American Journal of Epidemiology

The Author 2007. Published by the Johns Hopkins Bloomberg School of Public Health.
All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.

Vol. 166, No. 3

DOI: 10.1093/aje/kwm078
Advance Access publication May 2, 2007

Original Contribution
Preconception B-Vitamin and Homocysteine Status, Conception, and Early
Pregnancy Loss

Alayne G. Ronnenberg1, Scott A. Venners2, Xiping Xu2, Changzhong Chen3, Lihua Wang4, Wenwei
Guang4, Aiqun Huang4, and Xiaobin Wang5,6
1

Department of Nutrition, University of Massachusetts Amherst, Amherst, MA.

Division of Epidemiology and Biostatistics, School of Public Health, University of Illinois, Chicago, IL.
3
Channing Laboratory, Brigham and Womens Hospital, Harvard Medical School, Boston, MA.
4
Institute for Biomedicine, Anhui Medical University, Hefei, Peoples Republic of China.
5
Mary Ann and J. Milburn Smith Child Health Research Program, Department of Pediatrics, Feinberg School of Medicine,
Northwestern University, Chicago, IL.
6
Childrens Memorial Hospital and Childrens Memorial Research Center, Chicago, IL.
2

Maternal vitamin status contributes to clinical spontaneous abortion, but the role of B-vitamin and homocysteine
status in subclinical early pregnancy loss is unknown. Three-hundred sixty-four textile workers from Anqing, China,
who conceived at least once during prospective observation (19961998), provided daily urine specimens for up to
1 year, and urinary human chorionic gonadatropin was assayed to detect conception and early pregnancy loss.
Homocysteine, folate, and vitamins B6 and B12 were measured in preconception plasma. Relative to women in the
lowest quartile of vitamin B6, those in the third and fourth quartiles had higher adjusted proportional hazard ratios of
conception (hazard ratio (HR) 2.2, 95% condence interval (CI): 1.3, 3.4; HR 1.6, 95% CI: 1.1, 2.3, respectively), and the adjusted odds ratio for early pregnancy loss in conceptive cycles was lower in the fourth quartile
(odds ratio 0.5, 95% CI: 0.3, 1.0). Women with sufcient vitamin B6 had a higher adjusted hazard ratio of
conception (HR 1.4, 95% CI: 1.1, 1.9) and a lower adjusted odds ratio of early pregnancy loss in conceptive
cycles (odds ratio 0.7, 95% CI: 0.4, 1.1) than did women with vitamin B6 deciency. Poor vitamin B6 status
appears to decrease the probability of conception and to contribute to the risk of early pregnancy loss in this
population.
China; chorionic gonadatropin; folic acid; homocysteine; nutritional status; pregnancy outcome; vitamin B 6;
vitamin B 12

Abbreviations: CI, condence interval; hCG, human chorionic gonadotropin; HR, hazard ratio; OR, odds ratio; PLP, pyridoxal
5#-phosphate (vitamin B6).

Several studies indicate that micronutrient deficiencies

may be common among some Chinese women (14). We previously reported a high prevalence of folate and vitamin B6
deficiencies in a group of young women from Anhui province
who were attempting to become pregnant (5). There is growing evidence that maternal micronutrient status contributes

to poor pregnancy outcome and pregnancy loss. Poor folate

status during pregnancy has been associated with numerous
adverse pregnancy outcomes (6, 7), including clinical spontaneous abortion (812). Suboptimal vitamin B6 status (11)
and elevated plasma concentrations of homocysteine, a
marker of poor folate or vitamin B12 status, also have been

Correspondence to Dr. Alayne Ronnenberg, Department of Nutrition, 209 Chenoweth Laboratory, University of Massachusetts, 100 Holdsworth
Way, Amherst, MA 01003 (e-mail: ronnenberg@comcast.net).

304

Am J Epidemiol 2007;166:304312

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

Received for publication April 22, 2004; accepted for publication February 2, 2007.

Preconception Vitamin B6 and Early Pregnancy Loss

MATERIALS AND METHODS

Study population

This study is part of a large prospective study of reproductive health conducted from 1996 to 1998 in textile mills
in Anhui, China. The study protocols were approved by the
human subjects committees of the affiliated Chinese institutions and by the Institutional Review Board of the Harvard
School of Public Health, and all women provided written
informed consent.
A detailed description of the study population and data
collection methods has been reported previously (16). Briefly,
the eligibility criteria for women were as follows: 1) fulltime employment; 2) aged 20
34 years; 3) newly married;
and 4) had obtained permission to have a child. All the
women were nulliparous. Women were excluded if they
were already pregnant before enrollment, had tried unsuccessfully to get pregnant for 1 year or more at any time in the
past, or planned to quit/change jobs or move out of the city
over the 1-year course of follow-up. None of the women had
medical conditions that predisposed them to early pregnancy loss, and the rates of fertility and pregnancy losses
in our study population were similar to those in previous
studies (1517).
Of 1,006 newly married women who were screened (more
than 90 percent of newly married women employed at the
mill), 961 met the eligibility requirements and agreed to
enroll. We excluded 386 enrolled women from this analysis
because they did not collect daily urine (n 121), did not
begin collecting urine soon after stopping contraception
(n 53), never stopped using contraception (n 95), became pregnant due to contraceptive failure (n 78), were
lost to follow-up (n 8), withdrew shortly after enrollment
(n 27), or had inadequate diary data (n 4). The 386
women excluded for the previous reasons were similar to the
remaining enrolled women (16). Of the remaining 575
women, 193 were excluded because they did not have baseline nutritional biomarker data, three because they reported
current smoking or alcohol use, and 15 because they never
conceived and so were not at risk for pregnancy losses. The
Am J Epidemiol 2007;166:304312

current analysis is based on 1,165 menstrual cycles from the

remaining 364 women after exclusion of 99 cycles in which
there was no intercourse reported in a daily diary. Compared
with the 364 women included in this analysis, the 211 women
excluded because of missing data on nutrition, smoking, alcohol, or failure to conceive were similar in terms of age,
height, weight, and body mass index (weight (kg)/height (m)2)
but were more educated (45 percent with a high school education or beyond vs. 30 percent).
Data collection procedures

Womens height and body weight in light clothing were

measured to the nearest 0.1 cm and 0.1 kg, respectively. An
interviewer administered a baseline questionnaire that collected historical data on reproductive history, sociodemographic characteristics, alcohol use, and environmental and
occupational exposures. A follow-up questionnaire was administered trimonthly for women who had not become pregnant and at the end of the first, second, and third trimesters
of pregnancy for women who became pregnant. If a woman
reported a missed or late period or had early signs/symptoms
of pregnancy, she was instructed to have a check-up at the
affiliated hospital and to provide a urine sample for hCG
assay. Once pregnancy was confirmed, the woman received
regular prenatal care, delivery services, and postnatal care
and was followed up by staff at the designated hospitals
according to standard clinical guidelines.
Beginning with the date that contraceptive use ended,
each woman kept a daily diary to record sexual intercourse,
vaginal bleeding, medication, and medical conditions. She
also collected a daily morning first-urine specimen for hCG
assay. Daily diary information and urine specimens were collected for up to 12 months or until a pregnancy was clinically
confirmed. The first of two consecutive days of bleeding
reported in the daily diary was defined as the first day of a
menstrual cycle. Women were monitored during ensuing
pregnancies (or up to 1 year after beginning to attempt pregnancy if no pregnancy occurred), and all pregnancy outcomes were recorded.
Major outcomes and method of evaluation

The outcomes in this study were defined as follows.

1. Conception: a conception detected by urinary hCG assay.
To distinguish normal variation from a true hCG rise due
to conception and to address missing hCG values, we
used Bayesian methods (18, 19) to model daily conception status among all the female subjects, including
female controls who were not at risk of conception. We
showed that this model was 100 percent sensitive and
specific for those cycles in which the conception status
was observable; that is, the probability of conception
was 0.0 in all control cycles and 1.0 in all cycles with
conception leading to clinical pregnancy (refer to
Laboratory assay of urinary hCG) (16).
2. Clinical pregnancy: any pregnancy that lasted 6 weeks
(42 days) or more after the onset of the last menstrual
period and was confirmed by hCG assay.

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

associated with increased risk of clinical spontaneous abortion (9, 13, 14).
In addition to losses that occur after a pregnancy is clinically recognized, as many as two thirds of losses occur
before a woman knows she is pregnant (15, 16). However,
no studies to date have examined the role of nutritional
status in early pregnancy loss. In the current study, we assessed the associations between preconception B-vitamin
and homocysteine status and the risks of conception, early
pregnancy loss, and clinical pregnancy in a prospective cohort of young Chinese women. To accurately measure conception and early pregnancy loss, we assayed human
chorionic gonadotropin (hCG) in daily urine samples collected from the time a woman stopped using contraception
until she became pregnant. We assayed vitamin biomarkers
in plasma samples collected at baseline, before women stopped using contraception, to assess micronutrient status during the period when they were attempting to conceive.

305

306 Ronnenberg et al.

3. Early pregnancy loss: pregnancy loss detected by urinary hCG assay (refer to definition of conception above)
occurring less than 6 weeks (42 days) after the onset of
the last menstrual period.
Nutritional analyses

Laboratory assay of urinary hCG

Urine specimens were stored at
20
C. Urinary hCG

levels were analyzed per batch by the immunoradiometric
assay developed by OConnor et al. (26) using a combination
of capture antibodies for the hCG free b subunit and hCG b
core fragment (i.e., B204) and the intact hCG molecule
(i.e., B109). This assay is highly sensitive and specific. The
lowest hCG concentration detectable by the assay was
0.01 ng per ml (1 mIU 0.2 ng). The cross-reaction of
the assay with either intact luteinizing hormone or the luteinizing hormone-free b subunit was less than 1 percent. All
urine specimens from each woman were analyzed and tested
during a single run of the assay. Each urine specimen during
the window from
10 to 5 days of a menstrual cycle was
assayed in duplicate. Discrepancies of more than threefold
between duplicate assays were presumed to result from
technical error, and the assay was repeated. For the remainder, the geometric mean of the replicates was used to summarize the results for each sample. Urinary creatinine was
measured by the Jaffe reaction described by Husdan and
Rapoport (27). All hCG values were normalized to creatinine values to adjust for urine concentration. For reference

Statistical analysis

We described important epidemiologic characteristics of

subjects using means and frequencies. We then grouped
plasma concentrations of homocysteine and each B vitamin
(B6, B12, and folate) in two ways: 1) ordinal quartiles of
equal sample sizes and 2) binary categories of normal and
abnormal (refer to definitions of elevated homocysteine and
deficient B vitamins in Nutritional analyses). We used
Cox proportional hazards methods (28) to estimate the hazards of conception and clinical pregnancy 1) in the upper
three quartiles of homocysteine and B vitamins relative to
the lowest quartile and 2) in normal relative to abnormal
binary categories (B vitamins) or abnormal relative to normal binary categories (homocysteine). Because we prospectively observed study women for up to 1 year until they
achieved a clinical pregnancy, women who had early pregnancy losses prior to clinical pregnancy remained in our
cohort and might have had more than one observed conception. We considered the cycles after early pregnancy losses
to be the first cycles of new attempted conceptions. We used
robust variance estimates (29) to accommodate nonindependence of hazards in multiple attempted conceptions from the
same woman. Parameter estimates were similar when we
modeled all attempted conceptions versus only the first observed conception from each woman. There were nine rightcensored attempted conceptions in which conception did not
occur before the end of follow-up and 17 right-censored
clinical pregnancy attempts. Each woman had a maximum
of one observed clinical pregnancy, and the time to clinical
pregnancy was calculated from the beginning of the first
conception attempt. We presented these models with and
without adjustment for womans age, body mass index, history of pregnancy, self-reported stress, working shifts, occupational noise and dust exposures, education, and
husbands age, smoking, and alcohol drinking. In addition,
adjusted models for each B vitamin included the other two
remaining B vitamins.
We next limited our analysis to conception cycles and
used logistic regression to investigate the relative odds of
early pregnancy losses using the same quartile and binary
plasma concentration categories as above for homocysteine
and B vitamins. We first examined early pregnancy losses by
use of only the first observed conception. We then repeated
the analyses using all observed conceptions and estimated
standard errors using generalized estimating equations to
accommodate correlations in pregnancy losses among conceptions from the same women (30). Each model was analyzed with and without adjustment for the same covariates
as used in the proportional hazards models.
RESULTS

This report includes 364 women who conceived at least

once during prospective observation, had adequate daily
Am J Epidemiol 2007;166:304312

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

At the time of the initial interview, nonfasting blood

samples were collected via venipuncture into 10-ml
ethylenediaminetetraacetic acid-treated tubes. The blood
was centrifuged, and plasma was obtained and stored at

20
C until shipped on dry ice to the Harvard School of
Public Health, where it was stored at
70
C prior to nutritional analyses. Frozen samples were then transported to the
Jean Mayer USDA [US Department of Agriculture] Human
Nutrition Research Center on Aging, Tufts University,
Boston, Massachusetts, where plasma concentrations of homocysteine, folate, and vitamins B6 and B12 were measured.
The total homocysteine concentration in plasma was determined by use of a method described by Araki and Sako (20).
Plasma folate and vitamin B12 were determined by radioimmunoassay using a commercially available kit from the
BioRad Diagnostics Group (Hercules, California). Plasma
vitamin B6 (as pyridoxal 5#-phosphate (PLP)) was measured
by the tyrosine decarboxylase apoenzyme method (21).
Plasma vitamin concentrations were compared with published reference values to determine the proportions of
women with biochemical evidence of vitamin deficiency, defined as less than 6.8 nmol/liter (3 ng/ml) for folate (22), less
than 30 nmol/liter of PLP for vitamin B6 (23), and less than
258 pmol/liter (350 pg/ml) for vitamin B12 (24, 25). There is
no standard definition of elevated homocysteine. For these
analyses, we defined elevated homocysteine as a plasma
concentration of 12.4 lmol/liter or greater, which is consistent with the cutoff used in a previous analysis in this
cohort (22).

values, we determined levels of hCG from 67 nonconception cycles of 37 control women who were married but using
contraception (n 4), not married (n 23), or married but
not cohabitating with their husbands (n 10) (16).

Preconception Vitamin B6 and Early Pregnancy Loss

Am J Epidemiol 2007;166:304312

TABLE 1. Characteristics of 364 nulliparous female textile

workers in Anhui, China, who conceived at least once during
prospective observation, 19961998
Mean

Standard
deviation

Age (years)

24.9

Height (m)

1.58

1.5
0.05

Weight (kg)

49.1

6.0

Body mass index (kg/m2)

19.8

2.1

Age at menarche (years)

14.8

1.4

Average cycle length (days)

29

Vitamin B6 (nmol/liter)

39.3

12.8

Vitamin B12 (pmol/liter)

363

122

Folate (nmol/liter)

9.8

3.9

Homocysteine, lmol/liter

8.6

3.4

No.

Middle school or below

258

71

High school

104

29

College or above

002

01

Education

Self-reported level of stress in life

None or low

213

59

Moderate

144

40

High

007

02

Consumed tea

176

48

No use of vitamin supplements

357

98

Worked a rotating shift

346

95

Exposed to passive smoking

from husband

207

57

No history of a previous pregnancy

312

86

Conception (in 1,165 menstrual

cycles)

486

42

Early pregnancy loss (in 486

conceptions)

139

29

Vitamin B6

085

23

Vitamin B12

064

18

Folate

071

20

Homocysteine

038

10

Overall prevalence of

Abnormal* concentrations of

Median
(25th, 75th percentile)

Time from biomarker assessment

to stopping contraception (days)

52 (18, 167)

Time to rst conception

(menstrual cycles)

2 (1, 3)

Time to clinical pregnancy

(menstrual cycles) (n 346)

2 (1, 4)

* Abnormal plasma vitamin concentrations (deciencies) were

dened as follows: folate, <6.8 nmol/liter; vitamin B12, <258 pmol/
liter; and vitamin B6 (as pyridoxal-5#-phosphate), <30 nmol/liter.
Elevated plasma homocysteine was dened as 12.4 lmol/liter.

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

diary and hCG data, and for whom preconception vitamin

and homocysteine concentrations were available (table 1).
This was a young, generally lean cohort of newly married
women who had obtained permission to have a baby. Most
women had no more than a middle-school education, and
nearly all worked rotating shifts. Only about 2 percent of
women reported using vitamin supplements of any kind.
Although women in this group did not smoke or drink
alcohol, 57 percent were exposed to passive smoking, and
42 percent reported moderate to high life stress. Overall,
486 (42 percent) of 1,165 cycles resulted in conception;
139 (29 percent) of the 486 conceptions ended in early
pregnancy loss.
In general, B-vitamin deficiencies were common in this cohort. Plasma concentrations of PLP indicative of vitamin B6
deficiency were detected in 23 percent of the women, folate
deficiency was observed in 20 percent, 18 percent had biochemical evidence of vitamin B12 deficiency, and 10 percent
had elevated homocysteine. Overall, 43 percent of the
women were deficient in at least one vitamin, although just
9 percent were deficient in both folate and vitamin B6, 7 percent were deficient in both vitamins B6 and B12, and 5 percent were deficient in folate and vitamin B12. Only 3 percent
of women were deficient in all three vitamins. The total
homocysteine concentration was significantly inversely
correlated with both folate (r
0.17; p 0.001) and
vitamin B12 (r
0.14; p 0.009) and positively correlated with vitamin B6 (r 0.17; p 0.001). Significant
positive correlations (r 0.220.29; p < 0.001) also were
observed among the three vitamins.
Table 2 shows the relative hazards of conception for
women within different plasma concentrations of B vitamins and homocysteine. Relative to women in the first quartile of plasma vitamin B6, the relative hazards of conception
were higher for women in the third (adjusted hazard ratio
(HR) 2.2, 95 percent confidence interval (CI): 1.3, 3.4)
and fourth (adjusted HR 1.6, 95 percent CI: 1.1, 2.3)
quartiles. Women with a normal concentration of plasma
vitamin B6 had a higher hazard of conception (adjusted
HR 1.4, 95 percent CI: 1.1, 1.9) than did those with
vitamin B6 deficiency.
We assessed the relative odds of early pregnancy loss in
conception cycles of women within different plasma concentrations of B vitamins and homocysteine (table 3). Compared with women in the lowest quartile of plasma vitamin B6
(PLP: 30.4 nmol/liter), those in the fourth quartile
(PLP: 46.4 nmol/liter) had lower adjusted relative odds
of early pregnancy loss (odds ratio (OR) 0.5, 95 percent
CI: 0.3, 1.0). Relative to those with vitamin B6 deficiency
(PLP: <30 nmol/liter), women with normal vitamin B6 status had an adjusted odds of early pregnancy loss of 0.7
(95 percent CI: 0.4, 1.1). Compared with women in the
lowest quartile of plasma folate (folate: 7.2 nmol/liter),
those in the fourth quartile of plasma folate (folate: 11.8
nmol/liter) had adjusted relative odds of early pregnancy
loss that tended toward being lower (adjusted OR 0.6,
95 percent CI: 0.3, 1.2).
Table 4 shows that, relative to women in the lowest quartile
of vitamin B6 (PLP: 30.4 nmol/liter), women in the third
(38.346.3 nmol/liter: HR 2.0, 95 percent CI: 1.3, 3.0)

307

308 Ronnenberg et al.

TABLE 2. Relative hazards of conception by preconception plasma B-vitamin and homocysteine

concentrations in 364 nulliparous female textile workers in Anhui, China, who conceived at least once
during prospective observation, 19961998
Crude
No. of
women

No. of
conception
attempts*

Hazard
ratio

Adjustedy

95%
condence
interval

Hazard
ratio

95%
condence
interval

Vitamin B6 (nmol/liter)
13.030.4

91

133

30.538.2

91

118

1.3

0.9, 1.8

1.2

0.8, 1.6

38.346.3

91

134

2.2

1.4, 3.3

2.2

1.3, 3.4

46.489.0

91

110

1.6

1.2, 2.3

1.6

85

125

279

370

80.8277.8

91

115

277.9348.3

91

129

1.2

0.8, 1.7

0.9

0.6, 1.4

348.4434.5

91

128

1.1

0.8, 1.6

0.8

0.6, 1.2

1.0

0.7, 1.5

0.7

Vitamin B6 deciency (<30 nmol/liter)

Normal vitamin B6 (30 nmol/liter)

Referent

Referent

Referent
1.5

1.1, 2.0

1.1, 2.3
Referent

1.4

1.1, 1.9

Vitamin B12 (pmol/liter)

Referent

91

123

64

85

300

410

2.27.2

91

124

7.39.1

91

134

1.3

1.0, 1.8

1.1

0.8, 1.6

9.211.7

91

126

1.2

0.8, 1.7

1.1

0.7, 1.5

11.825.3

91

111

1.5

1.0, 2.1

1.3

71

96

293

399

3.26.7

91

128

6.87.9

90

114

0.7

0.5, 1.1

0.7

0.5, 1.0

8.09.5

92

136

0.8

0.5, 1.1

0.8

0.6, 1.2

9.627.8

91

117

0.9

0.6, 1.3

0.9

326

450

38

45

Vitamin B12 deciency (<258 pmol/liter)

Normal vitamin B12 (258 pmol/liter)

Referent
1.2

0.8, 1.7

0.5, 1.1
Referent

0.9

0.6, 1.3

Folate (nmol/liter)

Folate deciency (<6.8 nmol/liter)

Normal folate (6.8 nmol/liter)

Referent

Referent

Referent
1.4

1.0, 1.9

0.9, 1.9
Referent

1.2

0.9, 1.7

Homocysteine (lmol/liter)

Normal homocysteine (<12.4 lmol/liter)

Elevated homocysteine (12.4 lmol/liter)

Referent

Referent

Referent
1.4

0.9, 2.2

0.7, 1.4
Referent

1.5

1.0, 2.5

* Includes nine right-censored conception attempts in which conception did not occur before the end of follow-up.
y Models adjusted for womans age, body mass index, history of pregnancy, self-reported stress, working shifts,
occupational noise and dust exposures, education, and husbands age, smoking, and alcohol drinking. Models for
vitamin B6, vitamin B12, and folate adjusted for each of the other vitamins. Standard errors for both crude and
adjusted models were estimated to accommodate correlations in hazards of multiple conceptions from the same
woman.

and fourth (46.489.0 nmol/liter: HR 1.8, 95 percent

CI: 1.2, 2.7) quartiles had higher adjusted hazards of clinical
pregnancy. Relative to those with vitamin B6 deficiency
(PLP: <30 nmol/liter), women with normal vitamin B6 status had an adjusted hazard ratio of clinical pregnancy of 1.6
(95 percent CI: 1.1, 2.2). Compared with women in the
lowest quartile of plasma folate (folate: 7.2 nmol/liter),
those in the fourth quartile (folate: 11.8 nmol/liter) tended
to have a higher adjusted relative hazard of clinical pregnancy (adjusted HR 1.3, 95 percent CI: 0.9, 2.0).

We were unable to assess whether specific combinations

of B-vitamin deficiencies were associated with conception,
early pregnancy loss, or clinical pregnancy because of our
limited sample size. However, we did assess whether multiple deficiencies in general were associated with these outcomes. Being deficient in any one or two vitamins was not
significantly associated with conception, early pregnancy
loss, or clinical pregnancy. However, compared with women
who were not deficient in any of the three vitamins, women
who were deficient in all three vitamins (n 10) had
Am J Epidemiol 2007;166:304312

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

434.6828.3

Referent

Preconception Vitamin B6 and Early Pregnancy Loss

309

TABLE 3. Adjusted* relative odds of early pregnancy loss in conceptions by preconception plasma B-vitamin and homocysteine
concentrations in 364 nulliparous female textile workers in Anhui, China, who conceived at least once during prospective observation,
19961998
First observed conception only
No. of
conceptions

Early pregnancy
loss
No.

Odds ratio

All observed conceptionsy

95%
condence
interval

No. of
conceptions

Early pregnancy
loss
No.

126

43

34

Odds ratio

95%
condence
interval

Vitamin B6 (nmol/liter)
13.030.4

91

27

30

30.538.2

91

23

25

0.8

0.4, 1.5

117

29

25

0.7

0.4, 1.2

38.346.3

91

28

31

1.0

0.5, 2.0

134

45

34

1.0

0.6, 1.7

46.489.0

91

19

20

0.5

0.2, 1.1

109

22

20

0.5

0.3, 1.0

85

26

31

118

41

35

279

70

25

368

98

27

80.8277.8

91

20

22

111

26

23

277.9348.3

91

23

25

1.0

0.5, 2.1

126

40

32

1.3

0.7, 2.4

348.4434.5

91

30

33

1.8

0.9, 3.7

127

40

32

1.4

0.8, 2.6

434.6828.3

91

23

25

1.1

0.6, 2.4

122

33

27

1.1

0.6, 2.1

Vitamin B12 deciency

(<258 pmol/liter)

64

17

27

81

23

28

Normal vitamin B12

(258 pmol/liter)

300

79

26

405

116

29

2.27.2

91

27

30

119

36

30

7.39.1

91

28

31

1.0

0.5, 2.0

132

44

33

1.1

0.6, 1.9

9.211.7

91

23

25

0.8

0.4, 1.6

124

37

30

1.0

0.5, 1.8

11.825.3

0.7

0.3, 1.5

111

22

20

0.6

0.3, 1.2

92

28

30

394

111

28

127

38

30

Vitamin B6 deciency
(<30 nmol/liter)
Normal vitamin B6
(30 nmol/liter)

Referent

Referent
0.7

0.4, 1.3

Referent

Referent
0.7

0.4, 1.1

Referent

Referent
0.9

0.5, 1.7

Referent

Referent
0.9

0.5, 1.6

Folate (nmol/liter)
Referent

Referent

91

18

20

Folate deciency
(<6.8 nmol/liter)

71

22

31

Normal folate
(6.8 nmol/liter)

293

74

25

3.26.7

91

25

27

6.87.9

90

19

21

0.7

0.4, 1.5

112

25

22

0.7

0.4, 1.3

8.09.5

92

30

33

1.2

0.6, 2.3

131

46

35

1.2

0.7, 2.1

9.627.8

91

22

24

0.8

0.4, 1.6

116

30

26

0.8

0.5, 1.4

326

88

27

441

130

29

38

21

45

20

Referent
0.9

0.5, 1.6

Referent
1.0

0.6, 1.7

Homocysteine
(lmol/liter)

Normal homocysteine
(<12.4 lmol/liter)
Elevated homocysteine
(12.4 lmol/liter)

Referent

Referent
0.8

0.3, 1.8

Referent

Referent
0.7

0.3, 1.4

* All models adjusted for womans age, body mass index, history of pregnancy, self-reported stress, working shifts, occupational noise and dust
exposures, education, and husbands smoking. Models for vitamin B6, vitamin B12, and folate adjusted for each of the other vitamins. Estimated
parameters were similar in unadjusted models (not shown).
y Standard errors were estimated to accommodate correlations in pregnancy losses among conceptions from the same woman.

adjusted hazard ratios for conception of 0.3 (95 percent CI:

0.2, 0.5) and for clinical pregnancy of 0.2 (95 percent CI:
0.1, 0.6). We did not have sufficient statistical power to test
the odds of early pregnancy loss in those with three vitamin
deficiencies relative to those with none.
Am J Epidemiol 2007;166:304312

DISCUSSION

Our study examined the relations between preconception

B-vitamin status and conception, early pregnancy loss, and
clinical pregnancy in a prospective cohort of young Chinese

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

Vitamin B12 (pmol/liter)

310 Ronnenberg et al.

TABLE 4. Relative hazards of clinical pregnancy by preconception plasma B-vitamin and homocysteine
concentrations in 364 nulliparous female textile workers in Anhui, China, who conceived at least once
during prospective observation, 19961998
Crude
No. of
women*

Hazard
ratio

95%
condence
interval

Adjustedy
Hazard
ratio

95%
condence
interval

Vitamin B6 (nmol/liter)
13.030.4

91

30.538.2

91

1.4

1.0, 2.0

1.3

0.9, 2.0

38.346.3

91

1.9

1.3, 2.8

2.0

1.3, 3.0

46.489.0

91

1.8

1.2, 2.6

1.8

Vitamin B6 deciency (<30 nmol/liter)

Normal vitamin B6 (30 nmol/liter)

Referent

85
279

Referent

Referent
1.6

1.2, 2.2

1.2, 2.7
Referent

1.6

1.1, 2.2

Vitamin B12 (pmol/liter)

91

277.9348.3

91

1.0

0.7, 1.4

0.8

0.6, 1.2

348.4434.5

91

0.9

0.6, 1.3

0.7

0.5, 1.1

434.6828.3

91

1.0

0.7, 1.4

0.8

Vitamin B12 deciency (<258 pmol/liter)

Normal vitamin B12 (258 pmol/liter)

Referent

64
300

Referent

Referent
1.1

0.8, 1.5

0.5, 1.2
Referent

0.9

0.6, 1.3

Folate (nmol/liter)
2.27.2

91

7.39.1

91

1.1

0.8, 1.6

1.0

9.211.7

91

1.1

0.8, 1.6

1.0

0.7, 1.5

11.825.3

91

1.6

1.1, 2.3

1.3

0.9, 2.0

Folate deciency (<6.8 nmol/liter)

Normal folate (6.8 nmol/liter)

Referent

71
293

Referent

Referent
1.3

1.0, 1.8

0.6, 1.4

Referent
1.1

0.8, 1.6

Homocysteine (lmol/liter)
3.26.7

91

6.87.9

90

0.9

0.6, 1.2

0.8

0.6, 1.2

8.09.5

92

0.7

0.5, 1.1

0.8

0.5, 1.1

9.627.8

91

1.0

0.7, 1.4

1.0

Normal homocysteine (<12.4 lmol/liter)

Elevated homocysteine (12.4 lmol/liter)

Referent

326
38

Referent

Referent
1.5

0.9, 2.3

0.7, 1.4
Referent

1.6

1.0, 2.6

* Includes 17 right-censored clinical pregnancy attempts in which clinical pregnancy did not occur before the end
of follow-up.
y Models adjusted for womans age, body mass index, history of pregnancy, self-reported stress, working shifts,
occupational noise and dust exposures, education, and husbands age, smoking, and alcohol drinking. Models for
vitamin B6, vitamin B12, and folate adjusted for each of the other vitamins.

women who were attempting to become pregnant. We found

that higher preconception plasma vitamin B6 concentrations
were associated with reduced odds of early pregnancy losses
and higher probabilities of achieving conception and clinical pregnancy. In a previous report based on women from
this same cohort, we reported that the mean prepregnancy
plasma vitamin B6 concentration was significantly lower in
women whose pregnancies ended in a clinically recognized
spontaneous abortion than in those with livebirths (11). The
risk of clinical spontaneous abortion was more than twice as
high among women in the lowest vitamin B6 quintile compared with those in the highest. In a subsequent analysis of

livebirth outcomes (22), we also found that the risk of

preterm birth appeared to be about 50 percent lower among
women with adequate prepregnancy vitamin B6 status (OR
0.5, 95 percent CI: 0.2, 1.1) compared with women whose
vitamin B6 status was deficient. Previous studies have also
reported associations between maternal vitamin B6 status
and pregnancy outcome. In a case-control study, Wouters
et al. (14) found lower plasma concentrations of vitamin B6
(46 vs. 51 nmol/liter, p < 0.05) in women with histories of
recurrent spontaneous abortion compared with control
women. Goddijn-Wessel et al. (13) similarly reported lower
plasma vitamin B6 (42 vs. 53 nmol/liter, p < 0.05) among
Am J Epidemiol 2007;166:304312

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

80.8277.8

Preconception Vitamin B6 and Early Pregnancy Loss

Am J Epidemiol 2007;166:304312

In conclusion, we found that poor preconception vitamin B6

status was associated with increased risk of early pregnancy loss and reduced probabilities of conception and clinical pregnancy in a prospective cohort of young Chinese
women. This study underscores the potential importance
of micronutrient status at the time of conception on pregnancy outcome.

ACKNOWLEDGMENTS

This study is supported in part by grants R01 HD32505

and R01 HD41702 from the National Institute of Child
Health and Human Development; grants R01 ES08337, P01
ES06198, and R01 ES11682 from the National Institute of
Environmental Health Sciences; and grants 20-FY98-0701
and 20-FY02-56 from the US March of Dimes Birth Defects
Foundation. S. A. V. was supported by grant K01 ES12052
from the National Institute of Environmental Health
Sciences.
Conflict of interest: none declared.

REFERENCES
1. Zhang ZW, Qu JB, Moon CS, et al. Nutritional evaluation of
women in urban areas in continental China. Tohoku J Exp Med
1997;182:4159.
2. Zhan S, Hu Y, Li L. A correlation study on homocysteine
metabolism in pregnant women and neural tube defects in
urban and rural areas. (In Chinese). Zhonghua Yu Fang Yi Xue
Za Zhi 1997;31:2214.
3. Ge KY, Chang SY. Dietary intake of some essential micronutrients in China. Biomed Environ Sci 2001;14:31824.
4. Gao X, Yao M, McCrory MA, et al. Dietary pattern is associated with homocysteine and B vitamin status in an urban
Chinese population. J Nutr 2003;133:363642.
5. Ronnenberg AG, Goldman MB, Aitken IW, et al. Anemia and
deficiencies of folate and vitamin B-6 are common and vary
with season in Chinese women of childbearing age. J Nutr
2000;130:270310.
6. Scholl TO, Hediger ML, Schall JI, et al. Dietary and serum
folate: their influence on the outcome of pregnancy. Am J Clin
Nutr 1996;63:5205.
7. Scholl TO, Johnson WG. Folic acid: influence on the outcome
of pregnancy. Am J Clin Nutr 2000;71(suppl):1295S303S.
8. Nelen WL, Blom HJ, Steegers EA, et al. Hyperhomocysteinemia and recurrent early pregnancy loss: a meta-analysis.
Fertil Steril 2000;74:11969.
9. Nelen WL, Blom HJ, Steegers EA, et al. Homocysteine and
folate levels as risk factors for recurrent early pregnancy loss.
Obstet Gynecol 2000;95:51924.
10. Hibbard BM. Folates and the fetus. S Afr Med J 1975;49:
12236.
11. Ronnenberg AG, Goldman MB, Chen D, et al. Preconception
folate and vitamin B(6) status and clinical spontaneous
abortion in Chinese women. Obstet Gynecol 2002;100:
10713.
12. George L, Mills JL, Johansson AL, et al. Plasma folate
levels and risk of spontaneous abortion. JAMA 2002;288:
186773.

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

women with placental abruption or infarction compared

with control women. Taken in their entirety, these observations suggest that maternal vitamin B6 status may influence
reproductive events throughout the entire course of pregnancy, from the time of conception through delivery.
The physiology underlying the relation between low
vitamin B6 status and early pregnancy loss is unknown,
although several biologically plausible mechanisms are possible. Vitamin B6-dependent coenzymes participate in over
100 reactions involved in the metabolism of amino acids,
lipids, nucleic acids, and glycogen (31). Vitamin B6 deficiency has also been associated with impairment of enzymes
involved in the structural integrity of arterial walls (32),
which could affect implantation and early placental development. Low maternal prepregnancy vitamin B6 status could
influence early gestational events through these well-known
functional roles of the vitamin. In addition, numerous studies
have documented associations between low vitamin B6 status
and inflammatory responses (3336), and inflammation has
been linked to early pregnancy loss (37, 38). Although we
did not assess biomarkers of inflammation in our study, previous studies in Chinese textile workers have shown that
exposure to cotton dust initiates both acute and chronic lung
inflammation (39). Work-related inflammatory exposures
coupled with low vitamin B6 status may have affected the
risk of early pregnancy loss in the female textile workers in
our cohort. The potential relation between vitamin B6 and
inflammation might also help to explain our unexpected
positive correlation between plasma concentrations of vitamin B6 and homocysteine, since vitamin B6 levels appear to
be lower in inflammatory states independent of homocysteine levels (34). Additional studies that include an independent marker of inflammation, such as C-reactive protein (34),
will be needed, however, to help clarify a potential interrelation among vitamin B6 status, inflammation, and early
pregnancy loss.
A strength of the current study was that we collected
consecutive daily urine samples throughout the entire study
period, which allowed us to assay hCG and accurately detect
conception and early pregnancy losses. We also assessed
vitamin status using plasma samples collected prior to when
women began attempting to become pregnant. A limitation
was that we based micronutrient status assessment on a single blood sample taken at enrollment. Because some women
continued to use contraception after enrolling, some pregnancy outcomes occurred quite distal to micronutrient status
assessment. It is possible, therefore, that the womens nutritional status changed between assessment and pregnancy
outcome, leading to potential misclassification. For example, although vitamin supplement use in this rural area of
China was very low, health-conscious women still might have
been more likely to begin vitamin supplementation when they
stopped contraception. Misclassification of micronutrient
status might explain why the association of vitamin B6 with
the hazard of conception appeared to be nonlinear. Another
limitation was that, out of 961 enrolled women, we excluded
597 from this analysis. Compared with the excluded women,
those included had similar age, height, weight, and body mass
index, but less education. Thus, our sample might not be
representative of the entire enrolled cohort.

311

312 Ronnenberg et al.

27.
28.
29.
30.
31.

32.
33.
34.

35.

36.
37.
38.

39.

urine: application to malignancies. Cancer Res

1988;48:13616.
Husdan H, Rapoport A. Estimation of creatinine by the Jaffe
reaction. A comparison of three methods. Clin Chem
1968;14:22238.
Cox DR. Regression models and life tables (with discussion).
J R Stat Soc (B) 1972;34:187220.
Lin DY, Wei LJ. The robust inference for the proportional
hazards model. J Am Stat Assoc 1989;84:10748.
Liang KY, Zeger S. Longitudinal data analysis using generalized linear models. Biometrica 1986;73:1322.
Leklem JE, Reynolds RE. Challenges and direction in the
search for clinical applications of vitamin B6. In: Leklem JE,
Reynolds RE, eds. Current topics in nutrition and disease. New
York, NY: Alan R Liss, 1988.
Levene CI, Murray JC. The aetiological role of maternal vitamin-B6 deficiency in the development of atherosclerosis.
Lancet 1977;1:62830.
Kelly PJ, Kistler JP, Shih VE, et al. Inflammation, homocysteine, and vitamin B6 status after ischemic stroke. Stroke
2004;35:1215.
Friso S, Jacques PF, Wilson PW, et al. Low circulating vitamin
B(6) is associated with elevation of the inflammation marker
C-reactive protein independently of plasma homocysteine
levels. Circulation 2001;103:278891.
Saibeni S, Cattaneo M, Vecchi M, et al. Low vitamin B(6)
plasma levels, a risk factor for thrombosis, in inflammatory
bowel disease: role of inflammation and correlation with acute
phase reactants. Am J Gastroenterol 2003;98:11217.
Chiang EP, Bagley PJ, Selhub J, et al. Abnormal vitamin B(6)
status is associated with severity of symptoms in patients with
rheumatoid arthritis. Am J Med 2003;114:2837.
Thellin O, Heinen E. Pregnancy and the immune system: between tolerance and rejection. Toxicology 2003;185:17984.
Kwak JY, Beer AE, Kim SH, et al. Immunopathology of the
implantation site utilizing monoclonal antibodies to natural
killer cells in women with recurrent pregnancy losses. Am J
Reprod Immunol 1999;41:918.
Li D, Zhong YN, Rylander R, et al. Longitudinal study of the
health of cotton workers. Occup Environ Med 1995;52:32831.

Am J Epidemiol 2007;166:304312

Downloaded from http://aje.oxfordjournals.org/ by guest on November 13, 2014

13. Goddijn-Wessel TA, Wouters MG, van de Molen EF, et al.

Hyperhomocysteinemia: a risk factor for placental abruption
or infarction. Eur J Obstet Gynecol Reprod Biol 1996;66:
239.
14. Wouters MG, Boers GH, Blom HJ, et al. Hyperhomocysteinemia: a risk factor in women with unexplained recurrent early
pregnancy loss. Fertil Steril 1993;60:8205.
15. Wilcox AJ, Weinberg CR, OConnor JF, et al. Incidence of
early loss of pregnancy. N Engl J Med 1988;319:18994.
16. Wang X, Chen C, Wang L, et al. Conception, early pregnancy
loss, and time to clinical pregnancy: a population-based prospective study. Fertil Steril 2003;79:57784.
17. Zinaman MJ, Clegg ED, Brown CC, et al. Estimates of human
fertility and pregnancy loss. Fertil Steril 1996;65:5039.
18. Dunson DB, Zhou H. A Bayesian model of fecundability and
sterility. J Am Stat Assoc 2000;95:105462.
19. Gelman A, Carlin J, Stern H, et al. Bayesian data analysis.
London, United Kingdom: Chapman and Hall, 1997.
20. Araki A, Sako Y. Determination of free and total homocysteine
in human plasma by high-performance liquid chromatography
with fluorescence detection. J Chromatogr 1987;422:4352.
21. Shin YS, Rasshofer R, Friedrich B, et al. Pyridoxal-5#-phosphate determination by a sensitive micromethod in human
blood, urine and tissues; its relation to cystathioninuria in
neuroblastoma and biliary atresia. Clin Chim Acta 1983;127:
7785.
22. Ronnenberg AG, Goldman MB, Chen D, et al. Preconception
homocysteine and B vitamin status and birth outcomes in
Chinese women. Am J Clin Nutr 2002;76:138591.
23. Leklem JE. Vitamin B-6: a status report. J Nutr 1990;
120(suppl 11):15037.
24. Lindenbaum J, Rosenberg IH, Wilson PW, et al. Prevalence of
cobalamin deficiency in the Framingham elderly population.
Am J Clin Nutr 1994;60:211.
25. Allen LH, Casterline J. Vitamin B-12 deficiency in elderly
individuals: diagnosis and requirements. Am J Clin Nutr
1994;60:1214.
26. OConnor JF, Schlatterer JP, Birken S, et al. Development of
highly sensitive immunoassays to measure human chorionic
gonadotropin, its beta-subunit, and beta core fragment in the