OUP user menu

Vitamin B12 insufficiency and the risk of fetal neural tube defects

J.G. Ray, H.J. Blom
DOI: http://dx.doi.org/10.1093/qjmed/hcg043 289-295 First published online: 1 April 2003


Background: Although maternal folate insufficiency is a risk factor for fetal neural tube defects (NTDs), there is controversy about whether vitamin B12 (B12) insufficiency is also associated with an increased risk of NTDs.

Aim: To investigate whether low maternal B12 is associated with an increased risk of fetal NTDs.

Design: Systematic review.

Methods: A systematic search of Medline between 1980 and October 2002, with an examination of the citations of all retrieved studies. Studies were included that: (i) used a cohort or case‐control design; (ii) included case mothers with a prior or current NTD‐affected pregnancy; (iii) assessed a group of unaffected ‘controls’; and (iv) measured the vitamin B12 status of all participants.

Results: Overall, 17 case‐control studies were included, mean sample size 33 cases and 93 controls. In 5/6, mean amniotic fluid B12 concentration was significantly lower in case mothers than in controls. Of 11 that measured maternal serum or plasma B12, three observed a significantly lower mean concentration in case mothers vs. controls, while five others found a non‐significant lower trend in the case group. One study observed a significantly higher mean concentration of maternal serum methylmalonic acid among the maternal cases, while another found a non‐significant lower mean concentration of plasma holo‐transcobalamin. Five studies estimated the risk of NTDs in relation to low B12 or B12‐related metabolic markers: it was significantly increased in three studies, with a non‐significant trend in the fourth.

Discussion: There seems to be a moderate association between low maternal B12 status and the risk of fetal NTDs. However, several design limitations, and the inclusion of few study participants, may have under‐represented this. A large observational study, using reliable and valid indicators of B12 status in early pregnancy, could best assess the association between B12 insufficiency and the risk of fetal NTDs.


Increased preconception intake of folic acid can significantly reduce the risk of fetal neural tube defects (NTDs).1 Although some countries have developed initiatives that promote maternal preconceptional use of folic acid tablets, and others have fortified their cereal grain foods with synthetic folates,2,,3 fetal NTDs continue to affect approximately 6 in every 10 000 pregnancies.3 Maternal folate impairment is a demonstrated risk factor for fetal NTD, but a number of investigators have pointed to abnormalities of the homocysteine (tHcy) pathway, evidenced by elevations in plasma total tHcy, as the primary site of the metabolic defect.4 Vitamin B12 (B12), as methylcobalamin, is the key cofactor for methionine synthase, co‐contributing with folate to the capacity for remethylation of tHcy. What remains unknown is the degree to which low maternal B12 status increases the risk of NTDs, the subject of this systematic review.


We performed a systematic literature search of Medline between 1980 and October 2002. The following unlimited search expression was used: ‘(Neural tube defect OR spina bifida OR anencephaly) AND (vitamin B‐12 OR cobalamin OR methylmalonic acid or holotranscobalamin)’. We also examined the references of all studies for other potential citations, but did not search for unpublished studies. Reports that met all of the following criteria were accepted: (i) use of a cohort or case‐control design; (ii) inclusion of case mothers, defined as women with a prior or current NTD‐affected pregnancy; (iii) assessment of a group of unaffected ‘controls’; and (iv) assessment of B12 status among all participants, via determination of serum B12, methylmalonic acid (MMA) or holo‐transcobalamin (holoTC), the complex formed between B12 and its transport plasma‐binding protein, transcobalamin.

Data abstraction

The abstracted study data included the study period, the basic characteristics of the case mothers and controls, the method used to measure B12,5,,6 whether the serum sample was obtained during pregnancy,7 and whether participants were taking a B12‐containing supplement at the time of specimen collection (Table 1). Using the data from each study, we presented the comparative mean vitamin concentrations among case mothers and controls (Table 2), as well as the odds ratio (OR) for a NTD in the presence of an abnormal level (Table 3). We evaluated whether folate levels or folic acid supplement use were factored into, or adjusted for, within this risk estimate, since folate insufficiency is a major risk factor for NTD.25 Statistical significance was set at p≤0.05. Because of dissimilarities in the design and participant characteristics of most studies, the data were not pooled.

View this table:
Table 1

General features of 17 case‐control studies of maternal vitamin B12 insufficiency and the risk of NTD

ReferenceCountry and time periodCases with an NTD‐affected pregnancyControls without an NTDWas use of B12 supplements excluded prior to specimen collection?
General descriptionWere other concomitant anomalies excluded?Period of specimen collectionGeneral descriptionPeriod of specimen collection
Molloy 19858Ireland 1981–1983Current NTD birthNo88% were <20 weeks gestationCurrent non‐NTD birth<20 weeks gestationNo
Gardiki‐Kouidou 19889UK 1982–1987Current NTD pregnancyNo15–23 weeks, gestationCurrent non‐NTD pregnancy15–23 weeks gestationNo
Weekes 199210US before 1992Current NTD pregnancyNo14–22 weeks gestationCurrent non‐NTD healthy infant14–22 weeks gestationNo (more than 37% took a supplement)
Economides 199211UK before 1991Current terminated NTD pregnancyNo14–21 weeks gestationCurrent terminated non‐NTD pregnancy14–21 weeks gestationNo
Mills 199212Finland 1983–1989Current NTD pregnancyNo96% were <17 weeks gestationCurrent non‐NTD pregnancy94% were <17 weeks gestationMostly (8% of cases and 15% of controls took a supplement)
Wild 199313UK before 1992Prior NTD pregnancyNo>12 months postartumPrior non‐NTD pregnancy>12 months postpartumNo
Kirke 199314Ireland 1986–1990Current NTD birth after 23 weeks gestationNo75% were <20 weeks gestationCurrent non‐NTD birth after 23 weeks gestation75% were <20 weeks gestationNo
Steegers‐Theunissen 199415Netherlands before 1993Recent NTD pregnancyNo>3 months postpartumRecent non‐NTD pregnancy>3 months postpartumYes
Steegers‐Theunissen 199616Netherlands before 1994Current NTD pregnancyNoSecond trimesterCurrent non‐NTD pregnancySecond trimesterYes
Adams 199517US before 1994Current NTD pregnancyNoSecond trimesterCurrent non‐NTD pregnancySecond trimesterNo
Wright 199518Northern Ireland before 1995Recent NTD pregnancyNo3–12 months postpartumRecent non‐NTD pregnancy3–12 months postpartumNo
Wald 199619International before 1993Current NTD pregnancyNoBefore pregnancyCurrent non‐NTD pregnancyBefore pregnancyYes
Van der Put 199720Netherlands before 1997Remote NTD pregnancyNoNot pregnant25% men and 75% women with no history of NTDNot pregnantNo
Steen 199821US before 1997Current NTD pregnancyNo14–18 weeks gestationCurrent non‐NTD pregnancy14–18 weeks gestationNo
Dawson 199922US before 1998Current NTD pregnancyYes15–20 weeks gestationCurrent non‐NTD pregnancy15–20 weeks gestationNo (38% took a supplement)
Wilson 199923Canada before 1999Prior NTD pregnancyYesNot pregnantHealthy women, no history of NTDNot pregnantYes
Afman 200124Netherlands before 2000Prior NTD pregnancyYesNot pregnantHealthy women, no history of NTDNot pregnantNo
  • NTD, neural tube defect.

View this table:
Table 2

Results of case‐control studies of the mean concentration of vitamin B12 in maternal amniotic fluid or serum

Analyte and its sourceReferenceNumber of participantsB12 assay usedMean (SD) or median (range) concentration, pmol/lp
NTD casesNon‐NTD controlsCasesControls
Amniotic fluid B12Gardiki‐Kouidou 198892665RIA110 (50–175)162 (100–339)0.04
Weekes 199210847RIA140 (90)600 (380)<0.001
Economides 199211824RIA92 (74–177)207 (81–760)<0.01
Steegers‐Theunissen 1995162731RIA379 (298)481 (278)NS
Steen 1998211563RIA111 (109–325)*398 (355–531)*0.02
Dawson 1999221129RIA226 (190)618 (285)<0.001
Serum or plasma B12Molloy 1985832384Micro219 (129–442)204 (55–738)NS
Economides 199211824RIA151 (63–188)170 (66–858)NS
Mills 19921289178RIA356 (119)384 (142)NS
Wild 1993132929RIA331 (201–732)361 (229–858)NS
Kirke 19931481247Micro179 (75–560)218 (65–559)0.001
Steegers‐Theunissen 1994164150RIA268 (100)264 (115)NS
Adams 19951733132Mass spec.297 (NA)319 (NA)NS
Wright 1995181515RIA148 (60)218 (122)0.004
Wald 1996191979Micro214 (NA)236 (NA)0.05
ven der Put 1997206094RIA245 (43–620)255 (64–580)NS
Afman 2001244673RIA220 (70–1000)220 (89–600)NS
Serum MMAAdams 19951733132Mass spec.0.13 (NA)0.10 (NA)0.004
Plasma holoTC**Afman 2001244673RIA41 (1–372)50 (0.5–186)NS
  • NTD, neural tube defect; B12, vitamin B12; NS, not significant; RIA, radioimmunoassay; Micro, microbiological assay; Mass spec., mass spectrometry; NA, not available; MMA, methylmalonic acid; HoloTC, holotranscobalamin. *95%CI. **Derived as: HoloTC=(plasma vitamin B12)−(holo‐haptocorrin).

View this table:
Table 3

Results of case‐control studies evaluating the risk of neural tube defects according to measures of maternal vitamin B12 status

ReferenceNumber of participantsAnalyte and its sourceComparisonRisk estimate
NTD casesNon‐NTD controlsOR (95% CI) for NTD, cases vs. controlsWas the OR adjusted for maternal serum or plasma folate level?Was the OR adjusted for maternal folic acid supplement use?
Molloy 1985832384Serum B12<185 pmol/l vs. ≥185 pmol/l0.9 (0.4–1.9)NoNo
Kirke 19931481247Plasma B12 and folate≤lower quartile vs. ≥upper quartile of both analytes5.4 (1.2–25.2)NoNo
Van der Put 1997206094Serum B12≤5th centile vs. >95th centile3.9 (1.3–11.9)NoNo
Afman 2001244673Plasma B12≤lower quartile vs. ≥upper quartile1.8 (0.6–5.2)NoNo
Plasma holoTC≤lower quartile us. ≥upper quartile of2.9 (0.9–9.2)NoNo
Adams 19951733132Serum MMA≥90th centile vs. <10th centile13.3 (2.7–65.5)YesNo
  • NTD, neural tube defect; B12, vitamin B12; OR, odds ratio; HoloTC, holotranscobalamin; MMA, methylmalonic acid.


The Medline search yielded 107 studies. Of these, 17 case‐control studies met the inclusion criteria, and contained original data (Table 1). The average study comprised 33 cases and 93 controls. Only five studies mentioned excluding most12 or all15,16,19,,23 individuals who were taking a B12‐containing supplement at the time of biochemical testing (Table 1). In 5/6 studies, there was a significantly lower mean amniotic fluid B12 concentration in case mothers than in controls,9–11,21,,22 with a trend in the same direction in the sixth study16 (Table 2). Among the controls in five of these studies, amniocentesis was performed for cytogenetic testing, either because of advanced maternal age9,10,16,,21 or for reasons not given.22 In the sixth study, amniotic fluid samples were obtained at the time of pregnancy termination, for social reasons.11

Of 11 studies that measured maternal serum or plasma B12, eight observed either a significant14,18,,19 or non‐significant trend11–13,17,,20 lower mean concentration among the case mothers vs. controls, while three others did not8,15,,24 (Table 2). There were no study features evident, such as the period of specimen collection (Table 1) or type of B12 assay used (Table 2), to explain these differences. For example, in five studies, participants were pregnant at the time of serum/plasma specimen collection,8,11,12,14,,17 while in six studies they were not,13,15,18–20,,24 but this did not seem to relate to whether there was an observed difference in mean serum B12 concentrations between case mothers and controls (Table 2).

As regards other measures, one study found a significantly higher mean concentration of serum MMA, a marker of B12 impairment, among maternal cases.17 In another case‐control study, a non‐significant lower mean concentration of plasma holoTC was observed24 (Table 2).

Five studies provided data on the estimate risk of NTD in relation to serum B12,8,14,20,,24 holoTC24 or MMA17 (Table 3). An Irish group of investigators found no difference in the risk of NTDs between 32 case mothers and 384 controls, comparing serum B12 concentrations below and above 185 pmol/l (OR 0.9, 95% CI 0.4–1.9),8 while another group of researchers observed a significant associated risk between the lowest and highest quartile concentrations of combined plasma B12 and folate among 81 case mothers and 247 controls (OR 5.4, 95%CI 1.2–25.2.14 In a study from the Netherlands, with 60 case mothers and 94 controls, there was nearly a threefold increased risk for NTDs for a maternal serum B12 concentration below the 5th centile vs. that above the 95th centile (OR 3.9, 95% CI 1.3–11.9)20 (Table 3). A smaller Dutch study found a non‐significant increased risk for fetal NTD upon comparing the lower and upper quartile concentrations of plasma B12 (OR 1.8, 95%CI 0.6–5.2), and a borderline significant risk for holoTC (OR 2.9, 95%CI 0.9–9.2).24 In a US study with 33 case mothers and 132 controls, there was an increased risk for NTDs in the presence of a maternal serum MMA concentration above the 90th centile, compared to that below the 10th centile (adjusted OR 13.3, 95%CI 2.7–65.5).17 In only the latter study was maternal serum folate adjusted for in the risk estimate, but no study considered maternal folate supplement use (Table 3). Finally, in a sixth study, a genetic polymorphism of maternal methionine synthase reductase (MTRR 66A→G)—an enzyme that activates cobalamin‐dependent methionine synthase—in combination with low maternal serum B12, was associated with a four‐fold increased risk of NTDs (OR 4.8, 95% CI 1.5–15.8).23


Some observational data suggest a moderately strong association between low maternal B12 status and the risk of fetal NTDs. The risk of NTDs was most pronounced when comparisons were made using more extreme cut‐points (e.g. 5th vs. 95th centiles) to define abnormal and normal levels.

Limitations to current knowledge

In addition to publication bias, these disparate findings can be explained by several factors. Since severe B12 insufficiency is rare among non‐vegan adults,26–,28 only mild reductions in B12 concentrations would be expected in a study sample of young women. Hence, the detection of a small significant difference in the B12 status of case mothers and controls might be difficult, considering that the average study had only 33 cases and 93 controls, and that the measurement of B12 or its metabolites was done using assays with less than optimal accuracy.29–,32 The collection of maternal specimens remote from the time of the index pregnancy, or many weeks after the period of embryonic risk for NTD formation, would also be expected to dilute any true relationship between B12 insufficiency and NTD risk. Similarly, measurement ‘contamination’ from recent B12 supplement use and the effect of specimen collection at various gestational periods of pregnancy33,,34 might further underestimate the true effect of B12 insufficiency on the risk of NTD. The failure of most studies to adjust for maternal folate status may have further confounded matters, since B12 insufficiency may simply be a marker of concomitant folate impairment, while, on the other hand, high folate concentrations may mask the haematological signs of B12 insufficiency.

Nearly all studies reviewed herein found a lower concentration of amniotic fluid B12 among the maternal cases. But is the B12 status of the fetus accurately represented by that found in amniotic fluid, or is maternal serum B12 better? Among 76 women who underwent testing at approximately 17 weeks gestation, plasma B12 (320±130 pmol/l) was significantly lower than that measured in amniotic fluid (650±420 pmol/l), although both were highly correlated.35 Term newborns have B12 concentrations at least twice as high as their mothers, but the greatest concentration is found within the placental intervillous space.36 It seems that maternal B12 crosses the placenta into the fetal circulation in modest amounts,37 increasing with advanced gestational age,38 but perhaps, only once placental tissue stores have been saturated.39 Moreover, in pregnant rats, intrinsic factor‐cobalamin receptor (cubulin) activity declined 15‐fold in the visceral yolk sac membranes, but increased almost 20‐fold in the placental membranes, between the times of 14 and 19 days of gestation.40 Since concentrations of maternal serum or amniotic fluid B12 appear to be in a dynamic state of change throughout pregnancy, they may not accurately reflect that found within the early developing embryo, and thus, the respective NTD risk.

Ramifications for future research

This review does not resolve whether a deficiency of B12 or a related elevation in tHcy contributes to the formation of NTDs.4 To better understand whether maternal B12 insufficiency is a risk factor for NTD, a large observational study is needed. This study should use a reliable and valid indicator of B12 status, perhaps serum MMA,32 obtained in early pregnancy, and should adjust for maternal folate concentration. Such data may be of great clinical importance: Like folic acid, orally administered B12 appears to be a safe, simple, and inexpensive vitamin. If periconceptional maternal B12 deficiency too can be shown to increase the risk of NTDs in a consistent manner, with an attenuation of that risk with higher B12 intake, then there exists a rational basis for conducting a multicentre randomized controlled clinical trial comparing periconceptional B12 and folic acid supplements with folic acid alone. Until such data are made available, periconceptional folic acid remains the mainstay of NTD prevention, with consideration given to adding small doses of B12, such as that currently found in most multivitamin tablets.41


JGR is supported by a grant from the Physicians' Services Incorporated Foundation of Toronto, Ontario. HJB is an Established Investigator of the Netherlands Heart Foundation (D97.021).


  • Address correspondence to Dr J.G. Ray. e‐mail: jray515445{at}aol.com


View Abstract