P < 0.001). The majority of subjects in both groups had
normal deliveries or instrumental vaginal deliveries (Table
1). The distribution of different infant feeding practices was
also similar and no subject received a blood transfusion.
Rate of anaemia at different postpartum periods
At baseline 13=47 (27.7%) subjects had Hb < 10 g=dl. At the
same time period 57.7% of mothers had a Hb level of
< 11 g=dl and 74.5% < 12 g=dl. Seven of 47 (14.9%) subjects
Figure 1
Mean (s.d.)
Non-anaemic group
n
Pa
Mean (s.d.)
Hb (g=dL)
Baseline
6 weeks
3 months
6 months
13
13
12
10
9.1
12.5
12.6
12.6
(0.8)
(1.2)
(1.0)
(0.9)
34
34
28
24
11.5
13.2
13.2
13.2
(0.9)
(1.0)
(0.8)
(1.2)
< 0.001
0.03
0.03
0.17
MCV (f l)
Baseline
6 weeks
3 months
6 months
13
13
12
10
87.6
85.6
85.1
87.8
(7.0)
(6.5)
(5.4)
(5.9)
34
34
28
24
91.0
89.1
89.0
90.0
(5.4)
(4.9)
(3.9)
(3.6)
0.09
0.05
0.01
0.29
a
Tested by independent Student t-test.
showed a Hb < 12 g=dl at 6 weeks, 2=40 (5%) at 3 months and
2=34 (6.3%) at 6 months postpartum (Figure 1). Two subjects,
who were anaemic at baseline, were still anaemic (Hb
< 12 g=dl) at 3 months and 6 months postpartum.
Changes of CBC parameters at different postpartum
periods
he postpartum changes of Hb and MCV levels in the
anaemic and the non-anaemic groups are summarised in
Table 2. The anaemic group showed consistently lower Hb
levels as compared to the non-anaemic group throughout
the study and differences were signi®cant during the ®rst 3
months postpartum. No signi®cant differences in the MCV
level were observed between the anaemic group and the
non-anaemic group and a similar trend of MCV change was
found in both groups. However, the anaemic group showed
a lower MCV level than the non-anaemic group throughout
the study.
Postpartum changes in the percentage of subjects with different haemoglobin cut-off values.
European Journal of Clinical Nutrition
Postnatal iron status
SM Chan et al
542
Table 3 Iron status of anaemic group (haemoglobin < 10 g=dl at baseline) and non-anaemic group (haemoglobin 10 g=dl at baseline) at baseline and 6
weeks postpartum
Baselinea
Iron de®ciency anaemia
Anaemia with elevated iron storeb
Iron storage depletion
Iron de®ciency excluded
a
Hb (g=dl)
Ferritin (mg=l)
< 10
< 10
10
10
< 20
20
< 12
12
At 6 weeks postpartum
Anaemic
(n ˆ 12)
7
5
0
0
(58.3)
(41.7)
(0)
(0)
Non-anaemic
(n ˆ 32)
0
0
12
20
Hb (g=dl)
Ferritin (mg=l)
< 12
< 12
12
12
< 20
20
< 12
12
(0)
(0)
(37.5)
(62.5)
Anaemic
(n ˆ 13)
3
1
3
6
(23.1)
(7.6)
(23.1)
(46.2)
Non-anaemic
(n ˆ 34)
3
0
6
25
(8.8)
(0)
(17.7)
(73.5)
Three baseline data were missing due to some technical problems.
Elevated iron store due to either acute blood loss, infection or wound in¯ammation.
b
Percentage of iron de®ciency and iron de®ciency anaemia
at baseline and six weeks postpartum
The iron status of subjects is summarised in Table 3. At
baseline, seven subjects were found to have iron de®ciency
anaemia. An elevated iron store, due to acute blood loss,
infection or in¯ammation, was found in ®ve anaemic
subjects. About one-third of the non-anaemic subjects had
depleted iron stores. At 6 weeks postpartum, the iron status
of most of the subjects in both groups had returned to
normal (normal haemoglobin and serum ferritin levels).
Within the anaemic group, most subjects showed improved
iron status. However, 3=13 subjects still showed iron
de®ciency anaemia at 6 weeks postpartum. For the nonanaemic group, some subjects who had depleted iron stores
at baseline had returned to normal iron status (ie Hb level
12 g=dl and serum ferritin level 12 mg=l) by 6 weeks.
However three of these subjects, who were not anaemic
at baseline, had iron de®ciency anaemia at 6 weeks postpartum.
Food intake at different postpartum periods
Daily consumption of different food groups by the anaemic
group and the non-anaemic group are summarised in Table
4. Subjects in both groups showed similar daily consumption of poultry, egg, meat, ®sh, seafood and vegetables over
the whole study. In the early postpartum period, subjects
consumed two to threefold higher amounts of poultry and
eggs than the Hong Kong general population. The amount
then decreased gradually to a level similar to that of the
Hong Kong general population after 6 weeks postpartum. A
consistent amount of meat (nearly 100 g=day) was consumed
by the subjects throughout the study and this level was
comparable to that of the Hong Kong general population.
Table 4 Daily food intake (g=day) of the anaemic group (haemoglobin < 10 g=dl at baseline) and the nonanaemic group (haemoglobin 10 g=dl at baseline) at different postpartum periods
Anaemic
n
Non-anaemic
Mean (s.d.)
n
Mean (s.d.)
n
Meanb
Pa
Poultry
Baseline
2 weeks
6 weeks
3 months
6 months
12
13
13
12
11
153.8
169.3
67.1
46.7
61.1
(207.4)
(180.5)
(106.4)
(36.5)
(51.2)
33
32
34
29
27
103.1
185.6
73.2
51.7
50.7
(138.4)
(156.6)
(71.8)
(60.5)
(40.2)
1010
Ð
Ð
Ð
Ð
59
0.91
0.61
0.42
0.68
0.59
Egg
Baseline
2 weeks
6 weeks
3 months
6 months
12
13
13
12
11
56.3
50.4
28.6
16.3
26.6
(45.4)
(48.9)
(36.2)
(14.6)
(36.9)
33
32
34
29
27
44.9
55.6
35.5
26.8
23.5
(33.7)
(38.1)
(26.2)
(23.9)
(25.1)
1010
Ð
Ð
Ð
Ð
19
0.45
0.49
0.22
0.25
0.98
Meat
Baseline
2 weeks
6 weeks
3 months
6 months
12
13
13
12
11
98.8
117.4
120.0
89.4
99.6
(83.1)
(88.6)
(81.4)
(42.7)
(56.9)
33
32
34
29
27
108.9
89.7
105.8
103.4
126.0
(77.9)
(71.5)
(87.5)
(69.0)
(75.3)
1010
Ð
Ð
Ð
Ð
92
0.60
0.37
0.60
0.79
0.39
a
Tested by Mann ± Whitney U-test.
Mean value of 1010 (500 M, 510 F) Hong Kong adults (Woo et al, 1999).
b
European Journal of Clinical Nutrition
HK population
Postnatal iron status
SM Chan et al
Table 5 Daily iron and vitamin C intake of the anaemic group (baseline haemoglobin < 10 g=dl) and the non-anaemic group (baseline haemoglobin
10 g=dl) compared to the Hong Kong (HK) females and US Recommended Daily Allowance (RDA) at different postpartum periods
Anaemic
n
Mean (s.d.)
Median [IR]c
Iron intake (mg=day)
Baseline
12
14.7
13.8
2 weeks
13
19.2
14.7
6 weeks
13
23.5
13.1
3 months
12
14.1
13.9
6 months
11
22.3
14.0
(5.9)
[9.9 ± 19.9]
(13.5)
[10.4 ± 25.2]
(30.1)
[8.4 ± 21.0]
(5.9)
[9.1 ± 19.7]
(30.7)
[9.6 ± 17.9]
Vitamin C intake (mg=day)
Baseline
12
122.2 (107.5)
115.4 [27.3 ± 180.6]
2 weeks
13
99.4 (83.8)
89.8 [36.0 ± 164.1]
6 weeks
13
107.3 (77.1)
103.1 [35.5 ± 157.4]
3 months
12
89.1 (66.8)
80.0 [27.3 ± 141.3]
6 months
11
87.6 (58.0)
66.7 [39.6 ± 120.8]
HK females (25 ± 34 y)a
Non-anaemic
Percentage
of subjects
60% RDAd
n
83
33
92
32
69
34
75
29
73
27
75
33
62
32
77
34
67
29
64
27
Mean (s.d.)
Median [IR]c
15.0
14.0
18.9
15.0
16.4
15.2
18.8
14.8
19.2
15.9
(6.9)
[9.7 ± 17.8]
(10.9)
[11.6 ± 23.11]
(10.0)
[9.4 ± 22.0]
(12.8)
[10.3 ± 21.4]
(12.5)
[11.9 ± 21.0]
116.2
80.6
140.4
144.1
116.5
97.4
122.7
107.7
170.3
103.6
(99.1)
[49.2 ± 184.1]
(97.1)
[62.6 ± 179.9]
(82.0)
[52.9 ± 151.1]
(89.4)
[52.0 ± 180.9]
(213.6)
[66.0 ± 172.7]
543
Percentage
of subjects
60% RDAd
n
79
106
100
Ð
0.69
79
Ð
0.78
86
Ð
0.42
85
Ð
0.34
73
106
81
Ð
0.20
88
Ð
0.72
79
Ð
0.23
81
Ð
0.28
Mean (s.d.)
14.6 (5.3)
139.5 (79.9)
Pb
0.91
0.93
a
Source: Hong Kong Adult Dietary Survey 1995 (Leung et al, 1997).
Tested by Mann ± Whitney U-test.
c
Median and interquartile range [IR].
d
Source: US RDA 10th edition (National Research Council, 1989). US RDA for women aged 25 ± 50 y was used to compare with data from bottle-feeding
subjects, 60% of RDA for iron and vitamin C are 9 and 36 mg, respectively. US RDA for lactating women of ®rst 6 months lactation was used to compare
with data of all lactating (< 3 months or 3 months) subjects; 60% of RDA for iron and vitamin C are 9 and 57 mg, respectively.
b
In contrast, the subjects consumed less ®sh and seafood
during the early postpartum period. At baseline, mean daily
®sh and seafood intake was 29.2 and 9.6 g=day, respectively, for the anaemic and non-anaemic groups. The
amount then increased and remained at about 90 g=day in
both groups from 6 weeks to 6 months postpartum, which
was comparable to that of the Hong Kong general
population (93 g=day; details not shown). No signi®cant
differences in vegetables, legumes and fruit consumption
were observed between the non-anaemic and anaemic
groups. Mean consumption decreased from 425 to
236 g=day for the anaemic group and 352 to 280 g=day
for the non-anaemic group from baseline to 6 months
postpartum. This amount was low as compared to that of
the Hong Kong general population (530 g=day; details not
shown).
age, our subjects showed a comparable to higher daily iron
intake over the whole study. Iron intake in the majority of
subjects reached at least 60% ( 9 mg=day) of the US
Recommended Daily Allowance (RDA).
Although the mean vitamin C intake decreased from
around 122 to 90 mg=day in the anaemic group and
increased from about 120 to 170 mg=day in the nonanaemic group, no signi®cant differences were observed
between the two groups. Generally, both groups had a
lower vitamin C intake as compared with that of Hong
Kong females. However, more than two-thirds of subjects
in both groups reached 60% of the recommended daily
vitamin C intake.
Table 6 Predictors included in the stepwise regression model to predict
the rate of change in Hb level from baseline to 6 weeks postpartum
Nutrient intake at different postpartum periods
Daily iron and vitamin C intakes of subjects at different
postpartum periods are summarised in Table 5. Dietary
supplements were not common in our subjects. At baseline,
both groups of subjects showed a daily iron intake of about
15 mg. This level increased gradually in the ®rst 6 months
postpartum. Compared with Hong Kong females of similar
95% CI for B-value
Predictor
B-value
P-value
Lower
bound
Upper
bound
Baseline Hb
Baseline MCV
Baseline serum ferritin
ÿ0.458
0.046
0.016
0.000
0.003
0.020
ÿ0.586
0.016
0.003
ÿ0.330
0.075
0.029
r2
0.647
European Journal of Clinical Nutrition
Postnatal iron status
SM Chan et al
544
Determinants of the rate of change in Hb level from
baseline to 6 weeks postpartum
Results of stepwise regression indicated that baseline MCV
level and serum ferritin level were positively correlated
whereas baseline MCV level was negatively correlated
with the rate of change in Hb level in the ®rst 6 weeks
postpartum. These three variables explained about 65% of
the variability in the rate of Hb change during this period.
Other descriptive and dietary variables including dietary
intake of different nutrients and the consumption of the
traditional special dietary practices did not explain the
change in Hb in this period (Table 6).
Discussion
Data on postnatal iron nutritional status are limited in the
literature. Most published studies on this topic were conducted some time ago and focused on either biochemical
changes or dietary changes postpartum (Mbofung &
Atinmo, 1985; Murray et al, 1978; Powers et al, 1985).
Few studies have investigated determinants of anaemia
such as dietary factors during the postpartum period. The
present study provides both biochemical and dietary data of
the postpartum iron nutritional status of Hong Kong Chinese women, a population in whom nutritional deprivation
is uncommon.
There was limited information on appropriate cut-off
values to de®ne anaemia postpartum. In contrast, cut-off
values of haemoglobin level to de®ne anaemia are well
established for pregnant, non-pregnant and non-lactating
women. For pregnant women, haemoglobin cut-off levels
to de®ne anaemia in the ®rst, second and third trimesters
are 11, 10.5 and 11 g=dl, respectively (Institute of Medicine
(US), 1993; National Research Council, 1989). For nonpregnant and non-lactating women, a haemoglobin cut-off
of 12 g=dl is used to de®ne anaemia (Cook & Skikne, 1989;
Institute of Medicine (US), 1993). Previous studies have
used Hb cut-off values of either 10 g=dl (Lao et al, 1996),
11 g=dl (Singh et al, 1998) or 12 g=dl (Kuvibidila et al,
1994). Although it has been suggested that the haemodilution effect of pregnancy is still apparent at 6 ± 8 weeks
postpartum (Taylor & Lind, 1979), our ®ndings showed
that the mean Hb level of subjects had increased to
12 g=dl by 6 weeks postpartum and remained fairly
constant until 6 months postpartum. A Hb cut-off of
12 g=dl was chosen to de®ne anaemia at 6 weeks, 3 and 6
months postpartum, whereas a Hb cut-off of 10 g=dl was
used at baseline. Using this cut-off, 27.7% of our subjects
were anaemic at baseline. This rate is high as compared
with that (12%) reported in another Hong Kong study using
the same de®nition of anaemia (Lao et al, 1996). Differences in the study design may explain this discrepancy. In
Lao's study, maternal haemoglobin level and other biochemical markers of iron status were measured throughout
pregnancy and during the postpartum period. All subjects
included in the ®nal analysis had no blood disorders and
European Journal of Clinical Nutrition
were non-anaemic (Hb 10 g=dl) at or after 28 ± 30 weeks
of gestation. In our study, haematological data were only
checked during the ®rst antenatal booking visit which was
usually in the ®rst or the second trimester, as late booking is
uncommon. Only Hb and MCV levels were checked at this
®rst antenatal check-up and assessment of iron status
(serum iron, ferritin and total iron binding capacity)
would only be undertaken when antenatal anaemia was
suspected. Some of our subjects may have had normal Hb
levels at booking but then may have developed iron
depletion (iron de®ciency with or without anaemia)
during late pregnancy. Blood loss at delivery might have
further aggravated iron depletion and caused iron de®ciency anaemia in some of our subjects (Lao et al, 1996).
Our data showed that the anaemic group had greater
blood loss and a higher rate of primary postpartum haemorrhage than the non-anaemic group. As the antenatal Hb
level at booking did not differ signi®cantly between the two
groups, these results suggest that blood loss at delivery is
likely to be the major cause of postpartum anaemia. This
supports the ®ndings of a previous study (Lao et al, 1996).
The effect of greater amount of blood loss and a higher rate
of primary postpartum haemorrhage on the postnatal iron
status of the anaemic group can also be re¯ected by the
consistently lower Hb and MCV levels in the anaemic
group compared to the non-anaemic group throughout the
study. Moreover, our data also showed that two subjects
who were anaemic at baseline were still anaemic at 3 and 6
months postpartum.
Both physiological factors and dietary factors may
explain the postpartum change in the Hb and MCV
levels, and the iron status in our subjects. It has been
shown that physiological adaptation occurs after blood
loss at delivery with an increase in iron absorption in
postpartum women. Iron status of an individual is the
main factor affecting iron absorption (Monsen et al,
1978). Absorption increases as iron stores become depleted
and decreases as stores increase (Herbert, 1987). Iron
absorption in women at 10 weeks postpartum was found
to be double that of a non-pregnant control group (Hallberg, 1992). This may be one of the adapting mechanisms
of postpartum women to restore iron status after delivery.
Moreover, the slow rise in postpartum haematological
indices (Taylor, 1981) resulting from the gradual reduction
in the intrapartum blood volume (Ueland, 1976) suggests
that there is a continuous regulatory mechanism for women
to restore their postnatal iron status. In addition, menstruation is rare in the early postpartum period which will help
restore postnatal iron status after delivery.
The effects of different dietary components on iron
status or iron absorption have been studied extensively.
However, few studies have speci®cally focused on postpartum women. Consumption of meat, poultry and ®sh may
promote non-haem iron absorption (Engelmann et al, 1998;
Hallberg, 1987). The presence of organic acids such as
ascorbic acid, citric acid, malic acid and lactic acid as well
as the presence of alcohol also place an enhancing effect on
Postnatal iron status
SM Chan et al
non-haem iron absorption (Ballot et al, 1987; Bothwell
et al, 1989; Hallberg, 1982; Rossander-Hulthen & Hallberg, 1996). These dietary factors, which positively contribute to the iron status, are similar to those found in the
traditional special dietary practices, used in the Hong Kong
Chinese population. We therefore speculated that these
practices might be bene®cial for postnatal recovery of
iron status. A higher consumption of poultry and eggs,
and a comparable amount of meat consumption in our
subjects as compared with the Hong Kong general population was observed. The iron intake of the majority of our
subjects reached two-third of the RDA. This re¯ects that
the traditional special dietary practices provide a good
source of iron. However, whether the anaemic subjects
bene®t more from these dietary practices remains to be
identi®ed. There was no signi®cant correlation between the
rate of change in Hb level from baseline to 6 weeks
postpartum and the consumption of traditional special
dietary practices and other nutrient intakes, suggesting
that diet may have limited effect on the postnatal recovery
of iron status. Regression analysis thus suggested that the
postnatal recovery of iron status appeared to be related to
physiological factors rather than to dietary factors. However to exclude a role for traditional dietary practices in the
recovery of postnatal iron status, further study is needed
due to the limitation of relatively small sample size in this
study and the fact that there was wide variation in the
extent to which these traditional practices were followed by
different subjects. Ideally any bene®cial effect of these
practices should be studied using a randomised control
design, but conducting such a trial would be dif®cult to
undertake, not least because most Hong Kong Chinese
women still follow these practices to some extent.
Several limitations should be considered in the present
study. Although the measurement of both Hb and serum
ferritin levels is useful for diagnosing iron de®ciency
anaemia (Cook & Skikne, 1989), there are some drawbacks.
First, low serum ferritin concentration is a sensitive indicator of iron de®ciency only if there are no other concurrent
diseases. Serum ferritin concentration may be elevated in
patients with chronic diseases and in such patients, who also
have iron de®ciency, serum ferritin concentration is often
normal or elevated (Cook & Skikne, 1989). Iron de®ciency
anaemia may thus be under-diagnosed in these patients
(Cook et al, 1992; Institute of Medicine (US), 1993).
However, our exclusion criteria should have excluded any
such patients. Second, the method of estimating blood loss
at delivery in the present study was semi-quantitative.
However there is no reason to suspect that there was any
systematic bias in estimating blood loss in the anaemic and
non-anaemic groups.
Conclusions
Although the rate of anaemia decreased rapidly postpartum,
iron storage depletion were still found in some subjects at 6
weeks postpartum. Blood loss at delivery is likely to be the
main cause of postpartum anaemia. Although traditional
Chinese dietary practices in the postpartum period prescribe a high intake of poultry and eggs with a resultant
increased iron intake, our data suggested that the postnatal
recovery of iron status might be more related to physiological factors than dietary factors. However a role for these
dietary factors cannot be ruled out in view of the signi®cant
variation in the consumption of these food items and the
relatively small sample size of this study.
545
Acknowledgements ÐWe thank all the subjects for their commitment in the
study; Jannis WS Yau and Roselle HY Li for their assistance in recruiting
and scheduling subjects; and Peggo KW Lam for his valuable advice on
appropriate methods for statistical analysis.
References
Anonymous (1985): Summary of a report on assessment of the iron
nutritional status of the United States population. Expert Scienti®c
Working Group. Am. J. Clin. Nutr. 42, 1318 ± 1330.
Ballot D, Baynes RD, Bothwell TH, Gillooly M, MacFarlane, BJ,
MacPhail AP, Lyons G, Derman DP, Bezwoda WR & Torrance JD
(1987): The effects of fruit juices and fruits on the absorption of iron
from a rice meal. Br. J. Nutr. 57, 331 ± 343.
Beard J, Green W, Miller L & Finch C (1984): Effect of iron-de®ciency
anemia on hormone levels and thermoregulation during cold exposure.
Am. J. Physiol. 247, R114 ± R119.
Bothwell TH, Baynes RD, MacFarlane BJ & MacPhail AP (1989):
Nutritional iron requirements and food iron absorption. J. Intern.
Med. 226, 357 ± 365.
Chan SM, Nelson EAS, Leung SSF, Cheung PCK & Li CY (2000): Special
postpartum dietary practices of Hong Kong Chinese women. Eur.
J. Clin. Nutr. 54, 797 ± 802.
Church CF & Church HN (1975): Food Values of Portions Commonly
Used. Philadelphia: JB Lippincott.
Cook JD (1990): Adaptation in iron metabolism. Am. J. Clin. Nutr. 51,
301 ± 308.
Cook JD & Skikne BS (1989): Iron de®ciency: de®nition and diagnosis.
J. Intern. Med. 226, 349 ± 355.
Cook JD, Skikne BS & Baynes RD (1992): Screening strategies for
nutritional iron de®ciency. In Nutritional Anemias, ed. SJ Fomon &
S. Zlotkin. New York: Raven Press.
Cook JD, Skikne BS & Baynes RD (1994): Iron de®ciency: the global
perspective. Adv. Exp. Med. Biol. 356, 219 ± 228.
Dallman PR (1987): Iron de®ciency and the immune response. Am. J. Clin.
Nutr. 46, 329 ± 334.
Duthie SJ, Ven D, Yung GLK, Guang DZ, Chan SYW & Ma HK (1990):
Discrepancy between laboratory determination and visual estimation
of blood loss during normal delivery. Eur. J. Obstet. Gynecol 38,
119 ± 124.
Engelmann MD, Davidsson L, Sandstrom B, Walczyk T, Hurrell RF &
Michaelsen KF (1998): The in¯uence of meat on nonheme iron
absorption in infants. Pediatr. Res. 43, 768 ± 773.
Great Britain Panel on Dietary Reference Values of the Committee on
Medical Aspects of Food Policy (1991): Dietary reference values for
food energy and nutrients for the United Kingdom: report of the Panel
on Dietary Reference Values of the Committee on medical aspects of
Food Policy. London: HMSO.
Hallberg L (1982): Iron absorption and iron de®ciency. Hum. Nutr. Clin.
Nutr. 36, 259 ± 278.
Hallberg L (1987): Wheat ®ber, phytates and iron absorption. Scand. J.
Gastroenterol. 129(Suppl), 73 ± 79.
European Journal of Clinical Nutrition
Postnatal iron status
SM Chan et al
546
Hallberg L (1992): Iron balance in pregnancy and lactation. In Nestle
Nutrition Workshop Series: Nutritional Anemias, eds. SJ Fomon & S
Zlotkin. New York: Raven Press.
Hayashi RH (1998): Postpartum Hemorrhage and Puerperal Sepsis. In:
Essentials of Obstetrics and Gynecology, ed. NF Hacker & JG Moore.
Philadelphia, PA: WB Saunders.
Herbert V (1987): Recommended dietary intakes (RDI) of iron in humans.
Am. J. Clin. Nutr. 45, 679 ± 686.
Institute of Health (1980): Food Composition Table. Beijing: Chinese
People's Health Publishing Co.
Institute of Medicine (US) (1990): Nutrition During Pregnancy. Washington, DC: National Academy Press.
Institute of Medicine (US) (1993): Iron De®ciency Anemia: Recommended
Guidelines for the Prevention, Detection, and Management Among U.S.
Children and Women of Childbearing Age. Washington, DC: National
Academy Press.
Kuvibidila S, Yu LC, Ode DL, Warrier RP & Mbele V (1994): Assessment
of iron status of Zairean women of childbearing age by serum
transferrin receptor. Am. J. Clin. Nutr. 60, 603 ± 609.
Labbok M & Krasovec K (1990): Toward consistency in breastfeeding
de®nitions. Stud. Family Plann. 21, 226 ± 230.
Lao TT, Lee CP & Mak WP (1996): Postpartum anaemia is not related to
maternal iron status in the third trimester. Eur. J. Obstet. Gynecol. 64,
7 ± 10.
Leung SSF, Ho S, Woo J, Lam TH & Janus ED (1997): Hong Kong
Adult Dietary Survey 1995. Hong Kong: Chinese University of Hong
Kong.
Lozoff B, Klein NK, Nelson EC, McClish DK, Manuel M & Chacon ME
(1998): Behavior of infants with iron-de®ciency anemia. Child Devl. 69,
24 ± 36.
Mbofung CM & Atinmo T (1985): Zinc, copper and iron concentrations in
the plasma and diets of lactating Nigerian women. Br. J. Nutr. 53,
427 ± 439.
McCance RA, Widdowson EM, Paul AA & Southgate DAT (1978):
McCance and Widdowson's: the Composition of Foods. London: HM
Stationery Of®ce.
European Journal of Clinical Nutrition
Moffatt ME, Longstaffe S, Besant J & Dureski C (1994): Prevention of
iron de®ciency and psychomotor decline in high-risk infants through
use of iron-forti®ed infant formula: a randomised clinical trial. J.
Pediatr. 125, 527 ± 534.
Monsen ER, Hallberg L, Layrisse M, Hegsted DM, Cook JD, Mertz W,
Finch & CA (1978): Estimation of available dietary iron. Am. J. Clin.
Nutr. 31, 134 ± 141.
Murray MJ, Murray AB, Murray NJ & Murray MB (1978): The effect of iron
status of Nigerian mothers on that of their infants at birth and 6 months, and
on the concentration of Fe in breast milk. Br. J. Nutr. 39, 627 ± 630.
National Research Council FaNB (1989): Recommended Dietary Allowances. Washington, DC: National Academy Press.
Powers HJ, Bates CJ & Lamb WH (1985): Haematological response to
supplements of iron and ribo¯avin to pregnant and lactating women in
rural Gambia. Hum. Nutr. Clin. Nutr. 39, 117 ± 129.
Rossander-Hulthen L & Hallberg L (1996): Dietary factors in¯uencing
iron absorption Ð an overview. In Iron Nutrition in Health and Disease,
ed. L Hallberg & NG Asp. London: Libbey.
Singh K, Fong YF & Arulkumaran S (1998): Anaemia in pregnancy Ð a
cross-sectional study in Singapore. Eur. J. Clin. Nutr. 52, 65 ± 70.
Taylor DJ (1981): Puerperal haematological indices. Br. J. Obstet. Gynaecol. 88, 601 ± 606.
Taylor DJ & Lind T (1979): Red cell mass during and after normal
pregnancy. Br. J. Obstet. Gynaecol. 86, 364 ± 370.
Tung TC, Huang PC & Li HC (1961): Composition of foods used in
Taiwan. J. Formos. Med. Assoc. 60, 973 ± 1005.
Ueland K (1976): Maternal cardiovascular dynamics. VII. Intrapartum
blood volume changes. Am. J. Obstet. Gynecol. 126, 671 ± 677.
US Department of Health (1972): Food Composition Table for Use in
South East Asia. Washington, DC: US Department of Health.
Viteri FE (1997): Iron supplementation for the control of iron de®ciency in
populations at risk. Nutr. Rev. 55, 195 ± 209.
Woo J, Leung SSF, Ho SC & Chan SM (1999): Is there a typical Chinese diet
and what are the health implications? Ecol. Food Nutr. 38, 491 ± 503.
Yip R & Dallman PR (1996): Iron. In Present Knowledge in Nutrition, ed.
EE Ziegler & JL Filer. Washington, DC: ILSI Press.