Hyperhomocyst(e)inemia is associated with increased risks of stroke, myocardial infarction, and thrombosis.1 Homocyst(e)ine action is poorly understood but it might injure the vascular endothelium by generating hydrogen peroxide and impairing basal nitric oxide production.2 Folic acid, vitamin B6, and B12 supplementation can decrease plasma homocyst(e)ine concentrations; therefore, hyperhomocyst(e)inemia might be a treatable risk factor for vascular disease.3
Endothelial cell dysfunction appears to be the common denominator for several pathophysiologic changes in preeclampsia.4 We recently reported elevated homocyst(e)ine levels in 20 pregnant nulliparas with preeclampsia.5 Based on the findings from our earlier study and a rapidly expanding body of evidence linking homocyst(e)ine to endothelial dysfunction, we studied a larger group of women to better assess the risk of preeclampsia with elevated homocyst(e)ine concentrations. We also examined homocyst(e)ine concentrations among eclamptic women.
Materials and Methods
This case-control study was conducted at Harare Maternity Hospital, a University of Zimbabwe Medical School-affiliated hospital, from June 1995 through April 1996. Harare Maternity Hospital is a referral hospital that is part of the greater Harare Maternity Unit, which comprises Harare Maternity Hospital and nine clinics located in Harare's high-density suburbs, and delivers 36,000infants annually.
Potential cases were identified by daily surveillance of labor and delivery log books and medical records. Study subjects were recruited during their postpartum stays in the hospital. Preeclampsia was defined as a sustained 15 mm Hg diastolic increase or a 30 mm Hg systolic increase in blood pressure. If first-trimester blood pressures were unknown, preeclampsia was defined as persistent (6 or more hours) blood pressure of at least 140/90 mm Hg. Proteinuria was defined as urine protein concentration of 0.1 g/L or more in at least two random specimens collected at least 4 hours apart. Eclampsia was defined as above with the additional stipulation of maternal seizures. Nulliparity was not a criterion for diagnosis for this investigation. All 200 eligible women who were approached agreed to participate, 37 of whom had eclampsia and 163 of whom had preeclampsia.
Control subjects were women with normotensive pregnancies. Each day during enrollment, they were numbered in the order in which they were admitted and delivered within 2 hours of a case, and were approached in the order in which they were identified by research personnel. Of 201 control subjects approached, 200 (99%) agreed to participate.
Blood samples were available for 191 cases and 191 controls. After excluding 20 cases and six control subjects with chronic hypertension diagnosed before pregnancy or during the first 20 weeks of the index pregnancy, 171 cases, 33 with eclampsia and 138 with preeclampsia, and 185 normotensive control subjects were studied. This investigation was approved by the Medical Research Council of Zimbabwe and the Human Subjects Committee of the University of Washington Medical Center.
A structured interview questionnaire, administered during participants' postpartum hospital stays, was used to collect information on maternal sociodemographic, medical, reproductive, and lifestyle characteristics. Maternal anthropometric measurements (height, weight, and midarm circumference) were made during participants' postpartum hospital stays. Prepregnancy weight is difficult to obtain in developing countries, and because women who delivered at the study hospital were unlikely to be able to report prepregnancy weights, we elected a priori to use maternal midarm circumference, measured at delivery, as our primary measure of maternal prepregnancy adiposity. Women with midarm circumferences greater than 25 cm (the median value for normotensive controls) were classified as obese. Body mass index was calculated as weight (in kilograms) divided by height (in meters) squared. Information on maternal weight gain during pregnancy was not available for the study population.
Blood samples collected in 10-mL tubes with ethyl-enediamine tetraacetic acid, 12 to 72 hours postpartum, were immediately transported in a cooler with ice to the Reproductive Biology Laboratory, Parirenyatwa Hospital, University of Zimbabwe Medical School. At the laboratory, plasma was separated and erythrocytes were washed three times in a standard manner.6 Washed erythrocytes and plasma were divided into 1.0- to 2.0-mL aliquots, placed in cryovials, and stored at −70C. Plasma homocyst(e)ine concentrations were measured by high-performance liquid chromatography and electrochemical detection, as described previously, with minor modifications.7 The average interassay coefficient of variation was 7.2%. All laboratory analyses were masked to case or control status.
We examined frequency distributions of maternal sociodemographic characteristics, medical, and reproductive histories, according to case or control status. To estimate the relative association between varying concentrations of homocyst(e)ine and risk of preeclampsia, we categorized each subject according to quartiles determined by distribution of plasma homocyst(e)ine concentrations among normotensive controls and calculated odds ratios (ORs).8 Using the lowest quartile as the reference group, ORs and their confidence intervals (CI) were estimated. The Mantel extension test for linear trend in proportions9 was used in univariate analyses to test for a linear component of trend in risks between preeclampsia and homocyst(e)ine concentrations. Logistic regression was used to calculate maximum likelihood estimates for coefficients, and their standard errors were used to calculate ORs and CIs, adjusted for confounders.10 In multiple logistic regression models, significance for monotonic trends was assessed by treating the four quartiles as continuous variables after assigning scores as their values.10 To assess confounding, we entered variables into a logistic regression model one at a time, then compared the adjusted and unadjusted ORs. Final logistic regression models included covariates that altered unadjusted ORs by at least 10%. Effect modification was evaluated by stratified analyses and by including appropriate interaction terms in logistic regression models. Results of analyses stratified according to severity of disease (ie, eclampsia and preeclampsia) were sufficiently different to merit reporting OR estimates for the two case groups separately. All reported P values are two-tailed, and CIs were calculated at the 95% level.
Table 1 shows the characteristics of study participants. Median plasma homocyst(e)ine concentrations were 33% and 21% higher in the eclampsia (12.54 μmol/L) and preeclampsia (12.77 μmol/L) groups, respectively, compared with the control group (9.93 μmol/L, Kruskal-Wallis test P < .001 for comparison of each case group versus control group). The ORs of eclampsia increased significantly across increasing quartiles of concentrations of maternal plasma homocyst(e)ine (P < .001 unadjusted for linear trend in risk across quartiles). Women in the highest quartile had a 6.2 times higher risk of eclampsia than women in the lowest quartile (OR 6.16, 95 % CI 1.69, 22.40). After adjusting for potential confounding by maternal age, adiposity, parity, gestational age at delivery, twin pregnancy, residence, and use of prenatal vitamins, the ORs between extreme quartiles decreased slightly (adjusted OR 6.03, 95% CI 1.48, 24.52, P = .019 adjusted for linear trend in risk). Compared with women in the lowest quartile of plasma homocyst(e)ine, women in the highest decile (11 eclampsia cases and 20 control subjects with plasma homocyst(e)ine concentrations at or above 14.5 μmol/L) had a 9.2 times higher risk of eclampsia (unadjusted OR 9.17, 95% CI 2.31, 36.36). After controlling for maternal age, adiposity, parity, gestational age at delivery, twin pregnancy, residence, and use of prenatal vitamins, women with the highest decile, compared with women in the lowest quartile, remained at very high risk of developing eclampsia (adjusted OR 18.48, 95% CI 2.40, 142) (Table 2). The wide 95% CI is evidence that the association is statistically unstable, and should be interpreted with caution.
We also calculated OR estimates for preeclampsia in relation to increasing levels of maternal plasma homocyst(e)ine concentrations (Table 2). After adjusting for confounding factors, women in the highest quartile (median 13.9 μmol/L) had a 4.6 times higher risk of preeclampsia than women in the lowest quartile (adjusted OR 4.57; 95% CI 2.11, 9.88). Compared with women in the lowest quartile of plasma homocyst(e)ine, women with concentrations at or above 14.5 μmol/L (the highest decile), had a 7.4 times higher risk of preeclampsia (adjusted OR 7.39, 95% CI 2.81, 19.41).
Tables 3 and 4 summarize the results from analyses of the combination of maternal adiposity and elevated homocyst(e)ine concentrations (ie, upper quartile of the distribution) on the occurrence of eclampsia and preeclampsia, respectively. Compared with nonobese women without homocyst(e)ine elevations, obese women with elevated homocyst(e)ine had a 7.9 times higher risk of preeclampsia (adjusted OR 7.91, 95% CI 3.65, 17.15). The higher risk of preeclampsia associated with obesity and elevated homocyst(e)ine concentrations was greater than the sum of the increased ORs for each factor considered independently, thus providing evidence of a greater-than-additive effect between the two characteristics and preeclampsia. Evidence for an interaction between maternal adiposity and elevated homocyst(e)ine was even stronger for eclampsia (Table 3), although inference is hampered by the relatively small sample.
Tables 3 and 4 also summarize the results from analyses of the interactive effect of parity (nulliparous versus multiparous) and elevated homocyst(e)ine concentrations on the occurrence of eclampsia and preeclampsia. Compared with multiparas without homocyst(e)ine elevations, nulliparas with elevated homocyst(e)ine levels had a 12.9 times higher risk of preeclampsia (adjusted OR 12.90, 95% CI 5.28, 31.53). The excess risk of preeclampsia associated with nulliparity and elevated homocyst(e)ine concentrations was greater than the sum of the excess risks for each factor considered independently, thus providing evidence of a greater-than-additive effect between the two characteristics and preeclampsia.
Our case-control study showed that immediately postpartum, homocyst(e)ine concentrations in eclamptic and preeclamptic women were significantly higher than those of normotensive women. Reference values for postpartum homocyst(e)ine concentrations are limited,11 so we calculated ORs for eclampsia and preeclampsia for different quartiles of homocyst(e)ine concentrations. The sharpest increases in ORs for eclampsia and preeclampsia were in the fourth quartile (above 11.9 μmol/L). Women with homocyst(e)ine concentrations at or above 14.5 μmol/L were 7.39 times more likely to have their pregnancies complicated by preeclampsia compared with women whose concentrations were less than 7.3 μmol/L. The association between preeclampsia and elevated homocyst(e)ine is stronger or similar to associations reported previously for preeclampsia and other factors, such as chronic hypertension, antiphospholipid syndrome, prepregnancy obesity, diabetes mellitus, and nulliparity.12
Homocyst(e)ine is metabolized to excitatory amino acid neurotransmitters, such as homocysteic acid and cysteine sulfinic acid, which can cause seizures and excitotoxic neuronal death in rats.13 Elevated homocyst(e)ine levels in susceptible pregnant women might also contribute to eclamptic convulsions.
We previously reported that mean (± standard deviation) homocyst(e)ine levels were higher in 20 pregnant nulliparas with preeclampsia late in gestation (mean gestational age 35 weeks) than in normotensive pregnant women (8.66 ± 3.05 versus 4.99 ± 1.11 μmol/L).5 The relatively large size of the present study allowed us to further assess ORs for preeclampsia and eclampsia, adjust for potential confounders, and assess the association of nulliparity and adiposity, respectively, with homocyst(e)ine and nulliparity, preeclampsia and eclampsia. Limited information on maternal use of vitamins and other nutritional supplements, however, hindered a full assessment of their effect on risk of preeclampsia from elevated plasma homocyst(e)ine. Our study was retrospective, so we cannot determine whether higher levels of homocyst(e)ine preceded preeclampsia or whether differences can be attributed to preeclampsia-related alterations in maternal homocyst(e)inemetabolism.
Our subjects were not uniformly fasting at the time of blood collection. Results from previous studies of non-pregnant subjects conflict regarding the effects of food intake on homocyst(e)ine levels, showing no difference, slight decreases, or slight increases.14 Although variations in fasting status might have introduced increased variability in maternal plasma homocyst(e)ine concentrations, nonfasting values might not be clinically relevant or practical in studies of pregnant and lactating women. Differential misclassification of homocyst(e)ine concentrations is unlikely because all laboratory analyses were done without knowledge of participants' disease status.
Obesity and nulliparity are well known but poorly understood risk factors for preeclampsia.12 Our results are compatible with the hypothesis that homocyst(e)ine interacts with obesity and nulliparity to increase the risk of preeclampsia, but additional information from similar studies is needed before any firm conclusions can be drawn. Although obesity alone was not a risk factor for eclampsia in our study, there was a suggestion that obesity and homocyst(e)ine interacted to increase the risk of eclampsia.
Although preeclampsia is characterized clinically by maternal high blood pressure, proteinuria, and edema, women with the disorder are also more likely than normotensive pregnant women to have metabolic disturbances similar to those of nonpregnant women with coronary heart disease. Metabolic disturbances consistently noted in preeclampsia include hypertriglyceridemia,15 excessive lipid peroxidation or oxidative stress,16 insulin resistance,15 sympathetic nervous system overreactivity,17 plasma elevations of proinflammatory cytokines,18 and an imbalance in thromboxane and prostacyclin in favor of vasoconstriction.19 In placentas of preeclamptic women compared with normotensive women, pathologic changes were similar to those seen in atherosclerosis.20 There was also considerable overlap between epidemiology of preeclampsia and coronary heart disease because obesity, sedentary life style, history of diabetes, and chronic hypertension are risk factors for both disorders. The strong association between elevated homocyst(e)ine levels and preeclampsia and eclampsia is consistent with that pattern. Potential public health efforts to reduce coronary heart disease by lowering plasma homocyst(e)ine concentrations might also modify risks of preeclampsia and eclampsia. More information is needed to determine whether periconceptional homocyst(e)ine concentrations can identify women at high risk of developing preeclampsia or eclampsia to evaluate genetic and nongenetic determinants of pregnancy-associated elevations in homocyst(e)ine; and to evaluate whether homocyst(e)ine-lowering agents, use of nutritional supplements, and manipulation of food choices are associated with decreased incidence and severity of preeclampsia and eclampsia.
1. Malinow MR. Hyperhomocyst(e)inemia. A common and easily reversible risk factor for occlusive atherosclerosis. Circulation 1990;81:2004–6.
2. Stamler JS, Osborne JA, Jaraki O, Rabbani LE, Mullins M, Singel D, et al. Adverse vascular effects of homocysteine are modulated by endothelium-derived relaxing factor and related oxides of nitrogen. J Clin Invest 1993;91:308–18.
3. Malinow M. Homocyst(e)ine and arterial occlusive diseases. J Intern Med 1994;236:603–17.
4. Roberts JM, Taylor RN, Goldfien A. Clinical and biochemical evidence of endothelial cell dysfunction in the pregnancy syndrome preeclampsia. Am J Hypertens 1991;4:700–8.
5. Rajkovic A, Catalano PM, Malinow MR. Elevated homocyst(e)ine levels with preeclampsia. Obstet Gynecol 1997;90:168–71.
6. Rose HG, Oklander M. Improved procedure for the extraction of lipids from human erythrocytes. J Lipid Res 1965;6:428–31.
7. Malinow RM, Sexton G, Averbuch M, Grossman M, Wilson D, Upson B. Homocyst(e)inemia in daily practice: Levels in coronary artery disease. Coronary Artery Dis 1990;1:215–20.
8. Hsieh CC, Maisonneuve P, Boyle P, Macfarlane GJ, Robertson C. Analysis of quantitative data by quartiles in epidemiologic studies: Classification according to cases, noncases, or all subjects. Epidemiology 1991;2:137–40.
9. Mantel N. Chi-square tests with one degree of freedom: Extensions of the Mantel-Haenszel procedure. J Am Stat Assoc 1963;58:690–700.
10. Rothman K. Modern epidemiology. Boston, Massachusetts; Little, Brown, 1986.
11. Anderson A, Hultberg B, Brattstrom L, Isaksson A. Decreased serum homocysteine in pregnancy. Eur J Clin Chem Clin Biochem 1992;30:377–9.
12. American College of Obstetricians and Gynecologists. Hypertension in pregnancy. ACOG technical bulletin no. 219. Washington DC: American College of Obstetricians and Gynecologists, 1996.
13. Kim J, Koh JY, Choi D. L-Homocysteate is a potent neurotoxin on cultured cortical neurons. Brain Res 1987;437:103–6.
14. Ueland PM, Refsum H, Stabler SP, Malinow MR, Andersson A, Allen RH. Total homocysteine in plama or serum: Methods and clinical applications. Clin Chem 1993;39:1764–79.
15. Kaaja R, Tikkanen MJ, Viinnikka L, Ylikorkala O. Serum lipoproteins, insulin, and urinary prostanoid metabolites in normal and hypertensive pregnant women. Obstet Gynecol 1995;85:353–6.
16. Walsh SW. Lipid peroxidation in pregnancy. Hypertens Pregnancy 1994;13:1–32.
17. Schobel HP, Fischer T, Heuszer K, Geiger H, Schmieder RE. Preeclampsia—a state of sympathetic overactivity. N Engl J Med 1996;335:1480–5.
18. Kupferminc MJ, Peaceman AM, Wigton TR, Rehnberg BA, Socol ML. Tumor necrosis factor-alpha is elevated in plasma and amniotic fluid of patients with severe preeclampsia. Am J Obstet Gynecol 1994;170:1752–9.
19. Fitzgerald DJ, Entman SS, Mulloy K, FitzGerald GA. Decreased prostacyclin biosynthesis preceding the clinical manifestation of pregnancy-induced hypertension. Circulation 1987;75:956–63.
20. Robertson WB, Khong TY, Brosens I, De Wolf F, Sheppard BL, Bonnar J. The placental bed biopsy: Review from three European centers. Am J Obstet Gynecol 1986;155:401–12.