| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Endocrinology and Metabolism |
1 Perinatal Epidemiology Research Unit, Departments of Obstetrics and Gynaecology and Pediatrics, Dalhousie University, Halifax, Nova Scotia, Canada; Departments of 2 Pathology, 3 Obstetrics and Gynaecology, 4 Medicine, Dalhousie University, Halifax, Nova Scotia, Canada; 5 University of Ottawa and the Ottawa Hospital, Ottawa, Ontario, Canada.
aAddress correspondence to this author at: IWK Health Centre, 5850/5980 University Ave., Halifax, Nova Scotia B3K 6R8. Fax 902-470-7190; e-mail l.dodds{at}dal.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We conducted a prospective cohort study between 2002 and 2005. We measured tHcy and serum folate in blood samples from pregnant women (<20 weeks’ gestation) and collected detailed pregnancy information through a questionnaire and medical record review.
Results: Of the 2119 women included in the study, 103 had a pregnancy loss, 115 had gestational hypertension, 65 had preeclampsia, and 129 had an SGA infant. Subjects with increased tHcy concentrations were at increased risk of pregnancy loss [relative risk (RR) 2.1, 95% CI 1.2–3.6] or preeclampsia (RR 2.7, 95% CI 1.4–5.0) than subjects with lower tHcy concentrations, but increased tHcy concentration was not associated with increased risk of developing GH or having an SGA infant.
Conclusion: The finding of high tHcy in early pregnancy as a risk factor for pregnancy loss and preeclampsia is consistent with a hypothesis that increased tHcy results in abnormalities of the placental vasculature.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Homocysteine is formed during the metabolism of dietary methionine, found largely in animal protein. Increases in plasma tHcy are usually caused by nutritional deficiencies or genetic defects in the enzymes responsible for homocysteine metabolism. Folate, vitamin B12, and vitamin B6 are required for homocysteine metabolism, and deficiencies can result in increased tHcy concentrations (14). A mutation in the 5,10-methylenetetrahydrofolate reductase gene (MTHFR)2 has been associated with increased tHcy concentrations, especially in concert with deficient folate status (15).
An association between increased tHcy concentration and placenta-mediated pregnancy outcomes could be causal, or increased tHcy may be a marker, through an association with risk factors of a disease process that has already been initiated. A causal mechanism has been hypothesized: high homocysteine acts on the blood vessel walls, resulting in changes to the endothelial cells and ultimately endothelial dysfunction, especially within the placental vasculature (1)(13). These placental vascular changes are thought to be related to recurrent pregnancy loss, preeclampsia, and placenta abruption.
The main emphasis of this study was to examine the relationship between tHcy concentrations in early pregnancy and the subsequent development of gestational hypertension (GH) or preeclampsia. In addition, we also examined the relationship between tHcy concentrations and pregnancy loss and small for gestational age (SGA) infants. The ability to identify women at risk for these pregnancy conditions early in pregnancy would increase the potential for prevention strategies.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
20 weeks’ gestation at recruitment were excluded. A brief food recall questionnaire was administered, and participants provided blood samples. Blood for tHcy measurement was collected in K2EDTA VacutainerTM tubes (Becton Dickinson). Specimens were stored at 4 °C immediately after collection, transported to the laboratory within 30 min, and centrifuged at 3000g, 4 °C, for 10 min to separate serum, which was stored at –70 °C until analysis. We measured plasma tHcy concentrations by use of fluorescence polarization immunoassay on the Abbot AxSym analyzer using manufacturer’s reagents (Abbot Diagnostics). This assay has an analytical range of 1–50 µmol/L and precision (CV) of 5.5% at 7.4 µmol/L, 6.2% at 13.5 µmol/L, and 5.4% at 25.9 µmol/L. We measured serum folate on Beckman Coulter Access II or DXi immunoassay analyzers using manufacturer’s reagents. The assay has analytical linearity of 1– 45 nmol/L and precision (CV) of 9.4% at 3.4 nmol/L, 5.0% at 9.3 nmol/L, and 5.6% at 22.0 nmol/L.
During their 20th week of pregnancy, participants completed a questionnaire that included information on maternal and paternal age, education level, family income, prepregnancy weight, maternal height, smoking habits, chronic medical conditions, pregnancy history, physical activity prepregnancy and during the first 20 weeks of gestation, and caffeine intake during pregnancy. After delivery, we reviewed medical records to obtain detailed information on the pregnancy, including prenatal ultrasound data, maternal blood pressure readings, urinary protein findings, use of medications for hypertension before and during pregnancy, maternal weight at the time of delivery, antenatal hospital admissions, gestational age at delivery, and infant birth weight, sex, and outcome.
Criteria used to identify subjects affected by GH and preeclampsia were based on guidelines developed by the Canadian Hypertension Society (16). GH was defined as hypertension after midpregnancy (antenatal diastolic blood pressure 90 mm Hg after 20 weeks’ gestation in 2 or more readings at least 4 h apart) without proteinuria; preeclampsia was defined as hypertension after midpregnancy with proteinuria (1 protein reading on urine dipstick analysis or quantitative urine protein of 300 mg in a 24-h period). SGA was defined as the bottom 10th percentile of birth weight for each week of gestation and by sex, according to Canadian standards (17). Pregnancy loss was defined as spontaneous fetal death at any point in the pregnancy.
We decided a priori to define increased tHcy as concentrations 90th percentile based on the tHcy distribution among the referent group of normotensive subjects whose infants were live-born and not SGA (i.e., those who did not develop any of the study outcomes). Because the concentration of tHcy changes over the course of pregnancy (18), the 90th percentile was determined for each 2-week interval of gestational age from 4 to 20 weeks of gestation.
We selected potential confounders based on factors noted in previous research to be related with preeclampsia or the other outcomes. Prepregnancy weight was categorized as obese (yes or no), and total caffeine ingestion was categorized in quartiles. Obesity was defined as body mass index (BMI), calculated as weight in kilograms (kg) divided by height in meters squared (m2), 30 kg/m2. Caffeine content for coffee, tea, and soft drinks was assigned based on published data (19), and subjects with the total caffeine ingestion in the top quartile for the first 5 months of pregnancy were contrasted with those below the top quartile. Although we considered serum folate concentration a potential confounder for each outcome, the final models were also run without including folate concentration. Because folate concentrations may change during pregnancy, low folate was defined according to the bottom-quartile value during each 2-week gestational age interval among women without any of the study outcomes.
We calculated rates of pregnancy loss using the total number of study subjects for the denominator, whereas only live births were used for the denominator for all other outcomes. Relative risk (RR) and 95% CIs were generated for tHcy exposure and each outcome using logistic regression. We used a backward stepwise approach to identify confounding variables for each outcome. All potential confounding factors associated with either tHcy exposure or the outcome at P 
0.20 were entered into a starting multivariate model. The factor with the highest P value was removed from the model, and the RR associated with homocysteine and the outcome was compared between the full and reduced models. If the relative change in the RR for tHcy was <5%, the variable was excluded, and the procedure was repeated. Sensitivity, specificity, and likelihood ratio tests were calculated for high tHcy (90th percentile) and all outcomes.
We performed all analyses by use of Statistical Analysis Software version 8.2 (SAS Institute). The IWK Research Ethics Board approved all aspects of the study design, and written consent was received from all participants. The funding organization (Canadian Institutes of Health Research) had no role in the design or conduct of this study.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
shows the numbers of women recruited, excluded after recruitment, and with each outcome, plus the reasons for the 81 women not included in the analysis. Women with no follow-up data included those who moved and delivered elsewhere and women whose blood sample was inadequate for measuring tHcy. Of the 2119 women included in the analysis, 103 (4.9%) had a pregnancy loss after recruitment, of which 11 were stillbirths beyond 20 weeks’ gestation. Among the 2016 women with a live birth, 115 (5.7%) had GH, 65 (3.2%) had preeclampsia, and 129 (6.4%) had an SGA infant.
|
Table 1
shows the values of the mean, median, and 90th percentile cutoff of tHcy concentration for each gestational age group among subjects in the referent group (e.g., normotensive subjects whose infants were live-born and not SGA). tHcy concentrations tended to decrease as gestational age increased. Table 1
also shows the mean, median, and 25th percentile cutoff for serum folate by gestational age group. Folate concentrations tended to be lower in the later gestational age groups (up to 20 weeks).
|
Table 2
presents the distribution of maternal and pregnancy characteristics by tHcy exposure among the subjects with a live birth. Serum folate concentrations in the bottom quartile were significantly associated with tHcy concentration in the top decile. Other factors statistically significantly related to tHcy concentration included marital status, family income, total pregnancy weight gain, smoking before and during pregnancy, secondhand smoke exposure, caffeine ingestion, prepregnancy folic acid use, and change in paternity.
|
Table 3
shows the adjusted models for each outcome; results are also shown without adjustment for serum folate. The adjusted RRs associated with tHcy concentration in the top decile were statistically significant for pregnancy loss (RR 2.1, 95% CI 1.2–3.6) and preeclampsia (RR 2.7, 95% CI 1.4–5.0). The RRs associated with tHcy and the outcomes were essentially unchanged in models not adjusted for folate. The risks associated with high tHcy and SGA (RR 1.5, 95% CI 0.9–2.6) and GH (RR 0.6, 95% CI 0.3–1.2) were not statistically significant. Fourteen women with preeclampsia also had an SGA infant. When these 14 women were excluded from the SGA group, the relative risk of SGA associated with high tHcy was 1.2, 95% CI 0.7–2.1. Adjustment for food consumption in the 6 h before blood collection did not alter the results for any of the outcomes.
|
When tHcy was analyzed in quartiles, the adjusted RRs (and 95% CIs) for preeclampsia in quartiles 2–4, relative to the lowest quartile, were 0.6 (0.3–1.3), 0.9 (0.4–1.8), and 1.4 (0.7–2.7), respectively. In the analysis of pregnancy loss according to quartiles, the adjusted RRs (and 95% CIs) in quartiles 2–4, relative to the lowest quartile, were 2.2 (1.0–4.7), 2.3 (1.1- 5.0), and 4.6 (2.3–9.3), respectively.
When we stratified the data by gestational age at diagnosis of preeclampsia, the RR of increased tHcy associated with diagnosis of preeclampsia did not differ according to gestational age at diagnosis. Among women diagnosed before 35 weeks’ gestation, the RR associated with increased tHcy was 2.1 (95% CI 0.7–6.4), and for women diagnosed at 35 weeks’ gestation or later, the RR was 2.4 (95% CI 1.2–4.9). The results were also compared between women who had their blood tested before 12 weeks’ gestation with those whose blood was tested at 12 weeks’ gestation or later. For preeclampsia, the magnitude of the RRs was similar among women whose tHcy was measured before and after 12 weeks’ gestation. The RR associated with the top decile of tHcy was 3.0 (95% CI 1.0–9.1) and 2.3 (95% CI 1.1–5.0) for women whose blood was sampled at <12 and 12 weeks’ gestation, respectively.
Among women with serum folate concentrations in the bottom quartile, the RR of preeclampsia associated with high tHcy was 5.6 (95% CI 1.7–18.0), whereas among women with folate concentrations above the 25th percentile, the RR of preeclampsia associated with high tHcy was 1.8 (95% CI 0.8–4.0). However, the interaction term between folate concentration and tHcy concentration was not statistically significant.
Concentrations of tHcy in the top 10th percentile for gestational age were evaluated as a screening test for each of the outcomes. The sensitivities were 18.4%, 6.1%, 21.5%, and 15.5% for pregnancy loss, GH, preeclampsia, and SGA, respectively. Specificities for each outcome were about 89.5%. The likelihood ratio for a positive test (i.e., tHcy in the top 10th percentile for gestational week) was 2.0 for preeclampsia, 1.7 for pregnancy loss, 1.5 for SGA, and 0.6 for GH.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
tHcy concentrations decrease as pregnancy progresses (18), so increased tHcy was defined according to the gestational age–specific distribution of tHcy among the normotensive women in this study and was set, a priori, at the top decile. In a post hoc analysis, we also evaluated the relationship between the top quartile of tHcy and the outcomes. We did not observe a statistically significant increase in risk of preeclampsia when increased tHcy was defined as the top quartile, and there was no evidence of a dose–response relationship according to quartile of tHcy concentration. This is consistent with other prospective studies of preeclampsia and tHcy (6)(20), but small sample sizes in most studies make it difficult to identify a dose–response relationship. We did observe a dose–response relationship between quartiles of tHcy concentration and pregnancy loss, however. A continuous dose–response relationship has been described between tHcy concentrations and ischemic heart disease, where much larger sample sizes are common (21)(22). Alternatively, the dose response effects seen with ischemic heart disease may arise because of prolonged periods of exposure, usually decades.
Our study also found that an increase in tHcy concentration before 12 weeks’ gestation, well before the clinical manifestation of preeclampsia, was associated with increased risk of preeclampsia. It is unclear whether the relationship between increased tHcy and preeclampsia is causal or whether high tHcy is a marker for other pathology or risk factors related to preeclampsia. Increased tHcy concentrations are thought to promote vascular endothelial cell injury or dysfunction (14)(23), features of preeclampsia (24).
We did not find that high tHcy increased the risk of SGA infants, which is consistent with the findings of a study by Infante-Rivard et al. (25) However, whereas our RR was moderately, but not significantly increased (RR = 1.4), a statistically significant association in the opposite direction was found by Infante-Rivard (25). It has been suggested that preeclampsia is associated with a greater degree of endothelial dysfunction than SGA (26), which could explain the stronger association found between tHcy and preeclampsia than with tHcy and SGA. Our data support this assumption, since the RR of SGA alone (preeclamptic women excluded) associated with high tHcy was smaller than the RR for SGA that included women with preeclampsia.
Folate is required for the conversion of tHcy to methionine; without adequate folate, tHcy concentrations increase. Research has shown that increasing dietary folic acid results in a proportional reduction in plasma tHcy (27)(28). In Canada, flour and grain products have been fortified since 1998, which is consistent with the relatively high folate concentrations observed in this study (e.g., the median folate concentration observed in our cohort was 37.6 nmol/L and the lowest value was 8.6 nmol/L). In our cohort, folate was not an independent risk factor for preeclampsia or GH. Although the interaction term was not significant, the study was underpowered to determine whether folate concentration modified the effect of high tHcy on preeclampsia risk.
Although we found tHcy concentration to be associated with a doubling or more of risk of pregnancy loss and preeclampsia, high tHcy concentration does not possess the qualities of being a good screening test for predicting pregnancy loss or preeclampsia later in the pregnancy. Whereas the specificity of the test in our study was reasonable (about 90% for each of the outcomes), the sensitivity of the test was only 22% for preeclampsia and 18% for pregnancy loss. Very few studies have reported the test properties of tHcy concentrations before 20 weeks’ gestation as a screen for adverse pregnancy conditions. A recent study from Turkey found sensitivity and specificity to be 56% and 94% (for preeclampsia), respectively, for tHcy concentrations >6.2 µmol/L (which corresponded to their 95th percentile) from samples collected between 15 and 19 weeks’ gestation (3). It is difficult to explain why our observed sensitivity (22%) with a lower cutoff (90th percentile) was substantially lower than in the Turkish population, which also included nulliparous and multiparous women. Our inclusion of women who were tested earlier in pregnancy does not explain the differences, since our rates of sensitivity and specificity were unchanged when the cohort was restricted to those tested between 15 and 19 weeks. We compared the characteristics of true positives and false negatives among the subset of women with preeclampsia to determine a profile of the women who had low tHcy concentrations but had a diagnosis of preeclampsia. The group of false negatives were more likely to have folate concentrations above the 25th percentile and were somewhat more likely to be >35 years of age, be married (or common-law married), and have a female fetus (results not shown). Unfortunately, the data lacked statistical power for us to draw conclusions from this comparison.
GH and preeclampsia are generally thought to be separate diseases, but it has been suggested that the two conditions may represent different severities of a common disorder (29). It has been shown that about 20% of women with GH will progress to preeclampsia, with a higher proportion becoming preeclamptic with earlier diagnoses of GH (30). The concordance of some risk factors suggests a similar disease process in the two conditions (29). In our study, a different effect of increased tHcy concentration between GH and preeclampsia suggests the possibility of different pathologic processes in GH and preeclampsia. Because preeclampsia may be associated with reduced creatinine and uric acid clearance, increased tHcy concentrations in preeclamptic women may be explained by evolving renal dysfunction and creatinine elevation (31). Hence, these results cannot rule out the possibility that preeclampsia is a progressive form of GH.
Several limitations of this study should be noted. The small number of cases with the study outcomes limited the study power, especially in our ability to conduct subgroup analyses and to examine potential interactions. Because abruptio placenta is thought to be associated with endothelial dysfunction and vascular pathology, an association with tHcy would have provided further support for a role for increased tHcy and placenta-mediated pregnancy conditions. However, we were unable to analyze the relationship between tHcy and abruptio placenta because we did not have a sufficient number of cases of this outcome in our cohort. Previous research suggests that tHcy concentrations are higher after a period of fasting (32). Because the participants in this study were recruited when they came to the laboratory for their prenatal blood work, they were not required to have fasted before their blood test. However, it is unlikely this would have had a large impact, since the results were unchanged after adjusting for food consumption in the 6 h before blood collection.
This cohort was identified from women who had prenatal blood collected at the only obstetric hospital in Halifax County, which is only one of several options for blood collection services in Halifax County. In our setting of publicly funded healthcare, the decision to come to this centre for prenatal blood draws is largely dictated by convenience. The high recruitment rate (95%) of women who were invited to participate supports the generalizability of the findings from this study. Thus, although the cohort is not population-based, we believe it is representative of the Nova Scotia population.
This prospective cohort study has confirmed the findings of an association between increased tHcy in early pregnancy and pregnancy loss and the development of preeclampsia. These both represent multifactorial conditions, and it is unlikely that one marker will successfully predict pregnancy loss or the onset of preeclampsia. A question for future research is whether supplemental folic acid will reduce the risk of pregnancy loss or preeclampsia, as a consequence of a reduction in tHcy.
|
|
Acknowledgments |
|---|
Financial Disclosures: None declared.
Acknowledgments: We are grateful for the assistance of the study coordinators, Adelia Trenchard and Anne Spencer, and the Laboratory personnel at the Izaak Walton Killam Health Centre.
|
Footnotes |
|---|
2 Human genes: MTHFR, 5,10-methylenetetrahydrofolate reductase (NADPH).
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  B. Novakovic, M. Sibson, H. K. Ng, U. Manuelpillai, V. Rakyan, T. Down, S. Beck, T. Fournier, D. Evain-Brion, E. Dimitriadis, et al. Placenta-specific Methylation of the Vitamin D 24-Hydroxylase Gene: IMPLICATIONS FOR FEEDBACK AUTOREGULATION OF ACTIVE VITAMIN D LEVELS AT THE FETOMATERNAL INTERFACE J. Biol. Chem., May 29, 2009; 284(22): 14838 - 14848. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
